SEMIT2023 Full - Conference - Book - 07 12 2023 .

Table of Contents
Publisher and Indexing .............................................................................................................. 2

Journals ...................................................................................................................................... 3
Scientific Sponsors..................................................................................................................... 5
Media Sponsors.......................................................................................................................... 6
About SEMIT 2023.................................................................................................................... 7
AYBU Preface ........................................................................................................................... 9
Conference Report ................................................................................................................... 10
Conference Organization ........................................................................................................ 13
Conference Chairs .................................................................................................................... 20
Conference Coordinators ......................................................................................................... 20
Editorial Committee Members ................................................................................................. 21
Steering Committee Members ................................................................................................. 22
Scientific Committee Members ............................................................................................... 23
Executive Team Members ....................................................................................................... 26
Keynote Speakers..................................................................................................................... 27
Workshops ............................................................................................................................... 33
Panels ....................................................................................................................................... 40
Conference Program ................................................................................................................ 42
Abstracts Proceedings .............................................................................................................. 80
Full/Extended Abstract Papers Proceedings .......................................................................... 178
Important Note: This book consists of all papers of SEMIT 2023, excluding the papers going to be
published by Springer. Those papers will be published in a booklet by Springer.
Page 1
Publisher and Indexing
Page 2
Journals
Central European Journal of
Operations Research
(ISSN: 1613-9178, 1435-246X)
Indexing: SCIE – JCR (IF:1.7)

Special Issue Title:
"Persons and Freight Transportation in the age of
Sustainable and Collaborative Logistics"
Operations Research Perspectives

(ISSN: 2214-7160)
Indexing: Scopus- JCR (IF: 2.5)
Special Issue Title:

"Resilient, Sustainable and Smart Supply Chain
Management in the Healthcare Industry: Application
of Data Analytics and Optimization Techniques"
International Journal of Supply

and Operations Management
(ISSN: 23831359, 23832525)
Indexing: SCOPUS
Journal of Turkish Operations

Management
(ISSN: 2630-6433, 2630-6433)
Indexing: EBSCO - TRDizin
Page 3
Iranian Journal of management
Studies
(ISSN: 2008-7055, 2345-3745)
Indexing: Web of Science - Scopus
Researcher
(ISSN: 2717-9494, 2717-9494)
Indexing: EBSCO
Journal of Optimization and

Decision Making
(ISSN: 2822-6410)
Indexing: Google Scholar
Page 4
Scientific Sponsors
International University of The University of Texas at Dallas, Prague University of

Rabat, Morocco USA Economics and Business,
Czech Republic
Kent Business School, UK University of Sfax, Tunisia University of Sfax, Tunisia

Faculty of Economics and
Management
Sunway University, Malaysia Ankara Science University, Turkey Manisa Celal Bayar
University, Turkey
Manufacturing Engineering Abou Bekr Belkaid Tlemcen Suleyman Demirel University,

Laboratory of Tlemcen, Algeria University, Algeria Turkey
Halic University, Turkey Kohat University of Science & The Laboratory of Operations
Technology, Pakistan Research, decision and
Control, Tunis University,
Tunisia
Page 5
Media Sponsors
The conference team would like to thank to Laboratoire de Recherche

Opérationnelle, de Décision et de Contrôle de processus (LARODEC),
Institut Supérieur de Gestion, Université de Tunis for submitting the
papers and holding a good quality panel, while using the conference Bronze
promotional package.
LARODEC (Director: Nahla Ben Amor), the Laboratory of Operations Research, Decision,
and Control, is a prominent research institution affiliated with Tunis University in Tunisia.
LARODEC focuses on a wide array of both theoretical and practical research domains.
LARODEC is dedicated to conducting research in various fields, including:
• Optimization and Operations Research: Exploring mathematical and algorithmic
approaches to optimize complex systems and decision-making processes.
• Decision Support: Developing methodologies and tools to aid individuals and
organizations in making informed decisions.
• Process Control and Quality: Investigating techniques for monitoring and enhancing the
quality of processes in diverse industries.
• Random Models and Computer Science: Studying probabilistic models and their
applications in computer science and related areas.
• LARODEC's mission is to contribute to the advancement of knowledge in these research
domains, and promote interdisciplinary collaboration between academia and industry.
Over the years, LARODEC has made significant contributions to research and has
established itself as a hub for scientific excellence. Its researchers have published
numerous papers in reputable journals and have undertaken projects with both national
and international partners.
LARODEC actively collaborates with various academic and industrial institutions, fostering
partnerships that drive innovation and address complex challenges.
The laboratory also plays a pivotal role in nurturing the next generation of researchers and
professionals by providing a dynamic and stimulating academic environment for students
pursuing advanced degrees.
Page 6
About SEMIT 2023
International Conference on Science, Engineering Management and Information Technology
2023 is a gathering of faculty members, scholars, educators, industry experts, and university
students from all around the world. SEMIT 2023 aims to provide a platform to transfer
information, experience, and scientific and research findings of the recent theoretical and
practical achievements of engineering management and IT in both research and industry. The
conference is an extraordinary opportunity to encourage experts to respond to current national
and international engineering management and IT issues and to enhance university-industry
collaboration. In addition to paper presentations and invited lectures keynote speech, the
International Conference on Science, Engineering Management and IT 2023 is also going to
provide useful workshops and many other programs, the information of which is presented on
its website (https://semit2023.refconf.com/). The conference warmly welcomes all researchers,
lecturers, and students, as well as managers and experts in the field of engineering management
and IT.
The Conference language is English and the papers could be in English and Turkish. The
selected English papers will be published in Springer's CCIS book series or will be sent to the
specified journals on the conference website, which are indexed in the well-known indexing
databases, such as Scopus.
With the increasing development in the field of Industrial Engineering and Information
Technology, RefConf organization decided to hold the next International Conference on
Science, Engineering Management and Information Technology (SEMIT 2023) in cooperation
with Ankara Yildirim Beyazit University (AYBÜ).
Presenting numerous papers in 24 specified panels, scientific and practical lectures by keynote
speakers and prominent professors, as well as holding workshops, attracted participants from
different countries to the conference.
Conference Topics
All submissions falling within the broad area of Science, Engineering Management and
Information Technology are welcome. In addition to research papers, case studies would be
perfectly welcome. Full papers and abstracts are accepted for presenting in the conference and
publishing in the conference proceeding. The conference topics include, but not limited to the
following subjects:
• IT and EM based case studies of manufacturing/ service industries (including
automotive, food, tourism, petroleum, healthcare, insurance and banking, energy, etc.)
• E-government, E-commerce, E-learning
• Marketing and E-marketing for resources management
• Data science, big data, data mining and knowledge management in EM
• Decision making and support systems in an uncertain environment and risk
management
• Industry 4.0, supply chain 4.0, and logistics 4.0
• Supply chain management (green SCM, sustainable SCM, agile SCM, JIT SCM,
global SCM, etc.)
• Optimization and decision making: methods and algorithms
Page 7
• Inventory control, production planning and scheduling
• Applied soft computing in engineering management
• Metaheuristic algorithms and applications
• Quality and productivity management and tools
• Project management
• Blockchain in engineering management
• Artificial intelligence and expert systems
• Digital city
• Internet of things (IoT)
• Other fields of study related to EM and IT
The International Conference on Science, Engineering Management and Information
Technology was held on February and September 2022. SEMIT 2022-Feb. had the
support of 35 international scientific committee members from 19 countries and it had a
total of 138 papers. The other conference (SEMIT 2022- Sep.), was held with the support
of 36 international scientific committee members from 22 countries. A total of 123 papers
had been submitted and 66 ones were accepted for presentation in the conference and their
abstracts were published in the conference book.
Page 8
AYBU Preface
We are honored to present the proceedings of the INTERNATIONAL CONFERENCE ON
SCIENCE, ENGINEERING MANAGEMENT, AND INFORMATION TECHNOLOGY
(SEMIT 2023) to the program's authors and delegates. We hope that it will be a useful, exciting,
and inspiring resource for all readers.
SEMIT 2023 aims to provide a platform for discussion engineering management and
information technology concerns, challenges, and possibilities. The rapidly growing scope and
evolution of engineering management has resulted in new issues and questions. We feel the
papers submitted for SEMIT 2023 provide important tools to address these issues.
The reaction to the request for papers was very positive, with submissions coming from Turkey,
Iran, and other countries. Unfortunately, according to the results of the reviews and our capacity
restrictions, a few manuscripts from prestigious institutions were not accepted. We'd like to
offer our gratitude and appreciation to all of the reviewers who assisted us in reviving some of
our SCI, Scopus and TRdizin papers. We also like to express our gratitude to the members of
the organizing team for their efforts.
We wish all of the attendees of the next SEMIT a beautiful and productive conference, as well
as our international visitors a pleasant stay in Ankara.
Prof. Dr. Hasan Okuyucu

Dean of Faculty of Engineering and Natural Sciences
Page 9
Conference Report
Prof. A. Mirzazadeh,
Kharazmi University, Tehran, Iran
RefConf organization holds International Conference on

Science, Engineering Management and Information Technology
(SEMIT 2023) in cooperation with Ankara Yildirim Beyazit
University to keep pace with the development of Industrial
Engineering and Information Technology .SEMIT 2023 is a good opportunity to encourage
experts to respond to current engineering management and IT issues and to enhance
university-industry collaboration. Presenting numerous papers in 24 specified panels,
scientific and practical lectures by keynote speakers and prominent professors, as well as
holding workshops, attracted participants in the conference.
One of the features of the conference was the benefit of scientific support from 15
universities and associations, including International University of Rabat, The University
of Texas at Dallas, Prague University of Economics and Business, University of Kent,
University of Sfax and its Faculty of Economics and Management, Sunway University,
Ankara Science University, Manisa Celal Bayar University, Manufacturing Engineering
Laboratory of Tlemcen, Abou Bekr Belkaid Tlemcen University, Suleyman Demirel
University, Halic University, Kohat University of Science & Technology, and The
Laboratory of Operations Research, decision and Control of Tunis University. Selected
papers will be published by seven scientific journals: 1) Central European Journal of
Operations Research (CEJOR), 2) Operations Research Perspectives (ORP), 3)
International Journal of Supply and Operations Management (IJSOM), 4) Journal of
Turkish Operations Management (JTOM), 5) Iranian Journal of management Studies
(IJMS), 6) Researcher Journal and 7) Journal of Optimization and Decision Making
(JODM). The valuable presence of three keynote speakers and two opening and closing
speeches from France, Poland, Turkey, Mexico, and Vietnam in addition to the
participation of more than 300 authors, are among the other features that have given
credence to this conference.
The International Conference on Science, Engineering Management and Information

Technology, is held having the support of 36 international scientific committee members
from 17 countries. A total of 409 papers have been submitted in SEMIT 2023, 133 papers
of which were accepted for presentation in the conference that their abstracts are published
in this book and their full version will be published by conference full papers proceedings
and journals.
Page 10
Conference Committees
20
16
12
8
4
0
Figure 1. Geographical Diversity of Conference Committees
Comparing the current conference with the previous SEMITs (2022- Feb. and 2022- Sep.)
shows considerable increase in the number of papers accepted for presentation at the
conference.
Presented Papers
132
150 93
69
100
50
0
SEMIT 2022- SEMIT 20222- SEMIT 2023
Feb. Sep.
Figure 2. SEMIT 2023 Compared to SEMITs 2022
The countries’ share in the presented papers includes about 32% from Turkey and the rest were
from other countries in the world according to the chart below.
Other Countries’ Participation

20%
15%
10%
5%
0%
Figure 3. Coutnries’ Participation in SEMIT 2023
Page 11
Selected English papers will be published in Springer's CCIS book series. Articles published
in CCIS will be indexed in the following databases:
• Scopus
• SCImago
• EI-Compendex
• DBLP
• Google Scholar
• Mathematical Reviews
CCIS volumes are also submitted for the possible inclusion in Web of Science, which will be
evaluated by Calrivate Analytics.
We hope that this conference could have been a step towards promoting the Engineering
Management and Information Technology as well as the multidisciplinary fields
related to the two main fields of EM and IT. At the end, it is worth thanking all the
participants in holding this event. This success is due to the motivation, perseverance and
unremitting efforts of dear friends and colleagues at Ankara Yildirim Beyazit University
and other universities (board members and students), panel chairs, steering committee, the
scientific committee, reviewers, keynote speakers, workshop speakers, executive teams, panel
members, presenters, attendees and others.
Page 12
Conference Organization
Conference Chairs: Prof. A. Mirzazadeh
Prof. Mete Gundogan

Yildirim Beyazit University, Ankara, Turkey
Ankara Yildirim Beyazit University Prof. Dr. Ali Cengiz KÖSEOĞLU

Rector
Prof. Hasan Okuyucu

Dean of Faculty of Engineering and Natural Sciences
Dr. Babek D. Erdebilli,

AYBU, IE Department Board Member
Conference Coordinators: Leyla Chehrghani

(Planning Manager)
Dr. Zohreh Molamohamadi

(Scientific Coordinator)
Dr. Roya Soltani

(Development Manager)
Khatam University, Tehran, Iran
Steering Committee Members Prof. A. Mirzazadeh

Kharazmi University, Iran
Assoc. Prof. Dr. Babek Erdebilli

Yildirim Beyazit University, Turkey
Prof. Janny M.Y. Leung

University of Macao
President of IFORS (International Federation of Operational
Research Societies)
Prof. Gerhard Wilhelm Weber

Poznan University of Technology, Poland
Prof. Kok Lay Teo

JIMO Editor in Chief
Sunway University, Malaysia
Curtain University, Australia
Prof. Ruben Ruiz Garcia

University of Polytechnics, Spain
Prof. Sırma Zeynep Alparslan Gök

Suleyman Demirel University, Turkey
Page 13
Prof. Chefi Triki
University of Kent, UK
Dr. Yavuz Selim Ozdemir

Ankara Science University, Turkey
Dr. Ibrahim Yilmaz

Dr. Abdullah Yildizbasi

Editorial Committee Members: Prof. Gerhard Wilhelm Weber

Poznan University of Technology, Poznan, Poland
Dr. Arpan Kumar Kar

Indian Institute of Technology Delhi, New Delhi, India
Dr. Erfan Babaee Tirkolaee

Istinye University, Istanbul, Turkey
Prof. Elif Kılıç Delice

Ataturk University, Erzurum, Turkey
Dr. Zohreh Molamohamadi

Dr. Mustapha Oudani

International University of Rabat, Rabat, Morocco
Dr. Eloisa Macedo

University of Aveiro, Portugal
Dr. Tatiana Tchemisova

University of Aveiro, Aveiro, Portugal
Dr. Dinh Tran Ngoc Huy

International University of Japan, Japan
Scientific Committee Members: Prof. Kathryn Stecke

University of Texas at Dallas,
Naveen Jindal School of Management, USA
Prof. Josef Jablonsky

University of Economics and Business, Prague, Czech Republic
Prof. Bernardo Almada-Lobo

University of Porto, Portugal
Prof. Taicir Moalla Loukil

University of Sfax, Tunisia
President of The Tunisian Operational Research Society
Prof. Stefan Pickl

Bundeswehr University, Germany
Page 14
Prof. Alexandre Dolgui
IMT Atlantique (former Ecole des Mines de Nantes), France
Prof. Leopoldo Eduardo Cárdenas-Barrón

School of Engineering and Sciences, Tecnológico de Monterrey,
Mexico
Prof. Jie Sun

National University of Singapore
Prof. Ergun Eraslan

Ankara Yildirim Beyazit University, Turkey
Prof. Mehmet Kabak

Gazi University, Turkey
Prof. Mete Gundogan

Dr. Sankar Kumar Roy

Vidyasagar University, Midnapore, West Bengal, India
Prof. Sadia Samar Ali

King Abdul Aziz University, Saudi Arabia
Dr. Michael G. Kay

North Carolina State, University, USA
Dr. Aybike Özyüksel Çiftçioğlu

Manisa Celal Bayar University, Turkey
Dr. Burçin Kaplan

Istanbul Aydin University, Turkey
Dr. Paulina Golinska

Poznan University of technology, Poland
Prof. safa Bhar Layeb

University of Tunis EI Manar, Tunisia
Dr. Hakan Gultekin

Sultan Qabus University, Oman
Dr. Eren Ozceylan

Gaziantep University, Turkey
Dr. Nasr Hamood Mohamed Al-Hindi

Sultan Qaboos University, Oman
Dr. Sujan Piya

University of Sharjah, United Arab Emirates
Dr. Sita Ram Sharma

Chitkara University, India
Dr. Abdullah Yildizbasi

Page 15
Dr. Maher Agi
Rennes School of Business, France
Dr. Vijai Kumar Gahwalat

National Institute of Food Technology Entrepreneurship and
Management, India
Dr. Baris Bulent Kirlar

Suleyman Demirel University, Turkey
Dr. Marwa Hasni

University of Sfax, Tunisia
Dr. Mohammed Khalil

Halic University, Turkey
Dr. Peren Jerfi Janatalay

Dr. Yavuz Can

Prof. M. M. Nadakatti
KLS Gogte Institute of Technology, India
Dr. Kamran Yeganegi

Islamic Azad University, Zanjan Branch, Iran
Dr. M. Irfan Uddin

Kohat University of Science and Technology, Pakistan
Dariusz Jacek Jakóbczak

Koszalin University of Technology, Koszalin, Poland
Dr. Sadeq Damrah

Australian University – Kuwait
Executive Team Members: Sadaf Soltani

BSc in Information Technology, IT Expert in Devops
University, Iran
Ahmet Süha Hancıoğlu

Ankara Yıldırım Beyazıt University, Turkey
Dr. Gül Uslu

Asist. Prof., Sinop University, Turkey
Aslı Ece Acar Filizci

Melda Kevser
Mert Eyüboğlu
Yıldırım Beyazıt University, Turkey
Page 16
Alperen Oncel
Nurullah Güleç
Seyed Masoud Pourmohammadi Sani

Master Student, Ankara Yıldırım Beyazıt University, Turkey
Emine Nur Nacar

Kağan Özyurt
Opening and Closing Sessions Keynote Speakers:
Keynote Speaker (Opening Ceremony): Prof. Alexandre Dolgui

Fellow of IISE, Distinguished Professor
Head of Automation, Production and Computer Sciences
Department, France
Editor-in-Chief, International Journal of Production Research
Keynote Speaker (Closing Ceremony) Prof. Gerhard-Wilhelm Weber

Poznan University of Technology, Poland,
EURO Conference Advisor,
Chair of IFORS Developing Countries Online Resources,
Honorary chair of some EURO and international working groups,
Institute of Applied Mathematics, METU, Ankara, Turkey
Keynote Speakers: Prof. Sırma Zeynep Alparslan Gök

Department of Mathematics, Faculty of Arts and Sciences,
Süleyman Demirel University, Isparta-Türkiye
Dr. Mostafa Hajiaghaei-Keshteli

Associate Professor (Research), Tecnologico de Monterrey,
Mexico
Prof. Anand Nayyar

Vice-Chairman (Research) and Director- IoT and Intelligent
Systems Lab, School of Computer Science-Duy Tan University,
Da Nang, Vietnam
Workshop Holders: Prof. Paulina Golinska-Dawson, D.Sc. Ph.D., prof PUT

Professor at the Poznan University of Technology (PUT),
Institute of Logistics, Poznan, Poland
Prof. Meenakshi Kaushik, professor, dean (research)

Trinity Institute of Innovations in Professional Studies,
Guru Gobind Singh Indraprastha University, Delhi, NCR, India
Pr. Safa Bhar LAYEB

LR-OASIS, National Engineering School of Tunis, University of
Tunis El Manar, Tunisia
Page 17
Dr. Marwa Hasni
LR-OASIS, National Engineering School of Tunis, University
of Tunis El Manar, Tunisia
Panel Chairs: Dr. Babek Erdebili

Dr. Sonia Nasri

LARODEC laboratory, University of Manouba, Tunisia
Prof. Hakan Gultekin

Sultan Qaboos University, Oman
Dr. Vijay Kumar Gahlawat

Assistant Professor, Department of Basic and Applied Sciences,
NIFTEM Kundli, India
Dr. Mazdak Khodadadi

University of Science and Culture, Tehran, Iran
Dr. Eloisa Macedo

University of Aveiro, Portugal
Dr. Erfan Babaee Tirkolaee

Istinye University, Istanbul, Turkey
Prof. Paulina Golinska-Dawson

Poznan University of Technology (PUT), Poznan, Poland
Prof. Rohit Bansal

Department of Management Studies Vaish College of
Engineering, Rohtak, India
Dr. Sita Ram Sharma

Chitkara University, India
Mr. Nguyễn Thành Luân

HUFLIT, Vietnam
Dr. Alireza Goli

Department of Industrial Engineering and Future Studies, Faculty
of Engineering, University of Isfahan
Prof. Sadia Samar Ali

Department of Industrial Engineering, Faculty of Engineering,
King Abdulaziz University, Saudi Arabia
Prof. Meenakshi Kaushik

Trinity Institute of Innovations in Professional Studies (Guru
Gobind Indraprastha University), India
Dr. Aybike Özyüksel Çiftçioğlu

Manisa Celal Bayar University, Turkey
Dr. Serap Ergun

Isparta University of Applied Sciences, Turkey
Page 18
Prof. Nazanin Pilevari
Islamic Azad University, West Tehran Branch, Iran
Dr. Mehmet Kabak

Gazi University, Turkey
Dr. Hamidreza Irani

University of Tehran, Iran
Dr. Simerjeet Singh Bawa

Himalayan Institute of Management, Himalayan Group of
Professional Institutions, Kala-Amb, Distt- Sirmour, H.P., India
Panel Co-chairs Dr. Wided Oueslati

Business Higher School of Tunis, Manouba University,
BESTMOD Laboratory, Tunisia
Prof. Rakesh Kumar

NUST, Windhoek, Namibia
Dr. Kumar Rahul

Mr. Mohit Malik

Dr. Nguyen Duc Hoai Anh

HUFLIT, Vietnam
Mr. Tri-Quan Dang

HUFLIT, Vietnam
Prof. Asif Akhtar

Assistant Prof. Aligarh Muslim Central University, India
Dr. Asad Rehman

Associate Professor, Department of Business Administration,
Aligarh Muslim University, Aligarh
Dr. Soheyl Khalilpourazari

Concordia University, Montreal, Canada
Dr. Hamed Nozari

Islamic Azad University, Dubai Branch, UAE
Dr. Chakir Aziza

Faculty of Law, Economics and Social Sciences, Hassan II
University, Casablanca, Morocco
Dr. Ashok Kumar

Chitkara University, Himachal Pradesh
Page 19
Conference Chairs
Prof. A. Mirzazadeh Prof. Mete Gundogan

Kharazmi University, Yildirim Beyazit University,
Tehran, Iran Ankara, Turkey
Conference Coordinators
Planning Manager Scientific Coordinator Development Manager
Leyla Chehrghani Dr. Zohreh Dr. Roya Soltani

Kharazmi University, Tehran, Molamohamadi Khatam University,
Iran Kharazmi University, Tehran, Iran
Tehran, Iran
Page 20
Editorial Committee Members
Prof. Gerhard Dr. Arpan Kumar Dr. Erfan Babaee Prof. Elif Kılıç Delice
Wilhelm Weber Kar Tirkolaee Ataturk University,
Poznan University of Indian Institute of Istinye University Erzurum, Turkey
Technology, Technology Delhi, Istanbul, Turkey
Poznan, Poland New Delhi, India
Dr. Zohreh Dr. Mustapha Oudani Dr. Eloisa Macedo

Molamohamadi International University University of Aveiro,
Kharazmi University, of Rabat, Portugal
Tehran, Iran Rabat, Morocco
Dr. Tatiana Tchemisova Dr. Dinh Tran Ngoc Huy

University of Aveiro, International University of
Aveiro, Portugal Japan, Japan
Page 21
Steering Committee Members
Prof. A. Mirzazadeh Dr. Babek Erdebilli Prof. Janny M.Y. Leung Prof. Gerhard
Kharazmi University, Yildirim Beyazit University of Macao Wilhelm Weber
Iran University, Turkey President of IFORS Poznan University of
(International Federation Technology, Poland
of Operational Research
Societies)
Prof. Kok Lay Teo Prof. Ruben Ruiz Prof. Sırma Zeynep Prof. Chefi Triki
JIMO Editor in Chief Garcia Alparslan Gök University of Kent, UK
Sunway University, University of Suleyman Demirel
Malaysia, Polytechnics, Spain University, Turkey
Curtain University,
Australia
Dr. Yavuz Selim Dr. Ibrahim Yilmaz Dr. Abdullah

Ozdemir Yildirim Beyazit Yildizbasi
Ankara Science University, Turkey Yildirim Beyazit
University, Turkey University, Turkey
Page 22
Scientific Committee Members
Prof. Kathryn Stecke Prof. Josef Jablonsky Prof. Bernardo Prof. Taicir Moalla
University of Texas at University of Almada-Lobo Loukil
Dallas, Naveen Jindal Economics and Porto University, University of Sfax,
School of Business, Portugal Tunisia
Management, USA Prague, Czech Republic President of The
Tunisian Operational
Research Society
Prof. Stefan Pickl Prof. Alexandre Dolgui Prof. Leopoldo Eduardo Prof. Jie Sun
Bundeswehr IMT Atlantique (former Cárdenas-Barrón National University of
University, Germany Ecole des Mines de School of Engineering Singapore
Nantes), Nantes, France and Sciences,
Tecnológico de
Monterrey, Mexico
Prof. Ergun Eraslan Prof. Mehmet Kabak Prof. Mete Gundogan Dr. Sankar Kumar
Yildirim Beyazit Gazi University, Yildirim Beyazit Roy
University, Turkey Turkey University, Turkey Vidyasagar University,
Midnapore,
West Bengal, India
Page 23
Dr. Sadia Samar Ali Dr. Michael G. Kay Dr. Aybike Özyüksel Dr. Burçin Kaplan
King Abdul Aziz North Carolina State Çiftçioğlu Istanbul Aydin
University, Saudi University, USA Manisa Celal Bayar University, Turkey
Arabia University, Turkey
Dr. Paulina Golinska Prof. safa Bhar Layeb Dr. Hakan Gultekin Dr. Eren Ozceylan
Poznan University of University of Tunis EI Sultan Qabus Gaziantep University,
technology, Poland Manar, Tunisia University, Oman Turkey
Dr. Nasr Hamood Dr. Sujan Piya Dr. Sita Ram Sharma Dr. Abdullah
Mohamed Al-Hindi University of Sharjah, Chitkara University, Yildizbasi
Sultan Qaboos United Arab Emirates India Yildirim Beyazit
University, Oman University, Turkey
Page 24
Dr. Maher Agi Dr. Vijai Kumar Dr. Baris Bulent Dr. Marwa Hasni
Rennes School of Gahwalat Kirlar University of Sfax,
Business, France National Institute of Suleyman Demirel Tunisia
Food Technology University, Turkey
Entrepreneurship and
Management, India
Dr. Mohammed Dr. Peren Jerfi Dr. Yavuz Can Prof. M. M. Nadakatti
Khalil Janatalay Halic University, KLS Gogte Institute of
Halic University, Halic University, Turkey Technology, India
Turkey Turkey
Dr. Kamran Dr. M. Irfan Uddin Dariusz Jacek Dr. Sadeq Damrah
Yeganegi Kohat University of Jakóbczak Australian University,
Islamic Azad Science and Koszalin University of Kuwait
University, Zanjan Technology, Pakistan Technology,
Branch, Iran Koszalin, Poland
Page 25
Executive Team Members
Sadaf Soltani Ahmet Süha Hancıoğlu Dr. Gül Uslu

BSc in Information Technology, IT Ankara Yıldırım Beyazıt Sinop University,
Expert in Devops University, Iran University, Turkey Turkey
Aslı Ece Acar Filizci Melda Kevser Akgün Mert Eyüboğlu Alperen Oncel
Ankara Yıldırım Ankara Yıldırım Ankara Yıldırım Ankara Yıldırım
Beyazıt University, Beyazıt University, Beyazıt University, Beyazıt University,
Turkey Turkey Turkey Turkey
Emine Nur Nacar Nurullah GÜLEÇ Seyed Masoud Kağan Özyurt

Ankara Yildirim Ankara Yıldırım Pourmohammadi Sani AnkaraYıldırım Beyazıt
Beyazit University, Beyazıt University, Ankara Yıldırım University, Turkey
Turkey Turkey Beyazıt University,
Turkey
Page 26
Keynote Speakers
Opening Ceremony Keynote Speaker
AI in Manufacturing: Example of a
Large European Project ASSISTANT
Prof. Alexandre Dolgui

Fellow of IISE, Distinguished Professor
Head of Automation, Production and Computer Sciences
Department, France
Editor-in-Chief, International Journal of Production Research
Short Biography
Prof. Alexandre Dolgui is a Fellow of IISE, a Distinguished Professor of Industrial and Systems
Engineering, and the Head of the Automation, Production, and Computer Sciences Department at
IMT Atlantique, campus in Nantes, France. His research focuses on manufacturing line design,
production planning and scheduling, and supply chain engineering. His main results are based on
exact mathematical programming methods and their intelligent coupling with automatic control
techniques, heuristics and metaheuristics algorithms. He has contributed to the theory of assembly
line balancing, combinatorial design of machining lines, process planning, supply chain
scheduling, lot sizing and replenishment planning under uncertainties as well as to the theory of
resilience and risk analysis in supply networks. He is the author of over 800 publications and
communications, numerous books and articles, the editor-in-chief of the International Journal of
Production Research, an area editor of Computers & Industrial Engineering, a member of the
editorial boards for several other journals, including the International Journal of Production
Economics, Fellow of the European Academy for Industrial Management, Member of the Board
of the International Foundation for Production Research, former Chair of IFAC TC 5.2
Manufacturing Modelling for Management and Control, and Member of IFIP WG 5.7 Advances
in Production Management Systems. Prof. Alexandre Dolgui has received numerous international
and French awards for his research: best PhD theses, best articles in leading journals, ... He is in
the short list of Highly Cited Researchers in the field of Engineering by Clarivate.
Page 27
Closing Ceremony Keynote Speaker
Optimal Management of Defined Contribution

Pension Funds under the Effect of Inflation, Mortality
and Uncertainty
Prof. Gerhard Wilhelm Weber

Poznan University of Technology, Poland,
EURO Conference Advisor,
Chair of IFORS Developing Countries Online Resources,
Honorary chair of some EURO and international working groups,
Institute of Applied Mathematics, METU, Ankara, Turkey
Short Biography
Prof. Gerhard-Wilhelm Weber is a Professor at Poznan University of Technology, Poznan,
Poland, at Faculty of Engineering Management. His research is on mathematics, statistics,
operational research, data science, machine learning, finance, economics, medicine, neuro-, bio-
and earth-sciences, development, cosmology and spirituality. He is involved in the organization
of scientific life internationally. He received Diploma and Doctorate in Mathematics, and
Economics / Business Administration, at RWTH Aachen, and Habilitation at TU Darmstadt
(Germany). He replaced professorships at University of Cologne, and TU Chemnitz, Germany.
At Institute of Applied Mathematics, Middle East Technical University, Ankara, Turkey, he was
a Professor in Financial Mathematics and Scientific Computing, and Assistant to the Director,
and has been a member of five further graduate schools, institutes and departments of METU.
G.-W. Weber has affiliations at Universities of Siegen (Germany), Federation University
(Ballarat, Australia), University of Aveiro (Portugal), University of North Sumatra (Medan,
Indonesia), Malaysia University of Technology, Chinese University of Hong Kong, KTO
Karatay University (Konya, Turkey), Vidyasagar University (Midnapore, India), Mazandaran
University of Science and Technology (Babol, Iran), Istinye University (Istanbul, Turkey),
Georgian International Academy of Sciences, at EURO (Association of European OR Societies)
where he is “Advisor to EURO Conferences” and IFORS (International Federation of OR
Societies), where he is member in many national OR societies, working groups, IFORS
Newsletter and IFORS Developing Countries Online Resources, at Pacific Optimization
Research Activity Group, etc. G.-W. Weber has supervised many MSc. and PhD. students,
authored and edited numerous books and articles, and given many presentations from a diversity
of areas, in theory, methods and practice. He has been a member of many international editorial,
special issue and award boards; he participated at numerous research projects; G.-W. Weber
received various recognitions.
Page 28
Keynote Speaker
Prof. Sırma Zeynep

Alparslan Gök
Department of Mathematics, Faculty of
Arts and Sciences, Süleyman Demirel
University, Isparta-Türkiye
Short Biography
Dr. Sırma Zeynep ALPARSLAN-GÖK was born in Ankara-Türkiye. She works as a full-time
professor at Department of Mathematics, Faculty of Arts and Sciences, Süleyman Demirel
University in Isparta-Türkiye. She holds PhD at Institute of Applied Mathematics in Scientific
Computing (Middle East Technical University, Türkiye) and BSc in Mathematics (Ankara
University, Türkiye). Her research interests are Cooperative Game Theory, Operations Research,
Optimization and Scientific Computing. She has co-authored research papers at various core
journals, book chapters and conference proceedings. She has educated several graduate students
and is currently working with the others. She leads as editor, guest editor and editorial board
member in several journals, journal special issues and book projects.
Cooperative Game Theory under Interval Uncertainty

and its Applications on Economics and Operations
Research Situations
(in memory of Prof. Dr. Stef Tijs)
Abstract
Cooperative game theory has been enriched recently with several models, which provide decision
making support in collaborative situations under uncertainty. Such models are generalizations of
the classical model related with the type of the coalition values. In the sequel the characteristic
functions are not real valued as in classical case, but they capture the uncertainty on the outcome
of cooperation in its different forms such as interval, grey and fuzzy uncertainty.
To incorporate uncertainty in cooperative game theory is motivated by the real world, where noise
in observation and experimental design, incomplete information and further vagueness in
preference structures and decision making play an important role. This causes a great
mathematical challenge which is approached and well understood in the case of interval
uncertainty.
This talk surveys the mathematical foundations of cooperative interval game theory and its
connection in economics and Operations Research situations.
Other types of cooperative games under uncertainty are also briefly mentioned.
Page 29
Keynote Speaker

Associate Professor (Research)
Tecnologico de Monterrey, Mexico
Short Biography
Dr. Mostafa Hajiaghaei-Keshteli is an Associate Professor at the Department of Industrial
Engineering, Tecnológico de Monterrey, Mexico. He is also the head of the research department
at the Center of Sustainable Smart Logistics (CLIS). He is one of the top 2% of scientists
worldwide, according to Elsevier-Stanford reports in 2021 and 2022. According to SciVal, he is
one of the top 5 in Supply Chain, and also according to Reseach.com, he is one of the top 8
scientists in Engineering in Mexico. Prof. Mostafa has developed famous metaheuristics, such as
Keshtel Alg., Red Deer Alg., Tree Growth Alg., and Social Engineering Optimizer. Besides, he is
an Associate Editor in the Journals of Applied Soft Computing (ASOC), Engineering Applications
of Artificial Intelligence (EAAI), and Expert Systems with Application (ESWA). He has a solid
industrial background in different industries as a General Manager, R&D and Project Manager,
and Consulter in Agriculture, Food, Automotive, and Engine industries.
Recent Advances in Agriculture Supply Chain Network

Design
Abstract
The agricultural business is rapidly evolving, driven by technological advancements, changing
consumer demands, and the need for sustainable practices. As we navigate the challenges and
opportunities of the 21st century, optimizing agriculture supply chains becomes crucial for
ensuring food security, reducing waste, considering sustainability and circular economy, and
maximizing efficiency. In this keynote address, we will delve into the recent breakthroughs in
agriculture supply chains that are revolutionizing the way we produce, distribute, and consume
food. By harnessing the power of cutting-edge technologies, data-driven insights, and innovative
practices, we are witnessing an unprecedented transformation that promises to reshape the future
of agriculture, especially agriculture supply chain network design.
Page 30
Keynote Speaker
Prof. Anand Nayyar

Vice-Chairman (Research) and Director- IoT and Intelligent
Systems Lab
School of Computer Science-Duy Tan University, Da Nang,
Vietnam
Short Biography
Dr. Anand Nayyar received Ph.D (Computer Science) from Desh Bhagat University in 2017 in
the area of Wireless Sensor Networks, Swarm Intelligence and Network Simulation. He is
currently working in School of Computer Science-Duy Tan University, Da Nang, Vietnam as
Professor, Scientist, Vice-Chairman (Research) and Director- IoT and Intelligent Systems Lab. A
Certified Professional with 125+ Professional certifications from CISCO, Microsoft, Amazon, EC-
Council, Oracle, Google, Beingcert, EXIN, GAQM, Cyberoam and many more. Published more
than 175+ Research Papers in various High-Quality ISI-SCI/SCIE/SSCI Impact Factor- Q1, Q2,
Q3, Q4 Journals cum Scopus/ESCI indexed Journals, 70+ Papers in International Conferences
indexed with Springer, IEEE and ACM Digital Library, 40+ Book Chapters in various
SCOPUS/WEB OF SCIENCE Indexed Books with Springer, CRC Press, Wiley, IET, Elsevier
with Citations: 10500+, H-Index: 55 and I-Index: 195. Member of more than 60+ Associations as
Senior and Life Member like: IEEE (Senior Member) and ACM (Senior Member). He has
authored/co-authored cum Edited 50+ Books of Computer Science. Associated with more than
600+ International Conferences as Programme Committee/Chair/Advisory Board/Review Board
member. He has 18 Australian Patents, 7 German Patents, 4 Japanese Patents, 33 Indian Design
cum Utility Patents, 7 UK Patents, 1 USA Patent, 3 Indian Copyrights and 2 Canadian Copyrights
to his credit in the area of Wireless Communications, Artificial Intelligence, Cloud Computing,
IoT, Healthcare, Drones, Robotics and Image Processing. Awarded 44 Awards for Teaching and
Research—Young Scientist, Best Scientist, Best Senior Scientist, Asia Top 50 Academicians and
Researchers, Young Researcher Award, Outstanding Researcher Award, Excellence in Teaching,
Best Senior Scientist Award, DTU Best Professor and Researcher Award- 2019, 2020-2021, 2022,
Distinguished Scientist Award by National University of Singapore, Obada Prize 2023 and many
more. He is listed in Top 2% Scientists as per Stanford University (2020, 2021, 2022), Ad Index
(Rank No:1 Duy Tan University, Rank No:1 Computer Science in Viet Nam) and Listed on
Research.com (Top Scientist of Computer Science in Viet Nam- National Ranking: 2; D-Index:
31). He is acting as Associate Editor for Wireless Networks (Springer), Computer
Communications (Elsevier), International Journal of Sensor Networks (IJSNET) (Inderscience),
Frontiers in Computer Science, PeerJ Computer Science, Human Centric Computing and
Information Sciences (HCIS), Tech Science Press- IASC, IET-Quantum Communications, IET
Page 31
Wireless Sensor Systems, IET Networks, IJDST, IJISP, IJCINI, IJGC, IJSIR. He is acting as
Managing Editor of IGI-Global Journal, USA titled “International Journal of Knowledge and
Systems Science (IJKSS)” and Editor-in-Chief of IGI-Global, USA Journal titled “International
Journal of Smart Vehicles and Smart Transportation (IJSVST)”. He has reviewed more than 2500+
Articles for diverse Web of Science and Scopus Indexed Journals. He is currently researching in
the area of Wireless Sensor Networks, Internet of Things, Swarm Intelligence, Cloud Computing,
Artificial Intelligence, Drones, Blockchain, Cyber Security, Healthcare Informatics, Big Data and
Wireless Communications.
Internet of Medical Things and Artificial Internet of

Medical Things: Revolutionizing Healthcare
Abstract
With the accelerated pace of life, people don’t want to spend more time on hospital registration,
procedure, and other trivial things. Internet of medical things can help it by applying internet in
medical staff, like the patient can make an appointment online, the personal information and the
very first disease description can be sent to hospital information system and assigned to a doctor,
which decreases time on waiting and personal information registration. For people living
somewhere far away from medical resources, remote medical treatment may be a good choice to
let the man away from distant puzzles. Even inside the hospital, the internet can be useful by
sending physical checking messages from doctor to specialist for a traffic injured patient. Internet
of medical things can be really helpful in our life, and we do need it.
The internet of medical things (IoMT), also known as healthcare IoT, describes almost any type
of internet-connected medical or wellness device. From smart stethoscopes to consumer
wearables, examples
of this trending development in healthcare tech continue to pop up.
In this lecture, a detailed conceptual methodology of concept and techniques deriving Internet of
Medical Things and connected terminologies will be elaborated. The lecture will also outline, case
studies, latest technologies and research orientations.
Page 32
Workshops
Prof. Paulina Golinska-Dawsons

D.Sc. Ph.D., prof PUT
Professor at the Poznan University of Technology
(PUT), Poznan, Poland
Short Biography
Prof. Paulina Golinska-Dawson, PhD, is a professor at the Poznan University of Technology
and a member of the Faculty of Management Engineering there. She received her habilitation
in the Civil Engineering and Transportation, and PhD in Industrial Engineering. She is graduate
of the "Executive Master of Business Administration" joint program at the University of
Warsaw and the University of Illinois at Urbana Champaigne (USA). She was an expert of the
European Commission for the STRIA – Strategic Transport Research and Innovation Agenda.
She was visiting scholar at the University of Luxembourg, where she participated in projects
related to sustainable development.
She has been conducting research in the area of corporate sustainability for more than 15
years, in particular assessing the maturity of implementing sustainable practices in
manufacturing and supply chain, and supply chain management in the Circular Economy.
She is the editor-in-chief of the book series at Springer Verlag entitled. "EcoProdution –
Environmental Issues in Logistics and Manufacturing." She serves as board member of the
International Federation of Logistics and SCM Systems. She is the author and co-author of
more than 140 scientific publications.
Product as a Service (PaaS) with Value Retention

Processes
Research Implications for Engineering Management
from Introducing Circular Business Models
Abstract
The circular economy (CE) concept is transforming the way in which companies design, offer,
deliver their products to customers. The purpose of circular business models (CBMs) is to
establish more resource-efficient approaches that allow reducing, reusing and recycling and
Page 33
therefore they support the transition towards a “zero waste” economy. Product as a service is a
CBM in which the manufacturer (or other supply chain participant) keeps the ownership of a
product and offers to customers access to its functionality with related services for some limited
period of time. After the PaaS contract is over, the product is subject to value retention processes
(reuse, repair, refurbish, or remanufacture) to make it suitable for the next use cycle. The aim of
these workshops is to discuss the challenges from industrial cases, and the reflections from the
ongoing international research project SCANDERE on scaling up applications of PaaS in
consumer markets. The research gaps in the area of Engineering Management and the needed
research directions will be discussed.
Page 34
Workshop
Dr. Meenakshi Kaushik

Professor, dean (research)
Trinity Institute of Innovations in Professional
Studies, Guru Gobind Singh Indraprastha
University, Delhi, NCR, India
Short Biography
Dr. Meenakshi Kaushik has more than 16 years of experience in teaching, training, research
and corporate in the area of HR & OB, Values & Ethics, gender issues. She is also a Motivational
and PDP Trainer and imparted several sessions on personality grooming in various reputed
management colleges. Currently she is working as a Dean (Research) in Trinity Institute of
innovations in Professional studies, GURU GOBIND SINGH INDRAPRASTHA
UNIVERSITY. Her Ph.D. thesis is entitled "A Study of Leadership Effectiveness in women
executives in Context of Delhi Based business organizations” has brought relevant theoretical
and empirical imperatives on leadership development strategies among women leaders and
executives. She has published several research papers in reputed journals SCOPUS /ABDC
/UGC Care journals with several citations on Google Scholar and ResearchGate. She is also one
of the editorial boards in few reputed journals UGC Care and Scopus. She is also supervising
and evaluating PhD thesis of Aligarh Muslim university (AMU, Central university) candidates
and awarded PhD degree to six AMU candidates.
She is continuously been invited as a speaker, Resource person and as a session chair in many
conferences, seminars, FDPs and workshops. Recently edited & authored a book on “Digital
Transformation: Recent Trends and Practices” by Himalaya Publication is available on Amazon,
Kindle and Himalaya Website that is widely appreciated and ranked globally.
Latest Digital Transformation & Digital Technology

Trends in the global business world
Abstract
The Covid-19 Pandemic, which is probably the most devastating crisis the world faced in the
last many decades, had a far-reaching impact on the social and business environment. Companies
had to quickly innovate to sustain and thrive in their business. It is only due to tremendous
development in the digital technology world, that the companies could quickly transition to
different ways of work and sustain their business. It has become very significant to introspect
Page 35
and examine the Digital transformation trends & practices and emerging opportunities &
challenges in front of us globally. The emergence of digital transformation has made every aspect
of business very simple and easy to connect with people all over the globe due to its
unprecedented speed and global accessibility. The fast-paced technology is making its
contribution in every area and sphere. Digitalization refers to the use of digital technologies to
bring forth change in business processes, strategies, and business models by providing new
avenues and other opportunities. According to the current scenario, organizations & industries
have a vision, of where exactly they want to go & seek various opportunities to digitize their
organizational processes, strategies, operations & business models according to the demands of
the dynamic, changing customer preferences and competitors ‘strategies & have a proper
understanding of competitive landscape, long-term goals, and discovery driven planning
(DDP).Organizations adopting this new business model to successfully survive in the present
cut-throat competitive & digitally challenged environment to survive & they need to adopt and
follow a discovery-driven digital transformation approach by applying digital technology in the
majority of their operations by becoming more customer friendly, by continuously interacting
with them and their knowledge about the products and services, preferences, digitized operations
and accordingly adopting the various business models. Organizations and industries shifted from
physical platforms to digital platforms to fulfill their requirements and contactless, remote
customer relationships, remote learning, and working from home (WFH). Organizations and
industries have understood the significance of digitalization very well and have taken the
initiative to undertake digital transformation to better engage and serve their people and to serve
the customers’ needs to achieve competitive advantage.
Page 36
Workshops
Prof. Safa Bhar Layeb Marwa Hasni

LR-OASIS, National Engineering School of
LR-OASIS, National Engineering
Tunis, University of Tunis El Manar, Tunisia
School of Tunis, University of Tunis
El Manar, Tunisia
Short Biography
Prof. Safa Bhar Layeb is currently a professor of industrial engineering and an active member of
the OASIS Laboratory at the National Engineering School of Tunis, Tunisia. Safa holds a
Polytechnic Engineering degree, a Master's degree in Mathematical Engineering, a Ph.D. in
Applied Mathematics, and a university habilitation (HDR) in Industrial Engineering. She is the
founding chair of the African Working Group in Health Systems and member of the Executive
committee of the African Federation of Operations Research Societies (AFROS). She has served
as guest editor and reviewer for over twenty international journals (ANOR, CAIE, RAIRO,…).
She is an active member of the editorial, organizing and scientific committees of several
international conferences. She has supervised over twenty Master and Ph.D. thesis in collaboration
with international institutes (Politecnico di Milano, Italy; Université de Moncton, Canada; Faculty
of Transportation Engineering and Vehicle Engineering, Hongrie; Université de Tlemcen, Algérie;
etc.). She has co-authored over thirty papers in reputable indexed journals, fifty papers in
international conferences and, twenty chapter books. Her teaching and research interests include
Operations research, optimization and data science based-approaches and their applications in
healthcare organizations, industrial engineering, logistics, and supply chain management.
Dr. Marwa HASNI is Assistant professor in Industrial Engineering at ISSIG Gabes – University
of Gabes. She holds a PhD in Industrial Engineering from National Engineering School of Tunis -
University of Tunis El Manar and member of the Laboratory OASIS, National Engineering School
of Tunis, University of Tunis El Manar, Tunisia since 2016. She also holds a diploma of Engineer
in Industrial Engineering and Logistics of the National Engineering School of Carthage-University
of Carthage.Her research focuses on the development and the analysis of forecasting techniques
applied to production systems and to financial and banking systems. The interest being non-
Page 37
parametric sampling techniques and those associated with the Machine learning and deep learning
approaches. Marwa HASNI has published and reviewed a number of research studies on the side
of several international journals namely: International Journal of Production Economics,
International Journal of Production Research, International Journal of Decision Sciences, Risk and
Management, Managerial and Decision Economics, International Journal of Economics, and
Strategic Management of Business Processes.
Supervised machine learning using the Scikit-learn

Library
Abstract
In this workshop, we will dive into the fascinating world of machine learning using the powerful
Scikit-learn library. Whether you are new to the field or looking to expand your knowledge, this
hands-on session will provide you with the essential skills to build and deploy machine learning
models. We will start by introducing the concept of supervised learning and exploring various
training techniques, including the popular Train-test-Split method. You will learn how to optimize
your models using cross-validation, validation curves, and GridSearch CV, ensuring the best
performance for your data. Data preprocessing plays a crucial role in creating accurate models, and
we'll cover various techniques to clean and transform datasets, including using the "impute"
module for handling missing data. Additionally, we will delve into variable selection techniques,
allowing you to identify the most relevant features for your models and enhance their performance.
The main goal of the workshop is to develop a pipeline for heterogeneous dataset processing and
apply a machine learning-based analysis using the Scikit-learn Library.
Page 38
Workshop
Prof. Dr. Servet SOYGÜDER

Ankara Yıldırım Beyazıt University, Faculty of
Engineering and Natural Sciences, Department of
Industrial Engineering
Short Biography
Prof. Dr. Servet SOYGUDER received the B.Sc. degree in 2001 from Mechanical Engineering,
M.Sc. degree in 2004 from Mechanical Engineering at Firat University and Ph.D. degree in 2009
from Mechanical Engineering at Firat University, Elazig, Turkey. His current research interests
are Robot Control, Legged Robot Control, Pronking Robot Control, Intelligent Control, PLC
Control, Mechatronic, Industrial Applications, Mechanism Design, Machine Theory and
Dynamic and PLC Control. He took first prize from III. and IV. International Convention
Competition in 2008-2009 years. He did the first Post-Doctoral in department of Computer
Engineering at Tennessee State University, Intelligent Robot Control Laboratory, Nashville, USA
in 2010. He worked as Associate Professor Dr. at Firat University between 2010 and 2015. He
has done the second Post-Doctoral and given Robotics and Kinematics and Dynamics of
Machinery lectures at department of Engineering Technology at Middle Tennessee State
University, Merfreesboro, USA in 2015-2016. Also, He worked as an Adjunct Faculty Professor
at department of Mechanical Engineering at Tennessee State University between 2016-2018.
Now, Prof. Servet Soygüder is a full time Professor at the Department of Industrial Engineering,
School of Engineering and Science, Ankara Yıldırım Beyazıt University (AYBU), Turkey. His
general interests include robotics, system dynamics and control, advanced control theory,
mechanism design, intelligent control, operations research, and mechatronics. Also, Dr.
Soygüder’s special research interests are biomedical robotics and autonomous mobile robots. He
is the coordinator of the project named "Digital Youth Life Health Platform (D-HP)" supported
by the EU.
“Dissemination of Digital Youth Life Health Platform

(DYL-HP)” Project Supported by the European Union
Abstract
DYL-HP is a digital platform and a mobile application that offer health and sports services live
and free of charge, with the awareness of healthy young life that all humanity needs during and
after the Covid-19 pandemic, as an alternative to established physical hospitals. Digital Youth
Life Health Platform (DYL-HP) is one of the most comprehensive projects of Ankara Yıldırım
Beyazıt University (AYBU), supported by the European Union (EU) as project 2021-2-TR01-
KA220-YOU-000049540 and the project coordinator is Prof. Dr. Servet SOYGÜDER.
Page 39
Panels
Page 40
Page 41
Conference Program
Page 42
“SEMIT 2023 is scheduled based on ANKARA time zone”
Join Zoom Meeting
Ahmet Süha Hancıoğlu
Room 1
Thursday (14.09.2023):
https://zoom.us/j/95581302146?pwd=bnBNNEZWWU04MlZDb1ZYRlVaRjhmQT09
Research Assistant Ankara Yıldırım
Friday (15.09.2023):
Beyazıt University, Türkiye https://zoom.us/j/99632448040?pwd=VENVN3ozK0E1d3pkTHBqbktkUWZsQT09
Join Zoom Meeting

Dr. Gül Uslu
Room 2
Thursday (14.09.2023):
Asist. Prof., Sinop University, Turkey https://zoom.us/j/98034979852?pwd=YWdUcGpwNzZwU2d5bHZ5ZjA3UjNyQT09
Friday (15.09.2023):
https://zoom.us/j/94184543947?pwd=RU50R0ZvczVwckFabE9vdWNDTkh5UT09
Aslı Ece Acar Filizci

PhD Student, Ankara Yıldırım Beyazıt Join Zoom Meeting
Room 3
University, Türkiye Thursday (14.09.2023):

https://zoom.us/j/99984608586?pwd=TzRDZWlrYzYyVXd4WlVTM0ZaTnRNZz09
Alperen Oncel
Friday (15.09.2023):
Ankara Yıldırım Beyazıt University, https://zoom.us/j/96489213745?pwd=aU9NZzZYUnNOSmlvWlBYTmJwM2wvZz09
Turkey
Join Zoom Meeting

Melda Kevser Akgün
Room 4
Thursday (14.09.2023):
PhD(c) Research Assistant, Ankara https://zoom.us/j/99942828456?pwd=L3FQaVUzajAzR2VHSmdoV0VZSUc2dz09
Yıldırım Beyazıt University Friday (15.09.2023):
https://zoom.us/j/99942828456?pwd=L3FQaVUzajAzR2VHSmdoV0VZSUc2dz09
Join Zoom Meeting

Mert Eyüboğlu
Room 5
Page
Thursday (14.09.2023):
Ankara Yıldırım Beyazıt https://zoom.us/j/92730159234?pwd=bVZ0bHJOc1VzT1JJL1IwZjd0MU45Zz09
43
University, Turkey Friday (15.09.2023):

https://zoom.us/j/92684582220?pwd=SFdySDM1S2p1MGFtMzBPWEc1MjJYQT09
First day: Thursday 14.09.2023
Location/time 9-10:30 10:45-12:15 12:15-13:30 13:30-15:00 15:15-16:45
Room 1 Opening 10:45-11:30 15:15-16:00
ceremony
Keynote speech Workshop Keynote speech
Welcome Speech Prof. Sırma Prof. Paulina Dr. Mostafa
Prof. Hasan Zeynep Alparslan Golinska-Dawson Hajiaghaei-
Okuyucu Gök Keshteli
11:30-12:15 Break time

Conference Report
Prof. A.
Mirzazadeh Workshop
Prof. Dr. Servet
Keynote speech SOYGÜDER
Prof. Alexandre
Dolgui
Paper Session Paper Session Paper Session
Room 2 Closed Dr. Babek Break time Dr. Mazdak Prof. Paulina
Erdebili Khodadadi Golinska-Dawson
Room 3 Closed Break time
Dr. Sonia Nasri Dr. Eloisa Macedo Prof. Rohit Bansal
Paper Session Paper Session
Room 4 Closed Prof. Hakan Break time Paper Session Dr. Sita Ram
Gultekin Sharma
Room 5 Closed Dr. Vijay Kumar Break time Dr. Erfan Babaee Paper Session
Gahlawat Tirkolaee
Page
44
Second day: Friday 15.09.2023
Location/time 9-10:30 10:45-12:15 12:15-13:30 13:30-15 15:15-16:45
Room 1 13:30-14:20
Workshop Workshop (Cont.) Closing Session
Keynote speech
Prof. Safa Bhar Prof. Safa Bhar Dr. Anand Nayyar Speech
Layeb Layeb Prof. Mete
& & Break time 14:20-15:00 Gundogan
Dr. Marwa Hasni Dr. Marwa Hasni
Workshop Keynote speech
Prof. Meenakshi Prof. Gerhard
Kaushik Wilhelm Weber

Room 2 Mr. Nguyễn Prof. Meenakshi Break time Prof. Nazanin Closed
Thành Luân Kaushik Pilevari
Paper Session
Room 3 Dr. Aybike Break time Closed
Dr. Alireza Goli Dr. Mehmet Kabak
Ozyuksel
Paper Session
Room 4 Prof. Sadia Break time Dr. Hamidreza Closed
Dr. Serap Ergun
Samar Ali Irani
Paper Session
Room 5 Dr. Simerjeet Paper Session Break time Paper Session Closed
Singh Bawa
Page
45
Date: Thursday, 14.09.2023 Room 1, link:
Time: 9:00-10:30 https://zoom.us/j/95581302146?pwd=bnBNNEZWWU04MlZDb1ZYRlVaRjhmQT09
Host: Ahmet Süha Hancıoğlu
Opening Ceremony
# Title Time Duration (minutes)
1 Opening Presentation 9:00-9:05 5
Prof. Okuyucu (Dean of Faculty of Engineering and Natural Sciences)
2 9:05-9:15 10
Welcome Speech
3 Clip: AYBU Introduction 9:15-9:20 5
Prof. Mirzazadeh (Conference Chair)
4 9:20-9:40 20
Conference Report
5 Clip: SEMIT 2023 Achievements 9:40-9:45 5
Prof. Dolgui (Editor-in-Chief, International Journal of Production Research)
6 9:45-10:25 40
AI in Manufacturing: Example of a Large European Project SSISTANT
7 Clip: IJSOM 10:25-10:30 5
Page
46
Room 1, link:
Date: Thursday, 14.09.2023
Keynote Speech https://zoom.us/j/95581302146?pwd=bnBNNEZWWU04MlZDb1ZYRlVaRjhmQT09
Time:10:45-11:30
Title:
“Cooperative Game Theory under Interval Uncertainty and its Applications on Economics and Operations
Research Situations (in memory of Prof. Dr. Stef Tijs)”
Keynote Speaker:
Prof. Sırma Zeynep Alparslan Gök
Department of Mathematics, Faculty of Arts and Sciences, Süleyman Demirel University, Isparta-Türkiye
Abstract:
Cooperative game theory has been enriched recently with several models, which provide decision making support in collaborative
situations under uncertainty. Such models are generalizations of the classical model related with the type of the coalition values. In
the sequel the characteristic functions are not real valued as in classical case, but they capture the uncertainty on the outcome of
cooperation in its different forms such as interval, grey and fuzzy uncertainty.
To incorporate uncertainty in cooperative game theory is motivated by the real world, where noise in observation and experimental
design, incomplete information and further vagueness in preference structures and decision making play an important role. This causes
a great mathematical challenge which is approached and well understood in the case of interval uncertainty.
This talk surveys the mathematical foundations of cooperative interval game theory and its connection in economics and Operations
Research situations.
Other types of cooperative games under uncertainty are also briefly mentioned.
Page
47
Room 1, link:
Workshop https://zoom.us/j/95581302146?pwd=bnBNNEZWWU04MlZDb1ZYRlVaRjhmQT09
Time:11:30-12:15
Title:
“DISSEMINATION OF DIGITAL YOUTH LIFE HEALTH PLATFORM(DYL-HP)” project supported
by the European Union (EU)
Speaker:
Prof. Dr. Servet SOYGÜDER
Ankara Yıldırım Beyazıt University, Faculty of Engineering and Natural Sciences, Department of Industrial Engineering
Abstract:
DYL-HP is a digital platform and a mobile application that offer health and sports services live and free of charge, with the awareness
of healthy young life that all humanity needs during and after the Covid-19 pandemic, as an alternative to established physical hospitals.
Digital Youth Life Health Platform (DYL-HP) is one of the most comprehensive projects of Ankara Yıldırım Beyazıt University
(AYBU), supported by the European Union (EU) as project 2021-2-TR01-KA220-YOU-000049540 and the project coordinator is
Prof. Dr. Servet SOYGÜDER.
Page
48
Room 2, link:
Session Code: SP01 https://zoom.us/j/98034979852?pwd=YWdUcGpwNzZwU2d5bHZ5ZjA3UjNyQT09
Time:10:45-12:15
Host: Dr. Gül Uslu
Title:
“Artificial Intelligence and Expert Systems”
Panel Chair:
Dr. Erdebili, Ankara Yıldırım Beyazıt University
Paper Code Authors Paper Title

1254-SEMIT2023 Ann Mary Babu A diagnosis model based on federated learning for lung cancer classification
Hasret Pınar Tipioglu; Hilal Classification of Monkeypox and HPV with Supervised Machine Learning Algorithms using k-
1014-SEMIT2023
Arslan mer
Osman Çaylı; Emre Yapay Zekâ ile B2B Kampanya Ve Sadakat Yönetimi
1286-SEMIT2023
Karadağ; Atınç Yılmaz
SAR Görüntülerinden Derin Öğrenme Tabanlı Hedef Tanıma
1199-SEMIT2023 Bilge Sinan; Emre Sümer
Osman Çaylı; Fuat Karip; Sporcular İçin Sakatlık Ve Performans Analizi Çerçevesi
1285-SEMIT2023
Atınç Yılmaz
1298-SEMIT2023 Naval Lawande Exploring the trends of Artificial Intelligence on Recruitment: A Bibliometric study
Page
49
Room 3, link:
Session Code: SP02 https://zoom.us/j/99984608586?pwd=TzRDZWlrYzYyVXd4WlVTM0ZaTnRNZz09
Time:10:45-12:15
Host: Aslı Ece Acar Filizci
Title:
“Advanced Intelligent Systems and Sustainable Development”
Panel Chair:
Dr. Sonia Nasri, LARODEC laboratory, University of Manouba, Tunisia
Panel Member:
Dr. Wided Oueslati, Business Higher School of Tunis, Manouba University, BESTMOD Laboratory, Tunisia

Solving On-Demand Transportation Problems Using an Intelligent Evolutionary Algorithm Based
1249-SEMIT2023 Sonia Nasri
on Hierarchical Objective Exchange
Amira Mzita; Sonia
Nasri; Hend Bouziri;
1273-SEMIT2023 A Variable Neighborhood Search for the Two Echelon Location Routing Problem
Wassila Aggoune-
Mtalaa
Nahla Ben Amor; Wael A Deep Implicit Coordination Graph-Based Soft Actor-Critic Approach in Multi-Agent
1322-SEMIT2023
Fathallah Reinforcement Learning
Wided Oueslati; hela Cloud computing model for handling medical big data: A mobile hospital pervasive healthcare
1039-SEMIT2023
limam; Sonia Nasri application
Malek ghanem; Raouia
1256-SEMIT2023 Ayachi; Wided A Hybrid Deep Learning Recommender System Based on Trust and Tags
Guezguez
1299-SEMIT2023 Wided Guezguez Combination of self training and active learning for sentiment analysis
Page
50
Room 4, link:
Session Code: SP03 https://zoom.us/j/99942828456?pwd=L3FQaVUzajAzR2VHSmdoV0VZSUc2dz09
Time:10:45-12:15
Host: Melda Kevser Akgün
Title:
“IT and EM Based Case Studies”
Panel Chair:
Prof. Hakan Gultekin, Sultan Qaboos University, Oman

The Negative Effects of Personal Protective Equipment Protective Masks on
1007-SEMIT2023 Doğa Martı; Doğa Martı
Users and Evaluation of Their Designs
Muthukumar R; Lalitha Ramakrisnan;

1402-SEMIT2023 Association between Demographic factors and Internet Banking Usage
Poongodai A; Krishnamacharyulu CSG
Paul Amaechi Ako (formerly Ozor);
Chukwuemeka Onyedinma Anyaoha; An Investigation on The Possibility of Retrofitting 3d Printers from Local Scrap
1077-SEMIT2023
Ndudim Henry Ononiwu; Eveth Nkiru E-Waste in Nigeria
Nwobodo-Anyadiegwu
Lucia Omolayo Agashua; Christopher Interlocking fired bricks as innovative waste Abstraction Proxies; Case study
1218-SEMIT2023
Ehizemhen Igibah Abuja Northern Nigeria
Kishori Kasat; Naim Shaikh; Venkatesh Omnichannel strategy adoption issues in an eCommerce company: A SAP-LAP
1126-SEMIT2023
Iyengar analysis
Sari Andayani; Endah Susilowati; Diah
Hari Suryaningrum; Evinda Dwi Nur Aini; Academic misconduct after the rapid transition to remote learning: A Gender and
1079-SEMIT2023
Wandah Nur Aliyyah; Singgih Alfiyahya; XYZ generations perspective
Naufan Rahmanda Tasri; Moch. Idris
Page
1395-SEMIT2023 Mustafa Yilmaz Chinese Postman Problem with Dynamic Start/End Point
51
Room 5 link:
Session Code: SP04 https://zoom.us/j/92730159234?pwd=bVZ0bHJOc1VzT1JJL1IwZjd0MU45Zz09
Time:10:45-12:15
Host: Mert Eyüboğlu
Title:
“Data Science, Big Data, Data Mining, and Knowledge Management in EM”
Panel Chair:
Dr. Vijay Kumar Gahlawat, Assistant Professor, Department of Basic and Applied Sciences, NIFTEM Kundli, India
Panel Members:
Prof. Rakesh Kumar, NUST, Windhoek, Namibia
Dr. Kumar Rahul, NIFTEM Kundli, India
Mr. Mohit Malik, NIFTEM Kundli, India
Abdullah Ateş; Salih Yilmaz; Comparison of Cost and Benefit Analysis of Active Leakage Control Applications in Water
1029-SEMIT2023
Mahmut Firat Distribution Systems
Salih Yilmaz; Abdullah Ateş; Benefit and Cost Analysis of Network and Service Connection Renewal in Distribution
1030-SEMIT2023
Mahmut Firat Systems for Sustainable Water Loss Management
Marzieh Nikoukar; Safanaz
1149-SEMIT2023 Review of Fake News Classification Methods Using Machine Learning
Heidari
Iryna Mysiuk; Roman Mysiuk;
Roman Shuvar; Volodymyr
1257-SEMIT2023 Designing a Data Pipeline Architecture for Intelligent Analysis of Streaming Data
Yuzevych; Anatolii Pavlenchyk;
Volodymyr Dalyk
1260-SEMIT2023 Ghazi Alkhatib Validation and Verification of Management Support Systems: A Framework
Fereidoon Bidollahkhany; Mehdi Designing a national cyber security strategic management model with the foundational data
1405-SEMIT2023
Basiri theory approach
Page
52
Room 1, link:
Workshop https://zoom.us/j/95581302146?pwd=bnBNNEZWWU04MlZDb1ZYRlVaRjhmQT09
Time:13:30-15:00
Title:
“Product as a Service (PaaS) with value retention processes research implications for
Engineering Management from introducing circular business models”
Speaker:
Professor at the Poznan University of Technology (PUT), Institute of Logistics, Poznan, Poland
Abstract:
The circular economy (CE) concept is transforming the way in which companies design, offer, deliver their products to customers. The
purpose of circular business models (CBMs) is to establish more resource-efficient approaches that allow reducing, reusing and recycling
and therefore they support the transition towards a “zero waste” economy. Product as a service is a CBM in which the manufacturer (or
other supply chain participant) keeps the ownership of a product and offers to customers access to its functionality with related services
for some limited period of time. After the PaaS contract is over, the product is subject to value retention processes (reuse, repair,
refurbish, or remanufacture) to make it suitable for the next use cycle. The aim of these workshops is to discuss the challenges from
industrial cases, and the reflections from the ongoing international research project SCANDERE on scaling up applications of PaaS in
consumer markets. The research gaps in the area of Engineering Management and the needed research directions will be discussed.
Page
53
Room 2, link:
Time:13:30-15:00
Host: Dr. Gül Uslu
Title:
“Safety and Reliability Engineering, System Engineering and System Safety, Safety in Industry 4.0”
Panel Chair:
Dr. Mazdak Khodadadi, University of Science and Culture, Tehran, Iran

Seyedeh Zahra Mir
The Impact of Internet of Things, Blockchain, and Artificial Intelligence on Occupational Health
1396-SEMIT2023 Mohammadzadeh Noudehi;
and Safety
Farnaz Bagheri Khameneh
Tihomir Atanassov
Improving the Visual Ergonomics of Computerised Workplaces Through the Use of Specialised
1214-SEMIT2023 Dovramadjiev; Darina
Eye-Rest Software
Dobreva; Ralitsa Zlateva
Sodabeh Orak; Hesam Proposing a Conceptual Model for Critical Success Factors in Establishment of Knowledge
1307-SEMIT2023 Varaee; Mazdak Khodadadi- Management System Using Interpretive Structural Modeling; Case Study of an Oil and Gas Well
Karimvand Drilling Company
Negar Mohabbatpour;
1308-SEMIT2023 Mazdak Khodadadi- Accident Analysis with Bow Tie Method; Case of British Petroleum Texas Refinery
Karimvand
Sajad Barkhordari; Mazdak Impact of Technological and Human Needs for Re-engineering Processes on Improving
1312-SEMIT2023
Khodadadi-Karimvand Productivity
Page
54
Page 55
Room 3, link:
Time:13:30-15:00
Title:
“Applied Soft Computing in Engineering Management”
Panel Chair:
Dr. Eloisa Macedo, University of Aveiro, Portugal

Serhii Lienkov; Alexander
Myasischev; Oleksandr
1081-SEMIT2023 Sieliukov; Vadim Development of Uavs on Pixhawk Family Controllers with Radio Interference Protection System
Ovcharuk; Serhii Frolov;
Mykola Kubiavka
1090-SEMIT2023 Şehmus Aslan An Efficient Hybrid Genetic Algorithm for Two-Sided Assembly Line Balancing Problems
Samet Tosun; İbrahim
1365-SEMIT2023 A Study on Measuring the Mental Workload of Instructors and Evaluating Significant Differences
Yilmaz
1138-SEMIT2023 Abdou Diop Optical Transmission Systems Incident Management in the Telecommunications Sector in Senegal
1052-SEMIT2023 Jozef Stašák Business Process Linguistic Modelling - Design – Process Design Part I
1229-SEMIT2023 Jozef Stašák Business Process Linguistic Modelling - Design – Process Design Part II
Page
56
Room 4, link:
Time:13:30-15:00
Title:
“Supply Chain Network Design; Considering Recent Trends in Sustainability and Circular Economy”

Sanae Larioui; Yasmine Carbon Footprint Calculation and Optimization Using ISO 14067: Development of a Product
1096-SEMIT2023
Benomar Carbon Footprint Calculator
1032-SEMIT2023 Issam Najati Adoption of the blockchain in the supply chain: highlighting the factors of failure
Servet Soygüder; Medium Voltage Distribution Cell Customer Orders Planning Model with Capacity Plan-Ning
1347-SEMIT2023
Abdulhamid Kayal Methods
Zahra Payvandi; Mehdi
1071-SEMIT2023 A location inventory model under vendor managed inventory contract
Seifbarghy
Selin Yalçın; Ertuğrul Prioritization of Freight Carrier Selection Factors with the Best Worst Method in Multi-Criteria
1379-SEMIT2023
Ayyıldız Decision Analysis
Page
57
Room 5, link:
Time:13:30-15:00
Title:
“Optimization, Decision making and support systems”
Panel Chair:
Dr. Erfan Babaee Tirkolaee, Istinye University, Istanbul, Turkey
Angie del Rocio Optimization models for production planning in the agro-industrial sector: A systematic review of the
1047-SEMIT2023
Sanchez Marquina literature
1188-SEMIT2023 Nurullah Yilmaz Space-Filling Curve Based Global Optimization Algorithm via Auxiliary Function
Asma Banu; Ayesha Employee-Engagement Level as a Predictor of Organizational Performance A Study of Information
1274-SEMIT2023
Khatun; Dinesh Kumar Technology Sector in Telangana, India
Seyed Ali Hosseini; A Reinforcement Learning-Based Fault Tolerant Control Design Approach using Double Q-Learning
1098-SEMIT2023
Karim Salahshoor Algorithm
Sedef Çali; Adil Comparing OECD Countries’ Logistics Performances by Comprehensive Distance Based Ranking
1369-SEMIT2023
Baykasoğlu Method
Page
58
Room 1, link:
Keynote Speech https://zoom.us/j/95581302146?pwd=bnBNNEZWWU04MlZDb1ZYRlVaRjhmQT09
Time:15:15-16:00
Title:
“Recent Advances in Agriculture Supply Chain Network Design”
Keynote Speaker:
Associate Professor (Research), Tecnologico de Monterrey, Mexico
Abstract:
The agricultural business is rapidly evolving, driven by technological advancements, changing consumer demands, and the need for
sustainable practices. As we navigate the challenges and opportunities of the 21st century, optimizing agriculture supply chains becomes
crucial for ensuring food security, reducing waste, considering sustainability and circular economy, and maximizing efficiency. In this
keynote address, we will delve into the recent breakthroughs in agriculture supply chains that are revolutionizing the way we produce,
distribute, and consume food. By harnessing the power of cutting-edge technologies, data-driven insights, and innovative practices, we
are witnessing an unprecedented transformation that promises to reshape the future of agriculture, especially agriculture supply chain
network design.
Page
59
Room 2, link:
Time:15:15-16:45
Host: Dr. Gül Uslu
Title:
“Quality, Productivity, and Project Management”
Panel Chair:
Poznan University of Technology (PUT), Poznan, Poland
A Systematic Analysis and Classification of the Effects of Human Fault In Unmanned Aviation
1005-SEMIT2023 İbrahim Akyürek
Vehicles - UAV Design and Ergonomics
Vildan Tezcan; Harika
1301-SEMIT2023 Ergün; Enes Öztürk; Güler The Quality Assessment of a Company in Defence Industry: A Case Study Approach
Zengin; Deniz Efendioğlu
Veronica Moreno; Agatha
1108-SEMIT2023 Application of Best Project Management Practices: Case Study Application in Social Projects
C da Silva Ovando
Aviva Cantika Alfatihanti;
Aviva Cantika Alfatihanti; Analysis of Investment Feasibility Using the Life Cycle Cost (LCC) Method in Implementing
1326-SEMIT2023
Bernadette Detty Green Retrofits in High-Rise Office Buildings
Kussumardianadewi
Page
60
Room 3, link:
Time:15:15-16:45
Title:
“Other Fields of Study Related to EM and IT”
Panel Chair:
Prof. Rohit Bansal, Department of Management Studies Vaish College of Engineering, Rohtak, India
Panel Members:
Dr. Chakir Aziza, Faculty of Law, Economics and Social Sciences, Hassan II University, Casablanca, Morocco
The Impact of Technostress and COVID-19 Stress on Employee Burnout: The Mediating
1180-SEMIT2023 Nour Tarabah; Ugur Sener
Role of Resilience
Alysha P Maulidina; Rachel
Anastasia Wijaya; Kimberly
1305-SEMIT2023 Comparative Study of Data Compression Algorithms: Zstandard, zlib & LZ4
Mazel; Maria Seraphina
Astriani
B O Ozero; Samuel Ogbonna
Empirical Analysis of Profit-Intended Vehicle Fleet Maintenance Schemes in Developing
1054-SEMIT2023 Enibe; Paul Amaechi Ako
Economies: Nigeria in Retrospect
(formerly Ozor)
1361-SEMIT2023 Trupti G. Thite; Sonal K. Jagtap Multiclass Arrhythmia Detection from 2D ECG of COVID-19 Patients
1243-SEMIT2023 Raouia Ayachi A New Category Based Collaborative Filtering Approach Using Trust Circles
Page
61
Room 4, link:
Time:15:15-16:45
Title:
“Neural Networks: Analysis, Modelling and Numerical Computations and Data Science”
Panel Chair:
Dr. Sita Ram Sharma, Chitkara University, India
Panel Members:
Dr. Shallu Jamwal, Jammu University
Dr. Ashok Kumar, Chitkara University, India

Advancing Anemia Diagnosis: Harnessing Machine Learning Methods for Accurate
1009-SEMIT2023 Pakize Sümeyye Söylemez
Detection
Enhancement Optical Fiber Data Rate by Using 8 port DWDM OF 256-QAM-OFDM DP

1203-SEMIT2023 Adnan Ali Al-Awadi
Technique
Hermenegildo da Conceição
1074-SEMIT2023 Alberto; Jean Marie Dembele; Review of Intrusion Detection Systems for SCADA: A Machine Learning Approach
Idy Diop; Alassane Bah
A hierarchical formulation for inverse data envelopment analysis: Cost efficiency point of
1076-SEMIT2023 Mohammad Moghaddas
view
Gayatri Mirajkar; Mukil
1124-SEMIT2023 Alagirisamy; Sagar Shinde; Image Processing in Toxicology: A Systematic Review
Lalit Garg
Page
62
Room 5, link:
Time:15:15-16:45
Title:
“Information Technology in Construction Industry: Recent Applications and Future Perspective”
Aqeel Abdulridhah Jbarh;
1190-SEMIT2023 Parametric Study of Shear Strength UHPC Tapered Beam
Nasser Hakeem Tu'ma
1033-SEMIT2023 Er. Aditya Singh The Present and Future Scope of Virtual Reality in the Construction Industry
1053-SEMIT2023 Er. Aditya Singh Mixed Reality and Its Future Scope in the Construction Sector
Emmanuel Chidiebere Eze;
Blockchain-Driven Smart Contracts: An Overview of Application Areas and Gap
1100-SEMIT2023 Ernest Effah Ameyaw; Bahriye
Identification in Construction Management Literature
Ilhan Jones
Benedict Mario Gilbert
Identification of Green Retrofitting Procurement and Permitting Processes in High-rise
Fernandes Sihaloho; Aviva
1324-SEMIT2023 Office Buildings in Jakarta Based on PerMen PUPR No.21 Year 2021 and GBC Indonesia
Cantika Alfatihanti; Bernadette
that Affects Project Time Performance
Detty Kussumardianadewi
1236-SEMIT2023 Rana Ahmed Performance Investigation of Heat Sinks with Different Perforations Diameters
Page
63
Date: Friday, 15.09.2023 Room 1, link:
Time: 9:00-10:30 & Workshop https://zoom.us/j/99632448040?pwd=VENVN3ozK0E1d3pkTHBqbktkUWZsQT09
10:45-12:15 Host: Ahmet Süha Hancıoğlu
Title:
“Supervised machine learning using the Scikit-learn Library”
Speakers:
Prof. Safa Bhar Layeb
LR-OASIS, National Engineering School of Tunis, University of Tunis El Manar, Tunisia
Dr. Marwa Hasni
LR-OASIS, National Engineering School of Tunis, University of Tunis El Manar, Tunisia
Abstract:
In this workshop, we will dive into the fascinating world of machine learning using the powerful Scikit-learn library. Whether you are new to the
field or looking to expand your knowledge, this hands-on session will provide you with the essential skills to build and deploy machine learning
models. We will start by introducing the concept of supervised learning and exploring various training techniques, including the popular Train-
test-Split method. You will learn how to optimize your models using cross-validation, validation curves, and GridSearch CV, ensuring the best
performance for your data. Data preprocessing plays a crucial role in creating accurate models, and we'll cover various techniques to clean and
transform datasets, including using the "impute" module for handling missing data. Additionally, we will delve into variable selection techniques,
allowing you to identify the most relevant features for your models and enhance their performance.
Page
64
Room 2, link:
Date: Friday, 15.09.2023
Session Code: SP13 https://zoom.us/j/94184543947?pwd=RU50R0ZvczVwckFabE9vdWNDTkh5UT09
Time: 9:00-10:30
Host: Dr. Gül Uslu
Title:
“Advancing Innovation and Collaboration in Science, Engineering, and Information Technology”
Panel Chair:
Mr. Nguyễn Thành Luân, HUFLIT, Vietnam
Panel Members:
Dr. Nguyen Duc Hoai Anh, HUFLIT, Vietnam
Mr. Tri-Quan Dang, HUFLIT, Vietnam

Büşranur Güdar; Dılsad
1148-SEMIT2023 Computational Analysis Wearable Glove Produced with FDM Technology
Akgumus Gok
1183-SEMIT2023 Aqeela N. Abed Improve Channel Capacity of TDMA
Maeva Ngameni; Haibing

1201-SEMIT2023 Seismic Design of Underground Structures’ Vulnerability: A Review
Cai Cai
Juan Carlos Chancusig Mesh Technology as an Alternative Connectivity Model for the Distribution of Packets in a
1239-SEMIT2023
Chisag WLAN Network: A Case Study in the Hospital of Latacunga, Ecuador
Sameera Fernandes; Sheeja Cultivating Synergistic Global Leadership: Unifying Scientific Acumen for Effective Global
1313-SEMIT2023
M.S.; Swetha Parivara Problem Solving
Page
65
Room 3, link:
Session Code: SP14 https://zoom.us/j/96489213745?pwd=aU9NZzZYUnNOSmlvWlBYTmJwM2wvZz09
Time: 9:00-10:30
Host: Alperen Oncel
Title:
“Uncertain Decision Making”
Panel Chair:
Dr. Alireza Goli, Department of Industrial Engineering and Future Studies, Faculty of Engineering, University of Isfahan

1207-SEMIT2023 Suwin Sleesongsom Uncertainty Quantification Code Written in MATLAB
Dan Luo; Simon
Stochastic optimization methodology for production planning with uncertain demand and lead
1368-SEMIT2023 Thevenin; Alexandre
time based on the digital twin
Dolgui
Özge Daşdemir; Ergün
1064-SEMIT2023 An Integration of Fuzzy Logic and Risk Analysis in the Defense Industry
Eraslan
An approach to fuzzy distributed multi-objective linear programming for optimizing distribution
1349-SEMIT2023 Ibrahim Yilmaz
decisions in relief logistic operations
1061-SEMIT2023 Ibrahim Temam Ibrahim Sustainable Aviation Fuel Production and Aviation Sector
Page
66
Room 4, link:
Time: 9:00-10:30
Title:
“Industry 4.0, Supply Chain 4.0, and Logistics 4.0”
Panel Chair:
Prof. Sadia Samar Ali, Department of Industrial Engineering, Faculty of Engineering, King Abdulaziz University, Saudi Arabia
Kerem Elibal; Eren A Draft Model Proposal for Total Quality Management (TQM) Based Industry 4.0 Maturity
1015-SEMIT2023
Özceylan Model
Souad Lahmine; Fatima
1142-SEMIT2023 Applicability of conventional quality methods on quality 4.0
Bennouna
Melike Erdoğan; Zekeriya
1021-SEMIT2023 Konurhan; Melih Constructing the criteria in determining the product groups for Agriculture 4.0 applications
Yücesan; Muhammet Gül
Sadeq Damrah; Mahmoud
1036-SEMIT2023 Mathematical Modeling for Exploring Spread of Cyberattacks through IoT Devices
DarAssi; Yousef AbuHour
Mohit Malik; Vijay Kumar A Conceptual Framework for the Development of Data Acquisition and Integration System for
1017-SEMIT2023
Gahlawat; Rahul S Mor the Dairy Industry
Page
67
Room 5, link:
Session Code: SP16 https://zoom.us/j/92684582220?pwd=SFdySDM1S2p1MGFtMzBPWEc1MjJYQT09
Time: 9:00-10:30
Title:
“Marketing and E-marketing for Resources Management”
Panel Chair:
Dr. Simerjeet Singh Bawa, Himalayan Institute of Management, Himalayan Group of Professional Institutions, Kala-Amb, Distt- Sirmour, H.P., India

1182-SEMIT2023 Tanya Haqi; Ugur Sener Enhancing Repurchase Intentions of Online Music Platform Customers for the Middle East
Amir Hossein Raoof
The Impact of Firm-Generated Content on Industrial Customers' Purchase Intention through
1135-SEMIT2023 Ghotboddini; Sepehr
Brand Credibility
Mobayyen; Farinaz Tanoorian
A Study on the Cross-Level Impact Mechanism of High Involvement in Human Resource
1051-SEMIT2023 Ding Haoqing Management Strategy on Creativity of Core Members in Live Broadcasting Teams- The
Mediating Role of Organizational Innovation Atmosphere
The Impact of Industrial Cluster on SMEs Performance: Evidence from the Wood Industry
1060-SEMIT2023 Masud Ibrahim; Dora Yeboah
in Ghana
Masud Ibrahim; Dora Yeboah;
1222-SEMIT2023 Antecedents of Mobile Banking Apps adoption among Consumers in Ghana
Freeman Anthony Mensah
Page
68
Room 2, link:
Time: 10:45-12:15
Host: Dr. Gül Uslu
Title:
“Humanities and Social Sciences”
Panel Chair:
Prof. Meenakshi Kaushik, Trinity Institute of Innovations in Professional Studies (Guru Gobind Indraprastha University), India
Panel Members:
Prof. Asif Akhtar, Assistant Prof. Aligarh Muslim Central University, India
Dr. Asad Rehman, Associate Professor, Department of Business Administration, Aligarh Muslim University, Aligarh
1263-SEMIT2023 Meenakshi Kaushik Strengthening Women leadership in India & Globe in digital Era
Bassem Chaker; Sirine
1255-SEMIT2023 Staff scheduling in service industry considering qualifications
Bouaziz; Diala Dhouib
Alami Chentoufi
Impact of Gender Inequalities on Education in Morocco: An Economic Analysis of Barriers
1080-SEMIT2023 Mohammed; Ait Alla Aniss;
and Opportunities
Mamdouh Naoual
Role of Top Management Commitment and Information Technology Investment in Digital
1024-SEMIT2023 Pankaj Adatiya Tiwari
Transformation
Asif Akhtar; Shilpa Chauhan;
The critical factors of success of gamification in digital banking services: Using analytic
1272-SEMIT2023 Dr. Meenakshi Kaushik;
hierarchy process (AHP) approach
Ashish Gupta
Pankaj Kumar Gupta;
1293-SEMIT2023 Consumer Confidence on State Regulation of e-commerce in India
Sonakshi Singh; Manvi Gupta
Maria Anjum; Maria Anum;
1383-SEMIT2023 Humaira Kosar; Mariam Social Networks and Privacy Concerns: An Insight into IT Industry User Experience
Rehman
Page
69
Room 3, link:
Time: 10:45-12:15
Host: Alperen Oncel
Title:
“Advances in Engineering: Machine Learning Techniques”
Panel Chair:
Dr. Aybike Özyüksel Çiftçioğlu, Manisa Celal Bayar University, Turkey
Panel Member:
Dr. Soheyl Khalilpourazari, Concordia University, Montreal, Canada
Marwa Hasni; Ghada Ben
Amor; Safa Layeb Bhar; Fares
1111-SEMIT2023 Marketing Mix Modeling tools and methodologies: A comparative analysis
Ben Mahmoud; Assma
Esseghir
halil ibrahim ayaz; zehra
1131-SEMIT2023 Detection of Shilling Attack with Support Vector Machines Using Oversampling
kamışlı öztürk
1231-SEMIT2023 Aybike Özyüksel Çiftçioğlu Advancing Structural Integrity Assessment through Machine Learning in RC Columns
Başak Zengin; Pınar Usta;
1248-SEMIT2023 Fuzzy Logic Methods for Determining the Mechanical Behaviour of Masonry Walls
Ozge Onat
Ihtisham Ul Haq; Hamza
Harnessing Advanced AI Techniques: An In-depth Analysis of Machine Learning Models
1062-SEMIT2023 Ahmad Khan; Zubair Ahmad
for Improved Diabetes Prediction
Khan; Muhammad Anas
Machine Learning in Fault Diagnosis of Electromechanical Devices using Vibration
1104-SEMIT2023 Binnur Demir Erdem
Measurements
Page
70
Room 4, link:
Time: 10:45-12:15
Title:
“Artificial Intelligence Innovations”
Panel Chair:
Dr. Serap Ergun, Isparta University of Applied Sciences, Turkey
Enhancing Traffic Flow Prediction in Urban Areas through Deep Learning and Probe
1010-SEMIT2023 Serap Ergun
Information: A Comparative Study
A Multi-Agent Deep Q-Network Approach for Optimizing Signal Control in Multi-Intersection
1011-SEMIT2023 Serap Ergun
Road Environments for Improving Traffic Flow in Urban Areas
1013-SEMIT2023 Bayram Dündar A simulating annealing algorithm for the maximum diversity social-distancing problem (MDPs)
Affan Alkim; Emir Duhan An Artificial Intelligence-Enhanced UAV Configuration for Object Detection, Instant
1281-SEMIT2023
Güler Communication, and Real-Time Information Transmission
Maria Seraphina Astriani;
Raymond Bahana; Arif
Augmented Reality Immersive World with Hologram Special Effect in Early Childhood
1130-SEMIT2023 Priyono Susilo Ahmad;
Education
Andreas Kurniawan; Lee
Huey Yi
Sameera Fernandes; Sheeja Synergizing Global Leadership Competencies with Artificial Intelligence and Expert Systems:
1406-SEMIT2023
M.S. A Multidisciplinary Approach for the Future of Management
Page
71
Room 5, link:
Time: 10:45-12:15
Title:
“Artificial Intelligence: The Future of Everything”
1008-SEMIT2023 Fatma Üzel; Hilal Arslan Energy Consumption Forecasting using Time Series Analysis Methods
Allam Fatima Zohra; Latifa
Implementation and Comparison of LBP Descriptor and Hough Transform applied to Facial
1398-SEMIT2023 Hamami; Hicham Bousbia
Recognition
Salah
1116-SEMIT2023 Harinandan Tunga Multi-Objective Optimization based Efficient Healthcare Management System
Cameron Guthrie; Samuel
1325-SEMIT2023 Fosso-Wamba; Maciel M Adoption of generative artificial intelligence: a cross industry comparison
Queiroz
Exploring the Elastic, Electronic and Optical properties of cubic ZrCoAs half – Heusler
1156-SEMIT2023 Lynet Allan
Semiconductor: An ab-initio Study
Saqib Waris; Linta Khalil;
Syd Mustafa Hassan Naqvi;
1157-SEMIT2023 Kamran Liaquat Bhatti; Performance Enhance and Management of Renewable Based Microgrid with backup Fuel Cell
Muhammad Zahid Shafique;
Malik Hamza Attiq
Page
72
Room 1, link:
Keynote Speech https://zoom.us/j/99632448040?pwd=VENVN3ozK0E1d3pkTHBqbktkUWZsQT09
Time: 13:30-14:20
Title:
“Internet of Medical Things and Artificial Internet of Medical Things: Revolutionizing Healthcare”
Keynote Speaker:
Prof. Anand Nayyar
Vice-Chairman (Research) and Director- IoT and Intelligent Systems Lab
School of Computer Science-Duy Tan University, Da Nang, Vietnam
Abstract:
With the accelerated pace of life, people don’t want to spend more time on hospital registration, procedure, and other trivial things. Internet of
medical things can help it by applying internet in medical staff, like the patient can make an appointment online, the personal information and the
very first disease description can be sent to hospital information system and assigned to a doctor, which decreases time on waiting and personal
information registration. For people living somewhere far away from medical resources, remote medical treatment may be a good choice to let the
man away from distant puzzles. Even inside the hospital, the internet can be useful by sending physical checking messages from doctor to specialist
for a traffic injured patient. Internet of medical things can be really helpful in our life, and we do need it.
The internet of medical things (IoMT), also known as healthcare IoT, describes almost any type of internet-connected medical or wellness device.
From smart stethoscopes to consumer wearables, examples
of this trending development in healthcare tech continue to pop up.
In this lecture, a detailed conceptual methodology of concept and techniques deriving Internet of Medical Things and connected terminologies will
be elaborated. The lecture will also outline, case studies, latest technologies and research orientations.
Page
73
Room 1, link:
Workshop https://zoom.us/j/99632448040?pwd=VENVN3ozK0E1d3pkTHBqbktkUWZsQT09
Time: 14:20-15:00
Title:
“Latest Digital Transformation & Digital Technology Trends in the global business world”
Speaker:
Prof. Meenakshi Kaushik
Trinity Institute of Innovations in Professional Studies, Guru Gobind Singh Indraprastha University, Delhi, NCR, India
Abstract:
The Covid-19 Pandemic, which is probably the most devastating crisis the world faced in the last many decades, had a far-reaching impact on the
social and business environment. Companies had to quickly innovate to sustain and thrive in their business. It is only due to tremendous
development in the digital technology world, that the companies could quickly transition to different ways of work and sustain their business. It
has become very significant to introspect and examine the Digital transformation trends & practices and emerging opportunities & challenges in
front of us globally. The emergence of digital transformation has made every aspect of business very simple and easy to connect with people all
over the globe due to its unprecedented speed and global accessibility. The fast-paced technology is making its contribution in every area and
sphere. Digitalization refers to the use of digital technologies to bring forth change in business processes, strategies, and business models by
providing new avenues and other opportunities. According to the current scenario, organizations & industries have a vision, of where exactly they
want to go & seek various opportunities to digitize their organizational processes, strategies, operations & business models according to the
demands of the dynamic, changing customer preferences and competitors ‘strategies & have a proper understanding of competitive landscape,
long-term goals, and discovery driven planning (DDP).Organizations adopting this new business model to successfully survive in the present cut-
throat competitive & digitally challenged environment to survive & they need to adopt and follow a discovery-driven digital transformation
approach by applying digital technology in the majority of their operations by becoming more customer friendly, by continuously interacting with
them and their knowledge about the products and services, preferences, digitized operations and accordingly adopting the various business models.
Organizations and industries shifted from physical platforms to digital platforms to fulfill their requirements and contactless, remote customer
relationships, remote learning, and working from home (WFH). Organizations and industries have understood the significance of digitalization
very well and have taken the initiative to undertake digital transformation to better engage and serve their people and to serve the customers’
needs to achieve competitive advantage.
Page
74
Room 2, link:
Time: 13:30-15:00
Host: Dr. Gül Uslu
Title:
“Inventory Control, Production Planning and Scheduling”
Panel Chair:
Prof. Nazanin Pilevari, Islamic Azad University, West Tehran Branch, Iran
Elif Kılıç Delice; Elif A Multi Criteria Decision Making Based Integrated Approach for Determining Production
1404-SEMIT2023
Akviran Priority in Demand Management
1195-SEMIT2023 Şerife Özkar Performance analysis for the inventory system with a two-stage service process
Büşra Alakoç; Aslı Bektaş
Moment-Based Approximation Formula for Variance of Semi-Markovian Random Walk
1225-SEMIT2023 Kamışlık; Tülay Kesemen;
Process
Tahir Khaniyev
Aslı Bektaş Kamışlık; Hacer
Inventory Model of Type (s,S) with Heavy Tailed Demand and Heavy Tailed Interference of
1226-SEMIT2023 Hilal Hasançebi; Tülay
Chance
Kesemen; Tahir Khaniyev
Zeynep Yüksel; Dursun Emre
1283-SEMIT2023 Multi-Depot Vehicle Routing Problem with Drone Collaboration In Humanitarian Logistic
Epcim; Suleyman Mete
Page
75
Room 3, link:
Time: 13:30-15:00
Host: Alperen Oncel
Title:
“Multi-Criteria Decision-Making”
Panel Chair:
Dr. Mehmet Kabak, Gazi University, Turkey

Ezgi Aktas Potur;
1246-SEMIT2023 A Bibliometric Analysis of Multi-Criteria Decision-Making Techniques in Disaster Management
Mehmet Kabak
Murat Ata; Mete
1355-SEMIT2023 Food Distribution in Times of Famine
Gündoğan
Zeynep Melike Sever;
1360-SEMIT2023 Determining the Location for Conquest in a Computer Game Using MCDM Methods
Babek Erdebilli
Yavuz Selim Özdemir;
Bank Performance Evaluation During Covid-19 in Terms of Customer Perspective Using Fuzzy
1321-SEMIT2023 Babek Erdebilli; Ender
MCDM
Sevinç
Babek Erdebilli;
1400-SEMIT2023 Supplier Selection in the context of Industry 4.0 using hybrid DEA-SMART Method
Parmida Bahreini
Page
76
Room 4, link:
Time: 13:30-15:00
Title:
“Smart Marketing”
Panel Chair:
Dr. Hamidreza Irani, University of Tehran, Iran
Panel Member:
Dr. Hamed Nozari, Islamic Azad University, Dubai Branch, UAE
Mostafa Esmaeili Mahyari;
Hamid Reza Irani; Vahid
1103-SEMIT2023 Investigating the Social and Conceptual Structure of AI in Marketing: A Bibliometric Analysis
Nourmandi Pour; Hamed
Nozari
Prashant Chaudhary; Prabha Hashtag and Marketing Campaign on Twitter: From the Spectrum of Smartphone Industry
1117-SEMIT2023
Kiran; Sarika Sharma Perspective
Mohsen Mazaheriasad; Exploring the Influence of Entrepreneurial Culture on SMEs Performance through Marketing
1377-SEMIT2023
Maryam Jamali Capabilities and Social Media Usage
1363-SEMIT2023 vuslat erat; Babek Erdebilli E-Shopping Sites Preference Analysis with Multi-Criteria Decision-Making Methods
Hümeyra Akar; Babek
1362-SEMIT2023 Category Sort with TODIM Method, an Application for Retail Industry E-Commerce Site
Erdebilli
Page
77
Room 5, link:
Time: 13:30-15:00
Title:
“Advancement and Future of Technologies”

Amrut P Bhosale; Sachin
Dissecting the Economic and Environmental Compatibility of Battery Electric and Conventional
1330-SEMIT2023 Mastud; Muzammil
Vehicles in India
Bepari; Ketaki Bhosale
Hamed Atyia Soodi,
Optimization of Controller Parameter Multilevel Converter Based STATCOM for Reactive Power
1352-SEMIT2023 Yousif A. Hamad, and
Compensation
Ardam H. Mohammed
Gözde Saribaş; Emine
1023-SEMIT2023 Avşar Aydin; Ahmet Analysis of 3 Different Hexagonal Microstrip Antenna Designs for Breast Cancer Detection Studies
Aydin
Kasım Şahin Dönmez; Numerical Investigation of The Effect of The Use Of Multiple Propeller Systems On Flight
1102-SEMIT2023
Metin Uzun Performance In Unmanned Aerial Vehicles
1170-SEMIT2023 Salah Oudjertli A Study on the Physical Properties of Elaborated Graphene
Page
78
Date: Friday, 15.09.2023 Room 1, link:
Time: 15:15-16:45 https://zoom.us/j/99632448040?pwd=VENVN3ozK0E1d3pkTHBqbktkUWZsQT09
Closing Ceremony
# Title Time Duration (minutes)
1 Closing Presentation 15:15-15:20 5
The First PhD Thesis Competition of ODSIE 2023
2 Dr. Erfan Babaee Tirkolaee Speech (Istinye University, 15:20-15:35 15
Istanbul, Turkey)
Prof. Gundogan Speech (Conference Chair)
3 15:35-15:50 15
Concluding Remark
Prof. Weber Speech (Poznan University of Technology, Poland)
Optimal Management of Defined Contribution Pension
4 15:50-16:30 40
Funds under the Effect of Inflation, Mortality and
Uncertainty
5 Clip: SEMIT 2023 16:30-16:35 5
6 Appreciation 16:35-16:45 10
Page
79
SEMIT2023
Abstracts Proceedings
Page 80
SEMIT2023
Table of Contents
SP01: Artificial Intelligence and Expert Systems ................................................................. 86

A diagnosis model based on federated learning for lung cancer classification ................................................... 86
Department of Computer Science & Engineering Indian Institute of Information Technology Kottayam PALA,
Kottayam, Kerala, India ...................................................................................................................................... 86
Classification of Monkeypox and HPV with Supervised Machine Learning Algorithms using k-mer .............. 86
B2B Campaign and Loyalty Management with Artificial Intelligence .............................................................. 87
SAR Görüntülerinden Derin Öğrenme Tabanlı Hedef Tanıma........................................................................... 87
Injury and Performance Analysis Framework for Athletes ................................................................................ 88
Exploring the trends of Artificial Intelligence on Recruitment: A Bibliometric study ....................................... 89
SP02: Advanced Intelligent Systems and Sustainable Development .................................. 90
Solving On-Demand Transportation Problems Using an Intelligent Evolutionary Algorithm Based on
Hierarchical Objective Exchange ....................................................................................................................... 90
A Variable Neighborhood Search for the Two Echelon Location Routing Problem .......................................... 90
A Deep Implicit Coordination Graph-Based Soft Actor-Critic Approach in Multi-Agent Reinforcement
Learning .............................................................................................................................................................. 91
Cloud Computing Model for Handling Medical Big Data: A Mobile Hospital Pervasive Healthcare Application
............................................................................................................................................................................ 92
A Hybrid Deep Learning Recommender System Based on Trust and Tags ....................................................... 92
Combination of Self Training and Active Learning for Sentiment Analysis ...................................................... 93
SP03: IT and EM Based Case Studies ................................................................................... 94
The Negative Effects of Personal Protective Equipment Protective Masks on Users and Evaluation of Their
Designs ............................................................................................................................................................... 94
Association between Demographic factors and Internet Banking Usage ........................................................... 95
An Investigation on The Possibility of Retrofitting 3d Printers from Local Scrap E-Waste in Nigeria ............. 95
Interlocking Fired Bricks as Innovative Waste Abstraction Proxies; Case Study Abuja Northern Nigeria ....... 96
Omnichannel strategy adoption issues in an eCommerce company: a SAP-LAP analysis................................. 97
Academic Misconduct after the Rapid Transition to Remote Learning: A Gender and XYZ Generations
Perspective .......................................................................................................................................................... 98
Chinese Postman Problem with Dynamic Start/End Point ................................................................................. 99
SP04: Data Science, Big Data, Data Mining, and Knowledge Management in EM ........ 100
Comparison of Cost and Benefit Analysis of Active Leakage Control Applications in Water Distribution
Systems ............................................................................................................................................................. 100
Benefit and Cost Analysis of Network and Service Connection Renewal in Distribution Systems for
Sustainable Water Loss Management ............................................................................................................... 101
Review of Fake News Classification Methods Using Machine Learning......................................................... 101
Designing a Data Pipeline Architecture for Intelligent Analysis of Streaming Data ........................................ 102
Validation and Verification of Management Support Systems: A Framework ................................................ 102
Designing a National Cyber Security Strategic Management Model with the Foundational Data Theory
Approach .......................................................................................................................................................... 103
Page 81
SEMIT2023
SP05: Safety and Reliability Engineering, System Engineering and System Safety, Safety
in Industry 4.0 ........................................................................................................................ 104
The Impact of Internet of Things, Blockchain, and Artificial Intelligence on Occupational Health and Safety
.......................................................................................................................................................................... 104
Improving the Visual Ergonomics of Computerised Workplaces Through the Use of Specialised Eye-Rest
Software ............................................................................................................................................................ 104
Proposing a Conceptual Model for Critical Success Factors in Establishment of Knowledge Management
System Using Interpretive Structural Modeling; Case Study of an Oil and Gas Well Drilling Company ........ 105
Accident Analysis with Bow Tie Method; Case of British Petroleum Texas Refinery .................................... 106
Impact of Technological and Human Needs for Re-Engineering Processes on Improving Productivity ......... 106
A Study on Measuring the Mental Workload of Instructors and Evaluating Significant Differences .............. 107
SP06: Applied Soft Computing in engineering Management ............................................ 108
Development of UAVS on Pixhawk Family Controllers with Radio Interference Protection System ............. 108
An Efficient Hybrid Genetic Algorithm for Two-Sided Assembly Line Balancing Problems ......................... 109
Optical Transmission Systems Incident Management in the Telecommunications Sector in Senegal ............. 109
Business Process Linguistic Modelling- Design- Process Design Part I .......................................................... 110
Business Process Linguistic Modelling - Design – Process Design Part II ...................................................... 110
SP07: Supply Chain Network Design; Considering Recent Trends in Sustainability and
Circular Economy ................................................................................................................. 112
Carbon Footprint Calculation and Optimization Using ISO 14067: Development of a Product Carbon Footprint
Calculator.......................................................................................................................................................... 112
Adoption of the Blockchain in the Supply Chain: Highlighting the Factors of Failure .................................... 113
Medium Voltage Distribution Cell Customer Orders Planning Model with Capacity Plan-Ning Methods ..... 113
A Location Inventory Model under Vendor Managed Inventory Contract ....................................................... 114
Prioritization of Freight Carrier Selection Factors with the Best Worst Method in Multi-Criteria Decision
Analysis ............................................................................................................................................................ 114
SP08: Optimization, Decision Making and Support Systems ............................................ 116
Optimization Models for Production Planning in the Agro-Industrial Sector: A Systematic Review of the
Literature .......................................................................................................................................................... 116
Space-Filling Curve Based Global Optimization Algorithm via Auxiliary Function ....................................... 116
Employee -Engagement Level as a Predictor of Organizational Performance :A Study of Information
Technology Sector in Telangana, India ............................................................................................................ 117
A Reinforcement Learning-Based Fault Tolerant Control Design Approach using Double Q-Learning
Algorithm.......................................................................................................................................................... 117
Comparing OECD Countries’ Logistics Performances by Comprehensive Distance Based Ranking Method 118
SP09: Quality, Productivity, and Project Management ..................................................... 119
A Systematic Analysis and Classification of the Effects of Human Fault in Unmanned Aviation Vehicles -
UAV Design and Ergonomics .......................................................................................................................... 119
The Quality Assessment of a Company in Defence Industry: A Case Study Approach ................................... 120
Application of Best Project Management Practices: Case Study Application in Social Projects ..................... 120
Analysis of Investment Feasibility Using the Life Cycle Cost (LCC) Method in Implementing Green Retrofits
in High-Rise Office Buildings .......................................................................................................................... 121
SP10: Other Fields of Study Related to EM and IT ........................................................... 122
Page 82
SEMIT2023
The Impact of Technostress and COVID-19 Stress on Employee Burnout: The Mediating Role of Resilience
.......................................................................................................................................................................... 122
Comparative Study of Data Compression Algorithms: Zstandard, zlib & LZ4 ................................................ 123
Empirical Analysis of Profit-Intended Vehicle Fleet Maintenance Schemes in Developing Economies: Nigeria
in Retrospect ..................................................................................................................................................... 124
Multiclass Arrhythmia Detection from 2D ECG of COVID-19 Patients.......................................................... 124
A New Category Based Collaborative Filtering Approach Using Trust Circles ............................................... 125
SP11: Neural Networks: Analysis, Modelling and Numerical Computations and Data
Science .................................................................................................................................... 126
Advancing Anemia Diagnosis: Harnessing Machine Learning Methods for Accurate Detection .................... 126
Enhancement Optical Fiber Data Rate by Using 8 port DWDM OF 256-QAM-OFDM DP Technique .......... 126
Review of Intrusion Detection Systems for SCADA: A Machine Learning Approach .................................... 127
A Hierarchical Formulation for Inverse Data Envelopment Analysis: Cost Efficiency Point of View ............ 128
Image Processing in Toxicology: A Systematic Review .................................................................................. 128
SP12: Information Technology in Construction Industry: Recent Applications and
Future Perspective ................................................................................................................. 130
Parametric Study of Shear Strength UHPC Tapered Beam .............................................................................. 130
The Present and Future Scope of Virtual Reality in the Construction Industry ................................................ 130
Mixed Reality and Its Future Scope in the Construction Sector ....................................................................... 131
Blockchain-Driven Smart Contracts: An Overview of Application Areas and Gap Identification in Construction
Management Literature ..................................................................................................................................... 131
Identification of Green Retrofitting Procurement and Permitting Processes in High-rise Office Buildings in
Jakarta Based on PerMen PUPR No.21 Year 2021 and GBC Indonesia that Affects Project Time Performance
.......................................................................................................................................................................... 132
Performance Investigation of Heat Sinks with Different Perforations Diameters ............................................ 132
SP13: Advancing Innovation and Collaboration in Science, Engineering, and Information
Technology ............................................................................................................................. 134
Computational Analysis Wearable Glove Produced with FDM Technology ................................................... 134
Improve Channel Capacity of TDMA .............................................................................................................. 134
Seismic Design of Underground Structures’ Vulnerability: A Review ............................................................ 135
Mesh Technology as an Alternative Connectivity Model for the Distribution of Packets in a WLAN Network:
A Case Study in the Hospital of Latacunga, Ecuador ....................................................................................... 135
Cultivating Synergistic Global Leadership: Unifying Scientific Acumen for Effective Global Problem Solving
.......................................................................................................................................................................... 136
SP14: Uncertain Decision Making ....................................................................................... 137
Uncertainty Quantification Code Written in MATLAB ................................................................................... 137
Stochastic Optimization Methodology for Production Planning with Uncertain Demand and Lead Time Based
on The Digital Twin.......................................................................................................................................... 137
An Integration of Fuzzy Logic and Risk Analysis in the Defense Industry ...................................................... 138
An Approach to Fuzzy Distributed Multi-Objective Linear Programming for Optimizing Distribution
Decisions in Relief Logistic Operations ........................................................................................................... 139
Sustainable Aviation Fuel Production and Aviation Sector .............................................................................. 139
SP15: Industry 4.0, Supply Chain 4.0, and Logistics 4.0 ................................................... 141
Page 83
SEMIT2023
A Draft Model Proposal for Total Quality Management (TQM) Based Industry 4.0 Maturity Model............. 141
Applicability of Conventional Quality Methods on Quality 4.0 ....................................................................... 141
Constructing the Criteria in Determining the Product Groups for Agriculture 4.0 Applications ...................... 142
Mathematical Modeling for Exploring Spread of Cyberattacks through IoT Devices...................................... 142
A Conceptual Framework for the Development of Data Acquisition and Integration System for the Dairy
Industry ............................................................................................................................................................. 143
SP16: Marketing and E-marketing for Resources Management ...................................... 144
Enhancing Repurchase Intentions of Online Music Platform Customers for the Middle East ......................... 144
The Impact of Firm-Generated Content on Industrial Customers' Purchase Intention through Brand Credibility
.......................................................................................................................................................................... 144
A Study on the Cross-Level Impact Mechanism of High Involvement in Human Resource Management
Strategy on Creativity of Core Members in Live Broadcasting Teams- The Mediating Role of Organizational
Innovation Atmosphere..................................................................................................................................... 145
The Impact of Industrial Cluster on SMEs Performance: Evidence from the Wood Industry in Ghana .......... 146
Antecedents of Mobile Banking Apps adoption among Consumers in Ghana ................................................. 146
SP17: Humanities and Social Sciences................................................................................. 147
Strengthening Women leadership in India & Globe in digital Era ................................................................... 147
Staff Scheduling in Service Industry Considering Qualifications..................................................................... 147
Impact of Gender Inequalities on Education in Morocco: An Economic Analysis of Barriers and Opportunities
.......................................................................................................................................................................... 148
Role of Top Management Commitment and Information Technology Investment in Digital Transformation 149
The Critical Factors of Success of Gamification in Digital Banking Services: Using Analytic Hierarchy Process
(AHP) Approach ............................................................................................................................................... 149
Consumer Confidence on State Regulation of E-Commerce in India............................................................... 150
Social Networks and Privacy Concerns: An Insight into IT Industry User Experience.................................... 151
SP18: Advances in Engineering: Machine Learning Techniques ..................................... 152
Marketing Mix Modeling tools and methodologies: A comparative analysis .................................................. 152
Detection of Shilling Attack with Support Vector Machines Using Oversampling.......................................... 152
Advancing Structural Integrity Assessment through Machine Learning in RC Columns ................................ 153
Fuzzy Logic Methods for Determining the Mechanical Behaviour of Masonry Walls .................................... 154
Harnessing Advanced AI Techniques: An In-depth Analysis of Machine Learning Models for Improved
Diabetes Prediction ........................................................................................................................................... 154
Machine Learning in Fault Diagnosis of Electromechanical Devices using Vibration Measurements ............ 155
SP19: Artificial Intelligence Innovations ............................................................................. 156
Enhancing Traffic Flow Prediction in Urban Areas through Deep Learning and Probe Information: A
Comparative Study ........................................................................................................................................... 156
A Multi-Agent Deep Q-Network Approach for Optimizing Signal Control in Multi-Intersection Road
Environments for Improving Traffic Flow in Urban Areas .............................................................................. 156
A Simulating Annealing Algorithm for the Maximum Diversity Social-Distancing Problem (Mdps)............. 157
An Artificial Intelligence-Enhanced UAV Configuration for Object Detection, Instant Communication, and
Real-Time Information Transmission ............................................................................................................... 157
Augmented Reality Immersive World with Hologram Special Effect in Early Childhood Education ............. 158
Page 84
SEMIT2023
Synergizing Global Leadership Competencies with Artificial Intelligence and Expert Systems: A
Multidisciplinary Approach for the Future of Management ............................................................................. 158
SP20: Artificial Intelligence: The Future of Everything .................................................... 160
Energy Consumption Forecasting using Time Series Analysis Methods ......................................................... 160
Implementation and Comparison of LBP Descriptor and Hough Transform applied to Facial Recognition ... 160
Multi-Objective Optimization based Efficient Healthcare Management System ............................................. 161
Adoption of Generative Artificial Intelligence: A Cross Industry Comparison ................................................ 162
Exploring the Elastic, Electronic and Optical properties of cubic ZrCoAs half – Heusler Semiconductor: An ab-
initio Study........................................................................................................................................................ 162
Performance Enhance and Management of Renewable Based Microgrid with backup Fuel Cell .................... 163
SP21: Inventory Control, Production Planning and Scheduling ...................................... 164
A Multi Criteria Decision Making Based Integrated Approach for Determining Production Priority in Demand
Management ..................................................................................................................................................... 164
Performance Analysis for The Inventory System with a Two-Stage Service Process ...................................... 164
Moment-Based Approximation Formula for Variance of Semi-Markovian Random Walk Process ................ 165
Inventory Model of Type (s,S) with Heavy Tailed Demand and Heavy Tailed Interference of Chance .......... 166
Multi-Depot Vehicle Routing Problem with Drone Collaboration in Humanitarian Logistic .......................... 166
SP22: Multi-Criteria Decision-Making ............................................................................... 168
A Bibliometric Analysis of Multi-Criteria Decision-Making Techniques in Disaster Management ................ 168
Food Distribution in Times of Famine.............................................................................................................. 168
Determining the Location for Conquest in a Computer Game Using MCDM Methods .................................. 169
Bank Performance Evaluation During Covid-19 in Terms of Customer Perspective Using Fuzzy MCDM .... 169
Supplier Selection in the context of Industry 4.0 using hybrid DEA-SMART Method ................................... 170
SP23: Smart Marketing ........................................................................................................ 171
Investigating the Social and Conceptual Structure of AI in Marketing: A Bibliometric Analysis ................... 171
Hashtag and Marketing Campaign on Twitter: From the Spectrum of Smartphone Industry Perspective ....... 171
Exploring the Influence of Entrepreneurial Culture on SMEs Performance through Marketing Capabilities and
Social Media Usage .......................................................................................................................................... 172
E-Shopping Sites Preference Analysis with Multi-Criteria Decision-Making Methods ................................... 173
Category Sort with TODIM Method, an Application for Retail Industry E-Commerce Site............................ 173
SP24: Advancement and Future of Technologies ............................................................... 174
Dissecting the Economic and Environmental Compatibility of Battery Electric and Conventional Vehicles in
India .................................................................................................................................................................. 174
Compensation ................................................................................................................................................... 175
Analysis of 3 Different Hexagonal Microstrip Antenna Designs for Breast Cancer Detection Studies ........... 176
Numerical Investigation of the Effect of the Use of Multiple Propeller Systems on Flight Performance in
Unmanned Aerial Vehicles. .............................................................................................................................. 177
A Study on the Physical Properties of Elaborated Graphene ............................................................................ 177
Full Papers Proceedings ...................................................................................................... 178
Page 85
SEMIT2023
SP01: Artificial Intelligence and Expert Systems
A diagnosis model based on federated learning for lung cancer classification

Ann Mary Babu
Department of Computer Science & Engineering Indian Institute of Information Technology
Kottayam PALA, Kottayam, Kerala, India
Sivaiah Bellamkonda
Department of Computer Science & Engineering Indian Institute of Information Technology
Kottayam PALA, Kottayam, Kerala, India
Abstract
Lung cancer, causing a significant number of fa_x0002_talities annually, is a prominent contributor to
global mortality rates among individuals of both genders, with an estimated five million cases resulting
in death each year. The utilisation of Computed Tomography (CT) scans contribute valuable data for
the detection of lung cancer. The main objectives of this study are to diagnose malignant lung nodules
in the provided input lung images and to classify lung nodule based on its level of severity. This study
proposes a solution wherein locally computed updates are combined to stimulate the learning of a shared
model, while the training data remains dispersed throughout mobile devices. Federated Learning (FL)
is an advanced technology that operates in a collaborative and decentralized manner, with the primary
objective of preserving privacy. Its purpose is to address the obstacles posed by data silos and the
sensitivity of data. Therefore, the FL approach is employed for the classification of medical images in
collaborative and real-world circumstances.
Keywords: Lung Cancer, Deep Learning, Federated Learning, CT Images, X-Ray.
Classification of Monkeypox and HPV with Supervised Machine Learning Algorithms

using k-mer
Hasret Pınar Tipioğlu
Ankara University, Biology Department,
Hilal Arslan
Ankara Yıldırım Beyazıt University, Software Engineering Department
Abstract
According to the data of the World Health Organization, apart from Africa where it is endemic,
monkeypox cases have been seen widely in Europe and America since 2022. Monkeypox is sometimes
confused with warts because of the similarity of symptoms. For the correct diagnosis of the disease,
virus DNA is detected by molecular methods. For this purpose, in this study, HPV DNA causing warts
and monkeypox virus DNA causing monkeypox were tried to be classified by machine learning
algorithms. HPV DNA sequences compared in the study were obtained from NCBI, and monkey pox
virus DNA sequences were obtained from GISAID. The features used to train the models are k-mers
(2-mer and 3-mer) obtained from the DNA sequence of viruses. Then, classification was made with
MATLAB and supervised learning algorithms. According to the classification results, an accuracy of
Page 86
SEMIT2023
99% and above was obtained. As a result, these two diseases were successfully classified using virus
DNAs.
Keywords: Monkeypox, HPV, Machine Learning, K-Mer.
Yapay Zekâ ile B2B Kampanya Ve Sadakat Yönetimi

B2B Campaign and Loyalty Management with Artificial Intelligence
Osman Çaylı
VBT YAZILIM A.Ş, İstanbul, Türkiye
Emre Karadağ
Atınç Yılmaz
Bilgisayar Mühendisliği, İstanbul Beykent Üniversitesi, İstanbul, Türkiye
Abstract
E-ticaret kavramı bir işletmenin mal veya hizmetlerini internet veya farklı diğer elektronik ortamlar
üzerinden sunduğu ticaret modelidir. Bu modelde, şirketten şirkete (B2B) alanında da oldukça
gelişmeler görülmektedir. Son dönemlerde oyunlaştırma stratejileri e-ticaret alanında uygulanmaktadır.
Bu sayede müşteri bağımlığı, çalışan motivasyonu gibi kavramlarda artış sağlanmaktadır. Günümüz
rekabetçi ortamında yapay zeka teknolojileri uygulandığı her sektörde avantaj getirebilmektedir. Bu
çalışmada, yapay zeka yöntemlerinden karar ağacı yöntemi kullanılarak B2B alanında distribütörlerin
sadakat programlarında etkinliğinin arttırılması amacı ile uygulama geliştirilmiştir. Çalışma
neticesinde, yapay zeka yöntemleri desteği ile kurgulanan müşteri sadakat programı ile işletmelerin
rekabet avantajı elde edeceği gösterilmiştir. Yapay zeka yöntemleri ile modellenen program ile müşteri
memnuniyeti ve deneyimi arttırılarak sürdürülebilirlik de sağlanacağı ortaya konulmuştur. Bulgular
neticesinde probleme uygulanan modeller arasında en yüksek doğruluğa %78 oranı ile rastgele orman
yöntemi uygulanarak elde edilmiştir.
Keywords: Yapay Zeka, Karar Ağacı, Rastgele Orman, Tahminleme.
SAR Görüntülerinden Derin Öğrenme Tabanlı Hedef Tanıma

Bilge Sinan
Havelsan
Emre Sümer
Başkent university
Abstract
Bilge Sinan1, Emre Sümer2 Özet— Sentetik açıklıklı radar (SAR), gün boyunca ve her türlü hava
koşullarında çalışabilen aktif ve yüksek hassasiyetli bir görüntüleme sistemidir. SAR görüntülerine
dayalı hedef tanıma, son on yılda geniş çapta araştırılmıştır. SAR görüntüleri, girişime karşı zayıf
dirençli ve tek renkli olduğundan optik görüntüler kadar net ve okunaklı değildir. Geleneksel SAR hedef
tanıma yöntemleri, görüntünün istatistiksel ve fiziksel özelliklerine dayalı olarak gerçekleştirilir.
Ancak, bu yöntemlerin SAR hedef görüntülerinin karmaşık ve değişken doğasına uyarlanması zordur.
Uzak çekim SAR görüntülerinde ilgilenilen küçük hedefleri gözle tespit etmek zaman alıcıdır ve pratik
değildir. Bu nedenlen SAR görüntüleri için otomatik hedef tanıma algoritmalarının geliştirilmesini
Page 87
SEMIT2023
büyük önem arz etmektedir. Son yıllarda yapılan çalışmalar incelendiğinde evrişimli sinir ağı tabanlı
hedef tanıma yaklaşımları, hedefin özelliklerini SAR görüntülerinden başarılı bir şekilde çıkarabilmekte
ve yüksek bir tanıma doğruluğu elde edebilmektedir. Bu çalışmada, MSTAR veri kümeleri kullanılarak
güncel bir nesne tespit modeli olan YOLO’nun SAR hedeflerini tanıma performansı araştırılmıştır.
Yakın bir tarihte geliştirilmiş olan YOLOv8’in, önceki YOLO versiyonları olan v3 ve v5 modelleriyle
MSTAR veri seti üzerindeki performansı karşılaştırılmıştır. Bu çalışmada radarın 15° ve 17° lik bakış
açısından alınan BMP2, BRDM2, BTR60, BTR70, D7, T62, T72, ZIL131, ZSU234, 2S1 hedeflerin
görüntüleri otomatik olarak tespit edilmeye çalışılmıştır. Sonuçlar incelendiğinde, YOLOv8
versiyonunda diğer YOLO versiyonlarına kıyasla daha kısa bir eğitim süresiyle daha yüksek bir hedef
tanıma doğruluğuna ulaşılabileceği gösterilmiştir.
Keywords: Evrişimli Sinir Ağı (CNN), Sentetik Açıklıklı Radar (SAR), Hedef Tanıma.
Sporcular İçin Sakatlık Ve Performans Analizi Çerçevesi

Injury and Performance Analysis Framework for Athletes
Osman Çayli
Fuat Karip
Atınç Yilmaz
Bilgisayar Mühendisliği, İstanbul Beykent Üniversitesi, İstanbul, Türkiye
Abstract
Spor, dünyanın önde gelen sektörlerinden biri haline gelmiştir. Bu nedenle sporcu sağlığı, sektör
kazanımlarını veya zararlarını direkt etkileyen en önemli etkenlerden biridir. Sporcuların fiziksel sağlık
durumlarının ve performanslarının analizi yapılarak olası sakatlıklar veya performans düşüklükleri
öngörüldüğünde risklerin azaltılması mümkün olabilmektedir. Geleneksel yöntemler ile sınırlı olarak
tahminler yapılabilse de makine öğrenmesi yöntemlerinin desteği ile sporcuların performansını
arttırmak ve potansiyel sakatlık risklerini ortadan kaldırabilmek için önlem alabilmek mümkün
olacaktır. Bu çalışmada, makine öğrenmesi yöntemlerinin kullanımı ile sporcuların sakatlanma ve
düşük performans risklerinin tahmin edilmesi için modelleme yapılmıştır. Ortaya konan model bir
yazılıma entegre edilmiştir. Sporculardan elde edilen eş zamanlı veriler üzerinden beslenen model ile
sakatlık ve düşük performans konularında acil tedbir alabilme imkânı sağlanmıştır. Bu sayede yatırım
yapılan sporcuların mevcut riskleri minimize edilecek daha iyi performans göstermelerine olanak
sağlanacaktır.
Keywords: Uzman sistemler, kural tabanı, Sporcu Yönetimi, İyileşme İzleme, Risk Analizi, Sakatlık
Analizi, Veri Yönetimi, Yapoy Zeka.
Page 88
SEMIT2023
Exploring the trends of Artificial Intelligence on Recruitment: A Bibliometric study

Naval Lawande
Symbiosis Institute of Management Studies (SIMS) Symbiosis International (Deemed University),
Pune, India
Abstract
Purpose: Artificial Intelligence (AI) as a field has recently evolved as a trans-formative force in the
field of recruitment. Numerous empirical, conceptual, and exploratory studies have been conducted that
indicate the novel ways organizations identify, select, and attract top talent. Nevertheless, no scientific
mapping or attempt to map the literature scientifically in the domain has been done in the past. Hence,
the paper intends to showcase the trends in the field of AI and Recruitment research through of
knowledge and conceptual structures. Methodology: An analysis using bibliometric tools in the field of
AI and recruitment was conducted. A sample of 1321 documents was extracted from the Scopus
database. This was based on a search strategy determined by the author. An application that is equipped
with a bibliometric package was in-stalled. This software enabled the analysis of the dataset and various
themes, patterns and trends were identified based on the authors, countries, and citations. Findings:
Results show that the domain of AI and Recruitment needs to be given direction. There is also a pressing
need for interdisciplinary research in this area. The article provides some crucial insights on areas that
need further inquiry. The conceptual and social network structure depict a upward trend in terms of this
area of research. There is a growing demand for Talent Acquisition practitioners and Recruiters with
AI skill sets. Research limitations/implications: The current paper only had Scopus Data-base at its
backdrop. Future researchers could use multiple databases such as Web of Science and conduct a
comparative study. A systematic literature re-view would widen the scope and help identify some of
the niche uncharted territories of Recruitment and artificial intelligence.
Keywords: Artificial Intelligence, Recruitment, Bibliometric Study.
Page 89
SEMIT2023
SP02: Advanced Intelligent Systems and Sustainable Development
Solving On-Demand Transportation Problems Using an Intelligent Evolutionary Algorithm

Based on Hierarchical Objective Exchange
Sonia Nasri
Higher Institute of management, Tunis University, Tunisia, LARODEC of ISG
Hend Bouziri
Higher School of Economic and Commercial Sciences Tunis, Tunis University, LARODEC of ISG,
Tunis, Tunisia
Wassila Aggoune-Mtalaa
Luxembourg Institute of Science and Technology, L-4362 Esch-sur-Alzette, Luxembourg
Abstract
The aim of the paper is to provide advanced transportation solutions that effectively address current and
anticipated customer needs. To tackle this goal, we address a bi-objective personalized on-demand
transportation problem using a novel Evolutionary Lexicographic Algorithm. This algorithm focuses
on the behavior of solutions when evaluating the two objectives with varying priorities. This method
employs specific optimization techniques and an efficient dominance sorting approach, incorporating
an intelligent candidate list to minimize computation time. A comparison between optimal and
approximate solution sets highlights the effectiveness of the bi-objective lexicographic approach. A
deeper analysis comparing our novel approach to existing methods in the literature emphasizes the
promising outcomes within a broad range of on-demand mobility real-life cases.
Keywords: Dial a Ride Problem, Lexicographic Optimization, Bi-Objective Formulation, Customer
Service Quality.
A Variable Neighborhood Search for the Two Echelon Location Routing Problem
Amira Mzita
Business Higher School of Tunis, Manouba University, Tunisia
Sonia Nasri
Higher Institute of management, Tunis University, Tunisia, LARODEC of ISG
Hend Bouziri
Higher School of Economic and Commercial Sciences Tunis, Tunis University, LARODEC of ISG
Tunis, Tunisia
Wassila Aggoune-Mtalaa
Luxembourg Institute of Science and Technology, L-4362 Esch-sur-Alzette, Luxembourg
Abstract
This work presents a Two-Echelon Location Routing Problem, a logistics challenge aimed at optimizing
facility locations and delivery routes while minimizing associated costs. It involves the efficient
distribution of goods, with a specific focus on reducing carbon emissions. This is particularly relevant
in sectors like construction, where environmental responsibility holds significant importance. In this
study, we design a construction logistics network that incorporates Micro-Consolidation Centers. These
intermediate hubs receive materials from suppliers, consolidate them, and then deliver them to
Page 90
SEMIT2023
construction sites. The main objective is to minimize the total cost, encompassing Micro-Consolidation
Centers opening costs, transportation expenses, and storage charges, while fulfilling demand
requirements. The presented problem introduces notable challenges that require novel algorithmic
approaches. Therefore, we propose a new Variable Neighborhood Search as a solution strategy, which
combines local search with perturbation operators. These operators include swap, insertion, and 2-opt
techniques that play a critical role in enhancing the algorithm's search capabilities and enabling it to
move beyond local optima.
Keywords: Construction Supply Chain, Optimization, Location-Routing Problems, Micro-
Consolidation Center, Transportation.
A Deep Implicit Coordination Graph-Based Soft Actor-Critic Approach in Multi-Agent

Reinforcement Learning
Fathallah Wael
Laboratoire de Recherche Opérationnelle, de Décision et de Contrôle de processus (LARODEC),
Institut Supérieur de Gestion, Université de Tunis
Nahla Ben Amor
Laboratoire de Recherche Opérationnelle, de Décision et de Contrôle de processus (LARODEC),
Institut Supérieur de Gestion, Université de Tunis
Abstract
The popularity of multi-agent deep reinforcement learning (MADRL) is growing rapidly with the
demand for large-scale real-world tasks that require swarmintelligence, and many studies have
improved MADRL from the perspective ofnetwork structures or reinforcement learning methods.
Despite these advancements, the practical implementation of MADRL in real-world scenarios remains
impeded by the meager sample efficiency of models and the exorbitant costs associated with data
collection.
Thus, improving sample efficiency has emerged as a crucial challenge in applying MADRL. In this
paper, we propose to integratea Deep Implicit Coordination Graph (DICG) into a soft actor-critic
learning algorithm (SAC), an off-policy actor-critic deep RL algorithm based on the maximum entropy
reinforcement learning framework. The actor's objective is to maximize the expected reward while also
maximizing entropy. This means that the actor strives to succeed at the task while acting as randomly
as possible, ensuring exploration of different strategies.
To evaluate our method (so-called DICG-SAC), we investigatea partially observable environment
known as the traffic junction environment, which serves as a simulation for autonomous driving. This
environment is designed to accommodate multiple vehicles (agents) driving on predetermined routes
within a grid structure that includes intersections. Experiments show a substantial improvement in both
performance and sample efficiency by applying a DICG-SAC approach, surpassing state-of-the-art
methods.
Keywords: Multi-Agent System, Deep Reinforcement Learning, Coordination, Graph Neural Network,
Soft Actor Critic.
Page 91
SEMIT2023
Cloud Computing Model for Handling Medical Big Data: A Mobile Hospital Pervasive
Healthcare Application
Wided Oueslati
Ecole Supérieure de Commerce de Tunis, University of Manouba, BESTMOD Laboratory
Hela limam
Institut Supérieurd 'Informatique, Université de Tunis El Manar, Laboratoire BestMod
Sonia Nasri
Business Higher school of Tunis, Manouba university, LARODEC laboratory, Institut Supérieur de
Gestion de Tunis University of Tunis
Abstract
The handling of big data by cloud computing is becoming a good practice in the management of
information systems and can be considered one of the promising and emerging technologies in
computing in many domains. Actually, big data is somewhat dependent on the cloud for the flexibility
and extensions it provides through virtual machines and aids big data to be accessible. In fact, Big data
are huge and heterogeneous as they come from different sources and of different types i.e. structured,
semi-structured, and unstructured so they require proper storage and high analysis capabilities. Cloud
computing seems like a suitable repository to store such data in a decision support system. In fact, data
stored in the cloud can be extracted and accessed easily since cloud technology allows access to big
data and computer resources from anywhere and at any time that a network connection is available. Big
data and cloud computing have opened many research possibilities in different sectors. Among those
are those engaged in the field of healthcare. The healthcare sector generates each year a high volume of
data to be analyzed to solve organizations’ problems, predict epidemics, detect diseases and reduce
healthcare costs. The aim of this paper is to shed the light on how cloud computing as an emerging
paradigm is improving the healthcare sector, especially in rural zones. To this end, we propose a
pervasive healthcare cloud solution to store medical big data collected by medical staff via mobile
devices and to deliver that data to the medical cloud center for processing, analysis, and decision-
making
Keywords: Emergency Management, Mobile Physician, Smart Connectivity, Hospital Mobility,
Location-Based Services, Healthcare Pervasive Systems, SAAS, Cloud Computing.
A Hybrid Deep Learning Recommender System Based on Trust and Tags

Malek Ghanem1
Raouia Ayachi1
Wided Guezguez1
1
Larodec laboratory ISG Tunis Tunisia
Abstract
With the rapid growth of e-commerce, e-learning, and social networks, recommender systems have
become popular and important tools used in many applications. They aim to aid customers making
decisions that correspond to users’ preferences. Collaborative filtering is within the effective
recommendation techniques that leverages past assessments to predict future interests for users.
However, this algorithm suffers from sparsity. It does not have enough information to learn and generate
predictions.
Page 92
SEMIT2023
One approach to improve the recommendation quality is exploring trust and tag information. By trust,
we mean a relationship between two users that could be linked to the similarity of their interests, while
tags correspond to summarizing the characteristics of items and reflecting users’ preferences through
tagging behaviors.
Nevertheless, these data are also sparse because users do not provide sufficient information about these
secondary resources. This makes the computational complexity of recommendation high, more
precisely the calculation of the similarity matrix.
In this article, our main idea is to extract latent features from trust and tag information and use a hybrid
deep learning approach in order to generate recommendations. In fact, the autoencoder is used to extract
latent features. Then, these latter are used as an input in a hybrid deep learning model, composed of an
embedding layer, RNN and a fully connected linear dense layer.
To assess the effectiveness of our method, we compared it to the state-of-the-art models using two
datasets Lastfm and Delicious. The obtained results are satisfactory.
Keywords: Recommender System, Deep Learning, Trust and Tags, Feature extraction.
Combination of Self Training and Active Learning for Sentiment Analysis

Wided Guezguez
Abstract
With the advancement of web technology and its growth, there is a huge volume of data present in the
web for internet users and a lot of data is generated too. Internet has become a platform for online
learning, exchanging ideas and sharing opinions.
In this research work, we were interested in finding a great method for Sentiment Analysis. We focus
on the problem issue of data selection for building a classification model. The problem is due to, firstly,
the extremely small labelled training set and, secondly, acquiring a large high-quality training set with
high overhead costs. A new approach was proposed combining the Active learning and Self-training
methods to solve those problems.
Our main objective is to propose an efficient approach for labeling in order to enrich our training set
with labeled data to make our sentiment analysis model more efficient and more accurate. For this aim,
we propose to use active learning combined with self training.
Keywords: Sentiment Analysis, Self Training, Active Learning.
Page 93
SEMIT2023
SP03: IT and EM Based Case Studies
The Negative Effects of Personal Protective Equipment Protective Masks on Users and
Evaluation of Their Designs
Doğa Martı
Ankara Yıldırım Beyazıt University, Institute of Science and Technology, Occupational Health and
Safety, Ankara, Turkey
Ergun ERASLAN
Safety, Ankara, Turkey
Abstract
Personal Protective Equipment (PPE) plays a crucial role in ensuring the safety and well-being of
workers in various industries. PPE refers to special clothing, equipment and accessories designed to
protect individuals from the potential hazards and risks present in their work environment. PPE protects
workers from occupational hazards, ensures compliance with legal requirements, increases employee
confidence and productivity, prevents occupational diseases, and reduces work-related injuries and
costs. Organizations demonstrate their commitment to the health, safety and well-being of their
workforce by prioritizing the provision and appropriate use of PPE. Protective masks from PPE have
also emerged as very important tools to prevent the spread of respiratory diseases and protect public
health. However, the design of these masks plays a crucial role in usability, comfort and effectiveness.
However, protective masks can create some negative effects on employees. Masks can cause irritation,
squeezing, breathing difficulties, etc. in people. can create awkward situations. In this study, the
determination of the negativities caused by the mask, and then the design suggestions that can be made
in order to prevent these negativities are expressed. It has been seen that by providing training on the
use of masks and producing protective masks with better designs, work accidents and occupational
diseases will decrease and a more comfortable and safe working environment will be created. By
emphasizing user-centered design approaches, iterative processes, and collaboration between
researchers, engineers, and manufacturers, protective masks can significantly improve usability and
effectiveness. Addressing design challenges related to fit, breathability, communication, comfort and
wearer satisfaction allows masks to effectively protect individuals while encouraging their acceptance
and adherence to wearing protocols. Consequently, the importance of innovations in protective mask
design cannot be underestimated. These developments improve the effectiveness, comfort and wearer
experience of masks by providing optimum protection against airborne pollutants and infectious agents.
By adopting and implementing these innovations, a significant contribution can be made to the
protection of public health and well-being.
Keywords: Protective Mask, PPE, Mask Design.
Page 94
SEMIT2023
Association between Demographic factors and Internet Banking Usage

R MuthuKumar
Department of Management Studies (Karaikal Campus) Pondicherry University Puducherry, India
Lalitha Ramakrishnan
Department of Management Studies (Karaikal Campus) Pondicherry University Puducherry, India
A Poongodai
CMR College of Engineering & Technology Kandlakoya, Hyderabad
Krishnamacharyulu
Sri Venkateswara University Tirupati, India
Abstract
Internet banking, an innovative way of banking delivery, plays an important and major role in the
banking industry worldwide particularly India. Almost all the banks in India provide such innovative
internet banking services. But, despite the banks’ efforts towards increasing internet banking customers,
it is yet to reach its desired levels and attract all customers. This study made an attempt to identify the
association between demographic factors and the internet banking usage among the internet banking
users and customers. Using convenience sampling 429 proper responses were collected from the
internet banking users in India. Descriptive analysis was used to understand the profile of sample
respondents. In addition, Chi-square analysis was used to find out the relationship between demographic
factors and the internet banking usage. From the analysis results, it was found that there is a significant
association between those two: demographic characteristics and internet banking usage.
Keywords: Demographic Factors, Internet Banking, Chi-Square Analysis, Descriptive Analysis,
Banking Services.
An Investigation on The Possibility of Retrofitting 3d Printers from Local Scrap E-Waste in

Nigeria
Paul Amaechi AKO (FORMERLY OZOR)
Department of Mechanical Engineering University of Nigeria Nsukka123
Chukwuemeka Onyedinma ANYAOHA
Ndudim Henry ONONIWU1
Eveth Nkiru Nwobodo-Anyadiegwu
Department of Quality and Operations Management, University of Johannesburg, South Africa
Abstract
Three-dimensional (3D) printing has been largely described as an additive manufacturing procedure
whereby materials are dropped, coupled, or hardened under computer control to build a desired object.
It may be done in layered sequence using grains of different materials or liquids. The category of
equipment used in achieving 3D printing are varied in presentation. A conventional 3D printer
comprises various units and subsystems, but uniformity is achieved in the mode of operation. While the
technology is fast spreading to many parts of the world, the attendant cost is still prohibitive, especially
in low-income economies. This places substantial hindrance in not only the use-advocacy but more
importantly in its enormous benefits of good aesthetic finish, perfect identity, high accuracy, and
precision of throughputs, among others. In this study, an attempt to utilize the ubiquitous electrical and
electronic waste (E-Waste) prevalent in Nigeria to construct a cost-effective 3D printing demonstration
Page 95
SEMIT2023
unit that can be used to introduce students to additive manufacturing is made. Using an experimental
research approach, the study was divided into three different phases; which include scrap gathering,
sorting, characterization, and specification in phase 1. Phase 2 involves the fabrication and assembly of
components, while programming and calibration of the printer were accomplished in the final phase.
The work reveals that the development of 3D printers from mostly local E-waste can drastically reduce
the cost of their foreign counterparts. The results show that electronic and electrical scraps which
constitute menacing waste management issues can be converted to wealth through deployment in 3D
printer fabrication.
Keywords: E-waste, 3D Printing, retrofitting, characterisation, waste management, desired result.
Interlocking Fired Bricks as Innovative Waste Abstraction Proxies; Case Study Abuja
Northern Nigeria
Christopher Ehizemhen Igibah
Department of Civil Engineering, Federal University Oye-Ekiti State Nigeria
Olugbenga Oludolapo Amu
Christopher A. Fapohunda
Lucia Omolayo Agashua
Department of Civil Engineering, Federal University of Technology Akure Ondo State, Nigeria
Abstract
The disposal of large amounts of red clay, settled below pavement layout will substantially increase the
engineering cost of construction and cause damage to the ecological environment. Thus, to assess
efficiency of the interlocking fired bricks as waste to green pavement material, its properties was
scrutinized. Clay samplings were obtained from 3 sites where the samplings were prominent in Abuja
– Nigeria, stabilized with 3 diverse waste; waste lactose (WL), eggshell ash (ESA) and waste paper ash
(WPA), before scrutinized by means of standard techniques. Results acquired were compared with
EURO and ACI standard as well as subjected to variability analysis. CBR and Atterberg limit test
showed that all the clay samples did not conform to the requirement that CBR should not less than 8%
and PI should not be more than 12%. But increase in waste additive, CBR increases from 2.12% to
117.41%, whereas both liquid limit and plasticity index reduces from 59.2% and 23.3% to 32.9% as
well as 11.2% respectively for Bwari site at optimum percentage and significant correlation of r = 0.96,
α = 0.05. Further scrutiny on each of the concrete samples cured for 3, 7, 14, 21 and 28 days,
demonstrated that the samples produced with eggshell ash, waste paper ash and waste lactose had the
highest compressive strength values, followed by those produced without waste paper ash. Likewise,
XRD, XRF and FTIR investigation reveals the loss on ignition value which in turn designates those
clay samplings has lesser carbonaceous substance and greater mineral matter quantities.
Keyword: Fired-Bricks; Pavement; Green Construction, Interlocking.
Page 96
SEMIT2023
Omnichannel strategy adoption issues in an eCommerce company: a SAP-LAP analysis

Kishori Kasat
Symbiosis School for Liberal Arts, Symbiosis International (Deemed University)
Naim Shaikh
Global Business School and Research Centre, Dr. D. Y. Patil Vidyapeeth (Deemed to be University
Venkatesh Iyengar
Christ College, Pun
Abstract
The omnichannel strategy is aimed at producing an integrated and seamless customer experience across
manifold channels (e.g., social media, physical stores, website portals, mobile applications, instant
messaging platforms, etc.). Adopting omnichannel in an eCommerce company could be challenging
and complex with several implementation issues. Existing literature reports various adoption issues;
however, these are often deliberated in isolation. Legacy systems, data silos, organizational silos,
inconsistent customer journeys and experiences, resource allocations, technological challenges,
workforce readiness, order fulfilment, inventory, security and privacy concerns, data analytics, and
many more could be those noteworthy challenges. There is a need for a robust implementation plan
and top management commitment at all levels of management. This paper aims to analyze a case study
of eCom1, a leading eCommerce firm in India, which could enlighten on the current status of
omnichannel adoption measures undertaken by them. A situation-actor-process (SAP) - learning-action-
performance (LAP) framework has been applied for the case study of eCom1. The situation
characterizes the current scenario, actors the participants influencing the situation, and the process refers
to the adoption strategies. Based on this SAP, different learning issues are analyzed, guiding
appropriate actions followed by performance outcomes intended for its stakeholders. A resource-based
view (RBV) of the firm is suggested to understand the tangible and intangibles associated with
omnichannel strategy adoption issues for the eCommerce company. The synthesis of SAP will lead to
LAP and fixes the gap suggesting numerous actions based upon learnings from the current situation,
actors, and processes. The SAP-LAP analysis is being used to explain omnichannel adoption and
implementation issues from a managerial context as applied in the eCom1 case. This study is novel and
valuable to eCommerce decision-makers who plan to implement omnichannel strategies and make them
effective and efficient.
Keywords: Adoption, Competitive Advantage, Ecommerce, India, Omnichannel Strategy,
Performance Measures, SAP-LAP Analysis, Resource-Based View (RBV).
Page 97
SEMIT2023
Academic Misconduct after the Rapid Transition to Remote Learning: A Gender and XYZ
Generations Perspective
Sari Andayani
Accounting Department, Faculty of Economics and Business, Universitas Pembangunan Nasional
Veteran Jawa Timur, Jalan Raya Rungkut Madya, East Java, Indonesia
Endah Susilowati
Diah Hari Suryaningrum
Evinda Dwi Nur Aini
Wandah Nur Aliyyah
Singgih Alfiyahya
Naufan Rahmanda Tasri
Moch. Idris
Abstract
The rapid transition to remote learning during and after the COVID-19 pandemic may impact academic
misconduct rates due to changes in learning and evaluation systems. Further understanding of the
impact of the rapid transition to distance learning from a gender and Generation XYZ perspective will
be essential to identify specific problems and challenges and develop effective solutions. Therefore, the
purpose of this present study is to examine the factors that influence academic misconduct from the
perspective of gender and XYZ generation after the rapid change. The factors we analyzed were the
accessibility of technology, the demands of family responsibilities, social support and technical support,
stress and mental well-being, and preferences for learning and communication styles. Our study is an
exploratory study using a descriptive quantitative approach. In December 2022, we conducted an online
survey by distributing questionnaires to accounting students and lecturers at three universities on the
Indonesian island of Java. The three universities represent West Java, Central Java, and East Java. A
total of 155 questionnaires were analyzed. Our findings are very important as they provide insight and
knowledge into the culture of academic integrity in the face of rapidly changing learning technologies.
We suggest conducting in-depth interviews with several respondents of students and lecturers to help
understand more deeply the factors that influence deciding to commit academic misconduct.
Keywords: Family Demands and Responsibilities, Learning Preferences and Communication Styles,
Social Support and Technical Support, Stress and Mental Well-Being, Technology Accessibility.
Page 98
SEMIT2023
Chinese Postman Problem with Dynamic Start/End Point

Mustafa Yilmaz
Ataturk University, Faculty of Engineering, Department of Industrial Engineering, Erzurum, Turkey
Abstract
The Chinese Postman Problem (CPP) is a variant of the arc routing problem, and it is a routing problem
that aims to return to the starting point with the shortest tour length by passing through all arcs at least
once in a network consisting of nodes and arcs. CPP has many real-life applications such as snow
removal and road maintenance work. In the literature, it is possible to see different CPP applications by
adding new constraints to the problem. In almost all of these studies, the starting point is determined as
a fixed point and it is aimed that the tour or tours to be created will start from this point and end at this
point. In this study, a new mixed integer model for CPP is proposed and it is aimed to determine the
most appropriate starting or ending point by the model. The proposed model was run on a small sample
network and the results showed that tours with shorter lengths were obtained.
Keywords: Chinese Postman Problem, Variable Start/End Points, Mathematical Model.
Page 99
SEMIT2023
SP04: Data Science, Big Data, Data Mining, and Knowledge Management in EM
Comparison of Cost and Benefit Analysis of Active Leakage Control Applications in Water
Distribution Systems
Abdullah Ateş
İnönü Üniversitesi, Bilgisayar Mühendisliği Bölümü, Malatya
Salih Yilmaz
Çankırı Karatekin Üniversitesi, İnşaat Mühendisliği Bölümü, Çankırı
Mahmut Firat
Inönü Üniversitesi, İnşaat Mühendisliği Bölümü, Malatya
Abstract
The fight against leaks and failures in water distribution systems has gained a serious importance with
the increasing drought, the inefficient use of water resources and the rapid contamination of existing
water resources. These studies are generally carried out with two different approaches, active leakage
control and passive leakage control. Although the passive control method, which covers the repair of
the reported faults, is less costly and easier, the vast majority of the faults that occur in the system are
caused by unreported faults. It is possible to detect, repair and control such faults with active leakage
methods. In order to gain the expected efficiency from these methods, it is very important to regularly
monitor the networks by dividing them into isolated areas, to estimate the amount of preventable
leakage, and to determine the leak location with devices and equipment such as ground microphone,
regional recorder, regional correlator within the framework of a certain systematic program. The regular
maintenance of these systematic studies in the implementation of the active leakage method also creates
serious costs for water administrations. In addition to the costs of creating and monitoring measurable
regions, the costs of devices and equipment with different characteristics, sensitivities and costs to be
used for the determination of the fault location should also be considered in the studies to be carried
out. In this study, a standard calculation structure is defined for the benefits and costs of the ground
microphone and regional correlator methods used to locate unreported leaks. By using the defined
calculation methodology, the benefits and costs for both methods are analyzed and compared in a
sample network. Obtained results; For the most efficient operation, the method to be selected according
to the general characteristics of the networks plays an important role.
Keywords: Water Distribution System, Leakage, Active Leakage Control, Leakage Detection.
Page 100
SEMIT2023
Benefit and Cost Analysis of Network and Service Connection Renewal in Distribution Systems
for Sustainable Water Loss Management
Salih Yilmaz
Çankırı Karatekin Üniversitesi, İnşaat Mühendisliği Bölümü, Çankırı
Abdullah Ateş
İnönü Üniversitesi, Bilgisayar Mühendisliği Bölümü, Malatya
Mahmut Firat
Inönü Üniversitesi, İnşaat Mühendisliği Bölümü, Malatya
Abstract
Reducing the rate of non-revenue water (NRW) in water distribution systems (WDSs) has become one
of the most important goals for utilities. Increasing water losses bring along social, economic and
technical difficulties, which accelerates the work to be done to reduce water losses. Although water
losses are combated by applying active and passive leakage methods, in some cases it may not be
economical for the utilities to manage the network under current conditions and to continue the efforts
to reduce water losses. The aging network and its equipment can now make it difficult to manage this
system. In these cases, the method of rehabilitating the whole or partial parts of the network is applied.
It is possible to transfer and monitor the network and its components to the GIS database, to reduce the
risk of failure and to have low operating costs, to prevent and reduce leaks in subscriber connections by
manufacturing service connections to the meter, to keep the GGS rate at a very low level, and to increase
subscriber satisfaction and service quality with the application of this method. In a distribution system,
in case the existing network conditions are preserved, current failure rates, network operation,
maintenance and repair costs, new water source search and energy costs, initial investment and medium
and long-term operating costs in case of network renewal should be taken into consideration and benefit-
cost analyzes should be made. Due to the high level of pipe material, labor and construction costs, a
detailed cost-benefit analysis and alternative solutions should be evaluated before the network renewal
approach is preferred. In this study, the current costs and benefits that will be encountered in case of
renewal of the entire network (network + service connection), only subscriber lines or only network
lines for a sample network are discussed. The results show that it is that the rehabilitation method to be
chosen for different network conditions has a serious effect on increasing the planned benefit.
Keywords: Water Distribution Network, Network Renewal, Cost Analysis, Benefit Analysis.
Review of Fake News Classification Methods Using Machine Learning

Marzieh Nikoukar
Information Technology Management at the Islamic Azad University, Science and Research Branch
Safanaz Heidari
Department of Public Administration, Miandoab Branch, Islamic Azad University, Miandoab, Iran
Abstract
Based on a review of articles, it can be understood that online social networks have become an effective
and important social platform for communication, exchanging opinions, and sharing information.
However, they also provide the possibility of rapid and widespread dissemination of false information,
which can lead to harmful effects on individuals or society. The World Wide Web contains data various
formats such as documents, videos, and audio files. Detecting and classifying online news unstructured
formats (such as news articles, videos, and audio files) is relatively challenging as it requires human
Page 101
SEMIT2023
expertise. In this regard, the aim of this research was to review the methods of classifying fake news
using machine learning. Through reviewing credible domestic and foreign articles, it was determined
that methods such as geometric deep learning, maximum entropy, deep neural network algorithms,
semi-supervised learning, etc., are efficient approaches for detecting and classifying fake news.
Keywords: Fake News Classification, Fake News Detection, Machine Learning, Deep Learning.
Designing a Data Pipeline Architecture for Intelligent Analysis of Streaming Data

Iryna Mysiuk
Ivan Franko National University of Lviv, 1 Unіversytetska Str., Lviv, Ukraine
Roman Mysiuk
Roman Shuvar
Volodymyr Yuzevych
Karpenko Physico-Mechanical Institute of the NAS of Ukraine, 5 Naukova Str., Lviv, Ukraine
Anatolii Pavlenchyk
Ivan Boberskyi Lviv State University of Physical Culture, 11 Kostiushka Str., Lviv, Ukraine
Volodymyr Dalyk
Lviv Polytechnic National University, 12 Stepana Bandery Str., Lviv, 79013, Ukraine
Abstract
The paper describes the process of creating a data pipeline architecture. based on several data sources
and processing through intermediate data stores and final business analytics. Described the chain of
interaction at each step of the tools from the stage of direct collection to analytics based on the processed
data. Data collected from social networks can be considered among the most popular streaming data.
Since the collection of information consists in working with a large amount of information from several
web pages in an automated way using the Selenium tool, the data warehouse is divided into several
separate modules. In the described implementation, the basis for data storage is Elasticsearch, Kibana,
and Logstash (EKL) technology stack. Analysis and classification of results is based on work with a
machine learning model and data visualization methods. The main results obtained during the design
of such a data pipeline architecture will allow us to highlight the vulnerabilities and advantages of the
development. The designed data pipeline architecture can be useful for designing data processing
processes for other subjects.
Keywords: Data Pipeline, Intelligent Analysis, Data Streaming, Business Analysis.
Validation and Verification of Management Support Systems: A Framework

Ghazi Alkhatib
The Hashemite University, Zarqa, Jordan
Abstract
Abstract: This paper develops a framework that integrates validation and verification (V&V) techniques
in software engineering to the different management support systems (MSS) as it relates to data and
systems organizational hierarchies. MSS consists of Decision Support System, Expert System (ES), and
Artificial Intelligent/Machine Learning Systems (AI\ML). The framework delineates different factors
Page 102
SEMIT2023
for successful integration, such as development methodologies, user involvement strategies and users’
knowledge and intelligence, knowledge types of explicit knowledge (EK) vs. implicit knowledge (IK),
and ease of implementing V&V. The paper then develops three strategies for incorporating V&V for
building and maintaining reliable systems. These include micro V&V during construction, perpetual
interactive V&V during the use of MSS through users’ feedback to immediately remove any
inconsistencies between the new knowledge and existing knowledge, and periodic execution of V&V
operations of MSS to uncover knowledge decay in systems over time intervals. The framework
postulates that the second V&V approach is more cumbersome to use. The following factors are deemed
relevant to successful V&V deployment in MSS system development and operation: system
development, such as incremental development, and Correctness-by-Construction; user involvement,
such as user-led design, and agile methodologies; and the timing of user involvement, such as during
development and operations. The framework postulated that integrating V&V increases in complexity
and risk of using the system as data get filtered through the hierarchical organization of MSS: from data
– information – knowledge – intelligence, as well the increased dependence of decisions on input from
user experiences and the external environment. In addition, tacit knowledge is used more extensively
as we move higher in MSS. As a result, perfuming V&V operations become less lucid. Finally, the
paper presents a simple general-purpose example of a life cycle model which demonstrates how the
three V&V approaches will be used during the main stages of MSS development: input knowledge,
validation, adding knowledge to the knowledge base, construct the system by testing the knowledge
base, validate, operational testing with cases, Knowledge dissemination, and verify the system though
user feedback after operation over time.
Keywords: Validation, Verification, Management Support Systems, Expert System, Decision Support
Systems, Artificial Intelligence, Machine Learning.
Designing a National Cyber Security Strategic Management Model with the Foundational Data
Theory Approach
Fereidoon Bidollahkhany
Faculty of Business Administration, Karabuk University, Karabuk. Türkiye
Mehdi Basiri
Tehran command and staff university, Iran
Abstract
The ever-increasing expansion of the use of cyberspace has led to many economic, social, political, and
cultural activities of people being pushed into this space. In such a situation, we will inevitably witness
the entry of a large amount of data and various information of government and public centers,
organizations, businesses, and even private data of individuals into cyberspace. Based on this,
governments seek to develop basic strategies and solutions in this field to protect information and
preserve data assets. This research aims to provide a strategic management model of cyber security at
the national level to protect cyber information. In terms of the purpose of the research, it is practical,
and the method of its implementation is data foundation. To achieve the goal of the research, documents
and scientific books related to the research topic were first analyzed using the qualitative content
analysis method and MAXQDA software. In the following, after extracting the basic concepts and
themes of the research paradigm model in five dimensions concluded” the background, causal and
intervening conditions, consequences, and strategies “, by referring to the number of 15 cyber security
experts and using the focus group, the research designed model by making corrections was confirmed.
Keywords: Model, Strategic Management, Cyber Security, Foundational Data Theory.
Page 103
SEMIT2023
SP05: Safety and Reliability Engineering, System Engineering and System Safety,
Safety in Industry 4.0
The Impact of Internet of Things, Blockchain, and Artificial Intelligence on Occupational

Health and Safety
Seyedeh Zahra Mir Mohammadzadeh Noudehi
Environment Management (HSE) Department, Faculty of Environment and Natural Resources,
Science and Research Branch, Islamic Azad University, Tehran, Iran
Farnaz Bagheri Khameneh2
Environment Management (HSE) Department, Faculty of Environment and Natural Resources,
Science and Research Branch, Islamic Azad University, Tehran, Iran
Abstract
Occupational health and safety (OH&S) covers the physical, mental and social well-being of employers
in all occupations. The rate of using technology is increasing rapidly. Internet of Things, Blockchain
and Artificial Intelligence are emerging technologies being applied for creating innovative solutions in
various sectors, such as health and safety at work, so the purpose of this article is to know the impact
of the different technologies on occupational health and safety risks on different workplaces. An
analysis of the literature was carried out to achieve this objective. In order to describe the role of
different technologies in health and safety workplaces, the authors have considered the role of the most
important technologies in this field. Knowing the trends of changes in the industry will help to
concentrate on the most important questions and solutions. This article will review the most important
trends in technologies related to occupational health and safety.
Keywords: Occupational Health and Safety, Internet of Things, Blockchain, Artificial Intelligence.
Improving the Visual Ergonomics of Computerised Workplaces Through the Use of Specialised
Eye-Rest Software
Tihomir Dovramadjiev
Technical University of Varna, Bulgaria
Darina Dobreva
Ralitsa Zlateva
Abstract
Abstract. This article focuses on improving the visual ergonomics of computer workstations through
the use of specialised eye-rest software, taking into ac-count human factors. Prolonged screen time at
computer workstations often leads to visual discomfort and strain, collectively known as Computer
Vision Syndrome (CVS). To address this problem, users/workers should be informed of the potential
of eye-rest software as a solution. This article highlights how specialised eye-rest software incorporates
techniques such as regular eye exer-cises, break reminders, screen adjustments and blue light filters to
reduce visual strain and promote eye comfort. It presents recent research on the effectiveness of eye-
rest software in reducing visual discomfort and CVS symptoms, taking into account the individual needs
Page 104
SEMIT2023
and preferences of users. The article also examines the challenges and opportunities associated with the
implementation and adoption of eye-rest software, including user acceptance, software compatibility
and organisational support. The potential benefits of ergonomic and eye-rest software in improving
productivity and employee well-being by optimising visual ergonomics are detailed. The leading
priorities in the article are to provide accessible digital resources to all stakeholders for visual health
prevention and to optimal ergonomics in human – computer interaction (HCI) when working with
digital workstations.
Keywords: Ergonomics, Workplaces, CVS.
Proposing a Conceptual Model for Critical Success Factors in Establishment of Knowledge

Management System Using Interpretive Structural Modeling; Case Study of an Oil and Gas
Well Drilling Company
Sodabeh Orak
Department of Engineering, Ale Taha Institute of Higher Education, Tehran, Iran
Hesam Varaee
Department of Engineering, Ale Taha Institute of Higher Education, Tehran, Iran
Mazdak Khodadadi-Karimvand
Department of Industrial Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
Abstract
One of the management concerns in every organization is the existence of optimal management and
required resources. Allocating resources in order to achieve success is one of the most complex
activities of the organization. Providing a conceptual model by identifying the key success factors for
the implementation of the knowledge management system is of particular importance. Interpretive
structural modelling provides a proper order and framework for such systems and provides decision
makers with a real picture of the impact of key factors in achieving the final goal. In this research,
identifying the type of communication and determining the levels of key success factors has been done
with the help of interpretive structural modelling. The communication matrix was completed by the
experts and then by summarizing the opinions and performing ISM analysis, the factors were placed in
12 levels. According to the conceptual model of key success factors in knowledge management,
leadership is the basis of the model and the origin of other factors and should be considered as the most
important key success factor in knowledge management. Considering the dependence of knowledge
management on communication between people and how to use information technology, solutions are
introduced that achieve the goal of knowledge management.
Keywords: Knowledge Management, Interpretive Structural Modelling, Critical Success Factors, Oil
and Gas Industry, Well Drilling.
Page 105
SEMIT2023
Accident Analysis with Bow Tie Method; Case of British Petroleum Texas Refinery
Negar Mohabbatpour
Department of Safety Engineering, Faculty of Engineering, University of Science and Culture,
Tehran, Iran
Tehran, Iran
Abstract
In process industries, the liquid overflow from tanks and other containers is one of the main causes of
accidents. This may cause catastrophic accidents and a lot of damage to the organization. One of
significant cases in such incidents is the incorrect operation of the liquid level indicator and the high
level alarm. The accident at the Texas British Petroleum refinery in March 2005 is one of the recent
cases, which killed 15 workers and injured 150 of them. The explosion was caused by the overfilling
of the raffinate splitter tower and a blowdown drum releasing hot hydrocarbons. The resulting vapour
cloud ignited, destroying the ‘ISOM’ unit. This accident cost British Petroleum Company billions of
dollars and it can be considered one of the worst industrial disasters to date. To prevent and control the
occurrence of these accidents, it is possible to use different types of hazard identification, risk
assessment and analysis methods. This action is for prioritizing risks and making decisions in order to
reduce risk to an acceptable level. There are several tools which organizations can used for risk
assessment. One of the common risk assessment tools is Bow Tie method. The bow tie method is a
really visual way of understanding the impacts of a hazard, the risk it presents, the consequences and
the controls that should be put in place. The research results show that this technique was one of the
most effective risk assessment methods. The purpose of this article is to analyze the British Petroleum
Texas Refinery explosion accident with the Bow Tie method.
Keywords: Process Safety Management, Risk Assessment, Accident Analysis, Bow Tie Method.
Impact of Technological and Human Needs for Re-Engineering Processes on Improving

Productivity
Sajad Barkhordari
Department of Management and Economy, Science and Research Branch, Islamic Azad University,
Tehran, Iran
Department of Industrial Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
Abstract
This research examines the impact of technological and human needs for reengineering processes in
improving productivity. The statistical population of this research is 400 employees of an Oil and Gas
Well Drilling Company who are aware of the subject of the research. The sample size is 196 people
according to the size of the population using Cochran's formula and the sampling method is random. In
the current research, library and field methods were used to collect data, and the data collection tool is
a standard 16-question questionnaire. In connection with validity, content validity was used and in
connection with reliability, Cronbach's Alpha method was used, and its value was 0.779 Obtained.
Kolmogorov-Smirnov test was used to check the normality, and in the next step, structural equations
and path analysis were used. SPSS and PLS software were used for calculations. The results showed
Page 106
SEMIT2023
that technological and human needs for re-engineering processes have an effect on improving
productivity.
Keywords: Technological and Human Needs, Re-engineering Processes, Productivity Improvement.
A Study on Measuring the Mental Workload of Instructors and Evaluating Significant

Differences
Samet Tosun
Department of Occupational Health and Safety, Ankara Yıldırım Beyazıt University, Ankara, Türkiye
İbrahim Yilmaz
Department of Industrial Engineering, Ankara Yıldırım Beyazıt University, Ankara, Türkiye
Abstract
Mental workload is considered as the difference between the information processing capacity necessary
to meet performance expectations and the capacity that is possible to perform within a specified time
period. It is important to measure mental workloads in order to determine whether today's employees
are at the desired level mentally. As a result of this situation, many methods have been developed so far
to determine the mental workload. In this study, it is aimed to evaluate the mental workload of academic
staff working in universities in Tokat, Ankara and Sivas and to determine whether the mental workload
differs significantly according to the titles of academic staff. During the evaluation of the collected data,
the SPSS 27 (Statistical Package for the Social Sciences 27) package program was used, and the data
were evaluated with content analysis. The structured interview questions used in the study were
prepared by Rubio et al. (2017) developed and presented to the literature by using the CarMen-Q Mental
Workload Scale. The measurement consists of 4 sub-dimensions and 29 items consisting of cognitive
workload, temporary workload, performance-related workload and emotional workload. The internal
consistency coefficient of the study was calculated as a=0.94, and it was determined that the highest
mental workload sub-dimension was cognitive workload and the lowest sub-dimension was temporary
workload. In addition, the study has the feature of being a study in which CarMen-Q, a mental workload
assessment method that has not been studied before in the literature, is used. It has been emphasized
that it is necessary to determine the factors that increase the mental workload of the instructors and to
carry out studies to reduce them.
Keywords: Ergonomics, Mental Workload, CarMen-Q.
Page 107
SEMIT2023
SP06: Applied Soft Computing in engineering Management
Development of UAVS on Pixhawk Family Controllers with Radio Interference Protection

System
Serhii Lienkov
Taras Shevchenko National University of Kyiv
Alexander Myasischev
National University "Lviv Polytechnic"
Oleksandr Sieliukov
School of Aerospace Xi'an Jiaotong University
Vadim Ovcharuk
Lviv Polytechnic National University
Serhii Frolov
The National Defence University of Ukraine named after Ivan Cherniakhovskyi
Mykola Kubiavka
Taras Shevchenko National University of Kyiv
Abstract
Research and creation, adjustment and testing of rotor-type unmanned aerial vehicles (UAVs) in the
conditions of countering radio interference were carried out in the work. The problem solved during the
construction of such a UAV is the provision of stable flight in the area of operation of systems that
create radio-electronic interference (hereinafter - REI). The designed UAV is capable of automatically
carrying a payload of 2.0–2.5 kg at a speed of 40–45 km/h for 25–30 min under the control of the
Pixhawk flight controller. The article indicates the expediency of using a mathematical model based on
the third version (order) extended Kalman filter algorithm to ensure flight stability in difficult
meteorological conditions, including strong gusty winds up to 12–13 m/s). The necessity of using
rejection filters for software suppression of the effect of vibrations of the rotor-motor group on the
operation of the flight controller is indicated, which is also important when flying in the areas of
operation of radio-electronic interference systems. They lead to additional vibrations of the UAV. An
experimental study of the flight of the designed copter was carried out according to the lua script, which
allows control in conditions of loss of communication with navigation satellites. It is shown that in this
case, the UAV returns to the starting point using an algorithm that works according to the readings of
the barometer and magnetometer. The effectiveness of the use of dual-band GPS receivers with the
installation of antenna screens to counteract REI even in the conditions of their close location from a
flying drone is experimentally shown. A script has been developed that allows the UAV to perform
stable flight in the event that the navigation commands are replaced by REI systems. The work is of
practical interest, since the designed multicopter can be used in the areas of system operation that create
strong radio-electronic interference.
Keywords: GPS receiver, БПЛА, simple RTK2B Lite, Arducopter, Ardupilot, Pixhawk 2.4.8, Notch
filter, Mission Planner, STM32F427, X11 HobbyWing.
Page 108
SEMIT2023
An Efficient Hybrid Genetic Algorithm for Two-Sided Assembly Line Balancing Problems
Şehmus Aslan
Abstract
An assembly line is a manufacturing process in which a product is assembled by a series of sequential
steps performed by workers or machines at different workstations along a production line. A one-sided
line has a linear layout with workers on one side, while in a two-sided assembly line, the product being
assembled moves along the two lines, and workers on each side perform specific tasks simultaneously.
This setup increases the efficiency and throughput of the assembly process as it allows for concurrent
operations on both sides. Even though most research efforts concentrate on one-sided assembly lines,
the significance of two-sided assembly lines is increasingly recognized, particularly for the assembly
of large-sized products like cars, buses, or trucks. Since two-sided assembly line balancing problem is
well known as NP hard, a hybric genetic algorithm (hGA) is developed to minimize the cycle time. A
problem specific local search technique is embedded into the GA loop to enhance the convergence
ability towards the optimal solution. The proposed hGA is evaluated using a selection of test problems
sourced from existing literature. The computational outcomes and statistical examination demonstrate
that the suggested approach produces encouraging outcomes.
Keywords: Assembly Line Balancing, Two Sided Assembly Line, Hybrid Genetic Algorithm,
Metaheuristics.
Optical Transmission Systems Incident Management in the Telecommunications Sector in

Senegal
Abdou Diop
Polytechnic High School (ESP/UCAD)
Ibrahima Ngom
Doctoral School of Mathematics and Computer Science
Idy Diop
Cheikh Anta DIOP University of Dakar (UCAD), BP 5005- Dakar - SENEGAL
Abstract
We present the maintenance of optical fiber in the telecommunications sector in Senegal. This survey
is the start of a project to set up a real-time incident management system for fiber optic transmission
networks. In concrete terms, there are several stages in fiber optic maintenance: first, the incident is
detected by the NOC (Network Operation Centre), then the incident is analyzed by the operator's fiber
optic maintenance supervisor or transmission back office. Then the fiber optic maintenance team is
activated to deal with the incident. Next, the team in charge of the incident launches remedial
maintenance operations on the optical fiber. Finally, the team measures performance after repairs and
draws up an incident report. This work firstly enables us to understand optical fiber maintenance. Then
to propose an automatic system based on AI. This automatic, real-time system will improve the
efficiency and productivity of incident management by enabling the operator to use artificial
intelligence (AI) applied to image recognition to identify and notify the appropriate team of incidents
such as a cut in the optical fiber. At the same time, such a system is also very useful for the telecoms
regulator, who can use images to check the quality of fiber optic installations and compliance with
standards by operators and fiber optic access providers.
Keywords: Fiber Optics, Incidents, AI, Image Recognition, Management, Transmission, Telecoms.
Page 109
SEMIT2023
Business Process Linguistic Modelling- Design- Process Design Part I

Jozef Stašák
Peter Schmidt
Jaroslav Kultán
Abstract
Abstract At present, the business process management (BPM) seems to be a matter of principal
importance within any firm, company and institution, while there is a need to know the business
process (BP) structure and functionality and to quantify and to model them, in order to provide BPM
correctly and efficiently. There are two business process (BP) modelling approaches as well, while the
first is based on ARIS methodology initiated by prof. Scheer and developed by Software AG and it
seems to be the standard for BP modelling applied round the world. The shareholders and managers
like describing BPs via text in natural language (TNL Text), which consists of logical sentences V(i)
and any logical sentence contains a set of appropriate linguistic and lexical units. The question is: “How
to quantify them and how they should be quantified and applied for BP modelling purposes?” There are
many approaches and methods which enable converting the TNL text to BP models and one of them is
based on so called linguistic set existence and a co-operation among them is regulated by Principal
Business Process Linguistic Equation (PBPL Equation), which creates basis for BP linguistic modelling
related to their structure, functionality, and metrics, as well as their information, knowledge based and
organization support. On the other hand, the PBPL Equation create basis for design and implementation
of the knowledge-based (expert) system application, which seems to be a suitable aid for BP managers
as well.
Keywords: Business Process Modelling, Standardized Approach, Linguistic Approach, TNL Text,
Linguistic Sets, PBPL Equation, Expert System, Aid for Managers.
Business Process Linguistic Modelling - Design – Process Design Part II

Jozef Stašák
Intercedu a.s. Pezinok, SR
Peter Schmidt
Faculty of Informatics, University of Economy, Bratislava
Jaroslav Kultán
Faculty of Informatics, University of Economy, Bratislava
Abstract
At present, the business process management (BPM) seems to be a matter of principal importance
within any firm, company, and institution, while there is a need to know the business process (BP)
structure and functionality and to quantify and to model them, in order to provide BPM correctly and
efficiently. Those aspects seem to be a subject of the previous article, which deals with business process
modelling (BPLM), while that contribution deals with problems related to BP information, knowledge-
based and organizational support, which should enable BP efficient functionality and performance,
while linguistic approach based on linguistic sets and Principal Business Process Linguistic Equation
(PBPL Equation) is applied for those purposes as well. The BP information support modeling is based
on existing of so-called reference databases (RDBs) together with common and SAP databases, while
BP information support model seems to be a relatively individual subsystem, which consists of (a)
Master file subsystem and (b) Reference database (RDB) and (c) Communication subsystem, which
Page 110
SEMIT2023
provides interconnection among master files, RDBs and adequate SAP databases, however there is a
component, which provides interconnection among sematic networks within appropriate knowledge
base operating within BP knowledge-based support and organizational support as well, while the BP
knowledge-based and organizational support algorithms are based on PBPL Equation semiquantitative
solution too.
Keywords: Business Process Modelling, Information Support, Knowledge-Based Support,
Organizational Support, Modelling Algorithms PBPL Equation Semiquantitative Solution.
Page 111
SEMIT2023
SP07: Supply Chain Network Design; Considering Recent Trends in Sustainability

and Circular Economy
Carbon Footprint Calculation and Optimization Using ISO 14067: Development of a Product
Carbon Footprint Calculator
Yasmine Benomar
Laboratory System control and decision, ENSIT engineering school, Tangier, Morocco
Sanae Larioui
Abstract
This study presents a comprehensive research endeavor focused on calculating and optimizing the
carbon footprint of products using the internationally recognized ISO 14067 standard. A cutting-edge
product carbon footprint calculator was developed to facilitate streamlined and precise assessments.
The scope of analysis encompassed the cradle-to-gate phase, spanning from raw material extraction to
product manufacturing at the facility.
The product carbon footprint calculator allows for the accurate determination of a product's emissions,
providing results with dashboards for every phase (raw material manufacturing, transportation, and
product manufacturing). Furthermore, the study identifies opportunities for reducing carbon emissions
across various departments. It proposes strategic measures to minimize the impact while promoting
cost-effective, eco-conscious practices.
The implementation of solar panels was a successful achievement in reducing the overall carbon
footprint to meet energy requirements. This adoption resulted in an astonishing 72.78% optimization in
energy consumption carbon footprint and associated cost reduction, demonstrating the potential of
renewable energy solutions in reducing the environmental impact.
The research introduces a new strategy to optimize logistics, costs, and resource utilization through the
adoption of returnable packaging. Enhancing product packaging efficiency led to a substantial reduction
in waste generation and facilitated sustainable resource management. Moreover, this study offers a
global approach to address the critical challenge of environmental sustainability throughout a product's
lifecycle. The implementation of ISO 14067-based carbon footprint calculations and the product carbon
footprint calculator provide a strong foundation for eco-conscious practices within industries.
This research on carbon footprint calculation and optimization using ISO 14067 highlights the
importance of embracing sustainable practices in product development and manufacturing. The
proposed strategies for emissions reduction, renewable energy adoption, and returnable packaging
introduce a new era of product development. By utilizing the product carbon footprint calculator,
industries can accurately assess and reduce their environmental impact, paving the way for a greener
and more sustainable future.
Keywords: Carbon Footprint, ISO 14067, Product Carbon Footprint Calculator, Eco-Conscious
Practices, Organizational Sustainability.
Page 112
SEMIT2023
Adoption of the Blockchain in the Supply Chain: Highlighting the Factors of Failure
Issam Najati
HECF BUSINESS school
Abstract
The supply chain is inherently complex, involving many stakeholders, complex networks and
multifaceted processes (Mentzer et al., 2001). Applying blockchain technology in this context requires
careful consideration of various factors that can influence the success or failure of implementation
initiatives (Dasaklis et al., 2022; Saberi et al., 2019). Previous studies have addressed some challenges
faced in blockchain logistics projects, but a comprehensive analysis of success and failure factors is still
lacking (Bai et al., 2020; Dasaklis et al., 2022; Fosso Wamba et al., 2020; Hald & Kinra, 2019; Helo &
Hao, 2019; Katsikouli et al., 2021; Sternberg et al., 2021; Tiscini et al., 2020). This exploratory study
aims to provide an analysis of the failure factors associated with the application of blockchain in supply
chains, through in-depth interviews with experts, project managers and technology providers, also
taking advantage of strategic alignment theory (Henderson & Venkatraman, 1993). The results reveal
important failure factors, including a mismatch between technology, business and organizational
strategies; inadequate governance structures; insufficient technological capacities; and limited
collaboration among stakeholders.
Keywords: Supply Chain, Blockchain, Strategic Alignment, Failure Factors, Governance.
Medium Voltage Distribution Cell Customer Orders Planning Model with Capacity Plan-Ning
Methods
Servet Soygüder
Ankara Yıldırım Beyazıt Üniversitesi, Mühendislik ve Doğa Bilimleri Fakültesi, Ankara, Türkiye
Abdulhamid Kayal
Ankara Yıldırım Beyazıt Üniversitesi, Fen Bilimleri Enstitüsü, Mühendislik ve Teknoloji Yönetimi
Anabilim Dalı, Ankara, Türkiye
Abstract
In a company serves electrical sector, where multi mass production lines and various product types are
being manufactured, achieving proper capacity planning for customer order engineering team is
targeted in order to enable the company respond customers on time efficiently and in a productive way.
Data and information of previous two years are used besides the input from regional sales team and
organization to foreseen the upcoming sales volume in 2023. As the company is using SIOP (Sales,
Inventory and Operation Planning) mechanism both SIOP and Rough Cut Capacity Planning method
are utilized to figure out proper tool plays vital role to trigger customer order engineering team capacity
sizing and planning on time.
Keywords: Capacity Planning, Supply Chain Management, Inventory and Operations Planning,
Integrated Business Planning.
Page 113
SEMIT2023
A Location Inventory Model under Vendor Managed Inventory Contract

Zahra Payvandi
Department of Industrial Engineering, Faculty of Engineering, Alzahra University, Tehran, Iran
Mehdi Seifbarghy
Department of Industrial Engineering, Faculty of Engineering, Alzahra university, Tehran, Iran
Abstract
Vendor managed inventory (VMI) is an integrated approach that allows vendor orders on behalf of
retailers under a contractual agreement. Failure to comply with the upper stock limit will result in
levying penalty cost for the vendor which is agreed upon in a contractual agreement between vendor
and retailers. It also allows unequal shipment frequencies for the retailers. The purpose of this paper is
to develop a supply chain partnership strategy with single vendor, multiple retailers and multiple
distribution centers under VMI mode. We assume that, although the location of the vendor and the
retailers are known, that of the distribution centers are undetermined so the problem is to determine the
location of the distribution centers and decide on the best method for assigning retailers to DCs in order
to reduce the total cost. The model is formulated as a mixed integer non-linear problem that reduces the
joint relevant inventory cost. Also, we show that developed optimization problem can be solved
heuristically. In addition, we developed a numerical example to indicate the efficacy of the proposed
model. The example is solved by the usage of PYOMO, Python-based open-source optimization
modeling language. Furthermore, we use Bonmin, an open-source code for solving general mixed
integer non-linear problem, as the solver. Results indicate that if variability in retailer’s demand and
cost parameters see an increase, the heuristic procedure makes greater cost savings.
Keywords: Vendor Managed Inventory, Location Inventory Model, Heuristic Procedure, PYOMO.
Prioritization of Freight Carrier Selection Factors with the Best Worst Method in Multi-
Criteria Decision Analysis
Selin Yalçın
Ertuğrul Ayyıldız
Abstract
Efficient freight carrier selection is pivotal to modern logistics and supply chain management,
demanding a systematic approach to navigate the complexity of factors such as cost, reliability,
sustainability, and collaboration. This study presents a comprehensive investigation into the
prioritization of these factors using the Best Worst Method within the framework of Multi-Criteria
Decision Analysis (MCDA). Drawing insights from literature, industry expertise, and stakeholder
perspectives, a criteria set is identified. These encompass economic considerations, service quality
metrics, environmental impact, collaborative capabilities, and delivery timeliness. The holistic
perspective captures the multifaceted nature of contemporary carrier selection decisions. This study
employs the Best Worst Method, a one of most effective MCDA approach, to establish a structured
evaluative framework. Expert judgments and pairwise comparisons aid in the identification of both
"best" and "worst" criteria for each factor, enabling the assignment of priority weights. This research
has profound implications for logistics practitioners, supply chain managers, and decision architects
engaged in carrier selection processes. Prioritization shows a clear path for well-informed decision-
making aligned with strategic goals and sustainability objectives. Moreover, the study enriches the
Page 114
SEMIT2023
theoretical understanding of carrier selection within a multifaceted criteria context. This study offers a
novel framework that elevates decision-making precision in carrier selection. It contributes to
operational efficiency, sustainability, and collaborative dynamics within contemporary logistics
paradigms.
Page 115
SEMIT2023
SP08: Optimization, Decision Making and Support Systems
Optimization Models for Production Planning in the Agro-Industrial Sector: A Systematic

Review of the Literature
Angie del Rocio Sanchez Marquina
Faculty of Mathematical Sciences, National University of San Marcos, Lima, Peru
Abstract
The constant growth of the agro-industrial sector has generated various challenges for companies, such
as high costs, low productivity, reduced profits, loss of customers, and insufficient production levels.
In response to these is-sues, companies have turned to the use of optimization models in their pro-
duction planning. In this context, the objective of this research is to identify the approaches used for
applying optimization models in the production planning of the agro-industrial sector. To achieve this
goal, a systematic review of scientific publications available in online databases was conducted,
considering a period of the last decade. The results of this review reveal that the Mixed Integer Linear
Programming and Linear Programming models are the most widely used in the agro-industrial sector
for production planning. The findings also indicate the versatility of these models as they can be adapted
to different product quantities, planning horizons, and uncertainty in production. It is worth noting that
cost optimization is a common objective in production planning, but few studies were found to
incorporate uncertainty into the model. This suggests a need for further research that delves deeper into
the uncertainty commonly present in the agro-industrial sector.
Keywords: Production Planning, Optimization Model, Mathematical Optimization, Agro-industrial,
Systematic Review.
Space-Filling Curve Based Global Optimization Algorithm via Auxiliary Function

Nurullah Yilmaz
Süleyman Demirel University, Isparta, Türkiye
Abstract
In this study, we deal with the unconstrained global minimization of multi-modal, continuously
differentiable functions. We develop a new global optimization method based on the reducing
dimension technique by using space-filling curves and auxiliary function approach. First, smooth space-
filling curves are considered in order to reduce the dimension of the problem. Second, a new
continuously differentiable auxiliary function is proposed and the theory about this new auxiliary
function is presented. Third, a new algorithm is developed depend on auxiliary function, space-filling
curves and local searches. Then, the numerical stabilization of the proposed algorithm is investigated
on some test problems. Moreover, the effectiveness of the proposed algorithm on test problems is
demonstrated by comparing our numerical results with the results obtained by some recently offered
methods and algorithms.
Keywords: Auxiliary Function, Space Filling Curves, Global Optimization.
Page 116
SEMIT2023
Employee -Engagement Level as a Predictor of Organizational Performance :A Study of

Information Technology Sector in Telangana, India
Asma Bano
Mittal School of Business, Lovely Professional University, Punjab, India
Dr. Ayesha Khatun
Symbiosis Law School, Nagpur, Symbiosis International University
Dr. Dinesh Kumar
Mittal School of Business, Lovely Professional University, Punjab, India
ABSTRACT
Employee performance, dedication, competitive advantage, etc. all stem in large part from the level of
engagement present inside a business. The current study is under taken with an objective to determine
the relationship engagement level and emotional, financial, and non-financial components of
Organizational Performance (OP) among Telangana Information Technology Employee and to
determine the elements that contribute to Employee Engagement (EE). It is used to find whether there
is any connection among three dimensions of organizational performance and the level of involvement
in one's job and one's organization. In light of the study's primary emphasis on the viewpoint of the
employee engagement, only minor adjustments to the engagement and performance questionnaires were
made in consultation with IT Company workforce like managers, assistant managers, administrators,
team leaders and employee. While this study demonstrates that employee engagement is favourably
related to emotional, financial, and nonfinancial components of organizational performance, it
contradicts previous research by finding that engagement is not substantially related to the nonfinancial
component of organizational performance. Employees' levels of involvement and dedication did not
vary greatly across the different departments, between employees of different types of working hours
and schedules, between employees with different designations, and between employees with different
degrees of experience, as revealed by this study.
Keywords: Engagement, Telangana Information Employees, Performance, Working Hours, Roles and
Responsibilities.
A Reinforcement Learning-Based Fault Tolerant Control Design Approach using Double Q-

Learning Algorithm
Seyed Ali Hosseini
Department of Automation and Instrumentation Engineering, Petroleum University of Technology,
Ahwaz, Iran
Karim Salahshoor
Department of Automation and Instrumentation Engineering, Petroleum University of Technology,
Ahwaz, Iran
Abstract
This paper presents a novel approach, the Reinforcement Learning Based Fault-Tolerant Control (RL-
FTC) strategy, designed to address sensor faults in control systems. Initially, a closed-loop system is
established using the Conventional Proportional-Integral-Derivative (PID) controller. To investigate the
system's robustness, a simulated sensor fault is intentionally introduced, and the corresponding results
are analyzed. In addition to the RL-FTC strategy, we propose the utilization of the Double Q-Learning
(DQL) algorithm to tackle unknown faults in the control system. This enhanced approach aims to
improve the performance of the controller under challenging conditions, ensuring reliable operation
Page 117
SEMIT2023
even in the presence of unexpected faults. To evaluate the effectiveness of both strategies, a
comprehensive study is conducted on a Continuous Stirred Tank Reactor (CSTR) system, a
representative benchmark for control applications. Through extensive experimentation, the results
demonstrate the superiority of the proposed RL-FTC strategy compared to the Conventional Q-Learning
(CQL) algorithm, particularly in critical and sensitive systems like the CSTR. The key contribution of
this research lies in the ability of the RL-FTC approach to enhance the speed of decision-making, an
essential aspect in managing time-sensitive systems. By effectively addressing sensor faults and
improving control performance, this method proves to be a promising solution for real-world
applications in fault-tolerant control systems. By combining theoretical analysis, simulation studies,
and practical experiments on the CSTR system, this paper offers valuable insights into the development
of robust and reliable control strategies, opening avenues for further research in the field of
reinforcement learning-based fault-tolerant control.
Keywords: Fault-Tolerant Control (FTC), Reinforcement Learning (RL), Conventional Q-Learning
algorithm, Double Q-Learning algorithm, Adaptive Control, Conventional PID-Controller.
Comparing OECD Countries’ Logistics Performances by Comprehensive Distance Based

Ranking Method
Sedef Çalı
Dokuz Eylül University, Graduate School of Natural and Applied Sciences, Department of Industrial
Engineering, Izmir, Türkiye
Adil Baykasoğlu
Dokuz Eylül University, Faculty of Engineering, Department of Industrial Engineering, Izmir,
Türkiye
Abstract
Logistics are of crucial importance for countries to engage in international trade and build their
economies. In 2007, the World Bank (WB) developed the logistic performance index (LPI) to assess
how well countries perform in terms of logistics. The LPI score is obtained by averaging the scores on
the six components that have been weighted equally. The core components of LPI are customs,
infrastructure, services, timeliness, tracking and tracing, and international shipments. However, it is not
realistic to consider that these components are of equal importance. Rather than taking the average of
the component scores, more sophisticated techniques should be performed to acquire the overall logistic
performance scores. Considering these shortcomings, this study aims to propose a new multi-criteria
decision-making (MCDM) analysis to evaluate the logistics performances of OECD countries by using
LPI data published in 2023. The proposed methodology determines the weights of LPI indicators thanks
to the criteria importance through intercriteria correlation (CRITIC) method, which is an objective
weighting approach. The rankings of the OECD countries are obtained by using the comprehensive
distance-based ranking (COBRA) method, which is a relatively new distance-based MCDM method.
The results show that the international shipment score is the most important indicator of LPI. The
ranking indicates that Finland is the best OECD country in logistics performance.
Keywords: MCDM, Logistics Performance Index, COBRA, CRITIC.
Page 118
SEMIT2023
SP09: Quality, Productivity, and Project Management
A Systematic Analysis and Classification of the Effects of Human Fault in Unmanned Aviation
İbrahim Akyürek
Ankara Yıldırım Beyazıt University, Mühendislik ve Doğa Bilimleri Fakültesi, Endüstri Mühendisliği
Bölümü İş Sağlığı ve Güvenliği Doktora Programı Ankara, Turkey
Ergün Eraslan
Ankara Yıldırım Beyazıt Üniversitesi, Mühendislik ve Doğa Bilimleri Fakültesi, Endüstri
Mühendisliği Bölümü, Ankara, Turkey
Abstract
In this study, first of all, the published literature on the human factor in unmanned aerial vehicles was
reviewed. In the next step, a content analysis was made to classify the studies. Studies in the literature
were systematically grouped into main and subgroups. After the classification, the literature was
discussed and missing areas were revealed and it was tried to shed light on future research. Content
analysis was conducted on an article in which 69 studies were examined in depth. In this process, some
common issues that the researchers focused on were tried to be determined. In the classification grouped
in five main groups, the focus is on UAV Design and Ergonomics. Human factors have an effect on the
design of the user interfaces and screens of unmanned aerial vehicles. The design and ergonomic
features of the operator's ground control station (GCS), which may cause human error, were
investigated. The effects of user interface and screen designs and ergonomic structures on variables
such as workload, situation awareness and decision making were investigated. Focusing on the
development of optimum design and ergonomic structures that will reduce the workload, increase
situational awareness, positively affect perception and help operators make the right decision. In
addition, unmanned aerial vehicles are designed to eliminate the negative effects of the human factor.
As a result of the developments in the design and ergonomic structure, it is aimed to ensure that the
unmanned aerial vehicles are used more efficiently, effectively and safely by the operators. Ergonomic
problems and solution proposals have been studied in issues such as the screen angle of the ground
control stations, temperature, posture in front of the screen, working hours. The design of a ground
control station is examined and ergonomic factors are reviewed. Suggestions were made on factors such
as the angle of the screens, sitting position, thermal comfort.
Keywords: Unmanned Aerial Vehicle, Ground Control Station, Design, Ergonomics.
Page 119
SEMIT2023
The Quality Assessment of a Company in Defence Industry: A Case Study Approach

Vildan Tezcan
Ankara Yıldırım Beyazıt University
Harika Ergün
Enes Öztürk
Güler Zengin
Deniz Efendioğlu
Abstract
In order to address the issue of sub-component loss faced by a company in the defense industry, this
article explores SPC procedures. The missing components were found to have passed quality checks,
and records were retained. To address the issue, methods including root cause analysis, pareto analysis,
and the PDCA cycle are recommended. The study also contains information about the organizational
structure of the business and the workflow. The loss of time due to the problem is emphasized. What
kind of lean production techniques can be applied has been determined. The paper suggests employing
quality assurance methods to identify the best option.
Keywords: PDCA, SPC, Lean Production, 5 Why, Pareto Analysis.
Application of Best Project Management Practices: Case Study Application in Social Projects
Verónica Moreno Ponce de León
Universidad Privada Boliviana, Av. Capitán Victor Ustáriz, km. 6.5, Cochabamba, Bolivia
Agatha Clarice da Silva-Ovando
Abstract
This article pretends to highlight the benefits of applying the best project management projects,
presenting the experience of Fundación Educación para el Desarrollo FAUTAPO. This foundation
works with Bolivia's social projects and shows how those contribute to improving execution times,
achievement of objectives, and knowledge generation. The paper carries out a bibliographical and
documentary review and a survey coordination meeting with agents involved in the projects to
complement the information gathered from historical data. The study uses the descriptive method to
illustrate the experience of implementing best project management practices to a social project executed
over three years. Adopting the best project management practices in a social context improves objective
achievement and compliance with deadlines. Project communication management and project
stakeholder management are vital aspects of highly complex projects. The transference of knowledge
and lessons learned has proven essential to reduce the learning curve. There was no access to budget
information, so achieving objectives and deadlines were taken as analysis variables. The paper enriches
the documentation of experiences of best practices implementation in social projects to promote their
use. It contributes to the review and documentation of the experience gained through applying best
project management practices to a social project developed in a Latin American context.
Keywords: Project Management, Best Practices, Project Planning, Social Projects, Project
Accomplishment.
Page 120
SEMIT2023
Analysis of Investment Feasibility Using the Life Cycle Cost (LCC) Method in Implementing
Green Retrofits in High-Rise Office Buildings
Aviva Cantika Alfatihanti
Civil Engineering Department, Faculty of Engineering, Universitas Indonesia, Kampus UI Depok,
Indonesia
Yusuf Latief
Indonesia
Bernadette Detty Kussumardianadewi
Indonesia
Abstract
As one of the world's largest carbon gas emitters, construction activities contribute significantly to
global warming. Various countries in the world have joined forces to achieve the Net Zero Emission
target by 2050, and Indonesia is also contributing by implementing green building concepts in existing
buildings, including office buildings. There are challenges in implementing green building concepts in
existing buildings due to the lack of public understanding of the investment costs supported by the high
costs required. This research aims to help provide an understanding of investment costs using the life
cycle cost (LCC) method by providing information on the value of investment feasibility through the
application of green retrofits that refer to the new certification by the government. This research
involved five experts and more than ten respondents to provide opinions on qualitative assessment. At
the same time, the regulation of the Minister of Public Works and Public Housing No. 21 of 2001 was
used. Investment cost simulation was carried out for ten years to get the results of profit and benefits.
Keywords: Life Cycle Cost, Investment Feasibility, Green Retrofit, Regulation of the Minister of
Public Works and Public Housing No. 21 of 2021, Office Building.
Page 121
SEMIT2023
SP10: Other Fields of Study Related to EM and IT
The Impact of Technostress and COVID-19 Stress on Employee Burnout: The Mediating Role
of Resilience
Nour El Hoda Tarabah
Business Management, Ibn Haldun University
Ugur Sener
Business Management, Istanbul Aydin University
Abstract
In today’s rapidly evolving work landscape, employees are challenged with many stressors that can
harm their job performance and overall well-being. The COVID-19 pandemic has put employees under
so many stressors compounding existing stressors faced by employees in many sectors. Compliance
with COVID-19 interventions has also accelerated the use of technological advancements which has
also put employees under pressure to adapt to these technologies to thrive in their jobs especially since
organizations are heavily relying on technology to carry out their daily tasks. Thus, a modern disease
caused by technological advancements “Technostress” can also occur in employees. Such stressors
might lead to burnout if the necessary resources are not built and fostered. In this matter, resilience can
play an important role in mediating these stressors. Guided by the theoretical framework of the Job
Demands-Resources (JD-R) model, this study explores how job demands (technostress and COVID-19
stress) and job resources (resilience) influence the context of these diverse industries. It employs a
survey adopted from existing scales about each variable to gather data from white-collar workers in the
real estate, medical tourism, and education sectors in Istanbul, and mediation analysis was carried out.
COVID-19 Stress and Technostress had a direct impact on Burnout. However, resilience could only
mediate the relationship between COVID-19 Stress and Burnout. The findings of this study provide
theoretical and practical implications to foster a resilient workforce that can create a sustainable and
supportive work environment that can benefit employees and employers.
Keywords: Technostress, COVID-19 Stress; Employee Burnout; Resilience; Mediation Analysis.
Page 122
SEMIT2023
Comparative Study of Data Compression Algorithms: Zstandard, zlib & LZ4
Alysha P Maulidina
Computer Science Department, School of Computing and Creative Arts, Bina Nusantara University,
Jakarta, Indonesia
Rachel Anastasia Wijaya
Jakarta, Indonesia
Kimberly Mazel
Jakarta, Indonesia
Maria Seraphina Astriani
Jakarta, Indonesia
Abstract
Data compression continues to grow more important in the era of rapid communication and transfer of
data. Compres- sion algorithms allow this to happen, but many different options are available. This
comparative study investigates three lossless compression algorithms: Zstandard, zlib, and LZ4. This
study has a unique aim to provide insight into the performance of each algorithm to identify their
strengths and weaknesses in different applications. Previous studies have been done, however, this study
evaluates the latest modern options for lossless compression algorithms with a larger variety of data
types and measured benchmarks. The algorithms are tested with a corpus consisting of different files;
text-based, image, PDF, CSV, and JSON. The tests are done a set number of times and the average is
taken to increase reliability. These benchmarks used to evaluate the algorithms are the compression
ratio, size savings, compression speed, decompression speed, and respective space complexities.
Through these values, it was found that zlib had the highest compression ratios, despite its generally
slower compression speed. Zstandard’s compression ratios and size savings are slightly below zlib,
however, it has higher speeds. LZ4 has the worst compression ratios out of the three, but has the highest
speeds due to its low overhead, making it suitable for small microprocessors. Each algorithm has its
strengths and weaknesses, but zlib had the best overall performance in this study. Further research can
be done to improve the generalizability and reliability of the findings.
Keywords: Data Compression, Algorithm Analysis, Compression Algorithm, Zstandard, Zlib, LZ4.
Page 123
SEMIT2023
Empirical Analysis of Profit-Intended Vehicle Fleet Maintenance Schemes in Developing

B.O Ozero
University of Nigeria, Nsukka (Mechanical Engineering Department/Faculty of Engineering,
Nsukka, Nigeria)
Samuel Ogbonna Enibe
Nsukka, Nigeria)
Paul Amaechi Ako (formerly Ozor)
Department of Quality and Operations Management
Faculty of Engineering and Built Environment, University of Johannesburg, P. O. Box 534 Auckland
Park, South Africa
Abstract
Economic uncertainties, galloping inflation, and the overarching demand for customer safety as well as
enterprise profitability underscore the importance of proper analyses to evaluate the maintenance
decisions of service-producing systems. Whilst the maintenance planning of profit-intended equipment
can be done on the principles of minimum cost, appropriate models that strike a good balance between
reliability and cost can lead to the best maintenance policy for such systems. In this work, the existing
maintenance procedure for a typical commercial vehicle fleet Company in Nigeria, called X-tours, for
confidentiality is studied. The Company’s prevalent operation processes and control, maintenance
costs, the effect of climatic variations, and inflation were obtained through document analysis,
observation, and use of interviews with the opinion leaders. The resulting data provided the foundation
for an empirical investigation, formulation, and application of appropriate quantitative models that
minimize the operation costs as well as determine the useful life span of the vehicles. The results show
that a commercial vehicle, typical of the Toyota Hiace bus, operated by the studied company should be
replaced after about 5 years. The result also shows that within this period, the vehicle would have fully
recouped the capital invested in it.
Keywords: Empirical Analysis, Vehicle Fleet Maintenance, Minimum Cost, quantitative models,
Useful life.
Multiclass Arrhythmia Detection from 2D ECG of COVID-19 Patients

Trupti G. Thite; Sonal K. Jagtap
Abstract
Abstract: Electrocardiogram (ECG) is a graphical representation of the electrical activities occurring in
the heart. This graph is widely used in the medical field for monitoring purposes, allowing doctors to
diagnose arrhythmias. In many cases, cardiologists detect and diagnose arrhythmias by inspecting short-
time ECG recordings visually. An arrhythmia is an abnormal heart condition that is characterized by an
irregular heartbeat rate. The utilization of ECG trace images offers several benefits in automatic
diagnosis of the condition. The ECG trace image dataset can be easily obtained using smartphones,
making it readily available even in low-resource countries. According to the literature survey, there is
a lack of research on 2D image classification. Hence, our research focuses on the detection of multiclass
arrhythmias from 2D ECG signals. As the use of advanced convolutional neural networks allows for
the efficient processing of 2D datasets without extracting features manually, we propose the
Page 124
SEMIT2023
implementation of the deep convolutional neural network algorithm to achieve improved classification
accuracy in comparison to 1D signals.
Keywords: ECG, COVID -19, Multiclass Arrhythmia, Image, Deep CNN.
A New Category Based Collaborative Filtering Approach Using Trust Circles

Raouia Ayachi
Larodec laboratory ISG Tunis Tunisia, Tunisia
Abstract
Nowadays, the growth of data volume in the web has become more and more important. This
phenomenon is called information overload. A user, in this case, finds himself unable to select the right
choice over this large quantity of information that could be in his round of interest. That's why,
recommender systems were appeared to help users to select the right items that go with his preferences
at the right time. There are many approaches for making recommendation. Collaborative filtering, the
most popular approach that has been used in recommendation systems. This approach is based on the
historic of ratings of neighbours that seem similar to the target user and have the same attitude. Despite,
this approach makes good results but it suffers from some issues namely: data Sparsity, the cold start
problem and scalability.
In Online Social Networks, dealing with only similarities between users cannot provide a pertinent
recommendation result. Recent works have proposed models that incorporate trust values between users
into collaborative filtering recommendation in order to make an accurate and reliable recommendation.
These models have incorporated trust between users and consider that the user trusts the same set of
friends in different categories of items. In reality, there are users who are trustworthy in some categories
of items and not in others. For facing these problems cited above, we propose a memory-based
collaborative filtering approach based on trust circles of users. The new framework is composed of
three phases, namely: Construction of trust circles, Prediction process and Recommendation process.
In the first phase, we will make the inference from the general web of trust and build circles of trusted
friends that belong to a specific sub-category, then we compute trust values between users in the sub-
category. In the Prediction phase, we will exploit these trust values to refine the prediction and make it
more accurate and reliable. Then, we will make the recommendation based on the predicted values, and
we will recommend the top-k items which have the highest values. We choose Epinions dataset to
validate our proposed framework. Our approach outperforms state of the art methods.
Keywords: Recommendation System, Collaborative Filtering, Trust Circles, Categories of Items.
Page 125
SEMIT2023
SP11: Neural Networks: Analysis, Modelling and Numerical Computations and

Data Science
Advancing Anemia Diagnosis: Harnessing Machine Learning Methods for Accurate Detection
Pakize Sümeyye Söylemez
Ankara Yıldırım Beyazıt University, Ankara, Turkey
Hilal Arslan
Ankara Yıldırım Beyazıt University, Ankara, Turkey
Abstract
Anemia occurs when there are insufficient healthy red blood cells to transport adequate oxygen to the
organs and tissues in the body. Timely detection of anemia and appropriate treatment strategies are
crucial to avoid complications. Machine learning methods have gained popularity in diagnosing anemia,
particularly in scenarios with limited medical experts and laboratory resources. In this study, various
machine learning algorithms, including a k-nearest neighbor, support vector machines, decision trees,
adaptive boosting, and random forest, were employed to diagnose anemia patients and determine the
specific types of anemia. The dataset used for diagnosis comprised laboratory information from 5,553
anemia patients and 9,747 healthy individuals. Among the anemia patients, there were 4,182 cases of
iron deficiency-related anemia, 1,019 cases of hemoglobin insufficiency, 199 cases of B12 deficiency,
and 153 cases of folate deficiencies. A set of 24 characteristics, including B12 levels, blood cell counts,
iron levels, and gender, were utilized to determine the presence of anemia and determine the appropriate
treatment approach. It is worth noting that the study focused on the diagnosis of anemia, a condition
where there is a deficiency of healthy red blood cells. The decision tree algorithm proved to be
particularly effective in this context, outperforming other methods such as support vector machines
(96.3 % accuracy), k-nearest neighbor (87.7 % accuracy), and neural networks (98 % accuracy). By
leveraging the decision tree algorithm and optimizing its hyperparameters through grid search, this
study successfully achieved a remarkable accuracy of 100 in diagnosing anemia. This result is
particularly significant given the potential risks associated with misdiagnosing anemia, as it may lead
to complications such as hemochromatosis, which can result in organ damage or failure.
Keywords: Anemia, Decision Tree, Support Vector Machine, k-Nearest Neighbor, and Neural
Network.
Enhancement Optical Fiber Data Rate by Using 8 port DWDM OF 256-QAM-OFDM DP

Technique
Adnan Ali K. Al-Awadi
Department of Communication Engineering, Najaf Technical College Engineering, Al- Furat Al-
Awsat Technical University, Najaf, Iraq
Abstract
The primary objective of this study is to augment the data transmission rate within an optical fiber
system, effectively catering to the increasing requirements of network advancement, data center
accessibility, and the swift transmission and reception of high-speed data. To achieve elevated data
rates, we employ a combination of 256 QAM OFDM DP with DWDM. This strategic amalgamation
Page 126
SEMIT2023
permits a larger number of subcarriers and consequently, a heightened data rate. This advancement
enables us to surpass the limitations posed by existing systems, presenting a more effective solution
capable of satisfying the rising demands of both users and the network's data volume.
We propose and execute the simulation of a cutting-edge optical system, designed for exceptional
performance, encompassing high-capacity data transfer and minimal bit error rates. This simulation
employs Opti system simulation tools V19 and focuses on simulating DWDM utilizing the 256 QAM
OFDM DP modulation technique, implemented through MATLAB SIMULINK design. A series of
signal processing operations are performed to generate and manipulate signals within the system. The
system's design parameters, including the number of ports, subcarriers, FFT size, cyclic prefix ratio,
pilot carriers, data rate, fiber length, and sampling rate, are systematically configured. Random data and
OFDM modulation are generated for each port and polarization, and the resulting modulated signals are
visually represented on an Optical Spectrum Analyzer (OSA) to illustrate their wavelengths.
Following the modulation process, the signals undergo a simulated fiber propagation, fiber attenuation,
phase noise, and dispersion effects to the modulated signals. Subsequently, the received signals are
subjected to filtration using an FIR filter equipped with pre-established coefficients.
In conclusion, the system offers a comprehensive solution. The analysis of outcomes affirms that
adopting this approach results in a noteworthy augmentation of the system's capacity, resulting in a
throughput of 28.5 Tbsp., a substantial enhancement when compared to the available 2 Tbsp. systems.
Consequently, the system is well-equipped to support a substantial number of concurrent users and
effectively accommodate applications with demanding bandwidth requirements. The integration of DP
configuration fortifies the system's resilience in both transmission and reception, ameliorating the
impact of polarization-induced fading and elevating signal quality within challenging propagation
conditions. Furthermore, the incorporation of advanced interference mitigation techniques further
heightens the system's performance, particularly when dealing with co-channel and adjacent channel
interference.
Keyword: DWDM, DP, BER, 256 QAM, ISI, DCF, PILOT CARRIER, OFDM.
Review of Intrusion Detection Systems for SCADA: A Machine Learning Approach

Hermenegildo da C. Aberto
Institut de Recherche pour le Développement (IRD-Hann)
Jean Marie Dembele
Ecole Supérieure Polytechnique (ESP/UCAD)
Idy Diop
Alassane Bah
Abstract
Protecting networks and systems from unauthorized access and cyber threats has become increasingly
critical, and Intrusion Detection Systems (IDS) play a crucial role in their achievement. IDS is used in
Industrial Control Systems (ICS) like Supervisor Control and Data Acquisition (SCADA) to ensure the
safety of the operations of Critical Infrastructures such as power grids, water treatment, gas plants, etc.
This paper evaluates various IDS models designed for SCADA protocols such as DNP3, IEC 61850,
and Modbus. The aim is to provide a concise evaluation of the current situation in ML-based IDS and
propose afterward a suitable IDS model for Africa, specifically Mozambique and Senegal, where power
grid infrastructure is lacking. We explore the various IDS based on Ma-chine Learning (ML)
Page 127
SEMIT2023
techniques, including Decision Trees, Random Forest, k-Nearest Neighbors (k-NN), Support Vector
Machines (SVM), Deep Neural Networks, and others. The paper analyzes the performance of IDS using
ML and discusses the strengths, limitations, and challenges associated with the models. It concludes
that the reviewed IDS based on the ML models had a remarkable performance. It proposes future
directions to tackle the challenges and overcome the limitations of evaluating the models.
Keywords: Intrusion Detection Systems (IDS), Industrial Control Systems (ICS), SCADA, Critical
Infrastructures, DNP3, IEC 61850, Modbus, Decision Trees, Random Forest, k-Nearest Neighbors (k-
NN), Support Vector Machines (SVM), Deep Neural Networks.
A Hierarchical Formulation for Inverse Data Envelopment Analysis: Cost Efficiency Point of
View
Mohammad Moghaddas
Faculty of Basic Sciences, Department of Mathematics, Imam Hossein University, Tehran, Iran
Abstract
In this paper, for the first time, a new bi-level inverse data envelopment analysis (DEA) model is
introduced, which develops classical inverse DEA models from a cost efficiency standpoint. This bi-
level inverse DEA model provides a new perspective for managers dealing with the inverse DEA
problem and cost efficiency simultaneously in a hierarchical structure. Using the Karush-Kuhn-Tucker
(KKT) optimality conditions, the bi-level inverse DEA model is reduced to a one-level mathematical
program with complementarity constraints (MPCC). In the following, based on optimization theory and
some concepts of ordinary differential equations, a powerful neural network is built to solve this one-
level mathematical programming problem. The proposed model guarantees not only fixed technical
efficiency but also unchanged cost efficiency during the process of input estimation associated with a
perturbed output. Finally, a numerical example is described in detail to demonstrate the good
performance of the proposed method.
Keywords: Inverse Data Envelopment Analysis, Bi-Level Optimization, Cost Efficiency.
Image Processing in Toxicology: A Systematic Review

Gayatri Mirajkar
Electronics and Telecommunication Engineering, Arvind Gavali College of Engineering, Satara, India
Mukil Alagirisamy
Faculty of Engineering, Lincoln University College, Malaysia
Sagar Shinde
Department of Electronics and Telecommunication Engineering, Dr. D. Y. Patil Institute of
Technology, Pune, Maharashtra, India
Lalit Garg
Department of Computer Information Systems, Faculty of Information & Communication Technology
University of Malta, Msida- MSD 2080, Malta
Abstract
Cellular imaging has proven to be the key to identifying and application of biomarkers for tracking cell
fate and drug activity in vitro. The aim here is to improve the comprehension of the mechanism of
action of newly developed drugs. Automated microscopy and image processing have made high-
Page 128
SEMIT2023
throughput imaging studies, which quantify cellular morphological responses on a large scale, a
valuable tool in the early stages of drug development (Rezvani et al., 2022). Recently, multi-well assay
plates with high-throughput imaging systems have been used to take photomicrographs and document
cell morphology over time (Tandon et al., 2022). These systems provide numerous findings that must
be time-consuming and subjectively interpreted by humans (Tandon et al. 2022). Designing a step-by-
step approach for detecting toxicity in assay images requires the automatic detection and classification
of thousands of assay images. The steps involved in this process involve image acquisition with high-
throughput microscopy images followed by image processing techniques and analysis. Deep Learning
(DL) techniques have recently entered the scenario due to the wide range of drug-induced toxicity-
linked nuclei organization patterns (Jimenez-Carretero et al. 2018). We present a systematic review of
the different image analysis techniques for toxicity detection in cell assay images. In addition, the
review also focuses on the most prominent applications of DL to classify cell assay images as healthy,
or toxicity-affected.
Keywords: Deep Learning, Image Analysis, Convolutional Neural Networks, Assay Images, Toxicity
Affected.
Page 129
SEMIT2023

Future Perspective
Parametric Study of Shear Strength UHPC Tapered Beam

Aqeel Abdulridhah Jbarh
Department of Civil Engineering, University of Misan, Maysan, Iraq
Nasser Hakeem TU'MA
Abstract
This study introduces a verification of the structural confidence of the recently developed tapered beam
of new trends of tapered using a numerical analysis by ABAQUS program. The comparisons of
numerical results with the existing of previous experimental results, in scope of ultimate shear strength,
load-displacement relationships and other structural aspects. They showed a very good agreement
between than converges. Proposed rates ranged from 0.96 to 1.13. Parametric study that related to the
developed Ultra High Performance Concrete beams, is included with many structural parameters such
that crown length, simulations of the tapered beam by multiple semi-scale factor, steel reinforcement
strength, provided steel ratio (longitudinal and transversal). It noted that raising up the longitudinal
reinforcement ratio induces an increase in the yield and ultimate strength of beams. Shear testing on
steel concrete beams with Steel fiber ratio 2% by volume yielded positive findings in terms of the
influence of the shear reinforcement ratio on shear strength. The greater the transverse reinforcement
ratio, the more cracks occurred with a smaller openings width was noticed. Also, the shear cracking
under greater load increase when the shear reinforcement was increased. When the scale factor 2.4 was
increased, an increase in shear strength was observed at 348%. The crown's length is increased from 0
to 200 mm, the shear strength is increased by 10%.
Keywords: Shear Strength, Scale Factor, Reinforcement Ratio, Ultra High-Performance Concrete,
Crown Length.
The Present and Future Scope of Virtual Reality in the Construction Industry
Er. Aditya Singh
Amrita Vishwa Vidyapeetham, Coimbatore Campus, India
Abstract
Virtual Reality is developing and spreading all over the world at a rapid rate, its applications are being
increasing with time. In the recent years, the presence of virtual reality can be observed in the
construction industry as well. With the technological advancements in the recent decades, the
traditional way of construction is being transformed slowly and steadily, due to which the construction
industry is no longer separate from advanced technologies like virtual reality. In this paper, the focus
on virtual reality was given and its significance in the construction sector at present. Further, the
potential of virtual reality to change as well as transform the construction industry in the future, is
covered in the paper. Some research gaps in the current studies in addition to graphical analysis based
on the obtainable data from different sources were performed to support the study.
Page 130
SEMIT2023
Keywords: Virtual Reality, Construction 4.0, Quality Control, Construction Safety, Sustainability.
Mixed Reality and Its Future Scope in the Construction Sector

Er. Aditya Singh
Amrita Vishwa Vidyapeetham, Coimbatore, Tamil Nadu, India
Abstract
Mixed Reality is developing and spreading all over the world at a rapid rate, its applications are being
increasing with time. In the recent years, the presence of mixed reality can be observed in the
construction industry as well. With the technological advancements in the recent decades, the traditional
way of construction is being transformed slowly and steadily, due to which the construction industry is
no longer separate from advanced technologies like mixed reality. In this paper, the focus on mixed
reality was given and its significance in the construction sector at present. Further, the potential of
mixed reality to change as well as transform the construction industry in the future, is covered in the
paper. Some research gaps in the current studies in addition to graphical analysis based on the obtainable
data from different sources were performed to support the study.
Keywords: Mixed Reality, Construction 4.0, Quality Control, Construction Safety, Sustainability.
Blockchain-Driven Smart Contracts: An Overview of Application Areas and Gap Identification

in Construction Management Literature
Emmanuel Chidiebere Eze
Department of Architecture and Built Environment, Northumbria University Newcastle, Newcastle-
upon-Tyne
Ernest Effah Ameyaw
upon-Tyne
Bahriye Ilhan Jones
upon-Tyne
Abstract
The dynamics in industry 4.0 have continued to transform construction contract execution. Blockchain
and smart contract technologies represent the future of construction contracts delivery in the built
environment, although, they are still in the infancy in the construction sector. While studies on the
potentials of Blockchain-based smart contracts in construction have continued to increase as evident in
the extant literature, there is yet to be a consensus on the most crucial application areas in construction.
This emerging technology requires more attention; therefore, this study presents the result of a critical
review of blockchain-enabled smart contracts studies in construction. Forty documents were obtained
from the Scopus and Web of Science databases, and they formed the basis for the review and analysis.
The top application areas of smart contract technology were identified and discussed. Furthermore, the
application areas that require to be explored are refurbishment contracts, consultancy contracts, long-
term PPP projects, circular economy, and waste management. The adoption of blockchain-based smart
contracts in these areas will improve sustainable construction targets and enhance the diffusion of
blockchain-enabled smart contracts among key built environment experts.
Page 131
SEMIT2023
Keywords: Blockchain; Construction Contract; Critical Review; Smart Contract; Construction

Projects.
Identification of Green Retrofitting Procurement and Permitting Processes in High-rise Office

Buildings in Jakarta Based on PerMen PUPR No.21 Year 2021 and GBC Indonesia that Affects
Project Time Performance
Benedict Mario Gilbert Fernandes Sihaloho
Civil Engineering Department, Faculty of Engineering, Universitas Indonesia, Depok Indonesia
Yusuf Latief
Bernadette Detty Kussumardianadewi
Abstract
The concept of green buildings or green buildings is highly needed in the present era, considering the
drastic decline in global climate conditions. However, its development is greatly hindered because the
implementation of green building concepts is primarily focused on new constructions, while almost
two-thirds of the world's buildings are already built. Therefore, the objective of this research is to
enhance the Performance of Green Retrofitting Implementation Time, which is expected to accelerate
the growth of existing green buildings in Indonesia by identifying the Procurement and Permitting
Processes of Green Retrofitting in High-Rise Office Buildings in Jakarta based on PERMEN PU NO
21 TAHUN 2021 and Green Building Council Indonesia (GBCI) through work breakdown structure
(WBS) along with their high-risk activities and improvement strategies using a preventive approach
that Influences the Performance of Green Retrofitting implementation time. This research analysis
involves 15 respondents' data on risks in the procurement and permitting processes, which will be
processed using a qualitative risk approach. Based on the results of literature studies, expert validation,
and respondent questionnaires, a total of 83 activities in the procurement and permitting processes of
green retrofitting were identified, 214 risk indicators that affect the performance of green retrofitting
schedule, along with 56 prevention strategies for the high-risk situation, aiming to accelerate the growth
of existing green buildings in Indonesia.
Keywords: Green Retrofitting, High Rise Office Building, WBS, Procurement, Permitting, Schedule
Performance.
Performance Investigation of Heat Sinks with Different Perforations Diameters

Rana A. J. AL-Luhaibi
Mechanical Department/ Engineering College /Tikrit University/ Tikrit, Iraq
Ibrahim T. Nazzal
Mechanical Department/ Engineering College/Tikrit University /Tikrit, Iraq
Abstract
In this paper, the perforated heat sink was studied and compared with the non-perforated heat sink
experimentally. Four rectangular heat sinks were investigated to present this comparison; all heat sinks
used in this study have the same dimensions. Three of them are perforated and the fourth heat sink is
solid. The perforated heat sinks have 12 circular holes in each fin distributed in two rows (each row has
six holes) with different holes’ diameter: first heat sink 3 mm the diameter (PHS-Ø3mm), the second
Page 132
SEMIT2023
heat sink is 3.5 mm called (PHS-Ø3.5mm), and the third one its hole diameter is 4 mm (PHS-Ø4mm).
The experiment was carried out using forced convection device in the heat transfer laboratory. After
obtaining, analyzing and discussing the results, it was found that the perforated heat sink is better at
dissipating heat than the non-perforated one. Also, the PHS-Ø3mm has the highest heat transfer
enhancement (14.21%).
For the weight reduction, this paper proved that perforating the fins reduces the weight of the heat sink,
and the larger the perforation diameter, reduce more mass of heat sink. The highest reduction achieved
by PHS-Ø4mm.
Key words: Forced Convection, Heat Sink, Perforated Heat Sink, Perforation Diameter.
Page 133
SEMIT2023
SP13: Advancing Innovation and Collaboration in Science, Engineering, and

Information Technology
Computational Analysis Wearable Glove Produced with FDM Technology

Dilşad AKGÜMÜŞ GÖK
Istanbul Aydin University, Department of Mechanical Engineering, Istanbul, Turkey
Büşranur GÜDAR
Istanbul Aydin University, Department of Mechanical Engineering, Istanbul, Turkey
Abstract
Additive manufacturing is an emerging technology that offers the advantage of minimal wastage and
rapid prototyping, making it ideal for customized designs suitable for a wide range of applications.
Among the various types of additive manufacturing, Three-dimensional (3D) printing is particularly
prominent, and extensive research has been done on using different polymers for 3D printing
applications. Flexible manipulators are a category of robotic systems susceptible to parasitic elastic
effects. These versatile robotic systems find applications in healthcare, manufacturing, and aerospace
industries. They are commonly used in flexible skeleton systems, surgical robots, gloves, and wearable
mechanisms. The pre-sented research focuses on the comprehensive analysis of the body, considering
the integration of mechanical and electronic resources as communicative systems. This unique
perspective allows the body to be conceptualized as a design project that encompasses both fashion
design and engineering principles. Consequently, wearable devices attached to the body demand
specific technical considerations. The main objective of this study is to develop a functional prototype
of a glove using 3D printing technology (Fused Deposition Modeling - FDM).
Keywords: Wearable glove, FDM Technology, Analysis.
Improve Channel Capacity of TDMA

Aqeela N. Abed
College of Medicine, University of al- Iraqia, Iraq
Abstract
This paper gives time division multiple access analyses (TDMA) procedures and their crossovers with
the multiplexing of time division (TDM) for execution of Worldwide Portable the satellite
communications. In the satellite communication frameworks, and particularly in satellite systems,
numerous clients are dynamic at the similar time. The synchronous communications issue between
numerous multipoint or single versatile disciple clients, in any case, may be unraveled by utilizing the
plans of multiple access technique. Subsequently the frameworks assets as the power of transmitting
and constrained of the bandwidth, to utilize the fitting of channels by a total cost and to make a
distinctive multiple access to these channels. This produces an issue of partition and summation for
signals within gathering parts and the transmission, separately. By choosing the current issue comprises
within this improvement of transmission in arrange to isolate signals from various clients unmistakably
on the gathering portion of orthogonal channels. The present work is appropriated to study and calculate
the system capacity of TDMA/FDMA for multiple beam satellites by performing analysis among
multiple accesses. The system capacity could be increased by using the time division multiple access
Page 134
SEMIT2023
(TDMA) techniques by various different variables which have affected the channel capacity. The
number of mobile users has been increasing day by day with no or little increase in RF spectrum
allocation, TDMA isn't bandwidth limited. The system of digital capacity TDMA is marked in the
present work. The improvement of capacity would be in order tomes from 5-10 of FM analog in the
absence of adding any cell sites. For instance, the TIA standard in North America displays about 50
Erlang/km2 with a distance of 3 Km between site to site. and then, the TDMA standard allowable faster
handoff mechanism. Therefore, can conclude from the present work that the digital air interface of
cellular TDMA which used to build networks of individual communications. Thereby, the technique of
TDMA is the key to providing effectively the methods of channel allocation and handover.
Keywords: TDMA, capacity, Multiple Access, Channel Capacity.
Seismic Design of Underground Structures’ Vulnerability: A Review

Monkam Ngameni Huguette Maeva
Anhui University of Science and Technology, China
Haibing Cai
School of Civil Engineering and Architecture
Abstract
This paper reviews underground structures' seismic susceptibility, concentrating on design and
mitigation. Modern infrastructure requires underground structures, including tunnels, subway stations,
and storage facilities. Their underground placement makes seismic design and earthquake performance
difficult. Ground motion characterization, soil-structure interaction, structural response, and failure
mechanisms particular to underground structures are covered in this review of seismic susceptibility.
Geotechnical conditions, structural types, and construction methods affect their vulnerability, according
to the paper. Seismic stability design and analysis measure underground structures' dynamic response
to seismic loads. The objective is to avoid seismic forces from collapsing structures. Seismic stability
design includes advanced analytical and numerical methods like finite element analysis (FEA), discrete
element modeling (DEM), and boundary element method (BEM). It also examines modern seismic
vulnerability assessment methods and international design standards. This review also examines
seismic retrofitting and design innovations for underground structures. This paper synthesizes and
analyzes existing literature to understand underground structures' seismic vulnerability and encourage
seismic engineering research in this vital field.
Keywords: Stability Design: Infrastructure, Earthquake, Modeling, Engineering.
Mesh Technology as an Alternative Connectivity Model for the Distribution of Packets in a

WLAN Network: A Case Study in the Hospital of Latacunga, Ecuador
Juan Carlos Chancusig Chisag
Abstract
Currently, there is an incorrect use for some technologies that improve network traffic, which can affect
both the efficiency and packet transmission. The objective of this research is to optimize the packet
distribution of a WLAN network through the use of the top-down and McCabe methodologies. The
focus of mesh technology implementation is to establish an interconnected mesh among nodes with the
switch, performing IP configuration on each of them. The results of using mesh technology based on
Page 135
SEMIT2023
the top-down and McCabe methodologies in WLAN packet distribution led to a 32% of decrease in
packet delivery time, a 42% of reduction in the percentage of lost packets, a 42% of reduction in end-
to-end packet delay time, and finally, a 61% of increase in effective packet transfer, achieving the
purpose of the research.
Keywords: Wireless Network, Mesh Technology, Connectivity, Packet Distribution.
Cultivating Synergistic Global Leadership: Unifying Scientific Acumen for Effective Global
Problem Solving
Sameera Fernandes
Garden City Unviersity
Sheeja M.S.
Garden City Unviersity
Swetha Parivara
Garden City University
Abstract
In a rapidly evolving global landscape, leadership capable of addressing multifaceted challenges with
innovative solutions is imperative. This research paper explores the integration of scientific knowledge
with global leadership competencies to foster effective problem solving on a global scale. By analyzing
case studies, existing literature, and insights from successful leaders, this study illuminates the strategic
convergence of scientific acumen and leadership skills. The paper emphasizes the benefits of informed
decision-making, interdisciplinary collaboration, and ethical considerations derived from scientific
principles. It concludes by underscoring the pivotal role of cultivating leaders who can seamlessly
merge scientific insights with leadership prowess to navigate complex global issues.
Keywords: Global Leadership Competencies Scientific Integration Synergistic Leadership Effective
Problem Solving Multidisciplinary Leadership Global Challenges.
Page 136
SEMIT2023
SP14: Uncertain Decision Making
Uncertainty Quantification Code Written in MATLAB

Napat Suratemeekul
Department of Aeronautical Engineering, International Academy of Aviation Industry, King
Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand
Noppawit Kumkam
Suwin Sleesongsom
Abstract
This research presents examples of MATLAB code for a classroom of uncertainty quantification. The
codes are developed for educational purposes and newcomers or researchers. Demonstration codes are
presented in form of a general design conditions. Examples are chosen from mechanical components,
which are simply understood. The mechanical components are shafts, connecting rod, rear shaft
housing, and bolt etc. The quantification techniques are in the group of probabilistic such as Monte
Carlo Simulation (MCS) and Latin Hypercube Sampling (LHS). These examples can help the
newcomers to understand the difficulty in uncertainty quantify and extend to calculate reliability value
and probability of failure. Example computer codes are included in all parts.
Keywords: Uncertainty Quantify, Reliability, Monte Carlo Simulation, Latin Hypercube Sampling,
Education, MATLAB.
Stochastic Optimization Methodology for Production Planning with Uncertain Demand and
Lead Time Based on The Digital Twin
Dan Luo
IMT Atlantique, LS2N, France
Simon Thevenin
Alexandre Dolgui
Abstract
Nowadays, manufacturers are increasingly shortening production cycles and expanding product
assortments, which results in a volatile production environment characterized by customer demand
uncertainties, yields, production capacity, and lead times. To navigate these challenges effectively,
manufacturers require production planning tools capable of handling such uncertain parameters.
Relying on a single mathematical or simulation model with a sole objective is insufficient to offer
efficient and accurate decision support in complicated production scenarios. The advanced production
environment necessitates a comprehensive approach that considers various aspects and objectives. This
need becomes even more pronounced in the context of Industry 4.0, where research efforts increasingly
Page 137
SEMIT2023
focus on integrating advanced technologies to effectively address complex and uncertain production
environments. Therefore, we propose a novel stochastic optimization methodology for production
planning with uncertain demand and lead time, which combines simulation and Bayesian network
forecast methods. In contrast to the model for lot sizing with uncertain demand, we have integrated the
lot sizing model with the simulation model, the domain model, and the Bayesian network model to
predict the lead time and improve the lot sizing model with uncertain demand and lead time. The
production planning module provides various production plans as input for the simulation model, from
which we can determine accurate lead time and planning adherence. Furthermore, the feedback data
can be used to train the Bayesian Network model to predict the probability distribution of lead time. In
summary, our approach pioneers the handling of complex production environments that are
characterized by uncertainty. By combining stochastic optimization methodology with simulation and
Bayesian network methodologies, we present a reliable solution for manufacturers to optimize
production planning under uncertain demand and lead time factors.
Keywords: Stochastic Optimization, Machine Learning, Simulation, Uncertain Demand and Lead
Time.
An Integration of Fuzzy Logic and Risk Analysis in the Defense Industry

Özge Daşdemir
Industrial Engineering Department, Ankara Yıldırım Beyazıt University, Ankara, Türkiye
Ergün Eraslan
Industrial Engineering Department, Ankara Yıldırım Beyazıt University, Ankara, Türkiye
Abstract
In today's business environment, companies are increasing the importance they attach to risk assessment
studies in order to ensure occupational health and safety. Carrying out an effective risk analysis to
identify and prevent existing or potential hazards in businesses is of great importance. Especially in
sectors such as the Defense Industry, where strategic decisions are taken, protective and preventive
measures should be taken by making risk analyses. The objective of this study is to introduce a new
approach to conventional risk assessment by integrating fuzzy logic and the Fine-Kinney Method,
which is one of the risk analysis methods. The parameters used in the Fine-Kinney method were
evaluated within the framework of Fuzzy Logic and risk scores were found. Thanks to this integration,
uncertainties and risks were better understood, enabling more realistic and effective decisions to be
made. In addition, the methods were explained step by step and it was aimed to guide them so that they
can be used in similar problems. Through the integrated utilization of fuzzy logic and risk analysis, it
is possible to enhance the management of uncertainties and risks in the defense industry, leading to a
more flexible and effective decision-making process.
Keywords: Risk Analysis; Fuzzy Logic; Fine-Kinney Method; Multi-Criteria Decision-Making
(MCDM); Defense Industry.
Page 138
SEMIT2023
An Approach to Fuzzy Distributed Multi-Objective Linear Programming for Optimizing

Distribution Decisions in Relief Logistic Operations
Ibrahim Yilmaz
Department of Industrial Engineering, Ankara Yıldırım Beyazıt University (AYBU) Ankara, Turkey
Abstract
Humanitarian relief logistics (HRL) is generally thought to be one of the most important parts of a
disaster relief operation. Although there are many HRL network models, the majority of them do not
take uncertainty into account. However, after big disasters, the level of uncertainty makes it hard to
define and describe practical situations with precision. In this study, a fuzzy mixed integer multi-
objective (DFMM) mathematical model with a distributed decision-making strategy is proposed as a
way to improve how relief logistic operations decide how to distribute things. The proposed model
includes both forward and reverse flows, with multiple relief items in the planning time interval. The
DFMM approach is applied to take into account the fuzziness in the objectives, constraints, and reverse
rates. This paper contributes to the literature by 1) providing a framework that facilitates a fuzzy
distributed decision-making process; 2) proposing a novel model that optimizes distribution planning
for a multi-echelon, multi-product, multi-period, and multi-product HRL network; and 3) maximizing
the affected areas’ satisfaction levels by minimizing the sum of the maximum shortages in the affected
areas. A case study of the proposed optimization strategy for earthquake-related disaster planning in a
region of Turkey. In order to demonstrate the scalability and applicability of the proposed model,
computational results are demonstrated for various scenarios. Results indicate that decision-makers can
adjust the search direction to find a satisfactory solution in terms of disaster relief center location and
resource allocation. The proposed optimization strategy aims to minimize the sum of the maximum
shortages in the affected areas during earthquake-related disasters. This case study focuses on a region
in Turkey to showcase the effectiveness of the model. By presenting computational results for different
scenarios, it is evident that decision-makers have the flexibility to adjust their search direction and find
an optimal solution in terms of locating disaster relief centers and allocating resources efficiently.
Keywords: Humanitarian Relief Logistics, Fuzzy Multi-Objective Programming, Distributed Decision
Making, Mixed Integer Programming.
Sustainable Aviation Fuel Production and Aviation Sector

Ibrahim Temam Ibrahim
Ibn Haldun University
Abstract
Sustainable aviation fuels (SAF) present a feasible solution to decarbonize modern aviation. In contrast
to traditional jet fuels, SAFs are produced in a variety of ways, thereby making the choice of one of
these processes a complicated Multi Criteria Decision challenge that involves conflicting priorities. This
study evaluates SAF production processes using a multicriteria methodology, PROMETHEE-2 . Due
to the initial phase of SAF technology advancement and lack of data on these, several stakeholders in
aviation sector were enlisted to assist in the collection of data and preferences. The suggested
framework’s strength lies in its adaptability to suit the subjective opinions of diverse stakeholders,
selection of ranking system, and robustness of outcomes. This research utilizes a set of stakeholders in
a participative manner to rank 11 (A1 to A11) SAF production paths based on 24 parameters categorized
into social, environmental, economic, and technological evaluation criteria. Industry professionals were
given a form to rate SAF production methods according to a performance criterion. Data is validated
Page 139
SEMIT2023
using fuzzy TOPSIS and fuzzy VIKOR and PROMETHEE to reduce professionals’ judgmental
personal prejudice. The study shows that the optimal feedstock for SAF production is the direct
transition process of CO2 to SAF (A11) in the gasification or Fischer-T synthesis group.
Keywords: Aviation, Green Fuel, Sustainable Aviation Fuel, PROMETHEE-2.
Page 140
SEMIT2023
SP15: Industry 4.0, Supply Chain 4.0, and Logistics 4.0
A Draft Model Proposal for Total Quality Management (TQM) Based Industry 4.0 Maturity
Model
Kerem Elibal
BCS Metal Co., Gaziantep, Turkey
Eren Özcelyan
Industrial Engineering Department, Gaziantep University, Gaziantep, Turkey
Abstract
This paper aims to propose a draft model for Total Quality Management (TQM) based Industry 4.0
maturity model. “As the Quality Management Principles (QMPs) are globally accepted business
practices, an Industry 4.0 maturity model based on these principles should leverage the organization’s
efficiency and productivity within the context of Industry 4.0 technologies” is the main argument of
this study. Seven QMPs declared by International Standardization for Organization (ISO) and
recommended actions of these principles has been considered as factors; total 35 factors and maturity
statements of these factors has been constructed. This paper should be assessed as a draft framework
and should be developed in order to verify the proposed model as a future research.
Keywords: Maturity, Industry 4.0, TQM.
Applicability of Conventional Quality Methods on Quality 4.0

Souad Lahmine
USMBA, ENSA, LISA, FES, Morocco
Fatima Bennouna
USMBA, ENSA, LISA, FES, Morocco
Abstract
The constant evolution of information and communication technologies has given rise to a new
industrial era, known as "Quality 4.0." Quality4.0 is a modern approach to quality management that
leverages digital technologies and advanced analytics to improve and optimize manufacturing processes
(Jamkhaneh, 2022; Javaid et al., 2021). It is based on the principles of Industry 4.0, which emphasizes
the integration of cyber-physical systems, the Internet of Things (IoT) and data-driven decision-making
(Escobar et al., 2021). Quality4.0 aims to transform traditional quality management practices by
harnessing real-time data, artificial intelligence, and predictive analytics to achieve higher levels of
product quality, process efficiency and customer satisfaction (Lodi et al., 2022). This article explores
the concept of Quality4.0 and examines the applicability of conventional quality methods in the context
of this emerging paradigm. Conventional quality methods, such as Six Sigma, Total Quality
Management (TQM) and Lean, have been widely used and proven effective in optimizing quality in
traditional manufacturing environments. However, the advent of Quality4.0 introduces new challenges
and opportunities that require a reassessment of these established practices.
Page 141
SEMIT2023
Keywords: Quality 4.0; Industry 4.0; Artificial Intelligence (AI); Machine Learning (ML); Total
Quality Management (TQM); Six Sigma; Internet of Things (IoT); Cyber-Physical Systems; Statistical
Process Control (SPC).
Constructing the Criteria in Determining the Product Groups for Agriculture 4.0 Applications
Melike Erdoğan
Duzce University Department of Computer Engineering, Duzce, Turkey
Zekeriya Konurhan
Munzur University Department of Geography, Tunceli, Turkey
Melih Yücesan
Munzur University, Department of Emergency Aid and Disaster Management, 62000, Tunceli,
Turkey
Muhammet Gül
Istanbul University Department of Transportation and Logistics, Istanbul, Turkey
Abstract
With Agriculture 4.0, the use of techniques such as sensors, robots, artificial intelligence and machine
learning in agriculture has started. It is aimed to increase productivity in agriculture by reducing food
loss and waste through Agriculture 4.0. It is a critical decision to determine which products should be
handled first in order for Turkey to benefit from the advantages of Agriculture 4.0 as soon as possible
compared to developed countries in the field of agriculture. At this point, the problem of which factors
should be addressed in the determination of product & product groups arises. To handle this, in this
study, a multi-criteria analysis has been applied to prioritize the factors that should be considered in the
determination of the critical fruit and vegetable group for export, which should be considered as a
priority within the scope of Agriculture 4.0. In this context, a multi-criteria analysis has been carried
out by adopting Bayesian Best Worst Method BWM (B-BWM), which is an improved version of a
pairwise comparison-based BWM method and applied to group decisions. As a result of the analysis,
the most important and least important criteria to be used in determining which products or product
groups are more suitable for Agriculture 4.0 applications in Turkey and which should be invested in
priority have been determined.
Keywords: Agriculture 4.0, Bayesian Best Worst Method, Evaluation Criteria, MCDM.
Mathematical Modeling for Exploring Spread of Cyberattacks through IoT Devices

Sadeq Damrah
Australian University - Kuwait
Mahmoud DarAssi
Princess Sumaya University for Technology
Yousef AbuHour
Princess Sumaya University for Technology
Abstract
This proposed topic aims to explore the spread of cyberattacks, analyze defense techniques using
adequate security measures, and indicate the number of Internet of Things (IoT) devices that are
suspicious or vulnerable to attack. Additionally, this work will investigate the number of malicious IoT
devices and devices disabled due to the attack. We use a mathematical model to simulate these
Page 142
SEMIT2023
cyberattacks and analyze the reproduction number that measures the attack’s impact. We also explore
the proposed modulation’s theoretical analysis and stability. We will utilize open-source data for
parameter estimation to validate our work and solve the proposed model numerically using
mathematical programming tools.
Keywords: Mathematical Modeling, Iot, Cyber Security, Differential Equations.
A Conceptual Framework for the Development of Data Acquisition and Integration System for
the Dairy Industry
Mohit Malik
National Institute of Food Technology Entrepreneurship and Management (NIFTEM), Kundli
(Sonepat), India
Vijay Kumar Gahlawat
National Institute of Food Technology Entrepreneurship and Management (NIFTEM), Kundli
(Sonepat), India
Rahul S Mor
Business Systems & Operations, Faculty of Business and Law, University of Northampton, UK
Abstract
Data acquisition and integration are vital aspects of circular industrial initiatives and help industries to
develop balanced coordination with other stakeholders and commitment to circular standards. Small
and medium-sized enterprises hesitate to implement the necessary automation and information
collection process because of high investment issues. The dairy industry does have a growing need for
precise process information. The literature points toward the need for proper data management. This
research aims to develop a conceptual model for data acquisition and integration for the dairy industry.
A theoretical framework is proposed for data acquisition and integration in the dairy industry based on
Big Data Analytics (BDA) and the Internet of Things (IoT). The model will help monitor, collect,
manage, and analyze information flow in the dairy supply chain. It gives the industry direct access to
IoT technology for collecting real-time data on essential quality parameters like pH, temperature, etc.
The model will also help improve the data management process in the internal and external context of
data sharing. Implementation of such a system is necessary nowadays to improve food safety and
quality, especially in the Indian context.
Keywords: Big Data Analytics, Dairy Industry, Data acquisition and Integration, Internet of Things
(IoT), Optimization.
Page 143
SEMIT2023
SP16: Marketing and E-marketing for Resources Management
Enhancing Repurchase Intentions of Online Music Platform Customers for the Middle East
Tanya Haqi
Institute of Graduate Sciences, Istanbul Aydin University
Ugur Sener
Business Management, Istanbul Aydin University
Abstract
In the context of on-demand music platforms in Iraq, this study explores the concurrent mediation effect
of web design and e-satisfaction on the relationship between e-service quality and repurchase intention.
As on-demand music platforms are still relatively new in Iraq, it was deemed necessary to dive into the
customer's decision-making process and preferences when selecting the platform, they would like to
use. A parallel me-diation model was used to analyze the dataset gathered from internet users, especially
music listeners, across Iraq. The Literature found in different contexts indicates that e-service quality
significantly affects e-satisfaction, and that was proven to affect repurchase intention. Moreover, web
design and e-satisfaction both mediate the relationship between the quality of e-service and the
repurchase intention. Our results imply that to boost customer repurchase intention, online music
platforms in Iraq and the wider Middle East should pay attention to the quality of their e-services, web
design, and e-satisfaction consistent with the previous studies. This study adds to the body of knowledge
by examining a previously unresearched parallel mediation effect of web design and e-satisfaction on
the link between e-service quality and repurchase intention in the Iraqi environment, especially the
Music industry. Results may provide insight for the rest of the Middle East coun-tries where global
platforms compete with the local platform to increase their market shares.
Keywords: E-Service Quality, ESERVQUAL, Website Design, Repurchase Intention, On-Demand
Music Platforms.

Brand Credibility
Sepehr Mobayen
EMBA, Faculty of Economics & Administrative Science, Ferdowsi University, Mashhad, Iran
Amirhossein Raoof Ghotboddini
Farinaz Tanoorian
EMBA, Faculty of Management & Accounting, Shahid Beheshti University, Tehran, Iran
Abstract
Given that customers are the foundation of business survival, all businesses need to establish strong
relationships with their customers. Social media marketing activities can have a significant impact on
brand credibility and purchase intention by creating a two-way relationship and exchanging brand-
related information between customers and firms. This correlational-applied survey aimed to investigate
the effect of firm-created content (FCC) on customer purchase intention through the mediating role of
Page 144
SEMIT2023
brand credibility. The study population consisted of all construction companies in Mashhad (Iran),
among which 70 companies were selected as the sample. The reliability of the questionnaires was
assessed using Cronbach’s alpha and composite reliability methods, while the validity of the
questionnaires was examined using convergent validity and confirmatory factor analysis (CFA). The
results confirmed the reliability and validity of all items. Finally, the data were analyzed in Smart-PLS
and SPSS using the partial least squares structural equation modeling (PLS-SEM) method. The results
show that firm-created content has significant positive effects on both brand credibility and customer
purchase intention. On the other hand, a significant positive relationship was found between brand
credibility and customer purchase intention. Therefore, the mediating role of brand credibility in the
relationship between firm-created content and customer purchase intention was confirmed. The results
of this study can help companies and marketers better understand how firm-created content influences
consumer purchase intention for industrial products.
Keywords: Firm-Generated Content, Brand Credibility, Purchase Intention, Social Media Marketing,
User-Generated Content.

Management Strategy on Creativity of Core Members in Live Broadcasting Teams- The
Ding Haoqing
Innovation College, North-Chiang Mai University, Chiang Mai, Thailand
Guangxi Vocational College of Water Resources and Electric Power, China
Abstract
To study the influence of high-involvement human resource management (HRM) strategies on the
creativity of the core members of live teams, the article constructs a research model with “resource
security”, “typical model”, and “empowerment support” as independent variables and “creativity” as
the dependent variable in the organizational innovation climate. A total of 106 team anchor samples
and 506 individual anchor samples were collected through questionnaire research, and the data were
statistically analyzed using SPSS 26.0 and Mplus 8.3. The exploratory factor analysis and validation
factor analysis reveal that resource security, typical demonstration, and empowerment support play
mediating roles between high-involvement HRM strategies and creativity, while anchor traits play
moderating roles between them. Finally, specific recommendations are made in the recruitment of
anchors and the formation of live teams based on the study’s results.
Keywords: Creativity, Organizational Innovation Climate, High-Involvement Human Resource
Management.
Page 145
SEMIT2023
The Impact of Industrial Cluster on SMEs Performance: Evidence from the Wood Industry in
Ghana
Masud Ibrahim
Akenten Appiah-Menka University of Skills Training and Entrepreneurial Development
Dora Yeboah
GIMPA School of Business
Abstract
Purpose: Small and medium enterprises undertake their activities in a dynamic environment that is
affected by various processes that impact their business sustainability. This study assessed the effect of
industrial clusters on the performance of Small and Medium Enterprises (SMEs) in Ghana’s wood
industry. Design/Methodology: Using the structured questionnaire method, we surveyed 320 SME
wood manufacturers in three regions of Ghana. Results: Findings from this study revealed that the
factors of production, horizontal networking, innovation, and degree of network were positively related
to competitive advantage. This study further reveals a positive and significant relationship between
competitive advantage and firm performance. In addition, the findings of this study revealed that the
degree of network, horizontal networking, factors of production, and innovation indirectly influence
firm performance through competitive advantage. Implication: The managerial implications of this
study are also discussed. Originality/Value: This study contributes to the extant literature on cluster
innovation in a developing economy context.
Keywords: Industrial Cluster, Competitive Advantage, SMEs, SMEs Performance, Ghana.
Antecedents of Mobile Banking Apps adoption among Consumers in Ghana

Masud Ibrahim
Akenten Appiah-Menka University of Skills Training and Entrepreneurial Development
Dora Yeboah
GIMPA School of Business
Freeman Anthony Mensah
Akenten Appiah-Menka University of Skills Training and Entrepreneurial Development Department
of Management Studies
Abstract
Purpose: This study examined the determinants of mobile banking apps adoption among bank
customers in Ghana. A conceptual framework was framed from a revised Technological Acceptance
Model (TAM) model. A total of nine hypotheses were developed and tested using Structural Equation
Modelling Technique. Design/Methodology: The purposive sampling technique was employed to select
the various banks operating the mobile banking app and simple random sampling was used to sample
450 bank customers. Results: Findings from this study revealed that perceived ease of use and perceived
usefulness are influenced positively by design and perceived security. Also, perceived ease of use,
social influence and perceived security risk were found to have a strong influence on mobile banking
app adoption intention. Contribution: This study provides significant insights and implications for
banking organisations on how to ensure adoption and usage of their mobile banking apps.
Keywords: Mobile Banking App, Perceived Ease of Use, Perceived Usefulness, Perceived Security
Risk, Technology Acceptance Model.
Page 146
SEMIT2023
SP17: Humanities and Social Sciences
Strengthening Women leadership in India & Globe in digital Era

Dr. Meenakshi Kaushik
Trinity Institute of Innovations in Professional Studies, GGSIPU, India
Manu Kaushik
Maharaja Agrasen College, University of Delhi, India
Abstract
In India and globe the status of women has transitioned over a period of time from a housewife to a
businesswoman. Since women are an integral part of the society and globe, it is very important to
understand significantly the contribution of women in global economy & society. Developing
leadership skill-sets among women in digital transformation age is the need of the hour. National
education policy (NEP), several other governmental schemes& initiatives, continuous technological
advancements on a global perspective has set a clear vision and growing need of digital literacy for
Indian women. Social initiative education for all (EFA) has supported 5.5 million children in India and
half of them are girls. This research paper will be highlighting an urgent need for strengthening effective
leadership in women in India and globe. This will also cover various processes of building effective
leadership in women & factors affecting effective leadership in women, balancing work and personal
life and building resilience in the face of adversity by creating a supportive and collaborative work
environment by embracing digital transformation, digital literacy and technological competence in
women. This research paper will also give light on few important issues like empowering women
through digital skillsets by incorporating leadership strategies for women in digital age and their
significant role in driving digital transformation all over the globe. This conceptual paper will be based
on secondary data from various sources.
Keywords: Women Leadership, Sex-Role Orientation Digital Transformation, Digital Literacy,
Stereotyping, New Education Policy (NEP), Role Stress, Re-skill, Multiskilling, Upskilling.
Staff Scheduling in Service Industry Considering Qualifications

S. Bouaziz
University of Sfax, OLID Research Laboratory
B. Chaker
D. Dhouib
Abstract
Research has demonstrated that effectively managing human resources in the service sector has
significant advantages for managers and employees, par-ticularly when staff are in close proximity to
clients. Due to the variety of aspects that need to be taken into account, this scheduling challenge has
proven to be a difficult task, for even directors and professional planners. This paper deals with
personnel scheduling problems in the technical control field. In order to reduce workload and average
Page 147
SEMIT2023
premiums, jobs are distributed to inspectors over a time horizon. The solution complies with a variety
of restrictions, such as the maxi-mum tolerated radiation dose exposure limit, staff certifications, and
work hours. A new mathematical framework is developed using the goal programming ap-proach.
Results show a significant performance improvement in the scheduling operation.
Keywords: Personnel Scheduling, Qualification, Inspectors, Goal Programming.
Impact of Gender Inequalities on Education in Morocco: An Economic Analysis of Barriers and

Opportunities
Mohammed Alami Chentoufi
Laboratory of Mathematical Modeling and Economic Computation, Associated to the Laboratory of
Theoretical and Applied Economics Research, Faculty of Economics and Management, University
Hassan I, Settat, Morocco
AIT ALLA Aniss
Laboratory of Research in Management and Development, Faculty of Economics and Management,
University Hassan I, Settat, Morocco
Naoual Mamdouh
Laboratory of Mathematical Modeling and Economic Computation, Faculty of Economics and
Management, University Hassan I, Settat, Morocco
Abstract
Gender inequalities in the field of education pose a major issue that continues to challenge development
efforts in Morocco. Despite significant progress made in the education sector over the past decades,
persistent gender disparities hinder equitable access to quality education. This study aims to examine
the economic impact of these gender inequalities on the Moroccan education system, focusing on the
barriers that impede girls' access to education and the potential opportunities associated with promoting
gender equality. Drawing on an exhaustive review of recent economic literature, this research explores
the socio-economic determinants of gender inequalities in education in Morocco. It also examines the
economic consequences of these disparities, both in the short and long term, with an emphasis on their
impact on productivity, economic growth, and poverty reduction. The study employs a rigorous
quantitative methodology, analyzing empirical data from various national and international sources. An
advanced econometric model is applied to assess the extent of the impact of gender inequalities on
educational performance and the country's economic outcomes. Through this analysis, we hope to
provide relevant insights for policymakers, researchers, and development stakeholders concerned with
promoting gender equality and improving the education system in Morocco. The findings will
contribute to guiding public policies aimed at strengthening equitable access to education and creating
an enabling environment for the development of all individuals, regardless of their gender.
Keywords: Gender Inequalities, Education, Economic Development, Morocco.
Page 148
SEMIT2023
Role of Top Management Commitment and Information Technology Investment in Digital

Transformation
Pankaj Adatiya Tiwari
St Francis Institutions, Bangalore, Karnataka, India
Abstract
Purpose – In recent times, enterprises are more focused on information technology (IT) investment to
transform digitally. Conversely, the reasons to invest in IT are lacking in literature though it can enhance
the overall business. Based on the resource-based view (RBV), this research paper explores the
influence of IT infrastructure and investments on the digital transformation of organizations considering
the context of the digital transformation strategy. Furthermore, this research paper examines the
moderating role of digital transformation strategy on the relationships between top management
commitment, IT infrastructure, and digital transformation in business. Design/methodology/approach
– Through a questionnaire survey of small and medium enterprises, 127 sample data were collected,
and the ordinal regression method was used to test the hypothesis and the conceptual framework.
Findings – Digital transformation strategy moderates the relationship between top management
commitment, IT infrastructure, and organizational digital transformation. Moreover, top management
commitment and IT infrastructure have a significant positive effect on digital transformation.
Originality/value – This research paper investigates the moderating role of digital transformation
strategy in the relationship between top management commitment, IT infrastructure investment, and
digital transformation performance. As a result, the research paper supplements substantially the body
of knowledge on business success, IT business value, digital transformation, and strategic management.
Also, the findings can help improve managers’ perceptions of IT business value and provide theoretical
guidance on deriving digital transformation performance with the help of IT infrastructure investments.
Keywords: Information Technology Investment, Infrastructure, Digital Transformation Strategy, Top
Management Support, Top Management Commitment, Digital Transformation.
The Critical Factors of Success of Gamification in Digital Banking Services: Using Analytic
Hierarchy Process (AHP) Approach
Asif Akhtar
Associate Professor, Department of Business Administration, Aligarh Muslim Central University,
India
Shilpa Chauhan
Meenakshi Kaushik
Trinity Institute of Innovations in Professional Studies (Guru Gobind Indraprastha University), India
Ashish Gupta
Abstract
Purpose: The game elements in serious gamification are well acclaimed and have been experimented
with in various contexts, however, the literature seldom noticed any effort to prioritize the gamification
elements in the context of digital banking platforms. The current study is written with the purpose to
identify, evaluate and prioritize the critical factors of the success of gamification elements in digital
banking services, based on the Indian perspective.
Design/Methodology/Approach: The study examines critical success factors of gamification elements
through a literature review and expert opinion in the banking context. It utilizes analytical hierarchy
Page 149
SEMIT2023
process (AHP) to arrive at the relative importance of elements using a pair-wise comparison survey
instrument. The study evaluates the intensity of criteria, namely assessment, challenge, interactivity,
aesthetics, functionality, and control.
Findings: The survey was conducted, and analysis was done using the AHP approach. The findings
identified functionality as the most suitable game element for gamification in digital banking.
Limitations: This study assumes that the criteria for classification of gamification elements are
mutually exclusive but in reality, they are overlapping each other particularly in banking context. The
study suggests the application of analytical network process (ANP) to mitigate this issue.
Practical implications: The study highlights the factors that information technology teams and
management of banks must foresee while introducing gamification elements into their digital platform.
Originality: This study is the first of its kind to acknowledge the significance of elements in the
universe of gamification when applied to the banking industry.
Keywords: Game Elements, AHP, Digital Banking.
Consumer Confidence on State Regulation of E-Commerce in India

Pankaj Kumar Gupta
Department of Management Studies Jamia Millia Islamia New Delhi
Sonakshi Singh
Department of Management Studies Jamia Millia Islamia New Delhi
Manvi Gupta
Institute of Law and Finance, Goethe Universität, Frankfurt, Germany
Abstract
Transformation of business from physical space to cyber space in zeal to capture market share by large
number of organizations has perpetuated the regulators globally to reframe and redefine the regulation
to protect the rights of consumers and ensure their privacy and security. In many countries of the world
the legal protection available to the aggrieved consumers is insufficient to protect their interest and
build a healthy business environment. Protection of consumers in not only regulatory, but also ethical.
India has witnessed various changes in regulation of e-commerce to align with the global e-business
environment. In India, major changes in legislation dealing with consumer protection and competition
have been made in the recent times. However, various studies in India, Europe and other emerging
markets have indicated deficiencies in regulation of e-commerce that shakes the confidence of
consumers. Our paper attempts to explore the confidence of consumers in the India’s regulatory
framework during various stages of e-commerce interface. We have conducted a survey of 450
consumers that have experienced the application of state regulation in one or other form ranging from
the issues of advertisements, payments systems to that of consumer protection and remedial measures.
The results of Structural Equation Modelling analysis shows that the confidence of consumer in state
regulation in India is low that essentially requires major changes in regulation and dynamic control over
the e-commerce transactions. We also find a significant number of unfair trade practices that need to be
checked on part of regulator – Competition Commission of India. In addition, there are a host of issues
relating to data privacy and information security.
Keywords: Consumer Protection Act, Cyber Fraud, E-Commerce, Regulation.
Page 150
SEMIT2023
Social Networks and Privacy Concerns: An Insight into IT Industry User Experience
Maria Anjum
Department of Computer Science, Lahore College for Women University, Lahore, Pakistan
Maria Anum
Humaira Kosar
Mariam Rehman
Department of Information Technology, Government College University, Faisalabad, Pakistan
Abstract
Information systems in the form of social networks provide a virtual place where people share ideas,
information, personal feelings and experiences with family, friends and colleagues. This virtual
environment provides an opportunity to develop social relationships with people around the world and
make sustainable virtual communities. The sustainability of these virtual communities is critical
especially in the time of globalization and increased adoption of virtual work environments to access
and utilize best resources around the globe. This has encouraged academics and practitioners to study
user behavior in relation to social networks and has now become an important research area. Various
studies have been conducted to understand social networks and user behavior in the context of
personality traits, cultural differences, adoption of technology, gender differences, usage patterns,
employee’s performances, etc. However, these studies largely focus on user intentions in adoption of
technology and there is little evidence that information systems are analysed from a perspective of
actual use (which considers user’s experience over time) to form a trustworthy and sustainable social
network and virtual community. Therefore, in this research efforts are made to analyse ‘actual use’ of
virtual environments and the factors that affect towards the sustainability of such platforms. The
research has proposed a conceptual framework and evaluation is carried out through statistical analysis.
The findings show that actual experience is positively related to ease of use, self-disclosure,
performance expectancy, and hedonic expectancy. However, perceived risk was not found significant
towards actual use of social networks.
Keywords: User Behavior, Sustainability, Information systems, Social Networks, Social Networking
Sites, SNS, UTAUT, Structured Equation Modeling (SEM).
Page 151
SEMIT2023
SP18: Advances in Engineering: Machine Learning Techniques
Marketing Mix Modeling tools and methodologies: A comparative analysis

Ghada Ben Amor
National Engineering School of Tunis University of Tunis El Manar, Tunis, Tunisia
Safa Bhar Layeb
LR Oasis, National Engineering School of Tunis University of Tunis El Manar, Tunis, Tunisia
Marwa Hassni
LR OASIS
National Engineering School of Tunis University of Tunis El Manar, Tunis, Tunisia
Fares Ben Mahmoud
MASS Analytics Tunis, Tunisia
Assma Essghir
MASS Analytics Tunis, Tunisia
Abstract
Due to the growing awareness of user privacy protection, the internet is witnessing a slow vanishing of
third-party cookies. Numerous advertisers and companies are shifting back towards the socalled
Marketing Mix Modeling method in order to answer key marketing measurement questions. Meta and
Google made use of this opportunity and developed open-source Marketing Mix Modeling tools that
are based on different multi-linear regression techniques. Meta’s Product being based on Ridge
regression and Google’s using Bayesian regression. This paper represents a brief comparison between
those two different softwares in addition to a third software named MassTer developed by the English
company “MASS Analytics”. The purpose is to assess the capabilities and performance of these tools.
This will be done through running a Marketing Mix Modeling project on the three softwares and
analyzing the obtained results. The comparison mainly focuses on the technical and statistical
performance observed through different statistical metrics in order to point out the strengths and
weaknesses of each software.
Keywords: Marketing Mix Modeling, Artificial Intelligence, Ridge Regression, Multiple Linear
Regression, Supervised Machine Learning.
Detection of Shilling Attack with Support Vector Machines Using Oversampling

Halil İbrahim Ayaz
Department of Industrial Engineering, Faculty of Engineering, Necmettin Erbakan University, Konya,
Turkey
Zehra Kamişli Öztürk
Eskisehir Technical University, Department of Industrial Engineering, Eskisehir, Turkey
Abstract
Recommender systems (RS) enable predicting the future behaviour of similar users by evaluating their
past behaviour. Recommender systems overcome information overload problems. Collaborative
Filtering (CF) provides quick recommendations for big databases. However, these models are
vulnerable to shilling or profile injection attacks. Shilling attacks damage recommender system
Page 152
SEMIT2023
mechanism with synthetically generated fake profiles. These attacks aim to increase or decrease the rate
of target item considering their attack types. In the current literature, enormous shilling attack detection
methods are proposed. This study aims to detect future shilling attack detection with a classification
approach. Thus, the robustness of RS is provided to future similar shilling attacks. This study uses
Support Vector Machines (SVM) to robust RS models. However, attack sizes are very small compared
to the dataset. A recently proposed Outlier Detection-Based Oversampling Technique (ODBOT) is used
to solve this class imbalance problem. Comparative results are given for with ODBOT and without
ODBOT models. Precision, recall, F1 measure and accuracy rates are presented for both models. The
results show that the SVM model with ODBOT outperforms most of the performance metrics.
Keywords: Recommender Systems, Shilling Attack, Support Vector Machines, Over-sampling.
Advancing Structural Integrity Assessment through Machine Learning in RC Columns

Aybike Özyüksel Çiftçioğlu
Department of Civil Engineering, Faculty of Engineering, Manisa Celal Bayar University, Manisa,
Türkiye
Abstract
This research presents a comprehensive investigation focused on the predictive modeling of shear
strength in rectangular reinforced concrete (RC) columns using machine learning techniques.
Addressing the intricate challenge posed by shear strength complexity is pivotal for assessing column
stability and structural integrity. The study curates an extensive dataset of 545 experimental
observations from diverse literature sources. Employing four distinct machine learning regression
models, the study rigorously evaluates their performance using established metrics including coefficient
of determination, root mean squared error, mean absolute error, and mean absolute percentage error.
The outcomes highlight the robust predictive capabilities of the selected machine learning models.
Notably, high coefficient of determination values, ranging from 0.936 to 0.939 (93.6% to 93.9%
explanatory capacity), underscore the models' remarkable ability to elucidate shear strength variations.
Noteworthy accuracy is observed, with the Random Forest model demonstrating a slight performance
advantage over the Decision Tree model. Additionally, the efficacy of the XGBoost and LightGBM is
highlighted, suggesting their potential for accurate shear strength estimation. The study validates the
precision of the model through scatter plots and error distribution plots, confirming accurate shear
strength predictions in various scenarios. This work highlights the potential of machine learning models
in improving accurate shear strength estimation, thus significantly contributing to the advancement of
structural engineering methodologies.
Keywords: RC Columns; Shear Strength; Concrete; Machine Learning.
Page 153
SEMIT2023
Fuzzy Logic Methods for Determining the Mechanical Behaviour of Masonry Walls
Başak Zengin
Kahramanmaras Istiklal University, Construction Technology, Kahramanmaras/Turkey
Pınar Usta
Department of Civil Engineering, Technology Faculty, Isparta University of Applied Scien es, Isparta,
Turkey
Ozge Onat
Graduate Institute, Isparta University of Applied Sciences, Isparta, Turkey
Abstract
The mechanical behavior of the wall, in-situ wall tests and numerical analysis are required along with
the material properties. With the support of smart learning techniques, the state of the walls was
estimated. This is possible to obtain healthy data to support the experiment and modeling. Using
mathematical tools such as Fuzzy Logic has proven to be helpful in solving com-plex engineering
problems, without it being necessary to reproduce the phenomenon studied, when the only information
available is composed of the parameters of the problem and desire results. To analyze the behavior of
the wall more accurately and in a short time, analyses were made using the fuzzy method, one of the
smart learning techniques, and compared with the data in the studies in which experimental analysis
was applied. The behavior of the wall, the ductility, and the energy capacity of the wall were tried to be
estimated. In the fuzzy, material parameters and wall load capacities that will affect the properties of
the wall are used as inputs. Thirty-five (35) data sets, experiments, and modeling data from different
studies were taken. Estimation results were compared with empirical results.
Keywords: Ductility, Energy Capacity, Fuzzy Logic, and Masonry Wall.
Harnessing Advanced AI Techniques: An In-depth Analysis of Machine Learning Models for

Improved Diabetes Prediction
Ihtisham Ul Haq
Department of Mechatronics Engineering, University of Engineering & Technology Peshawar,
Pakistan
Muhammad Anas
Department of Electrical Engineering, University of Gujrat, Hafiz Hayat Campus Gujrat, Pakistan
Hamza Ahmad Khan
Pakistan
Zubair Ahmad Khan
Pakistan
Abstract
By 2023, diabetes will have become a major public health problem in Pakistan, affecting an estimated
26.3% of the adult population. The rising death rate is attributable in part to diabetes-related
complications, highlighting the critical need for cutting-edge, trustworthy diabetes prediction systems
that enable early detection and treatment. Since machine learning (ML) models have the potential to
improve the accuracy of diabetes forecasts, this study is focused on exploiting these methods of artificial
intelligence (AI) for better predictions. As such, the importance of this research to the field of medical
informatics cannot be overstated. A wide range of ML and deep learning models, such as Support
Vector Machine (SVM), K-Nearest Neighbors (KNN), Logistic Regression, Naive Bayes, Random
Page 154
SEMIT2023
Forests, Decision Trees, and a novel ensemble learning approach, were tested and compared. The
models were evaluated on a conventional diabetes dataset. Our min-max normalization and other data
pre-processing procedures greatly improved model accuracy. At first, the models were only about 70%
accurate. However, after the pre-processing phase, we saw a dramatic improvement in performance. In
particular, Random Forests reached a 96.49 percent accuracy, while both Decision Trees and KNN
models hit 96.33 percent. To improve upon the prediction accuracy of the separate models, we used a
novel ensemble learning mechanism. This method demonstrates the efficacy of hybrid techniques in
AI-driven healthcare solutions by utilizing the strengths of different ML models to create a single
optimal model. Our re-search establishes a new paradigm for the field of illness prediction and has far-
reaching consequences for diabetes care and early intervention techniques. We anticipate that our
findings will encourage more investigation into the utility of cutting-edge AI techniques for improved
disease prediction, ultimately leading to a more secure and manageable future.
Keywords: Diabetes; Ensemble Learning; Medical Informatics; Machine Learning; Hybrid Models.
Machine Learning in Fault Diagnosis of Electromechanical Devices using Vibration

Measurements
Binnur Demir Erdem
Mechanical Engineering, Near East University, Nicosia, N. Cyprus, Via Mersin 10, Turkey
Mahmut Ahsen Savaş
Mechanical Engineering, Near East University, Nicosia, N. Cyprus, Via Mersin 10, Turkey
Kamil Dimililer
Electrical and Electronic Engineering, Near East University, Nicosia, N. Cyprus, Via Mersin 10,
Turkey
Abstract
Mechanical engineering is evolving from conventional engineering principles to contemporary
electronic mechanical engineering concepts as a result of the significant advancements in science and
technology. Mechanical engineering has entered a new stage of growth where the fusion of artificial
intelligence technology and mechanical engineering has become a hotspot due to new advancements
and breakthroughs. Artificial intelligence is usually seen from the perspective of mechanical and
electronic engineering. By using the powerful data processing capabilities of computers, can solve a
wide range of challenging issues and boost automation in mechanical and electronic engineering. The
fundamental problem with traditional mechanical fault detection is that machines can only identify
slight defects because technical conditions are so challenging and complex. The utilization process is
complicated as a result. This tendency necessitates the constant advancement of equipment fault
diagnosis technologies. The field of mechanical engineering has advanced to a new level of
development that includes artificial intelligence. The main goal is to categorize mechanical failure
analysis using computers' advanced data processing abilities to deal with a variety of difficult problems
in mechanical engineering. ML techniques will be used to analyze the acquired dataset. The Random
Forest algorithm produced results at the highest rate of 88.6%.
Keywords: Mechanical Engineering, Artificial Intelligence, Machine Learning, Random Forest.
Page 155
SEMIT2023
SP19: Artificial Intelligence Innovations
Enhancing Traffic Flow Prediction in Urban Areas through Deep Learning and Probe
Information: A Comparative Study
Serap Ergun
Isparta Applied Sciences University, Department of Computer Engineering, Turkey
Abstract
With the remarkable development of Intelligent Transportation Systems in recent years, the easy
collection of traffic information and various vehicle-related data has been made possible. The
availability of probe information allows for the acquisition of more extensive traffic data in addition to
observation information. In this paper, the traffic flow prediction method on urban roads using probe
information is considered. Accurate and real-time traffic information is deemed indispensable for the
deployment of high-performance intelligent transportation systems. Traffic flow, being a complicated
phenomenon, can be expressed in terms of its characteristics without prior knowledge of site-specific
features, thanks to the application of deep learning, which automatically acquires feature quantities.
Consequently, a traffic flow prediction model utilizing deep learning is investigated in this re-search.
Additionally, a comparison is made with other traffic flow prediction methods. The research aims to
improve prediction accuracy and contribute to the advancement of more efficient intelligent
transportation systems.
Keywords: Traffic Flow Prediction, Intelligent Transportation Systems, Probe Information, Deep
Learning.
A Multi-Agent Deep Q-Network Approach for Optimizing Signal Control in Multi-Intersection

Road Environments for Improving Traffic Flow in Urban Areas
Serap Ergun
Isparta Applied Sciences University, Department of Computer Engineering, Turkey
Abstract
In urban areas, the worsening of traffic congestion results in increased time and economic losses,
exerting a significant influence on our daily lives. In-appropriate signal switching on ordinary roads
stands out as a contributing factor to traffic congestion. The manipulation of parameters in general
signal control is conventionally estimated through human experiences, yielding suboptimal outcomes.
Consequently, the control of traffic lights and the enhancement of traffic behavior emerge as viable
approaches for mitigating traffic congestion. This research incorporates a multi-agent system with Deep
Q-Network method to address this pressing issue. In this urgent scenario, a signal control system that
aims to achieve appropriate parameter manipulation within a road environment featuring multiple
intersections is proposed. The findings presented in this manuscript offer valuable insights for tackling
traffic congestion effectively.
Keywords: Intelligent Transport Systems, Traffic Congestion and Control, Signal Switching, Multi-
Agent System.
Page 156
SEMIT2023
A Simulating Annealing Algorithm for the Maximum Diversity Social-Distancing Problem

(Mdps)
Bayram Dündar
Department of Industrial Engineering, Bartin University, Turkey
Abstract
During the pandemic, the implementation of the social distance rule has significantly mitigated the risk
of contracting the virus so that societies can continue their vital functions healthily. In this study, we
have considered the problem of assigning students to seats that are as far away from each other as
possible in a classroom considering the social distance. We have developed a Simulating Annealing
algorithm (SA) for the Maximum Diversity - Social Distancing (MDPs) problem in which the total
distance among a certain number of students is maximized. MDPs is classified as NP-hard problem.
With the SA approach, an optimal or near-optimal solution to the MDPs problem is obtained in a
reasonable time frame.
Keywords: Seat Assignment Problem, Maximum Diversity-Social Distancing Problem, Simulated
Annealing, Optimization.
An Artificial Intelligence-Enhanced UAV Configuration for Object Detection, Instant

Communication, and Real-Time Information Transmission
Emir Guler
Aerospace Engineering Department, Tarsus University, Mersin, Turkey
Tolga Ozcan
Mechanical Engineering Department, Engineering Faculty, Erciyes University, Kayseri, Turkey
Affan Alkim
Aerospace Engineering Department, Tarsus University, Mersin, Turkey
Erin Faysal
Mechanical Engineering Department, Engineering Faculty, Erciyes University, Kayseri, Turkey
Abstract
A UAV design that can detect targets individually or in surveillance flight missions in a swarm
formation is presented. For the mission in a swarm operation, the UAV is designed to communicate
with the members of the swarm and work together to achieve the common objective. The necessary
communication mechanism has been designed to adapt the UAV to the swarm, and it has been equipped
with the necessary hardware to perform the task in coordination. To communicate with other UAVs and
also to establish an effective communication with the ground control station, the interconnection and
bandwidth need of the swarm has been addressed with a proper network architecture. A PixHawk 4
flight controller, a Raspberry Pi 4 as auxiliary computer for general purpose computing power, and an
NVIDIA Jetson Nano developer kit for real-time object detection are used to provide flight control of
each individual in the designed architecture. For multi-hop communication, the transmission of packets
over the network is executed based on the 802.11s standard, thus ensuring that the images and videos
of the detected targets are transmitted to the ground station in real-time, uninterrupted and without
latency.
Keywords: UAV, Swarm, Communication, Information Transfer.
Page 157
SEMIT2023
Augmented Reality Immersive World with Hologram Special Effect in Early Childhood
Education
Maria Seraphina Astriani
Jakarta, Indonesia
Raymond Bahana
Jakarta, Indonesia
Arif Priyono Susilo Ahmad
Creative Advertising Program, Visual Communication Design Department, School of Design, Bina
Nusantara University, Jakarta, Indonesia
Andreas Kurniawan
Jakarta, Indonesia
Lee Huey Yi
Water Lily, Melaka, Malaysia
Abstract
Early childhood education is a critical stage in education that can act as the foundation for the
intelligence of an individual. In general, students of this age group are unable to remain seated and study
for long periods of time because they tend to get bored. They are more likely to be attracted to visual
information and enjoy interactive activities. This condition challenges the world of education to keep
developing learning methods that are able to help students to learn the materials. The combination of
Augmented Reality technology and hologram special effects is a promising solution because it allows
students to see an immersive world and allows them to interact with real-life physical objects while
learning. This solution is specially designed to make it easier for students to learn the materials without
using wearable devices.
Keywords: Augmented Reality, Hologram, Education, Early Childhood, Immersive.
Synergizing Global Leadership Competencies with Artificial Intelligence and Expert Systems: A
Multidisciplinary Approach for the Future of Management
Sameera Fernandes,
Sheeja M.S.
Swetha Parivara
Abstract
The advent of Artificial Intelligence (AI) and Expert Systems (ES) has ushered in a new era of
possibilities in various sectors, including management and leadership. This research aims to explore the
integration of these advanced technologies with global leadership competencies. The study employs a
multidisciplinary approach, incorporating data from fields such as computer science, management
studies, and organizational behavior. The primary objective is to understand how AI and ES can
augment and refine the skills required for effective global leadership.
Page 158
SEMIT2023
Utilizing a mixed-methods research design, the study will gather data through surveys, interviews, and
case studies involving executives and leaders in multinational corporations. These primary data sources
will be supplemented by an exhaustive review of existing literature on AI, ES, and global leadership
competencies. A key focus will be on identifying the synergies between technology and leadership
skills, and how these can be harnessed for organizational success.
The research aims to achieve several key objectives: to identify the core competencies required for
global leadership in the age of AI; to explore the capabilities of AI and ES in enhancing these
competencies; and to provide a framework for the effective integration of technology and leadership.
The study will culminate in actionable insights and recommendations for organizations aiming to
pioneer in the realm of technologically augmented global leadership.
The anticipated outcomes of this research are twofold. Firstly, it seeks to make a significant academic
contribution by bridging the gap between technology and leadership studies. Secondly, it aims to offer
practical guidelines for organizations to leverage AI and ES in their leadership development programs.
By creating a synergy between global leadership competencies and advanced technologies, this research
aspires to set a new standard for the future of management, grounded in empirical evidence and
multidisciplinary rigor.
Keywords: Global Leadership Competencies, Scientific Decision-Making, Comparative Study,
Empirical Data, Mixed-Methods Approach, Organizational Development, Cross-Industry Analysis,
Leadership Training, Actionable Insights.
Page 159
SEMIT2023
SP20: Artificial Intelligence: The Future of Everything
Energy Consumption Forecasting using Time Series Analysis Methods

Fatma Üzel
Ankara Yildirim Beyazit University, Ankara, Turkey
Hilal Arslan
Abstract
Developing technology depending on the expectations of people increases the demand for energy. In
addition, it is predicted that the need for energy will reach even higher levels with the increasing
population. However, with small measures that can be taken in energy consumption, more efficient use
of existing energy can be achieved. In particular, the excess energy consumed in households can be
reduced by individual measures. In this study, ML algorithms are used to estimate the consumption in
the next periods by considering the energy data consumed in a house located in Kecioren district of
Ankara city. According to the results obtained, the Time Series ARIMA model was used and it was
observed that it gave good results. In the future, it is thought that the proposed model will be applied to
businesses with larger energy consumption and thus contribute to energy efficiency.
Keywords: Time Series, ARIMA, Prediction.
Recognition
Dr. Allam Fatima Zohra
Signal and Communication Laboratory, Department of Electronics, National Polytechnic School, 10,
Avenue des Frères Oudek, El Harrach, 16200 Algiers, Algeria
Pr. Hamami Mitiche Latifa
Pr. Bousbia-Salah Hicham
Abstract
Facial recognition is an image processing technology that identifies or verifies the identity of an
individual based on their unique facial features. It uses algorithms to analyze features such as eye
distance, face shape, eye, nose, and mouth contours to create a digital signature specific to each person.
Our work focuses on the implementation and comparison of two descriptors, widely developed in the
literature, in this case, the Local Binary Patterns (LBP) and the Hough Transform. The objective is to
evaluate their respective performance in facial recognition in terms of accuracy, robustness to lighting
variations and processing speed.
Page 160
SEMIT2023
The two descriptors are implemented and applied on a specific database to extract relevant facial
features. The results obtained made it possible to determine the advantages and limits of each descriptor
in the field of facial recognition. We compared the performance of these two descriptors in terms of
precision, calculation speed and robustness.
Keywords: Facial Recognition, Local Binary Pattern, Hough Transform.
Multi-Objective Optimization based Efficient Healthcare Management System

Harinandan Tunga
Department of Computer Science & Engineering, RCC Institute of Information Technology
Surjendu Pal
Samarjit Kar
Department of Mathematics, National Institute of Technology Durgapur
Debasis Giri
Department of Information Technology, Maulana Abul Kalamazad University of Technology
Abstract
An efficient multi-objective healthcare system is formulated in a comprehensive manner which
maximizes the Quality of Service (QoS) and minimizes the cost of operation. This model is solved using
an elitist Non-Dominated Sorting Genetic Algorithm (NSGA II) and Multi-Objective Particle Swarm
Optimization. The proposed model is solved using two meta-heuristic algorithms namely, Non-
Dominated Sorting Genetic Algorithm (NSGA II) and Multi-Objective Particle Swarm Optimization
(MOPSO) to generate a Pareto-optimal front, and analysis is performed on the given data. The first
objective function focuses on minimizing the overall cost associated with various resources in
healthcare, including doctors, nurses, technicians, beds, rooms, and medical equipment. The second
objective function aims to maximize the quality of service provided, considering the expertise of
doctors, nurses and the availability of beds. The proposed method is implemented on the Python
working platform and its performance is verified. The performance of the proposed method is confirmed
by contrasting it with MOPSO and NSGA II algorithms in terms of approximate Pareto optimal results,
hypervolume (HV), generational distance (GD), inverted generational distance (IGD), space and spread.
As a result, the obtained results showed that the proposed method is efficient for improving quality of
service (QoS) and reducing the total cost of healthcare management system performance.
Keywords: Quality of Service, Multi-Objective Optimization, Healthcare Management System,
MOPSO, NSGA-II, Cost Minimization, Quality Maximization.
Page 161
SEMIT2023
Adoption of Generative Artificial Intelligence: A Cross Industry Comparison

Cameron Guthrie
TBS Business School
Samuel Fosso-Wamba
TBS Business School
Maciel M Queiroz
FGV EAESP
Abstract
Generative AI (Gen AI) is taking the world by storm. As the media reports on its latest achievements,
from passing tertiary entrance exams to composing music and poetry, managers are looking for the
business case to justify investment in these emerging technologies. This exploratory study reports on
the state of Gen AI adoption by companies across 19 industries. Based on questionnaire data collected
from 300 managers in the UK and the USA, we find that current Gen AI projects and maturity levels
differ by industry. Our study provides an early description of the Gen AI landscape, contrasts and
analyses industry differences, and concludes with several avenues for further research.
Keywords: ChatGPT, Artificial Intelligence, Technology Adoption, Generative AI.
Exploring the Elastic, Electronic and Optical properties of cubic ZrCoAs half – Heusler
Semiconductor: An ab-initio Study
Lynet Allan
Department of Physics, Faculty of Science and Technology, University of Nairobi, P.O.Box 30197-
00100 Nairobi Kenya
R.E Mapasha
Department of Physics, University of Pretoria, Private Bag x 20, Hatfield, Republic of South Africa
Winfred M. Mulwa
Department of Physics, Faculty of Science, Egerton University, P.O Box 536-20115 Egerton Kenya
Julius M. Mwabora
00100 Nairobi Kenya
Robinson J. Musembi
00100 Nairobi Kenya
Abstract
The electronic, mechanical, elastic, and optical properties of the ZrCoAs half-Heusler compound have
been systematically studied, with and without Spin Orbit Coupling (SOC) effects, using the plane wave
self-consistent field approach. The exchange correlation functional used in the study was Perdew-
Burke-Erzerhof of the generalized gradient approximation (GGA-PBE). It was found that without SOC
effects, the results significantly deviated from experimental values. The incorporation of spin-orbit
effects leads to an increase in the Kohn-Sham bandgap. The electronic bandgap of the material was
found to be an indirect 0.85 eV gap between the high symmetry points of W and L without SOC effects.
However, this bandgap was increased to an indirect 1.095 eV gap between Г and L when SOC was
factored into the computations. The formation of the electronic bandgap was attributed to specific states,
namely, Co 3d, Zr 4s, and As 4p, for the conduction band, and Co 3d and As 4p for the valence band
without SOC effects. However, when SOC was taken into account, it coupled the orbital angular
Page 162
SEMIT2023
momentum of the electron with its spin, generating a splitting of the energy levels for degenerate states
and resulting in them having slightly different energies. As a result, some states became energetically
more favorable, and others less favorable. With SOC effects, Co 5d, Zr 8d, and As 4p, became dominant
in the conduction band, while Co 3d and As 4p dominated the valence band. The lattice constant also
decreased by 0.05% when SOC was factored into the computations, showing a much-improved
agreement with experimental observations. The material also exhibited ductility and mechanical
stability, making it promising for a variety of applications. Moreover, it was discovered that the
material's optical properties were excellent for photovoltaic applications, indicating a potential
application in solar energy conversion technology.
Keywords: DFT, Spin Orbit Coupling, Electronic, Mechanical, Elastic, And Optical Properties, Half
Heuslers ZrCoAs.
Performance Enhance and Management of Renewable Based Microgrid with backup Fuel Cell
Saqib Waris
Polytechnic institute australia (PIA)
Linta Khalil
National Fertilizer Corporation Multan (NFC)
Syd Mustafa Hassan Naqvi
Deakin University Waurn Ponds
Kamran Liaquat Bhatti
NFC Multan
Muhammad Zahid Shafique
University of Wollongong Australia
Malik Hamza Attiq
Pakistan Engineering Council
Abstract
Abstract Solar and wind energy have proved to be the major sources, with fast growth rates in the last
two to three decades, of renewable energy for electricity generation. This study focuses on several
studies of the new integration of wind energy from a permanent magnetic synchronous generation
system (PMSG) [1] with a grid-connected system. The aim is focused on extracting maximum power
from wind and solar energy system is a novel technique called incremental conductance method to get
maximum power point tracking (MPPT) [1], it gives faster responses to the changes in PV voltage and
current. In addition, wind and solar power supplies are integrated in an appropriate manner to always
guarantee the continuity of demand supplies. The overall operation of the wind generator is measured
by the speed calculation, basically, a rotor speed estimator that is sensor free and prevents all mechanical
sensors. Rotor speed has been estimated for boosters for the control of the turbine speed by keeping
voltage and current (input dc). Furthermore, under critical conditions of demand and PVA variation on
wind farm power, the test system is supported by fuel cell resources.
Keywords: Photo voltaic array (PVA), maximum power point tracking (MPPT), Permanent magnet
synchronous generator (PMSG), Synchronous reference frame (SRF), Photovoltaic (PV), Direct current
(DC).
Page 163
SEMIT2023
SP21: Inventory Control, Production Planning and Scheduling
A Multi Criteria Decision Making Based Integrated Approach for Determining Production
Priority in Demand Management
Elif Kılıç Delice
Atatürk University, Erzurum, 25240, Turkey
Elif Akviran
A101 Yeni Mağazacılık A.Ş., İzmir, Turkey
Abstract
In the demand management process; It is tried to balance the customers' needs with the company's
supply possibilities. Effective demand management starts with effective stock control and tracking of
customer orders and ends with customer satisfaction. Correctly determining production priorities is one
of the ways to increase customer satisfaction. In this context, a new approach has been developed to
determine the production priority needed in the demand management process. A new integrated
approach based on Multi-Criteria Decision Making (MCDM) is proposed for determining production
priorities. In the first stage of this approach, which consists of four stages, customer alternatives are
deter-mined. In the second stage, production priority criteria are determined depending on customer
characteristics. In the third stage; with LBWA and CRITIC methods, objective and subjective criteria
weights are calculated by taking into account expert opinions and real data. Then, using the objective
and subjective criterion weights, the integrated criterion weights are determined by the SOWIA method.
In the literature; there are few studies that determine production priority using the MCDM method. In
addition, no study in which quantitative and qualitative methods are evaluated in an integrated manner
and used in the ceramics industry has not been found in the literature. For this reason, it is thought that
this study will contribute to the literature.
Keywords: Production Priority, LBWA, CRITIC.
Performance Analysis for The Inventory System with a Two-Stage Service Process
Serife OZKAR
Department of International Trade and Logistics, Balıkesir University
Abstract
A queueing-inventory system with a two-stage service is examined. The service time in each stage
follows a phase-type distribution with different representations. In other words, the all service time of
the customer is a phase-type distribution by way of the convolution of the two phase-type distributions.
The system where customers arrive to the system according to a MAP has a single production facility.
Items are put into the inventory and each arriving customer demands an item from the inventory. When
the inventory level is positive, if an arriving customer finds the server idle gets into service immediately.
Served customer leaves the system and the on-hand inventory is decreased one unit of item at service
completion epoch. Otherwise, the customer enters into a queue of infinite capacity to be served under
the FCFS discipline. If there is at least a customer in the queue and the inventory level is reached to
zero, one item is immediately local purchased for no losing the waiting customer (for maintain customer
Page 164
SEMIT2023
goodwill). The production facility produces items according to an (s,S) policy. The production times
are exponentially distributed. The production is switched on when the inventory level depletes to s and
the production remains on until the inventory level reaches to the maximum level S. Once the inventory
level becomes S, the production process is switched off. The solution of the system in the steady-state
is achieved by using matrix-geometric method and illustrative numerical examples about the effect of
parameters on the performance measures and optimum production policy are performed.
Keywords: Queueing-Inventory, Markovian Arrival Process, Two-Stage Service, Convolution Of
Phase-Type Distributions, Matrix-Geometric Method.
Moment-Based Approximation Formula for Variance of Semi-Markovian Random Walk

Process
Büşra Alakoç
Karadeniz Technical University, Department of Mathematics, Trabzon, Turkey
Aslı Bektaş Kamışlık
Recep Tayyip Erdoğan University, Department of Mathematics, Rize, Turkey
Tülay Kesemen
TOBB University of Economics and Technology, Department of Industrial Engineering, Ankara,
Turkey
Tahir Khaniyev
TOBB University of Economics and Technology, Department of Industrial Engineering, Ankara,
Turkey
Azerbaijan State University of Economics, The Center of Digital Economics, Baku, Azerbaijan
Abstract
In this study approximation formula is presented for the variance of the semi-Markovian random walk
process X(t). Numerous research has addressed to similar moment problems, and asymptotic expansions
have been reached for the moments of this process. The most important feature that distinguishes this
study from other studies in the literature is that for the first time, approximation formulas were obtained
for the moments of the process instead of asymptotic expansions. First of all, approximation formulas
for the moments of boundary functional S_N(z) of the process were introduced in the sense of Kambo.
Then using these results, approximation formulas were obtained for the first two moments and the
variance of the process X(t). Moreover, a special example is given using the Monte Carlo simulation
method to examine the accuracy of the proposed approximation formulas.
Keywords: Semi-Markovian Random Walk, Moment-Based Approximation, Boundary Functional,
Monte Carlo Simulation.
Page 165
SEMIT2023
Inventory Model of Type (s,S) with Heavy Tailed Demand and Heavy Tailed Interference of
Chance
Aslı Bektaş Kamışlık
Recep Tayyip Erdoğan University, Department of Mathematics, Rize, TURKEY
Hacer Hilal Hasançebi
Recep Tayyip Erdoğan University, Department of Mathematics, Rize, TURKEY
Tülay Kesemen
Karadeniz Technical University, Department of Mathematics, Trabzon, TURKEY
Tahir Khaniyev
Tobb University of Economics and Technology, Department of Industrial Engineering, Ankara,
Turkey
Azerbaijan State University of Economics, The Center of Digital Economics, Baku, Azerbaijan
Abstract
In this study, a stochastic process X(t) that describes an inventory model of type (s,S) is considered with
heavy tailed demand and heavy tailed interference of chance. More precisely, in this study, it is assumed
that the demand and interference of chance random variables have the same Pareto distribution with the
same index 1<α<2. In the mentioned case, it is known that the Pareto distribution is heavy tailed with
infinite variance, and in the domain of attraction of an alpha-stable distribution with index 1<α<2. Under
this assumption, an asymptotic expansion for the ergodic distribution of the process X(t) is obtained
and the weak convergence theorem is proved.
Keywords: Inventory Model of Type (s,S), Renewal Reward Process, Heavy-Tailed Demand, Heavy-
Tailed Interference of Chance.
Multi-Depot Vehicle Routing Problem with Drone Collaboration in Humanitarian Logistic

Zeynep Yüksel
Gaziantep University, Faculty of Engineering, Department of Industrial Engineering, Gaziantep,
Turkey
Dursun Emre Epcim
Turkey
Süleyman Mete
Turkey
Abstract
Routing problems are used in many areas to obtain the most appropriate results in terms of time and
cost. An attempt is made to address the issue by formulating mathematical models that incorporate
multiple variables such as capacity, time, cost, and demand, tailored to the specific area of application.
Natural disasters are one of these applications. In natural disasters, especially time management is a
critical issue. For this reason, routing models play a crucial role in delivering aid to disaster victims and
transporting disaster victims to hospitals. In this study, a mathematical model is proposed to be applied
in post-disaster humanitarian aid logistics. The model, which aims to minimize the total distribution
time, also considers the distribution of perishable foods. Drones are integrated into this operational
framework to facilitate the dissemination of perishable commodities. Thus, a new mathematical model
for the multi-warehouse vehicle routing problem, which includes both truck-drone cooperation and
Page 166
SEMIT2023
perishable foods, has been introduced to the literature. The proposed model is solved with an illustrative
example using GAMS/CPLEX software and the results are tested.
Keywords: Drone-Truck Collaboration, Humanitarian Logistic, Multi-Depot Vehicle Routing
Problem, Perishable Commodities.
Page 167
SEMIT2023
SP22: Multi-Criteria Decision-Making
A Bibliometric Analysis of Multi-Criteria Decision-Making Techniques in Disaster

Management
Mehmet Kabak
Gazi University, Ankara, Türkiye
Ezgi Aktas Potur
Gazi University, Ankara, Türkiye
Abstract
Disaster management is a crucial discipline aimed at minimizing the potential damage stemming from
natural calamities and protecting communities against their possible impacts. To achieve effective and
beneficial solutions, a key fundamental tool that can be employed is the utilization of multi-criteria
decision-making techniques within this process. Disaster management is a multifaceted process that
involves significant uncertainties, risks, and intricacies. This complexity arises from the necessity to
make decisions not only based on cost-effectiveness but also taking into consideration the social,
environmental, and public health impacts. The utilization of multi-criteria decision-making techniques
in pre-disaster planning, intervention strategy development, and recovery initiatives empowers disaster
managers with a broader perspective, enabling comprehensive scenario analyses and contributing to
risk reduction efforts. The application of multi-criteria decision-making techniques contributes to the
formulation of enhanced and sustainable strategies in the field of disaster management. These
techniques should be acknowledged as crucial tools, aiming to bolster community resilience against
disasters and mitigate the detrimental impacts they entail. The primary objective of this paper is to
perform a bibliometric analysis of the extant academic literature concerning the application of multi-
criteria decision-making (MCDM) techniques in the field of disaster management. This study is
envisaged to provide a guideline for researchers in their future studies in the field of disaster
management.
Keywords: Disaster Management, Decision-Making, Bibliometric Analysis.
Food Distribution in Times of Famine

Mete Gündoğan
Ankara Yıldırım Beyazıt University, Department of Industrial Engineering, Ankara, Turkey
Murat ATA
TUBİTAK, Ankara, Turkey
Abstract
Whether or not "global warming" is real, which has been much on the agenda in recent years, "famine"
or "drought" are highly probable events. Whether they are the result of our evil deeds, our trials, or
natural events that occur from time to time, it is necessary to take precautions against famine and
drought. In addition to preventing the waste of resources, it will be necessary to use the limited resources
that are available, at least for a time, fairly and efficiently. At this point it will be necessary to solve the
problems of identifying, supplying and distributing needs. Water resources, dry food that can be stored
Page 168
SEMIT2023
for a long time, fresh food that can be grown quickly according to the season, health supplies and
medicines are vital. Clearly, the duration and predictability of shortages are also very important. In this
study, assuming that communication channels are open and developed, we will propose a solution for
the supply and distribution of such supplies in the medium term.
Keywords: Famine, Food Distribution, Resource Management.
Determining the Location for Conquest in a Computer Game Using MCDM Methods
Zeynep Melike Sever
Department of Industrial Engineering, TOBB ETU, Ankara, Turkey
Babek Erdebilli
Department of Industrial Engineering, Yıldırım Beyazıt University, Ankara, Turkey
Abstract
Within the scope of this project, research has been conducted on a computer game. The primary
objective of this study is to make a selection among the countries that will be conquered in the game.
During the process of making this selection, various Multiple Criteria Decision Making (MCDM)
methods such as SWARA, AHP, TOPSIS, ELECTRE, and VIKOR have been explained and applied.
The game developers meticulously examined historical details during the game's production phase in
order to determine relevant parameters that would accurately mirror reality. As a result, it can be
confidently asserted that the executed project maintains a high level of coherence. Moreover, owing to
the inherent structure of the game, which necessitates decision-making at each stage, it is exceptionally
well-suited for the application of MCDM methods. The result of this study reveal that Trebizond and
Byzantium emerge as the prominent alternatives for the conquest location. The alignment of the result
with a decision made in history adds a layer of realism to the issue, thereby contributing an original
piece of work to the existing literature. The decisions derived from the application of these methods
within the game have been meticulously compared and thoughtfully interpreted.
Keywords: MCDM, AHP, ELECTRE, TOPSIS, Location Problem.
MCDM
Yavuz Selim Özdemir
Ankara Science University
Babek Erdebilli
Ender Sevinç
Ankara Science University
Abstract
Industrial, economic, and financial concepts are affected by multiple factors. Analytical techniques that
must lead to logical conclusions are frequently used to solve such multi-criteria problems. Due to the
COVID-19 pandemic process, customers’ behaviors were affected. To evaluate a bank's dependability,
a set of evaluation criteria from the perspective of its clients can be used. Using multi-criteria evaluation
methods to regulate their importance to the review purpose, we can associate multidimensional and
Page 169
SEMIT2023
contradictory accurate criteria into a single universalizing value. The main objective of this study is to
examine public banks from the perspective of their customers using the fuzzy multi-criteria decision-
making approaches. In this study, the ranking of banks from the customer's viewpoint is determined
using the PROMETHEE, while the weight of the banks is inferred using the Fuzzy AHP method with
a group of experts. These techniques are favored due to their quick computation times and
straightforward calculations. The bank data set was provided from June 2020 to June 2022. The number
of bank clients, ATMs, branches, deposit rates, and ten more criteria were used to rank the banks from
the customer's perspective.
Keywords: Bank Performance Evaluation, Fuzzy AHP, PROMETHEE, Group Decision Making.
Supplier Selection in the context of Industry 4.0 using hybrid DEA-SMART Method
Babek Erdebilli
Parmida Bahreini
Abstract
In the era of Industry 4.0, choosing suppliers for online commerce is of utmost importance and calls for
the application of efficient, data-centric techniques. Businesses are under increasing pressure to improve
their supply chain management strategies and select the best suppliers in the online commerce
environment of Industry 4.0. Traditional approaches, however, sometimes don't include a thorough
assessment of suppliers across several dimensions. In order to close this gap, this paper proposes a new
approach that combines Data Envelopment Analysis (DEA) with the Simple Multi-Attribute Ranking
Technique (SMART). The first phase is applying DEA to determine how effective suppliers are using
the data that has been gathered. DEA offers a quantitative indicator of how efficiently providers convert
their inputs into outputs. This combination score enables rating suppliers while simultaneously
considering multi-attribute evaluation and quantifying efficiency assessment, then using a number of
different criteria, providers are evaluated using the SMART approach. The findings of this analysis help
to improve supplier selection procedures in the context of online commerce, which falls under the
purview of Industry 4.0.
Keywords: Supplier selection, Industry 4.0, Data Envelopment Analysis, Simple Multi-Attribute
Ranking Technique.
Page 170
SEMIT2023
SP23: Smart Marketing
Investigating the Social and Conceptual Structure of AI in Marketing: A Bibliometric Analysis

Mostafa Esmaeili Mahyari
Department of business Administration, College of Farabi, University of Tehran, Qom, Iran
Hamid Reza Irani
Department of business Administration, College of Farabi, University of Tehran, Qom, Iran
Vahid Nourmandi Pour
College of Farabi, University of Tehran, Qom, Iran
Hamed Nozari
Islamic Azad University, Dubai Branch, UAE
Abstract
The role of AI in marketing in becoming more and more important and is being widely applied in
various sector of marketing. This popularity encouraged many scholars to conduct research in this filed.
The aim of this study is investigating the social and conceptual structure of AI in marketing. To reach
this goal, along with descriptive analyses, keywords co occurrence and co authorship networks were
portrayed through VOS viewer software. Descriptive analyses introduce the highly cited papers,
authors, countries, universities and journals completed with the co authorship analysis showing the
social structure of this filed. Keywords co occurrence analysis also show the conceptual structure of
this filed in particular through temporal status highlights the recent research topics. Findings show that
AI is becoming an important tool for businesses to identify and understand the behavioral patterns and
needs of customers specifically through customer journey and customer experience. These technologies
not only assisted businesses but also helped customers in their decision making.
Keywords: Artificial Intelligence, AI, Marketing, Bibliometrics, Machine Learning, Internet Of
Things.
Hashtag and Marketing Campaign on Twitter: From the Spectrum of Smartphone Industry
Perspective
Prashant Chaudhary
School of Management (PG), Dr. Vishwanath Karad MIT World Peace University, Pune (India)
Prabha Kiran
Westminster International University, Tashkent
Sarika Sharma
Symbiosis Institute of Computer Studies and Research Symbiosis International (Deemed University)
Abstract
Social Media has completely revolutionised the way companies engage with its target segment. Earlier
it was advertisement, then companies paying billions to pay for ad positioning, and now it focuses on a
different type of campaign, known as ‘Hashtag Campaign’, focusing on the message and means with
which one is communicating with the target group. The present research is focused on sentiment
analysis of Indian Smart phone brands engagement on twitter with help of hashtag marketing campaign.
Page 171
SEMIT2023
A case study is also included for this purpose and five major phone brands were selected those having
highest engagement on twitter. The findings were based on analysis of close to 18000 tweets and 87
hashtags. The data was collected by developing twitter developer account and then mining tweets with
R language. After the collection of tweets, sentiment analysis was done with Microsoft Azure Machine
Learning tool. Based on the results it can be suggested that the companies top authorities need to engage
directly with the consumers so as to create confidence in the customers, creativity should not be missed
but entire focus should not be on creativity, the brands with the negative sentiment should focus more
on streamlining the delivery of the product and services, and companies need to focus on product
development for long term benefits and subsequent focus on campaigns, as giveaways strategy is good
for short term promotions.
Keywords: Hashtag, Social Media Marketing, Marketing Campaign, Sentiment Analysis, Smartphone
Industry, And Customer Engagement.
Exploring the Influence of Entrepreneurial Culture on SMEs Performance through Marketing

Capabilities and Social Media Usage
Mohsen Mazaheriasad
Islamic Azad University, South Tehran Branch
Maryam Jamali
University of Tehran, Iran
Abstract
This article examines the impact of entrepreneurial culture on the performance of small and medium-
sized enterprises (SMEs) in Tehran Province, Iran, while investigating the mediating role of marketing
capabilities and social media usage. Data was collected through a questionnaire consisting of 25
questions that were divided into four sections: "Entrepreneurial culture," "SMEs performance,"
"Marketing capabilities," and "Social media usage". Non-probability sampling was used to collect
responses from 80 SME owners or managers, and the validity and reliability of the data were verified.
The research utilized structural equation modeling with partial least squares (PLS) approach via
SmartPLS 4 software for data analysis. The results revealed that entrepreneurial culture has a positive
and significant impact on SMEs performance, marketing capabilities, and social media usage.
Additionally, marketing capabilities and social media usage also positively and significantly affect
SMEs performance. Moreover, social media usage positively affects marketing capabilities. Finally, the
study confirmed the indirect effect of entrepreneurial culture on SMEs performance through the
mediating roles of marketing capabilities and social media usage. The article concludes with practical
and research-based recommendations based on the research findings.
Keywords: Entrepreneurial Culture, SMEs Performance, Marketing Capabilities, Social Media Usage.
Page 172
SEMIT2023
E-Shopping Sites Preference Analysis with Multi-Criteria Decision-Making Methods

Vuslat Erat
TOBB University of Economics and Technology, Çankaya, Ankara
Babek Erdebilli
Department of Industrial Engineering, Ankara Yıldırım Beyazıt University, Ankara, Turkey
Abstract
With the developing technology, the use of the internet has become widespread. Today, people who do
not want to waste time prefer e-shopping sites. From this point of view, in this study, Turkey's five most
popular e-shopping sites (Trendyol, Hepsiburada, Amazon, Getir, Morhipo) were selected as
alternatives. Seven criteria were used in the study: site design, product variety, reliability, detailed
filtering, service quality, ease of use of the site, and price. As a method, MOOSRA, MOORA and
TOPSIS methods from Multi-Criteria Decision Making (MCDM) methods were selected and the results
were compared. While determining the criterion weights, the Pairwise Comparison method was used.
As a result of the study, it has been seen that the most important criterion is the price, and the least
important criterion is the site design. When the MOOSRA, MOORA and TOPSIS methods were
applied, it was concluded that the order of the alternatives was the same and Trendyol was the best e-
shopping site.
Keywords: Multiple Criteria Decision Making (MCDM), MOOSRA, MOORA, TOPSIS, E-Shopping,
Pairwise Comparison.
Category Sort with TODIM Method, an Application for Retail Industry E-Commerce Site
Ayşe Hümeyra Akar
TOBB University of Economics and Technology, Çankaya, Ankara
Babek Erdebilli
Department of Industrial Engineering, Ankara Yıldırım Beyazıt University, Ankara, Turkey
Abstract
Abstract In recent years, with the development of classical multi-criteria decision-making methods and
the production of various methods, it is frequently used to solve various real-life problems. TODIM
(Topsis-based multi-criteria decision making) is one of the methods that can be used in solving ranking
problems with its structure based on expectancy theory. With the effect of increasing online sales
volume, changing and developing world balances and pandemic, e-commerce has become
indispensable for companies. For this reason, taking part in the digital world has become indispensable
for companies. The easiest way to step into the digital world is through successful e-commerce sites.
The success of companies' e-commerce sites depends on the correct product category ranking to enable
users to navigate easily and improve the shopping experience. Therefore, strategically aligning product
categories is critical for improving user experience and increasing sales performance. In this context,
our aim is to use the TODIM method to rank product categories on e-commerce sites.
Keywords: TODIM, E-Commerce, Sorting Problem, MCDM.
Page 173
SEMIT2023
SP24: Advancement and Future of Technologies
Dissecting the Economic and Environmental Compatibility of Battery Electric and

Conventional Vehicles in India
Amrut P Bhosale
Rajarambapu Institute of Technology, Rajaramnagar, Uran Islampur, Maharashtra, India
Sachin A. Mastud
Veermata Jijabai Technological Institute, Matunga, Mumbai, India
Muzammil Bepari
Bharati Vidyapeeth's College of Engineering, Kolhapur, India
Ketaki Bhosale
D. Y. Patil College of Engineering and Technology, Kolhapur, India
Abstract
Fuel supplies for conventional vehicles are vulnerable to scarcity, which could ultimately lead to an
increase in fuel prices. There has been a realization regarding national energy security as a result of
these high gasoline costs, which further increase the overall cost of ownership. Additionally, the
emissions from burning conventional fuels make consideration of the already pressing environmental
issues necessary. On the other hand, because they have low running costs and no tailpipe emissions,
electric vehicles are being considered as a viable alternative to conventional automobiles. But when a
vehicle's whole life cycle is taken into account, the common sense belief that electric vehicles are
cheaper and emit no emissions may be misleading. Therefore, it is necessary to consider both the
economic and environmental elements of whether electric vehicles are a viable alternative to
conventional automobiles. In this article, a life cycle analysis—both economic and environmental—
between battery-electric and conventional automobiles is presented in the context of India. For financial
analysis, a Total Cost of Ownership (TCO) model is created to show how compatible battery-electric
vehicles are. OpenLCA software, which is based on the ReCePi 2016 technique, is used to conduct the
environmental analysis for all impact categories at both the mid-point and end-point levels. According
to the findings, electric vehicles are more expensive than conventional automobiles in India based on
current data and regulations. However, it is shown that electric vehicles have cost parity and can
sometimes even become more inexpensive than conventional automobiles by using specific optimizing
factors in sensitivity analysis. The results of environmental studies show that battery electric vehicles
emit fewer greenhouse gases (GHGs) than do conventional automobiles. Battery electric vehicles,
however, had less of an impact in ten of the eighteen impact categories that were examined, and they
even have a lower impact score at the end-point level.
Keywords: Electric Vehicles, Life Cycle Analysis, Environmental Impact of Electric Vehicles, Total
Cost of Ownership, Battery Electric Vehicles, Lithium Ion Batteries.
Page 174
SEMIT2023
Optimization of Controller Parameter Multilevel Converter Based STATCOM for Reactive

Power Compensation
Hamed Atyia Soodi
Collage of Information Technology, Imam Ja’afar Al-Sadiq University, Baghdad, Iraq
Yousif A. Hamad
Artificial Intelligence Laboratory, Siberian Federal University, Krasnoyarsk, 660074 Krasnoyarsk,
Russia,
Department of Computer Science, Collage of Computer Science and Information Technology,
University of Kirkuk, Kirkuk, Iraq
Ardam H. Mohammed
Abstract
Static Synchronous Compensator (STATCOM) is a self-tuning proportional-integral (PI) controller that
has been developed to handle the voltage and reactive power of Cascaded H-Bridges (CHB-
STATCOM) in order to improve the power system. The purpose of optimizing controller parameters
for a Multilevel Converter-based STATCOM (Static Synchronous Compensator) in the context of
reactive power compensation is to enhance the performance and efficiency of the STATCOM system
in regulating the power factor and voltage of an electrical system. This controller was introduced in this
article. This research presented that the STATCOM should have the capability to be operate both in
inductive and capacitive mode of operation and absorb or injects the desired amount of reactive power
into the grid. The reactionary power that is received or supplied contributes to the improvement of the
power system. The benefits of the self-tuning PI controller have been optimized to their full potential
through the use of Accelerated Particle Swarm Optimization (APSO) strategies. The STATCOM is
implemented into the electrical system, and its reaction to varying reactive and voltage levels is
analyzed as part of the process of determining whether or not the optimization techniques are successful.
Two controllers are now in operation at the STATCOM. The first controller is in charge of managing
the DC voltage (which is employ me to produce reference signal depended on differences in phase
angle), and the second controller is in charge of controlling the reactive power. Both controllers are in
use at the same time. (Which is what is utilized to produce the modulation index). The output of the
controller is then transferred to the phase change carrier, which is followed by pulse width modulation
in a sinusoidal pattern. Because of this, the power converter switch may now function properly.
(PSPWM).
Keywords. APSO; Reactive power compensation; STATCOM; PI control; Phase shift pulse width
modulation.
Page 175
SEMIT2023
Analysis of 3 Different Hexagonal Microstrip Antenna Designs for Breast Cancer Detection
Studies
Gözde Saribaş
Çukurova University, Department of Biomedical Engineering, Sarıçam, Adana, Turkey
Emine Avşar Aydin
Adana Alparslan Türkeş Science and Technology University, Department of Aerospace Engineering,
Sarıçam, Adana, Turkey
Ahmet Aydin
Çukurova University, Department of Biomedical Engineering, 01250, Sarıçam, Adana, Turkey
Abstract
Cancer, among various diseases that have claimed numerous lives throughout history, holds a prominent
place. Breast cancer has gained significant recognition and is frequently diagnosed. The early and
accurate detection of breast cancer plays a vital role in improving treatment outcomes. Extensive
research has been conducted to achieve early and accurate diagnosis, resulting in remarkable
advancements. Microwave Imaging has emerged as a popular method for detecting breast cancer, and
microstrip antennas have found widespread use in such studies due to their small size, compatibility
with other devices, and cost-effectiveness. This study focuses on the design and simulation of three
distinct octagonal microstrip antenna configurations specifically intended for microwave imaging
techniques. The CST Microwave Studio Suite program was employed for the design and simulation
process. The performance evaluation of the designed antennas was carried out through a three-stage
scenario sequence in the simulation environment. In the first scenario, two samples of the antenna model
were positioned facing each other, followed by measurements. The second scenario involved placing a
homogeneous breast model between the antennas from the first scenario. Subsequently, in the third
scenario, a tumor model was incorporated into the breast model. All three designed antennas exhibited
ultra-wideband characteristics within the 3-8 GHz range. Notably, the measurement values of the first
antenna, particularly in the 5-6 GHz frequency range, differed when compared to measurements taken
without the breast model. The presence of the breast model and tumor breast model caused a decrease
in the measured value. Similarly, the second antenna displayed changes in measurement values across
the antenna-breast model and the tumor breast model within the 5-7 GHz frequency range. Lastly, the
third designed antenna exhibited attenuation when exposed to microwave signals within the 5.5-7 GHz
frequency range. The measurement results with the breast model and the breast model with a tumor
were nearly identical, attributed to the similar dielectric properties of fibro glandular tissue and tumor
tissue. In conclusion, based on the analysis of the three simulation models created for the designed
antennas, it concluded that these antennas demonstrated satisfactory operational performance and are
suitable for breast cancer detection studies.
Keywords: Microwave Imaging, Breast Cancer, Microstrip Antenna, CST Simulation Program.
Page 176
SEMIT2023
Numerical Investigation of the Effect of the Use of Multiple Propeller Systems on Flight
Performance in Unmanned Aerial Vehicles.
Kasım Şahin Dönmez
department of aeronautical science and technologies, graduate student, Iskenderun Technical
University, 31200 Hatay, Türkiye
Metin Uzun
Department of Airframe and Powerplant Maintenance, Faculty of Aeronautics and Astronautics,
Iskenderun Technical University, 31200 Hatay, Türkiye
Abstract
The aim of this study is to try to solve the need for Fixed Propeller Unmanned Aerial Vehicles by using
multiple propellers in the horizontal and vertical directions without changing the electronic equipment
and propeller type with the changes in the total take-off weight. In unmanned aerial vehicles, energy
needs are met by using electrical power effectively with the help of a battery, which is stored as a fuel
system in many of them. There will be changes in all of the engine, battery and propeller systems that
the aircraft will need as a result of flight altitude, body weight and useful load information. The weight
and cost of electronic equipment do not increase proportionally to the amount of increase in aircraft
take-off weight. For example, if you want the battery capacity to increase by 2 times, the weight and
cost of the battery will increase more than 2 times. However, the required thrust, that is, the weight
increase in the aircraft, will cause the change of all electronic equipment. When the compatibility
between these working equipments is ensured, the problem can be solved by changing the number of
electronic equipment even if the weight of the aircraft changes. In this study, with the help of Ansys
Fluent, a certain distance was left between the propellers placed in horizontal and vertical directions as
double and triple of equal diameter, and the minimum distance of this distance and whether there was
a loss of efficiency in the use of multiple propellers were explained in detail. Here, multiple propeller
pairs were operated both in the same direction and in opposite directions, and distance and direction
analyzes were also carried out. In addition, the compatibility of the propeller pairs is shown graphically
and visually whether it is similar or not in different revolutions.
Keywords: Propeller, Unmanned Aerial Vehicles, Efficiency.
A Study on the Physical Properties of Elaborated Graphene

Oudjertli Salah
Research Center in Industrial Technologies (CRTI), Algeria
Abstract
Graphene is a highly promising material for nanotechnology applications in information technology
and computing. It has the potential for the development of various electronic devices, including
transistors, sensors, batteries, and supercapacitors, and also shows promise for use in quantum
computing due to its spintronics property. In this context, we have synthesized the graphene by
Chemically reduced graphene oxide (rGO), it's a method commonly used and cost-effective method for
synthesizing Graphene. Our studies have been conducted on the physical properties of elaborated
Graphene, including the calculation of electron mobility, the Hall Effect and the electrical conductivity.
These experimental measurements are crucial for evaluating the electronic characteristics of the
elaborated Graphene and determining its potential for electronic and optical applications.
Page 177
SEMIT2023
Full/Extended Abstract Papers Proceedings

This book consists of all papers of SEMIT 2023, excluding the
papers going to be published by Springer. Those papers will be
published in a booklet by Springer.
Page 178
SEMIT 2023
Table of Contents
SP01: Artificial Intelligence and Expert Systems ............................................................... 182
Classification of Monkeypox and HPV with Supervised Machine Learning Algorithms using k-mer ............ 182
Yapay Zekâ ile B2B Kampanya ve Sadakat Yönetimi ..................................................................................... 189
SAR Görüntülerinden Derin Öğrenme Tabanlı Hedef Tanima......................................................................... 201
Sporcular için Sakatlık ve Performans Analizi Çerçevesi ................................................................................ 209
Exploring the trends of Artificial Intelligence in Recruitment: A Bibliometric study ...................................... 217
SP03: IT and EM Based Case Studies ................................................................................. 232
Kişisel Koruyucu Donanimlardan Koruyucu Maskelerin Kullanicilardaki Olumsuz Etkileri Ve Tasarimlarinin
Değerlendirilmesi ............................................................................................................................................. 232
Interlocking Fired Bricks as Innovative Waste Abstraction Proxies; Case Study Abuja Northern Nigeria ..... 244
Chinese Postman Problem with Dynamic Start/End Point ............................................................................... 257
SP04: Data Science, Big Data, Data Mining, and Knowledge Management in EM ........ 262
Comparison of Cost and Benefit Analysis of Active Leakage Control Applications in Water Distribution
Systems ............................................................................................................................................................. 262
Benefit and Cost Analysis of Network and Service Connection Renewal in Distribution Systems for
Sustainable Water Loss Management ............................................................................................................... 272
Designing A National Cyber Security Strategic Management Model with the Foundational Data Theory
Approach .......................................................................................................................................................... 282
SP05: Safety and Reliability Engineering, System Engineering and System Safety, Safety
in Industry 4.0 ........................................................................................................................ 285
Accident Analysis with Bow Tie Method; Case of British Petroleum Texas Refinery .................................... 285
SP06: Applied Soft Computing in engineering Management ............................................ 297
A Study on Measuring the Mental Workload of Instructors and Evaluating Significant Differences .............. 297
Optical Transmission Systems Incident Management in the Telecommunications Sector in Senegal ............. 302
Business Process Linguistic Modelling - Design – Process Design Part I ........................................................ 312
SP07: Supply Chain Network Design; Considering Recent Trends in Sustainability and
Circular Economy ................................................................................................................. 325
Developing a Cradle-to-Gate Product Carbon Footprint Calculator with ISO 14067 ...................................... 325
Kapasite Planlama Yöntemleri İle Orta Gerilim Dağitim Hücre Müşteri Siparişleri Planlama Modeli ........... 338
A location inventory model under vendor managed Inventory contract ........................................................... 348
Prioritization of Freight Carrier Selection Factors with the Best Worst Method in Multi-Criteria Decision
Analysis ............................................................................................................................................................ 363
SP08: Optimization, Decision Making and Support Systems ............................................ 375
Optimization models for production planning in the agro-industrial sector: A systematic review of the literature
.......................................................................................................................................................................... 375
Comparing OECD Countries’ Logistics Performances by Comprehensive Distance-Based Ranking (COBRA)
Method .............................................................................................................................................................. 388
SP09: Quality, Productivity, and Project Management ..................................................... 400
İnsansiz Hava Araçlarinda İnsan Hatasinin Etkilerinin Sistematik Analizi Ve Siniflandirilmasi- İha Tasarimi
Ve Ergonomi ..................................................................................................................................................... 400
The Quality Assessment of a Company in Defence Industry: A Case Study Approach ................................... 410
Page 179
SEMIT 2023
Application of Best Project Management Practices: Case Study Application in Social Projects ..................... 419
Analysis of Investment Feasibility Using the Life Cycle Cost (LCC) Method in Implementing Green Retrofits
in High-Rise Office Buildings .......................................................................................................................... 436
SP10: Other Fields of Study Related to EM and IT ........................................................... 448
The Impact of Technostress and COVID-19 Stress on Employee Burnout: The Mediating Role of Resilience
.......................................................................................................................................................................... 448
EmpiricalAnalysis of Profit-Intended Vehicle FleetMaintenance Schemes in Developing Economies: Nigeria in
Retrospect ......................................................................................................................................................... 469
Future Perspective ................................................................................................................. 482
Parametric Study of Shear Strength UHPC Tapered Beam .............................................................................. 482
The Present and Future Scope of Virtual Reality in the Construction Industry ................................................ 496
Mixed Reality and Its Future Scope in the Construction Sector ....................................................................... 507
Identification of Green Retrofitting Procurement and Permitting Processes in High-rise Office Buildings in
Jakarta Based on PerMen PUPR No.21 Year 2021 and GBC Indonesia that Affects Project Time Performance
.......................................................................................................................................................................... 518
Performance Investigation of Heat Sinks with Different Perforations Diameters ............................................ 530
SP13: Advancing Innovation and Collaboration in Science, Engineering, and Information
Technology ............................................................................................................................. 542
Improve Channel Capacity of TDMA .............................................................................................................. 542
Seismic Design of Underground Structures’ Vulnerability – Review .............................................................. 557
SP14: Uncertain Decision Making ....................................................................................... 566
Stochastic optimization methodology for production planning with uncertain demand and lead time based on
the digital twin .................................................................................................................................................. 566
Sustainable Aviation Fuel Production and Aviation Sector .............................................................................. 570
SP15: Industry 4.0, Supply Chain 4.0, and Logistics 4.0 ................................................... 594
A Draft Model Proposal for Total Quality Management (TQM) Based Industry 4.0 Maturity Model............. 594
Prioritizing Criteria for Selecting Product Groups in Smart Agriculture Practices .......................................... 605
SP16: Marketing and E-marketing for Resources Management ...................................... 607
Enhancing Repurchase Intentions of Online Music Platform Customers for the Middle East ......................... 607
The Impact of Firm-Generated Content on Industrial Customers' Purchase Intention through Brand Credibility
.......................................................................................................................................................................... 609
A Study on the Cross-Level Impact Mechanism of High Involvement in Human Resource Management
Strategy on Creativity of Core Members in Live Broadcasting Teams - The Mediating Role of Organizational
Innovation Atmosphere..................................................................................................................................... 621
The Impact of Industrial Cluster on SMEs Performance: Evidence from the Wood Industry in Ghana .......... 635
SP17: Humanities and Social Sciences................................................................................. 653
Strengthening Women leadership in India & Globe in digital Era ................................................................... 653
Staff scheduling in the service industry considering qualifications .................................................................. 662
Impact of Gender Inequalities on Education in Morocco: An Economic Analysis of Barriers and Opportunities
.......................................................................................................................................................................... 673
SP18: Advances in Engineering: Machine Learning Techniques ..................................... 683
Classification-based Shilling Attack Detection using ODBOT ........................................................................ 683
Page 180
SEMIT 2023
Advancing Structural Integrity Assessment through Machine Learning in RC Columns ................................ 686
Fuzzy Logic Methods for Determining the Mechanical Behavior of Masonry Walls ...................................... 696
SP19: Artificial Intelligence Innovations ............................................................................. 708
Deep Learning and Probe Data for Urban Traffic Flow Prediction .................................................................. 708
Multi-Agent Deep Q-Network for Optimal Signal Control in Urban Intersections .......................................... 710
SP20: Artificial Intelligence: The Future of Everything .................................................... 712
Energy Consumption Forecasting using Time Series Analysis Methods ......................................................... 712
Implementation and Comparison of LBP Descriptor and Hough Transform applied to Facial Recognition ... 714
Multi-Objective Optimization based Efficient Healthcare Management System ............................................. 724
Exploring the Elastic, Electronic, Dynamical, and Optical properties of cubic ZrCoAs half – Heusler
Semiconductor: An ab-initio Study .................................................................................................................. 738
Performance Enhance and Management of Renewable Based Microgrid with backup Fuel Cell .................... 750
SP21: Inventory Control, Production Planning and Scheduling ...................................... 763
A New Integrated Criteria Weighting Approach for Determining Production Priority in Demand Management
.......................................................................................................................................................................... 763
Multi-Depot Vehicle Routing Problem with Drone Collaboration in Humanitarian Logistic .......................... 775
SP22: Multi-Criteria Decision-Making ............................................................................... 782
A Bibliometric Analysis of Multi-Criteria Decision-Making Techniques in Disaster Management and
Transportation in Disaster Conditions .............................................................................................................. 782
Food Distribution in Times of Famine .............................................................................................................. 802
Determining the Location for Conquest in a Computer Game Using MCDM Methods .................................. 817
Bank Performance Evaluation During Covid-19 in Terms of Customer Perspective Using Fuzzy MCDM .... 851
Supplier Selection in the Context of Industry 4.0 Using Hybrid DEA-SMART Method ................................. 860
SP23: Smart Marketing ........................................................................................................ 872
Investigating the social and conceptual structure of AI in marketing: a bibliometric analysis ......................... 872
Examining the Impact of Entrepreneurial Culture on SME Performance: The Role of Marketing Capabilities
and Social Media .............................................................................................................................................. 883
E-Shopping Sites Preference Analysis with Moosra Method ........................................................................... 897
Category Sort wıth TODIM Method, an Applıcatıon for Retaıl Sector E-Commerce Sıte ............................... 901
SP24: Advancement and Future of Technologies ............................................................... 905
Dissecting the Economic Feasibility and Life Cycle Assessment of Battery Electric and Internal Combustion
Engine Vehicles: A Case Study of India ........................................................................................................... 905
Compensatio ..................................................................................................................................................... 921
Numerical Investigation of The Effect Of The Use Of Multiple Propeller Systems On Flight Performance In
Unmanned Aerial Vehicles. .............................................................................................................................. 933
A Study on the Physical Properties of Elaborated Graphene ............................................................................ 945
Important Note: This book consists of all papers of SEMIT 2023, excluding the papers going to be
published by Springer. Those papers will be published in a booklet by Springer.
Page 181
SEMIT 2023
SP01: Artificial Intelligence and Expert Systems
Classification of Monkeypox and HPV with Supervised Machine Learning Algorithms using k-
mer
Hasret Pınar Tipioğlu1, * and Hilal Arslan2
1
Ankara University, Biology Department,
2
Ankara Yıldırım Beyazıt University, Software Engineering Department
*
Correspondence: hasretpinar@gmail.com
Abstract
According to the data of the World Health Organization, apart from Africa where it is endemic,
monkeypox cases have been seen widely in Europe and America since 2022. Monkeypox is sometimes
confused with warts because of the similarity of symptoms. For the correct diagnosis of the disease,
virus DNA is detected by molecular methods. For this purpose, in this study, HPV DNA causing warts
and monkeypox virus DNA causing monkeypox were tried to be classified by machine learning
algorithms. HPV DNA sequences compared in the study were obtained from NCBI, and monkey pox
virus DNA sequences were obtained from GISAID. The features used to train the models are k-mers
(2-mer and 3-mer) obtained from the DNA sequence of viruses. Then, classification was made with
MATLAB and supervised learning algorithms. According to the classification results, an accuracy of
99% and above was obtained. As a result, these two diseases were successfully classified using virus
DNAs.
Keywords: monkeypox, HPV, machine learning, k-mer
Introduction
Monkeypox (Mpox) is a viral illness caused by the monkeypox virus. Mpox be transmitted to humans
through physical contact with someone who is infectious, with contaminated materials, or with infected
animals.
Mpox was identified in monkeys from Singapore in Denmark in 1958 [1]. Mpox was first seen in
humans in 1970 in the Democratic Republic of the Congo in a 9-month-old baby. Thereafter, it
continued to occur in humans, mostly endemic in Africa [2]. On 2022, the World Health Organization
declared that Mpox has spread in more than one country outside of Africa (Canada, Australia, Israel,
and the United Arab Emirates) and a global outbreak occurred [3]. Mpox causes signs and symptoms
which usually begin within a week but can start 1–21 days after exposure. Symptoms are such as fever,
malaise, fatigue, headache, muscle aches, back pain, low energy, rash, and swollen lymph nodes
typically last 2–4 weeks, but may last longer in someone with a weakened immune system [3]. Mpox
causes a rash with lesions that might be described as bumps or pustules. A Mpox rash can resemble
chickenpox, herpes, warts or syphilis.
Mpox is sometimes difficult to diagnose because of its similarity to chickenpox, herpes, warts or
syphilis [4]. Correct diagnosis is important and necessary for the patient to start treatment as soon as
possible. For a correct diagnosis, a specialist doctor's examination is required. Sometimes, even a
specialist doctor's examination may be confused with other diseases and may not be enough to make
the diagnosis correctly. In order to avoid such confusion, it may be necessary to get help from machines
Page 182
SEMIT 2023
as well as human factors for correct diagnosis. DNA analysis, which is the gold standard method, is
used especially for the correct diagnosis of viruses. For DNA analysis in viruses, real-time quantitative
PCR (qPCR) is used to search for the virus to be detected in individuals at risk of infection through
genomic amplification [5].
Mpox is a DNA virus belonging to the Orthopoxvirus group of the family Poxviridae, which causes
Mpox and is closely related to smallpox [6]. The formation of warts is caused by Human papillomavirus
(HPV). HPV is a DNA virus belonging to the family papillomaviridae [7]. Like Mpox, HPV DNA is
tried to be detected by using qPCR in the diagnosis of HPV. Thus, viruses are diagnosed by their DNA.
Today, artificial intelligence algorithms are used in bioinformatics tools and methods developed for
genome classification. Machine learning is a branch of artificial intelligence that enables algorithms to
learn and build models from previous observations to make predictions about new independent data.
Because machine learning can learn from very complex datasets, it is increasingly applied in the natural
sciences, including bioinformatics for metagenomic sequenced data [8].
Therefore, in the study, classification was made with supervised machine learning algorithms to
distinguish between HPV16 (a type of HPV causes cervical cancer) genome sequences and Mpox
genome sequences in order to distinguish warts that are likely to be confused with Mpox in the clinic.
Complete genome sequences of Mpox and HPV obtained from open sources.
Thus, doctors will prevent the confusion of Mpox with warts, and patients will not lose time in terms
of treatment and diagnosis.
Chen et al. (2021) evaluated 640 PVs (Papiloma virus) using a combination of 3-mer (trimer)
distributions, genetic measurements, and phylogenetic algorithms. They found that HPVs that share a
single host (human) often have similar trimer distributions. As a result of the analysis, HPV types were
detected with high accuracy [9].
In other study, a new model was developed to evaluate the HPV16 lineage using machine learning tools.
A total of 645 HPV 16 genomes were analyzed using the Genome-Wide Association Study (GWAS),
which identified 56 lineage-specific Single Nucleotide Polymorphisms (SNPs). From the SNPs found,
training-test models were created using different algorithms such as Random Forest (RF), Support
Vector Machine (SVM) and K-nearest neighbor (KNN). For validation, a different set of HPV 16
sequences (n = 1,028), whose lineage was predetermined by Maximum Likelihood Tree (MLT), was
used. The RF-based model showed an accuracy of 99.5% in known strain samples, allowing precise
identification of the HPV 16 lineage. In addition, the RF model identified the ancestry of 273 samples
that MLT could not identify. It has also been shown in the study that the RF-based model is almost 40
times faster than the MLT [10].
Tian et al., a deep learning-based method was used to detect the HPV integration region. A total of 3608
known HPV integration sites were used to train the model, and 584 revised HPV integration sites were
used as the test dataset [11].
In a thesis study, machine learning was used for viral metagenome classification. Previously sequenced
viral genomes were used in the thesis. In this thesis, Random Forest, Feed-Forward Neural Networks
and Convolutional Neural Networks are used to create binary classifiers (virus/non-virus). An algorithm
based on Hidden Markov Models has also been used to identify and classify highly different viral
sequences in different viral families [12].
Raju et al. (2021), a hierarchical classification was made by developing a machine learning architecture
called VirusTaxo. VirusTaxo taxonomically classifies DNA and RNA viruses using the genome
sequence. VirusTaxo extracts k-mers from the genome sequence and determines the order, family, and
genus of a virus. The average accuracy of VirusTaxo for DNA viruses is 99% (order), 98% (family)
and 95% (strain), and 97% (order), 96% (family) and 82% (strain) for RNA viruses [13].
Alakus and Baykara (2022) classified HPV and Mpox DNAs with deep learning algorithm. An average
of 96.08% accuracy and 99.83% F1 score were obtained in the study [14]. Thus, these two diseases
have been effectively classified according to their DNA sequences. In this study, classifier models with
Page 183
SEMIT 2023
different structures were developed using the machine learning toolbox of the MATLAB programming
language.
Data and Methods

Viruses contain either DNA or RNA as genetic material. Viruses are separated DNA or RNA viruses
according to the genetic material they contain. Viruses are classified according to the taxonomic
ordering necessary to find their genetic relationship, ancestry, and origin. Taxonomic classification is
done for the correct classification of a virus. Traditionally, in classification; various phenotypic features
of viruses such as molecular composition, structure, proteins, host and pathogenicity are used [15]. With
the development of sequencing technologies, more viruses are classified according to their genomic
sequences rather than their phenotypic characteristics. Comparison of virus sequences using binary
sequence similarity and phylogenetic relationships, taxonomic analysis of new viruses [16].
In computational methods used for virus taxonomy, generally; It is based on genome structure similarity
and the homologous genes [17]. However, the homology-based methods used require high
computational processes and usually expert interpretation.
Three main approaches are used to classify viruses according to their genomic sequence structures. The
first is based on commonly used sequence alignment; similarity search methods and binary distance
based methods. The second is the phylogenetic-based approach. When classifying a new sequence here,
it is realigned with the existing set of reference sequences, and as a result, either a new phylogenetic
tree is extracted or the given sequence is placed in the existing tree. The third is the alignment-free
approach, which includes methods based on nucleotide correlations and sequence composition. It
converts genomic sequences or certain parts of these sequences into feature vectors and processes
accordingly [18].
In this study we conduct the k-mer methods which is the alignment-free approach. k-mers are short
nucleotide sequences of k length in a genome sequence.
For this purpose, the number of 2-mer and 3-mer in the whole genome sequence of each virus was
determined. These determined numbers were divided by the total number of nucleotides in the complete
genome of each virus and these values were used as features to be used for machine learning algorithms.
Dataset is labeled 1 for Mpox and 2 for HPVs.
2-mers: AA, AC, AG, AT, CA, CC, CG, CT, GA, GC, GG, GT, TA, TC, TG, TT
3-mers: AAA, AAC, AAG, AAT, ACA, ACC, ACG, ACT, AGA, AGC, AGG, AGT, ATA, ATC,
ATG, ATT, CAA, CAC, CAG, CAT, CCA, CCC, CCG, CCT, CGA, CGC, CGG, CGT, CTA, CTC,
CTG, CTT, GAA, GAC, GAG, GAT, GCA, GCC, GCG, GCT, GGA, GGC, GGG, GGT, GTA, GTC,
GTG, GTT, TAA, TAC, TAG, TAT, TCA, TCC, TCG, TCT, TGA, TGC, TGG, TGT, TTA, TTC,
TTG, TTT
HPV16 complete genome sequences for HPV were used and downloaded from NCBI's own website,
which is open source. 623 HPV 16 complete genome sequences were used. For Mpox, complete genome
sequences were used and downloaded from GISAID's own website, which is open source. Sequences
of 619 humanMpox complete genomes were used. The complete genome sequences of both virus types
are in fasta format.
Machine Learning Algorithms

Machine learning is actually the imitation of human learning ability by computers. In doing so,
algorithms used in machine learning make use of various mathematical equations. Algorithms learn
from the available data and infer the information to be predicted. Algorithms are used for many
operations such as classification and clustering, and they help to obtain accurate and fast inferences.
The purpose of classification; It is to ensure that the data at hand are assigned to different categories
(classes). Classification, also called pattern recognition, is one of the most well-known processes of
Page 184
SEMIT 2023
machine learning. Supervised learning algorithms are used for classification. Some of the machine
learning algorithms used for classification are; DT, ANNs, SVM is KNN [19, 20].
Decisicion Tree, the algorithm starts from the root node of the tree to predict the given class. This
algorithm compares the values of the root property with the record (real dataset) property and according
to the comparison it follows the branch and jumps to the next node. For the next node, the algorithm
again compares the attribute value with other child nodes. The process continues until it reaches the leaf
node of the tree [21].
Artifical Neural Network (ANN): The working principle of artificial neural networks is similar to
neurons in the human brain. It consists of neurons and their connections, as in the human brain. Artificial
neural networks are structures formed by connecting artificial nerve cells to each other. Artificial neural
networks are examined in three main layers; Input Layer, Intermediate (Hidden) Layers and Output
Layer.
Information is transmitted to the network through the input layer. They are processed in the intermediate
layers and sent from there to the output layer. What is meant by information processing is to convert
the information coming to the network into output by using the weight values of the network. In order
for the network to produce the correct outputs for the inputs, the weights must have the correct values.
The basic operation in Artificial Neural Networks; It is to calculate the parameters w (weight parameter)
and b (bias value) that the model will give the best score. Each neuron is calculated in the same way
and they are connected to each other in series or in parallel.
An artificial neuron consists of five parts;
1. Inputs: Inputs are data coming to neurons. Data from these inputs are sent to the neuron nucleus to
be collected, as in biological nerve cells.
2. Weights: The information coming to the artificial nerve cell is transmitted to the nucleus by
multiplying the weight of the connections they come from before reaching the nucleus through the
inputs. In this way, the effect of the inputs on the output to be produced can be adjusted.
3. Addition Function (Combination Function): Addition function is a function that calculates the net
input of an artificial neuron cell by adding the inputs multiplied by weights.
4. Activation function: A function that takes the weighted sum of all the inputs in the previous layer
and then generates an output value (typically non-linear) and passes it to the next layer. (e.g. ReLU or
sigmoid).
5. Outputs: The value output from the activation function becomes the output value of the cell. Although
each cell has multiple inputs, it has only one output. This output can be linked to any number of cells
[22].
K Nearest Neighbor (KNN) is one of the most used classification methods among supervised learning
methods. It is a non-parametric algorithm. For each sample, it searches for the nearest neighborhood
among all samples in the data set. And it requires a lot of memory space. The necessity of cost and
processing load that is directly proportional to the size of the data set is one of its disadvantages. KNN
algorithm steps:
1. A k neighborhood number and a distance metric are determined.
2. To classify the sample has k nearest neighbors.
3. Assignment is made to the class with the highest frequency among the classes of the k nearest
neighbors.
The distance measures used in the proximity calculation in the KNN algorithm for continuous variables
are Euclid, Manhattan and Minkowski. Euclid and Manhattan distance measures are a special case of
the Minkowski distance measure given in Equation (2). While the expression in equation (2)
corresponds to the Manhattan distance measure for p=1, it corresponds to the Euclidean distance
measure for p=2 [19, 20].
Page 185
SEMIT 2023
(𝑥, 𝑦) = (∑𝑛𝑖=1 |xi − yi |𝑝 ) 1/𝑝 (2)

Support Vector Machines (SVM) are based on statistical learning theory. High accuracy rate and no
overfitting problem are among the advantages. SVM is divided into linear and nonlinear. While SVM
tries to minimize the classification error, it also tries to maximize the margin to prevent over-learning.
These two objectives are controlled by the C parameter. The larger the C, the greater the importance
given to the classification error. 𝑥𝑖 data point, 𝑦𝑖 class information, 𝜉𝑖 the mathematical programming
model that emerges in Equation (3) is given to show the idle variable that will take a positive value if
the point i is on the wrong side of the hyperplane, the intersection point b and the normal vector of the
hyperplane 𝑤 [19, 20].
1
𝑚𝑖𝑛𝑖𝑚𝑖𝑧𝑒 2 ‖𝑤‖2 + C(∑i 𝜉 𝑖 ) (3)
s.t.
𝑤0 +𝑤 𝑇𝑥 𝑖 − 𝑏 ≥ 1− , 𝑖𝑓 𝑦 𝑖 = 1
𝑤0 +𝑤 𝑇𝑥 𝑖 − 𝑏 ≤ −1+ , 𝑖𝑓 𝑦 𝑖 = -1
In this study, classifier models with different algorithms were used from the MATLAB machine
learning toolbox. These models are the methods of Neural network, NB classifier, KNN algorithm, DT
algorithms, and SVM, whose details are given below. All numerical results are obtained using
MATLAB R2022b on an Intel processor under Windows 10 operating system.
Experimental result
HPV and Mpox complete genome sequences were classified using Matlab machine learning toolbox
algorithms. For this purpose, Fine Tree, Medium Tree, Coarse Tree, Linear Discriminant, Logistic
Regression, Gaussian Naive Bayes, Kernel Naive Bayes, Linear SVM, Quadratic SVM, CubicSVM,
Fine Gauss SVM, Medium Gaussian SVM, Coarse Gaussian SVM, Fine KNN, Medium KNN , Coarse
KNN, Cosine KNN, Cubic KNN, Weighted KNN, Subspace KNN, Boosted Trees, Bagged Trees,
Subspace Discriminant, Subspace KNN, Subspace Discriminant, Narrow Neural Network, Medium
Neural Network, Wide Neural Network, Bilayered Neural Network, Trilayered Neural Network, SVM
Kernel and Logistic Regression Kernel algorithms were used.
For algorithm validation, 10-fold cross validation was applied. Algorithms for 2-mer and 3-mers were
applied separately. Validation accuracy results are given in the Table 1 below.
Table 1. Machine learning algorithms for classification HPV and Mpox with 2-mer (Accuracy, F1-
Score, Precision, Recall)
2-mer
Model Type Accuracy F1-Score Precision Recall
Decision Tree 99,92 0,998 0,998 0,999
Logistic Regression 100,00 1 1 1
Naive Bayes 100,00 1 1 1
SVM 100,00 1 1 1
KNN 100,00 1 1 1
ANN 100,00 1 1 1
Page 186
SEMIT 2023
Table 2. Machine learning algorithms for classification HPV and Mpox with 3-mer (Accuracy, F1-
Score, Precision, Recall)
3-mer
Model Type Accuracy F1-Score Precision Recall
Decision Tree 100,00 1 1 1
Logistic Regression 100,00 1 1 1
Naive Bayes 99,88 0,997 1 0,995
SVM 100,00 1 1 1
KNN 100,00 1 1 1
ANN 100,00 1 1 1
Conclusion
In this study, complete DNA genome sequence of viruses that cause Mpox and warts were used. The
purpose is distinction was made between these two similar diseases by classifying the DNA genomes.
Various classification algorithms in machine learning using Matlab, were applied to our data set. For
determinig features, the number of 2-mer and 3-mers are subdivided into each virus genome length. For
2-mers, accuracy is provided to be 99% and above. Again for 3-mers, accuracy was obtained as 99%and
above, mostly 100%. When we compare 2-mers and 3-mers, it is seen that 3-mers give higher accuracy
results.
In this study, it has been shown that HPV and Mpox viruses are separated from each other with their
complete DNA genomes with high accuracy by machine learning methods. Machine learning methods
for virus classification are a promising bioinformatics tool in future studies for this purpose.
References
1. Magnus, P.V.; Andersen, E.K.; Petersen, K.B.; Birch-Andersen, A. A pox-like disease in cynomolgus
monkeys. Acta Pathol. Microbiol. Scand. 1959, 46, 156–176. [CrossRef]
2. Durski, K.N.; McCollum, A.M.; Nakazawa, Y.; Petersen, B.W.; Reynolds, M.G.; Briand, S.;
Djingarey, M.H.; Olson, V.; Damon, I.K.; Khalakdina, A. Emergence of Monkeypox-West and Central
Africa, 1970–2017. MMWR Morb. Mortal. Wkly. Rep. 2018, 67, 306–310. [CrossRef] [PubMed]
3. World Health Organization. Available online: https://www.who.int/news-room/fact-
sheets/detail/monkeypox ((accessed on 13 August 2023 ).
4. Centers for Disease Control and Prevention. Intervention Services for People with or Exposed to
Monkeypox. Available online: https://www.cdc.gov/poxvirus/monkeypox/health-
departments/intervention-services.html (accessed on 16 August 2023).
5. Argüelles, M. E. Á., Martínez, Z. P., Alba, S. R., Alba, J. M. G., Fernandez-Verdugo, A. M.,
González, I. C., ... & García, S. M. (2023). Detecting, Quantifying, and Isolating Monkeypox Virus in
Suspected Cases, Spain. Emerging Infectious Diseases, 29(7), 1465.
6. Di Gennaro, F.; Veronese, N.; Marotta, C.; Shin, J.I.; Koyanagi, A.; Silenzi, A.; Antunes, M.;
Saracino, A.; Bavaro, D.F.; Soysal, P.; et al. Human Monkeypox: A Comprehensive Narrative Review
and Analysis of the Public Health Implications. Microorganisms 2022, 10, 1633. [CrossRef] [PubMed]
7. Alp Avcı, G., & Bozdayı, G. (2013). İnsan papilloma virüsü. Kafkas J Med Sci 2013; 3(3):136–144
8. Ren, J., Song, K., Deng, C., Ahlgren, N. A., Fuhrman, J. A., Li, Y., Xie, X., Poplin, R. and Sun, F.
2020. Identifying viruses from metagenomic data using deep learning. Quantitative Biology, 8(1), 64-
77.
9. Chen, Z., Utro, F., Platt, D., DeSalle, R., Parida, L., Chan, P. K., & Burk, R. D. (2021). K-mer
analyses reveal different evolutionary histories of alpha, beta, and gamma
papillomaviruses. International Journal of Molecular Sciences, 22(17), 9657.
10. Asensio-Puig, L., Alemany, L. and Pavon, M. A. 2022. A Straightforward HPV16 Lineage
Classification Based on Machine Learning. Frontiers in artificial intelligence, 5.
Page 187
SEMIT 2023
11. Tian, R., Zhou, P., Li, M., Tan, J., Cui, Z., Xu, W., Wei, J., Zhu, J., Jin, Z., Cao, C., Fan, W., Xie,
W., Huang, Z., Xie, H., You, Z., Niu, G., Wu, C., Guo, X.,Weng, X., Tian, X., Yu, F. and Hu, Z. 2021.
DeepHPV: a deep learning model to predict human papillomavirus integration sites. Briefings in
Bioinformatics, 22(4), bbaa242.
12. Bzhalava, Z. 2020. Machine Learning and Data-Parallel Processing for Viral Metagenomics,
(Thesis for Doctoral Degree) Department of Laboratory Medicine Karolinska Institutet, 58
13. Raju, R. S., Nahid, A. A., Shuvo, P., & Islam, R. (2021). VirusTaxo: Taxonomic classification of
virus genome using multi-class hierarchical classification by k-mer enrichment. bioRxiv, 2021-04.
14. Alakus, T. B., & Baykara, M. (2022). Comparison of Monkeypox and wart DNA sequences with
deep learning model. Applied Sciences, 12(20), 10216.
15. Simmonds, P., Adams, M. J., Benkő, M., Breitbart, M., Brister, J. R., Carstens, E. B., ... & Zerbini,
F. M. (2017). Virus taxonomy in the age of metagenomics. Nature Reviews Microbiology, 15(3), 161-
168.
16. Raju, R. S., Al Nahid, A., Dev, P. C., & Islam, R. (2022). VirusTaxo: Taxonomic classification of
viruses from the genome sequence using k-mer enrichment. Genomics, 114(4), 110414.
17. Dougan, T. J., & Quake, S. R. (2019). Viral taxonomy derived from evolutionary genome
relationships. PLoS One, 14(8), e0220440.
18. Remita, M. A., Halioui, A., Malick Diouara, A. A., Daigle, B., Kiani, G., & Diallo, A. B. (2017). A
machine learning approach for viral genome classification. BMC bioinformatics, 18, 1-11.
19. Akgül, G., Çelik, A. A., AYDIN, Z. E., & ÖZTÜRK, Z. K. (2020). Hipotiroidi Hastalığı Teşhisinde
Sınıflandırma Algoritmalarının Kullanımı. Bilişim Teknolojileri Dergisi, 13(3), 255-268.
20. Özbay Karakuş, M., & Er, O. (2022). A comparative study on prediction of survival event of heart
failure patients using machine learning algorithms. Neural Computing and Applications, 34(16),
13895-13908.
21. Available online https://www.javatpoint.com/machine-learning-decision-tree-classification-
algorithm (accessed on 17 August 2023).
22. Available online: https://www.veribilimiokulu.com/yapay-sinir-agiartificial-neural-network-nedir/
(accessed on 17 August 2023).
Page 188
SEMIT 2023
Yapay Zekâ ile B2B Kampanya ve Sadakat Yönetimi

B2B Campaign and Loyalty Management with Artificial Intelligence
Osman Çaylı Ve Emre Karadağ

VBT YAZILIM AR-GE
VBT YAZILIM A.Ş.
İstanbul, Türkiye
Osman.Cayli@vbt.com.tr, Emre.Karadag@vbt.com.tr
Atınç Yılmaz
Bilgisayar Mühendisliği
İstanbul Beykent Üniversitesi
İstanbul, Türkiye
atincyilmaz@beykent.edu.tr
Özetçe
E-ticaret kavramı bir işletmenin mal veya hizmetlerini internet veya farklı diğer elektronik ortamlar
üzerinden sunduğu ticaret modelidir. Bu modelde, şirketten şirkete (B2B) alanında da oldukça
gelişmeler görülmektedir. Son dönemlerde oyunlaştırma stratejileri e-ticaret alanında uygulanmaktadır.
Bu sayede müşteri bağımlığı, çalışan motivasyonu gibi kavramlarda artış sağlanmaktadır. Günümüz
rekabetçi ortamında yapay zeka teknolojileri uygulandığı her sektörde avantaj getirebilmektedir. Bu
çalışmada, yapay zeka yöntemlerinden karar ağacı yöntemi kullanılarak B2B alanında distribütörlerin
sadakat programlarında etkinliğinin arttırılması amacı ile uygulama geliştirilmiştir. Çalışma
neticesinde, yapay zeka yöntemleri desteği ile kurgulanan müşteri sadakat programı ile işletmelerin
rekabet avantajı elde edeceği gösterilmiştir. Yapay zeka yöntemleri ile modellenen program ile müşteri
memnuniyeti ve deneyimi arttırılarak sürdürülebilirlik de sağlanacağı ortaya konulmuştur. Bulgular
neticesinde probleme uygulanan modeller arasında en yüksek doğruluğa %78 oranı ile rastgele orman
yöntemi uygulanarak elde edilmiştir.
Anahtar Kelimeler: Yapay zeka, karar ağacı, rastgele orman, tahminleme
Abstract
The concept of e-commerce is a business model in which a business offers its goods or services over
the internet or other electronic media. Significant developments are devoloped in the field of business
to business (B2B). Recently, gamification strategies have been applied in the field of e-commerce. In
this way, concepts such as customer dependency and employee motivation are increased. In today's
competitive environment, artificial intelligence technologies can bring advantages in every sector where
it is applied. In this study, an application has been developed with the aim of increasing the efficiency
of distributors in loyalty programs in the B2B field by using decision tree algorithm which is one of the
artificial intelligence method. As a result of the study, it has been shown that businesses will gain
competitive advantage with the customer loyalty program designed with the support of artificial
intelligence methods. With the program modeled with artificial intelligence methods, it has been
revealed that sustainability will be ensured by increasing customer satisfaction and experience. As a
result of the findings, the highest accuracy among the models applied to the problem was obtained by
applying the random forest method with a rate of 78%.
Keywords: Artificial intelligence, decision tree, random forest, prediction
Giriş (Introductıon)
E-ticaret, elektronik ortamlar üzerinden mal veya hizmet alım satımı için yapılan ticari faaliyetlerdir.
2000’li yılların başından itibaren internetin de kullanımının artması ile birlikte internet ekonomisi ve e-
ticaret kavramları yaygınlaşmaya başlamıştır [1]. İnternetin sunmuş olduğu erişebilirlik, kolaylık ve
özgürlük ile, zamandan ve coğrafi alandan bağımsız olarak alışveriş imkânı kullanıcılara ve ticaret
Page 189
SEMIT 2023
yapan firmalara sunulmuştur. E-ticaretin en büyük avantajları aşağıdaki şekilde listelenebilir:

• Global erişim
• Geniş müşteri kitlesi
• Düşük firma maliyetleri
• Kolaylık
• Hız
• Coğrafi ve zamandan bağımsızlık
Birden fazla işletmenin birbiri ile yaptığı ticari faaliyetlere şirketten şirkete (B2B) ticaret denmektedir.
Bu ticaret hareketleri işletmeler arasında mal veya hizmet alımının dışında birbirleri arasında kurulan
iş birliklerini de kapsamaktadır. B2B alanındaki aktörler toptancılar, distribütörler, üreticiler, kurumsal
şirketler olabilmektedir. İşletmelerin müşteri sadakatini oluşturması, iş süreçlerini güçlendirmek,
hizmet veya ürünlerinin satışlarının arttırmak, alımlara teşvik etmek ve rol alan çalışanların
motivasyonunu arttırmak için oyunlaştırma stratejileri uygulanmaktadır. B2B alanında oyunlaştırma,
distribütörlere, iç müşterileri, iş ortaklarına yapılabilmektedir. Bu sayede sadakatin arttırımı ile
süreklilik sağlanarak; hizmet veya ürünlerin alımlarının teşvik edilmesi dolayısıyla satış süreçlerinin
veriminin artması sağlanabilmektedir.
Bayii satış sistemleri, ürünlerin toptan satışının veya dağıtımının yapılması için bayi ile şirket arasında
süreçlerin yürütülmesini ve yönetilmesini sağlayan ticari yazılımlardır. Bu sistemler sayesinde coğrafi
sınırlılık ortadan kaldırılarak şirketin ürünlerinin satışını yapabilmesi için müşterilere kolayca ve çok
daha az maliyet ile ulaşması sağlanmaktadır. Bayii satış sistemleri ile stok yönetimi, sipariş yönetimi,
fiyatlandırma yönetimi, raporlama, müşteri yönetimi ve süreç optimizasyonu yapılabilmektedir.
Kampanya yönetimi ile hem şirket ile bayii arasında hem de bayii aracılığı ile şirketin son müşteriye
ürünlerini daha etkili, hızlı şekilde ulaşıp tanıtması sağlanmakta; bunun neticesinde de şirketin
karlılığının ve veriminin artışı elde edilebilmektedir. Şirketler, ürünlerinin satışının arttırılması için
müşterilere kişiselleştirilmiş fiyatlandırma, hediye veya farklı teklifler sunabilir. Ayrıca tekrarlayan
alışverişlerin olması durumunda özel kampanya tanımı veya indirim gibi davranışlar ile müşterilerin
şirkete aidiyetinin artması sağlanabilir. Kampanya yönetimi süreci dikkatli şekilde analiz edilmesi
gereken, organizasyonun doğru şekilde planlanmasından sonra iyi şekilde izlenmesine ihtiyaç duyulan
bir süreçtir. Doğru planlama ile sonuçlanan kampanya yönetimi, müşteri sadakatinin etkin şekilde
arttırılmasını sağlayarak süreçte yer alan tüm aktörlerin memnuniyeti değer olarak
kazandırabilmektedir.
Yapay zekâ, rasyonel davranışların benzetimini gerçekleştiren yöntemler topluluğudur. Bu
algoritmaların kullanımı ile geçmiş deneyimlerden faydalanarak daha önce karşılaşılmayan durumlar
için deneyimlerinden elde ettiği öğrenme ile genelleme yaparak, öngörüde bulunma, karar verme,
keşfedilmemiş bilgiye ulaşabilme durumları sağlayabilmektedir. Kurumlar için yeni müşteri kazanımı,
mevcut müşteriyi elinde tutmaktan çok daha maliyetli olmaktadır. Bu nedenle, rekabetçi bir pazarda,
müşteri sadakat programı kurumlar için oldukça önemli bir konumdadır. Yapay zeka yöntemlerinin
uygulanarak B2B alanı ile ilgili sistemlere entegre edilmesi ile kampanyaların kişiselleştirilmesi
sağlanarak kampanyaların çok daha etkin hale getirilmesi sağlanabilmektedir. Öncelikle B2B alanında
yapay zeka algoritmalarının kullanımı şirketlerin hizmet veya ürün sunduğu müşterileri daha iyi
tanıyabilmesine olanak sunmaktadır. Müşterilerin tüm davranışları takip edilip analiz edilerek
müşteriler kategorize edilerek benzer davranıştaki müşterilere uygun kampanya tanımları yapılarak
satış hedeflerine daha hızlı şekilde ulaşılabilmektedir. Bunun yanında yapay zeka modelinin öngörüsü
ile kişiselleştirilmiş kampanya tanımı yapılarak müşteriye özel spesifik öneriler ve kişiselleştirilmiş
fiyat teklifleri oluşturulabilir. Ayrıca sunulan hizmet ve ürünlerin fiyatlarını yine yapay zeka
yöntemlerinin sağladığı destek ile dinamik hale getirilebilmektedir. Bu sayede en iyi fiyatlamanın
yanında satış, kar, verimlilik ve müşteri memnuniyetinin arttırılması kazanılan değer olarak elde
edilebilmektedir. Yapay zeka yöntemlerinin B2B konusuna entegrasyonu ile kurumun karlılığı, verimi
ve müşteri elde etme oranı artışı gibi önemli sonuçlar gözlemlenerek şirkete fayda sağlanabilmektedir.
En iyi sürecin organize edilmesi için güvenilir veri ile çalışmak ve analiz süreçlerinin doğru planlanması
önemli bir kriterdir. Bununla birlikte verilerin elde edilmesi esnasında müşteri verilerinin
Page 190
SEMIT 2023
mahremiyetine dikkat edilmeli; bunun yanında siber güvenlik kavramlarının sürece dahli ile mevcut
veri ve elde edilen değerlerin korunmasının da sağlanması gereklidir.
Bu çalışmada, yapay zeka yöntemleri kullanılarak B2B alanında müşteri sadakat programlarının
geliştirilmesi amaçlanmaktadır. Bayii satış sistemleri geleneksel olarak, panelden veya kullanılan bir
ERP sistemi ile entegrasyon sağlanarak veriyi alarak hizmeti sunmak üzerine tasarlanmıştır. Önerilen
model ile, geleneksel bayi sistemlerinden farklı olarak, müşterilerin satın alma taleplerini arttırmayı
sağlayacak ürünlerin öne çıkartılması ve bunun yanında müşterilerin kurumlara bağlılığını arttırmayı
sağlayacak kişiselleştirilmiş kampanyaların oluşturulması çalışmanın yenilikçi yönleri olarak
hedeflenmektedir. Kampanya tanımları müşterilerin satın alma limitleri, yönelimlerine göre farklı
şekilde tanımlanacak; böylece oyunlaştırma mantığı ile müşterilerin hedefe ulaşma isteği ile firma
karlılığını arttıracaktır
Öncelikle çalışmada müşteri sadakat programının B2B alanındaki temel unsurları belirtilecektir. Bir
sonraki kısımda ise yapay zeka yöntemlerinden karar ağacı ve rastgele orman algoritmalarının desteği
ile müşteri davranışlarının analizi yapılacak ve müşteri sadakat programı modellemesi ortaya
konacaktır.
Benzer Çalışmalar
Literatürde, müşteri sadakat programı kapsamında yapılan güncel çalışmalar incelenmiştir.
Çalışmalarda genel olarak müşteri hareketlerinin ve eğilimlerinin tahmini için derin öğrenme ve makine
öğrenmesi algoritmalarının kullanıldığı görülmektedir. Bunun yanında çalışmaların bir kısmı mevcut
müşterilerin aidiyeti üzerine odaklanırken; diğer kısmı yeni müşteri kazanımı konusunda çalışmalarda
hedef alınmıştır.
Ayvaz ve ark. 2021 yılında yaptıkları çalışmada derin öğrenme yöntemleri kullanılarak kampanyalara
yönelik müşteri tepkilerinin tahmin edilmesi amaçlanmışlardır [2]. Uygulamada, GSM sağlayıcısından
elde edilen kampanya katılım verileri üzerinde çalışılmıştır. Otomatik olarak elde edilen özellik
tespitine göre manuel elde edilen özellik tespitlerinin daha başarılı olduğu ifade edilmiştir. Xiao ve ark.
2023 yılında yaptıkları çalışmada, e-ticaret platformlarındaki pazarlama kampanyaları ile platform
kullanıcıları arasındaki pazarlama kampanyalarının etki eğiliminin tahmini için model önerilmişlerdir
[3]. Bölgeye duyarlı karma (LSH- locality-sensitive hashing) algoritması ile tekrarlayan sinir ağı
(recurrent neural network -RNN) algoritmaları zaman serisi probleminin öğrenimi için çalışmada
kullanılmıştır. Sonuç olarak modelin bir pazarlama kampanyası bağlamında kullanıcı etkileşimi
davranışını tahmin edebildiğini ve kampanyanın etkisindeki eğilimlerdeki değişiklikleri dinamik olarak
öngörebildiğini göstermiştir. Tang ve ark. 2018 yılında yaptıkları çalışmada, pazarlama kampanyaları
kapsamında sosyal etkisini tahmin etmek için evrişimli sinir ağını kullanmıştır [4]. Kullanıcı özellikleri
ile kullanıcı potansiyelleri de çalışmada modellenirken dikkate alınmıştır. Margaris ve ark. 2016 yılında
yaptıkları çalışmada, kullanılan tavsiye algoritmasını geliştirmek için ürünlerin kalite özelliklerini ve
kullanıcılar ile farklı ürün kategorilerindeki satın alma davranışlarını etkileyen faktörleri dikkate alarak
kullanıcının duygusal etkisini tahmin ederek kullanıcı hedefli bilgilerin daha doğru bir şekilde yayılması
sağlanmıştır [5]. Modele bu özelliklerin dahil edilmesi, daha doğru tavsiyelere ve daha iyi kullanıcı
hedefli kampanyaların elde edilmesinin sağlanacağı belirtilmiştir. Bunun yanında çevrimiçi reklam
verimliliğinin de artacağı öngörülmüştür. Yaiprasert ve ark. 2023 yılında yaptıkları çalışmada, yiyecek
dağıtım işi için dijital pazarlamada makine öğrenimi algoritmalarından topluluk geliştirme
algoritmalarının kullanımına odaklanmışlardır [6]. Modelleme karar ağacı, naive bayes ve k-en yakın
komşu algoritmalarından faydalanılmıştır. Netice olarak karar ağacı ve k-en yakın komşu algoritmaları
%100 doğruluk oranına ulaşırken naive bayes yöntemi ile %97.175 doğruluk oranına ulaşılmıştır. Üç
makine öğrenimi algoritmasından oluşan modelleme ile zaman ve maliyetleri azaltarak gıda dağıtım
işinde dijital pazarlama stratejilerini başarılı bir şekilde iyileştirdiği gösterilmiştir. Van Leeuwen ve ark.
2022 yılında yaptıkları çalışmada, konaklama şirketlerinin hiyerarşik kümeleme analizi kullanılarak
algoritmik bir yaklaşım ile ham verilerden eyleme dönüştürülebilir tüm süreçteki adımları sağlamayı
amaçlamışlardır [7]. Müşteri odaklı yaklaşım ile müşteri profillerini gruplayarak pazarlama
departmanlarındaki süreçlerin ilerlemesi sağlanarak şirketin daha fazla kar elde edeceği çalışmada
öngörülmektedir. Sajjad ve ark. 2020 yılında yaptıkları çalışmada, sosyal medya platformlarından
Page 191
SEMIT 2023
Facebook üzerinden kullanıcıların sadakati ile Pazar etkileşimi ilişkilendirilmektedir [8]. Pazar
bağlılığı, yenilikçi bağlılık, işlevsel bağlılık, duygusal bağlılık ve toplumsal bağlılık olmak üzere dört
parametre aracılığıyla ölçülmüştür. Sosyal medya üzerinden temin edilen 309 yanıt ilk olarak basit
regresyon yöntemi ile modellenmiştir. Sonuç olarak modelin istatistiksel olarak anlamlı olduğu ve
müşteri sadakatinin yüzde 65’inin sağlandığı ifade edilmiştir. Ardından hiyerarşik regresyon ile tekrar
modelleme yapılmıştır. Sun ve ark. 2023 yılında yaptıkları çalışmada müşteri yaşam döngüsü değeri
teorisine dayalı olarak müşteri değeri ölçümü ve müşteri segmentasyonuna odaklanmışlardır [9]. İlk
olarak veri seçimi, veri ön işleme, veri dönüşümü ve bilgi keşfi gibi adımlar yapıldıktan sonra makine
öğrenimi algoritmaları ile müşteri ilişkileri yönetimi analiz modellerine dayalı olarak müşteri değeri
segmentasyonu uygulanmıştır. Son olarak ise gerçek bir e-ticaret platformunun gerçek müşteri işlem
verileri kullanılarak çalışmada önerilen müşteri segmentasyonu yönteminin geçerliliği ve
uygulanabilirliği doğrulanmıştır. Hsu ve ark. Çalışmalarında, kullanıcıların müşteri hizmetleri sohbet
botlarına olan memnuniyetini ve sadakatini tahmin etmek için sohbete dayalı yapay zeka modellemesi
önermişlerdir [10]. Önerilen model, müşteri hizmetleri chatbot’larına ilişkin algıları hakkında yanıt
veren 219 kullanıcıdan toplanan anket verileri kullanılarak değerlendirilmiştir. Bulgular, AI chatbot
hizmet kurtarma kalitesinin ve AI chatbot konuşma kalitesinin kullanıcı memnuniyetini önemli ölçüde
etkilediğini göstermektedir. Smaili ve ark. 2023 yılında ortaya koydukları çalışmada, belirli bir müşteri
tarafından satın alınan ürünlerin çeşitlendirilmesi düşünülerek yeni bir modelleme önermişlerdir [11].
Yenilik, frekans, Para ve çeşitlilik (Recency, Frequency, Monetary, Diversity – RFMD)’e
segmentasyon, bir müşterinin davranış kalıplarını tespit etmek için perakende pazarında uygulanmıştır.
Önerilen model, müşteri davranışının tahmin kalitesini arttırarak şirket tarafında olumlu yanıt
alınabilecek müşterilere yönelimi sağlayabilmektedir.
Yapay Zeka Yöntemleri

Yapay zeka, veri analizi, karar verme, dil anlama, örüntü tanıma, öğrenme gibi kompleks problemlerin
çözümü için kullanılan; son dönemlerde oldukça popüler olan yöntemler topluluğudur. En önemli güçlü
yönlerinden biri öğrenme yeteneğidir. Genel anlamda insan davranışlarının matematiksel olarak
benzetimi şeklinde tanılandırılabilir.
Yapay zekanın alt dalı olan makine öğrenmesi algoritmaları ise verilerden ilişki çıkartarak genelleme
yapabilme yeteneğine sahip olmaktadır. Makine öğrenmesi yöntemlerinin temel hedefi insan dahli
olmadan öğrenme kabiliyetini elde etmesidir. Öğrenme sürecinde, modellenmesi gereken konu ile ilgili
olan ve geçmişteki vakaları içerisinde barındıran veri setinden deneyim ve tecrübe elde edilmektedir.
Karar Ağaçları
Makine öğrenmesi yapay zeka alanında yer alan disiplinlerden biridir. Karar ağacı algoritmaları ise
karar bilgisi ve sonuç arasında ilişkiyi kurgulayarak modelleyen makine öğrenmesi yöntemidir. Karar
ağaçları hiyerarşik bir ağaç topolojisine sahiptir. Karar noktaları düğümleri oluştururken, alt kümelere
hiyerarşik olarak ayrılarak son düğümde sonuca ulaşılmaktadır. Düğümlerin öncelikleri bilgi kazancı
mantığına dayanarak belirlenmektedir. En fazla bilgi kazancının elde edildiği düğüm kök düğüm olarak
en önemli özellik olmaktadır.
Karar ağacı algoritmaları sınıflandırma ve regresyon problemlerinde sık tercih edilen yöntemlerdendir.
Bunun yanında graf yapısında olması, kolaylıkla modellenebilmesi ve görsel olarak probleme ait
ilişkileri görselleştirebilmesi sayesinde karar verme ile ilgili problemler için oldukça uygun ve iyi sonuç
üreten algoritmalardan biridir. Aşırı uyum gibi çok karşılaşılan problemlerin çözümü için ağacın mimari
derinliği, durma kriteri gibi değerlerin parametrik olarak uygun şekilde seçilmesi gerekmektedir.
Karar ağaçları algoritması müşteri sadakati ve kampanya yönetimi problemi için uygulanabilecek etkili
bir yöntemdir. Modelleme için kullanılacak olan veriler hazırlandıktan sonra karar ağacının girdileri
belirlenmesi gerekmektedir. Seçilen eğitim seti verileri ile karar ağacı modeli oluşturulurken; test
setinde bulunan veriler ile modelin güvenilirliği sınanmaktadır. Karar ağaçları yöntemleri ile
kurgulanan model ilgili zaman için etkili sonuçlar verse de modelin sürekli olarak revize edilerek güncel
halde kalması sağlanmalıdır.
Page 192
SEMIT 2023
Rastgele Orman
Rastgele Orman yöntemi topluluk tabanlı ve ağaç yapısı mimarisine sahip makine öğrenmesi
yöntemidir. Genelleme yeteneğini arttırabilmek ve daha güçlü karar verme yeteneği elde edebilmek için
birden fazla karar ağacının birleştirilmesi ile rastgele orman mimarisi meydana gelmektedir. Rastgele
orman yönteminde çalışma adımı olarak eğitim verilerini kullanarak rastgele alt kümeler üzerinden
karar ağacının meydana gelmesi sağlanmaktadır. Birden fazla oluşturulan karar ağacı sayesinde farklı
özelliklerin öğrenilmesi sağlandığından ötürü daha güçlü bir yapıya sahip olmaktadır. Tahminleme
safhasında ise meydana getirilen tüm karar ağaçları üzerinden farklı tahminler elde edilmektedir.
Sonuca karar verilirken en fazla sonucun elde edildiği sınıf üzerinden tahmin kararı rastgele orman
yöntemi tarafından verilmektedir. Regresyon problemlerinde ise sayısal sonuçların ortalaması alınarak
tahmin yapılmaktadır. Rastgele orman yöntemi kullanımı ile parametrelerin önem derecesinin
değerlendirilmesi, ezberleme önlemi gibi avantajlar elde edilerek etkin bir modelleme
yapılabilmektedir.
Rastgele orman yönteminin müşteri sadakati ve kampanya yönetimi konusunda uygulanması etkin bir
sonuç elde edilmesine yardımcı olacaktır. Mevcut olarak bulunan veri içerisinde yer alan müşteri
özellikleri, müşteri davranışları, satın alımlar, kampanya kullanımları gibi özellikler işlenerek
genelleme yeteneğini iyi seviyeye getirilmesi sağlanabilecektir. Bununla birlikte sistemin öngörü gücü
de etkinleşecektir. Rastgele orman yönteminin bir çok karar ağacı mimarisinden oluştuğundan ötürü
çoklu ağaçların birleştirilmesi neticesinde tek bir karar ağacı modeline göre daha başarılı ve kararlı
sonuçlar elde edilmesi sağlanacaktır. Ayrıca rastgele orman yönteminin müşteri davranışlarının
analizinde özellik seçimi konusunda destek olacağı da söylenebilmektedir. Bu sayede parametrelerin
minimize edilerek daha etkin özellik seçimi yapılarak hem doğruluk oranı hem de zaman maliyeti
konusunda iyileştirme yapılabilme imkanı oluşturulacaktır. Bunun yanında kötü, kaynağı güvenilir
olmayan, eksik veriler ile çalışmak rastgele orman yönteminin performansını oldukça düşürecektir. Bu
nedenle veri seçimi yöntemin modellenmesi için oldukça önemlidir. Ayrıca rastgele orman yöntemi
modellenirken belirlenen ağaç sayısı ve hiperparametre seçimleri modelin zaman performansını da
doğrudan etkilemektedir.
Naive Bayes
Naive Bayes yöntemi istatistiksel tabanlı bir makine öğrenme yöntemidir. Sınıflandırma problemlerinde
verisetindeki olasılıksal durumlar baz alınarak sisteme öngörü yapma yeteneği kazandırmaktadır.
Temel olarak Bayes Teorimi tabanlıdır. Bayes mantığında bir durumun olasılığı için durumun
sonuçlarını etkileyebilecek diğer durumların ihtimalleri de kullanılmaktadır. Naive Bayes yönteminde
ise sonuç için tüm özelliklerin etkisinin bağımsız olduğu varsayımına dayanılmaktadır. Gauss Naive
Bayes, Multinominal Naive Bayes ve Bernoulli Naive Bayes olarak çeşitli topolojileri bulunmaktadır.
Naive bayes istatistiksel tabanlı bir makine öğrenme yöntemi olduğundan ötürü kişiselleştirilmiş
kampanya yönetimi için uygun bir yöntem değildir. Nitekim, bulgular kısmında elde edilen sonuçlarda
en düşük doğruluk sonucuna Naive Bayes yöntemi ulaşmıştır. Naive bayes yöntemi ilgili problemin
tamamında etkin olmasa da hibrit bir yaklaşımda müşterilerin segmentasyonu veya kampanya yönetimi
için önemli özelliklerin seçimi konusunda önemli bir rol alabileceği öngörülmektedir.
Lojistik Regresyon
Lojistik regresyon yöntemi istatistiksel tabanlı olarak sınıflandırma problemlerinde uygulanan bir
makine öğrenmesi yöntemidir. Yöntemin odağı doğal logaritma ile modellenmiş ve dönüştürülmüş
olasılık oranıdır. Bunun yanında lojistik regresyon yöntemi ikili sınıflandırma problemlerine daha
uygun olmak ile birlikte çoklu sınıflandırma problemlerine de uygulanabilmektedir. Yöntem, 0-1
değerleri arasında giriş değerinin hangi sınıfa ait olabileceğine dair ihtimali sigmoid fonksiyon
kullanarak tahmin etmektedir. Lojistik regresyon yöntemi sınıflandırma problemlerine uygun finans,
sağlık, ekonomi gibi alanlardaki durumlar için uygun sonuçlar verebilmektedir.
Page 193
SEMIT 2023
Lojistik regresyon yöntemi istatistiksel bir yöntem olarak kişiselleştirilmiş kampanya yönetimi
problemine yatkın bir algoritmaya sahiptir. Fakat çalışmada tercih edilen ağaç topolojisine sahip olan
karar ağaçları ve rastgele orman yöntemleri kadar probleme uygulanmakta etkin sonuçlar vermeyebilir.
Bunun nedeni lojistik regresyon yönteminin müşterinin geçmiş satın alımları gibi bağımsız değişkenin
kampanya katılım durumu gibi bağımlı değişkenin üzerindeki etkisine odaklanmasıdır. Oysa,
uygulanması gereken problem çok daha fazla parametreye ve kompleks bir yapıya sahiptir.
Deneysel Çalışmalar
Müşteri sadakat programı modellemesinde yapay zeka yöntemlerinden karar ağacı, rastgele orman,
naive bayes ve lojistik regresyon yöntemleri kullanılmıştır. Şekil I’de çalışmanın blok diyagramı yer
almaktadır. Karar ağacı ve rastgele orman algoritmalarının tercih edilmesinin nedeni kurgulanan
sınıflandırma problemine ait verinin ağaç yapısına uygun yapıda olmasıdır. Ayrıca probleme uygun
yaklaşımın ortaya konması için iki farklı yaklaşıma sahip olan naive bayes ve lojistik regresyon
modelleri de modelde uygulanmıştır. Modelleme için .net ortamında C# programlama dili ile .net
predicate builder, ML.Net ve Accord.Net kütüphaneleri uygulanmıştır. Ayrıca KNIME programı
kullanımı ile karar ağacı modellenmesi için destek sağlanmıştır. .Net predicate builder kütüphanesi
LINQ (Dil Entegrasyonu Sorgusu) ifadelerinin meydana getirilmesi; bunun yanında karmaşık olan
sorgularda tümleştirme eylemlerinin yapılması için tercih edilmektedir. Bu sayede veri kaynakları ile
haberleşme sağlanmaktadır. Kullanılan ML.Net ve Accord.Net kütüphaneleri ise karar ağacı
topolojisinin meydana getirilmesi, ağın analiz edilmesi, test edilmesi ve uygulanması için çalışmada
uygulanmıştır.
Şekil I. KAMPANYA SADAKAT YÖNETİMİ BLOK DİYAGRAMI
Geliştirilen yapay zeka destekli B2B kampanya ve sadakat sistemi 4 bölümden oluşmaktadır. Bu
bölümler:
• Form başlığı, kullanıcının göreceği açıklamalar ve resimler.
• Kampanyanın yayınlanacağı tarihler, depo ve şirket seçimleri, hediye kuponu tanımlama, kampanya
yararlanma sayısı, ‘Tek Seçimli kampanya’ seçeneği ve bayi filtresi
• Kullanıcı sepetinin koşullarının belirlenmesi. Bu kısımda çoklu ürün grupları seçilerek (Ürün Kodu,
Marka, Alt Tipi vs.) kampanyanın geçerli olacağı ürünleri ve koşulları belirlenmektedir.
• İlk 3 maddeyi sağlayan kullanıcının sepet koşullarını sağlaması halinde, kullanıcıya verilecek
hediyeler/indirimlerin tanımlanacağı bölüm. Bu kısımda hediye verilecek ürün yada ürünlerin tanımı
yapılmaktadır. Ürüne yüzdelik olarak indirim yapabilir, fiyatını istediği kurdan sabitleyebilir, sepet
toplamına indirim uygulayabilir veya 3. Madde de tanımladığı ürün veya ürünlere indirim yapabilir.
Ortaya konan sistemde ilk olarak “Strategy pattern”(strateji model) ile kampanya yapısı
oluşturulmuştur. Her koşula bir arayüz tanımlanarak, performans açısından daha iyi sonuç üretmesinin
yanında yeni koşulların kolayca eklenebilmesi, kodların kolayca okunabilmesi için bu yöntem tercih
Page 194
SEMIT 2023
edilmiştir. Bunun yanında kampanyanın ürün ve bayi gruplarının eşleşmesi ile birlikte kompleks bir
yapı ortaya çıktığı için veritabanı tarafında maliyeti arttırdığı gözlemlenmiştir.
Bu nedenle ortaya çıkan fazla yükleme ve maliyet problemlerinden ötürü sistem predicate modeli ile
tekrar oluşturulmuştur. Bu yöntem kurgulandığında kullanıcı sepet alanına geldiğinde, kullanıcının
yararlanabileceği kampanyanlar veritabanından 20’şer 20’şer olarak çekilmektedir. Sistem içerisinde
kullanıcının sepetine bakılmaksızın, tanımlanan kampanyalarda tarih ve bayiye uygun olan
kampanyalar alınmaktadır. Daha sonra bu kampanyalar bir döngüye sokulmakta; ardından ise predicate
süreci devreye girmektedir. Sepet ve kampanya koşulları döngü içerisinde tespit edilir. Döngü içerisinde
sorgular hazırlanıp minimum seviyede veritabanına bağlantı sağlanır.
Eğer kullanıcının daha önce yaptığı siparişlere yapılmış bir kampanya var ise kullanıcının, belirlenen
tarihler arasında yaptığı siparişleri sap’den alınıp sorguya dahil edilmektedir. Kullanıcıya yansıyacak
kısmı içinde, ürünün fiyatı sabit ise ürünün yeni fiyatı yönetici tarafından belirlenen tutar ve kur
değerinde olur. Oran belirlenmiş ise belirlenen oran kadar indirim yapılır. Sonraki adımda ise kampanya
ile ilgili tüm durumlar ve olaylar belirlenen kurallar neticesinde karar ağacı mimarisinde
modellenmiştir.
Sistem kurgulandıktan sonra test edilmiştir. Kampanya Testlerinde, Ortalama 20 kullanıcı, 50
kampanya kaydı ile bir ürüne tek fiyat halinde yapılmış kampanyanın hesaplanıp kullanıcıya yansıma
süresi, ortalama 3-6 saniye arasında olduğu gözlemlenmiştir.
Çalışmada yazılım kullanılarak bayilik ve hizmet sağlanan firmaların yazılımdaki davranışlarından elde
edilen 60 adet örnek, veriseti olarak kullanılmıştır. Verisetinde bayi adı, ürün adı, ürünün depolandığı
yer, sepet ile ilgili koşul durumu, hediye kuponu durumu, katılan kampanya sayısı gibi 11 giriş
parametresi bulunmaktadır (Tablo I). Hedef parametre ise kampanya hakkı olup olmadığı bilgisini
veren alan olmaktadır. Kullanılan veriseti makine öğrenmesi yöntemlerinde kullanılmak üzere eğitim
ve test seti olarak ayrılmıştır. Verinin %70’i eğitim seti olarak; %30’u ise test seti olarak kullanılmıştır.
Farklı oranlarda eğitim seti ve test seti kullanılmış; fakat optimum sonuçlar bu oranda kullanıldığında
elde edilmiştir. Belirtilen veri seti incelendiğinde, uygulama öncesi/sonrası yeni sistem üzerinde, daha
fazla sayıda kampanyadan yararlanılmıştır. Yeni sistem sayesinde kullanıcının ihtiyacı ve popüler
ürünlerin tespiti gibi hususlar daha net tespit edilmiştir. Bu da kampanya sayılarında artışa neden
olmuştur. Nisan ayı ele alındığında, kampanya sayısı 3 tane daha fazladır. Cirolar incelendiğinde,
kampanya sayıları yakın olsa da, kur farkından dolayı cirolar daha fazla artmıştır. Bayiler genel olarak
Mayıs ayında kampanyalardan daha çok yararlanmaktadır. Mayıs ayı ve yaz sezonu satışların genel
olarak arttığı bir dönemdir. Uygulama sonrası kampanyalar gözlemlendiğinde Mayıs ayında,
kampanyalardan yararlanma sayısı daha yüksektir. Toplam Ciro olarak bakıldığında Haziran ayı
maksimum miktardadır. Burada döviz değişkeni bir kenara bırakılır ise, bayilerin daha önceden yaptığı
siparişler, uygulama sonrası da hesaba katılmıştır. Bu sayede bayilerin siparişleri, daha uzun bir zaman
aralığına denk gelmektedir. Bunun sonucunda siparişleri kaybolmayan bayilerin kampanyalara olan
ilgisini arttırmaktadır. Bu da kampanya sayılarında artışa neden olmuştur.
Çalışmada karar ağacı için ID3(Iterative Dichotomiser 3) algoritması uygulanmıştır. ID3 algoritması,
karar verme ile ilgili sınıflandırma ve regresyon problemlerindeki gücünden dolayı tercih edilmiştir.
Prensip olarak verilerin analizi ile birlikte ağacın düğümlerin iteratif olarak oluşturulması mantığı ile
çalışan bir algoritmadır. Müşteri sadakat yönetimi karar ağacı modellemesi için kullanılan parametreler
Tablo I’de gösterilmiştir. Modelde 11 giriş parametresi kullanılırken; kampanya hakkı parametresi
sonuç parametresi olarak değerlendirilmiştir. Çalışmada karar ağacı modelinde bölme kriteri Gini
endeksi ile yapılırken kullanılan modelin parametre değerleri için maksimum derinlik değeri 8,
minimum bölme değeri 2 ve minimum yapraktaki veri sayısı 2 olarak belirlenmiştir. Gini endeksi ile
oluşturulan ID3 karar ağacının kök düğümü “Sepet Toplam Fiyatı” olarak bulunmuştur (Şekil II). Karar
ağacı modeli test edildiğinde ise doğruluk oranı %73.2 olarak elde edilmiştir (Tablo II).
TABLO I. KARAR AĞACI MODELİ PARAMETRELERİ
Parametre
Parametre Adı
No
1 Tanımlı Bayi
2 Tanımlı Şirket
Page 195
SEMIT 2023
Parametre
Parametre Adı
No
3 Tanımlı Hediye Kuponu
4 Sepet Koşul Durumu
5 Bayi
6 Ürün
7 Ürünün Deposu
8 Alınan Ürün Adedi
9 Sepet Toplam Fiyatı
10 Hediye Kuponu
Kampanyadan Yararlanma
11
Sayısı
12 Kampanya Hakkı
Çalışmanın son kısmında kampanya yönetimi modellemesi için en uygun yöntemin karar ağacı
olduğunun ortaya konması adına farklı makine öğrenmesi yöntemleri aynı problem için uygulanmıştır.
Bu nedenle sık kullanılan makine öğrenmesi yöntemleri olan rastgele orman, naive bayes ve lojistik
regresyon yöntemleri de aynı durum için modellenmiştir. Rastgele orman yöntemi topluluk tabanlı olan
sınıflandırma ve regresyon problemleri için başarılı sonuçlar elde edilebilen makine öğrenmesi
yaklaşımlarından biridir. Birden fazla karar ağacının birlikte kullanımı ile daha güçlü model ortaya
konma prensibine dayanan algoritmaya sahiptir. Rastgele orman modeli için karar ağacının
bölünmelerinin değerlendirilmesi için gini endeksi tercih edilirken hiperparametre olarak kullanılan
karar ağacı sayısı 10, maksimum derinlik sayısı 8, minimum bölme değeri 2, minimum yaprak veri
sayısı 2 değerleri uygulanmıştır. Elde edilen rastgele orman modelinin ağaç görünümü Şekil III’de
gösterilmiştir. Naive bayes yöntemi olasılık tabanlı sınıflandırıcı makine öğrenmesi yöntemidir.
Şekil II. KARAR AĞACI KÖK DÜĞÜM
Page 196
SEMIT 2023
Çalışma mantığı bazı durumlarda gerçek dünya problemleri için yetersiz kalsa da uygun problemler için
hızlı ve etkin sonuçlar vermektedir. Lojistik regresyon algoritması ise genellikle ikili sınıflandırma
problemleri için kullanılan bir makine öğrenmesi yöntemidir. Algoritmanın amacı olasılık dağılımı
kullanarak girdi ve çıktı arasındaki ilişkinin modelini ortaya çıkartmaktır. Modellemede hızlı ve verimli
sonuçlar elde edilmesi sağlansa da uygun olmayan problemler için özellikler arasındaki olması gereken
ilişkiye tam önemi gösteremediğinden dolayı çok etkin sonuçlar vermeyebilmektedir. Uygulanan tüm
yöntemlerin performans karşılaştırılması Tablo II’de gösterilmiştir. Probleme uygulanan yöntemler
neticesinde karar ağacı yöntemi ile %73.2, rastgele orman yöntemi ile %78, naive bayes yöntemi ile
%59 ve lojistik regresyon yöntemi ile %61 doğruluk oranları elde edilmiştir. Sonuçlar irdelendiğinde
en etkin sonuçları karar ağacı ve rastgele orman yöntemlerinin ortaya koyduğu görülmektedir. Her iki
yöntemin de ağaç tabanlı olması probleme uygun olmasının en önemli nedenidir. Rastgele orman
yönteminin bir çok farklı ağaç yapısı ile daha güçlü bir model ortaya koyabilme yeteneği sayesinde
uygulanan yöntemler arasında en yüksek doğruluk değerine ulaşmıştır. Bunun yanında zamansal
performanslar incelendiğinde karar ağacı yöntemi rastgele orman yöntemine göre daha hızlı sonuç
üretmektedir.
Page 197
SEMIT 2023
Şekil III. RASTGELE ORMAN MODELİ

TABLO II. PERFORMANS TABLOSU
Performans Metrikleri
Doğruluk Cohen Kappa
Yöntem Adı
Oranı Katsayısı
Karar Ağacı 0,732 0,477
Rastgele Orman 0,78 0,523
Naive Bayes 0,59 0,34
Page 198
SEMIT 2023
Performans Metrikleri
Yöntem Adı
Oranı Katsayısı
Lojistik Regresyon 0,61 0,37
Konu ile ilgili gelecek çalışmalardaki hedef olarak farklı yöntemsel yaklaşımlar ve veri setinin daha
zenginleştirilmesi ile başarı oranının arttırılması hedeflenmektedir.
Sonuç ve Tartışma
Bayii satış sistemleri geleneksel olarak, admin panelden veya kullanılan bir ERP sistemi ile entegrasyon
sağlanarak veriyi alarak hizmeti sunmak üzerine tasarlanmıştır. Geleneksel bayi sistemlerinden farklı
olarak, müşterilerin satın alma taleplerini arttırmayı sağlayacak ürünlerin öne çıkartılması ve bunun
yanında müşterilerin kurumlara bağlılığını arttırmayı sağlayacak kişiselleştirilmiş kampanyaların
oluşturulması çalışmanın yenilikçi yönleri olarak ortaya konulmaktadır. Bu amaç doğrultusunda, bir
müşterinin sık şekilde satın aldığı ürünlerin dışında daha önce tercih etmediği fakat analizler neticesinde
de bu ürünü benzer müşterilerin satın aldığı ve firma için de karlılığı yüksek olan ürünler ürün
reyonlarında dinamik olarak konumlandırılması sağlanmaktadır.
Geliştirilen sistem ile kampanya tanımları müşterilerin satın alma limitleri, yönelimlerine göre farklı
şekilde tanımlanacak; böylece oyunlaştırma mantığı ile müşterilerin hedefe ulaşma isteği ile firma
karlılığını arttıracaktır. Çalışmanın Ar-Ge niteliği ise geleneksel bayi satış sistemlerindeki özelliklere
ek olarak karar ağacı yönteminin kullanımı ile firmaların satış miktarının ve karlılık oranının
arttırılmasını sağlayacak şekilde oyunlaştırma stratejisi ile kişiselleştirilmiş müşteri kampanyaları ve
ürün reyon yönetimi özelliklerinin sisteme entegre edilmesidir.
Çalışmanın bilimsel katkısı ise kişiselleştirilmiş kampanya tahmini problemi için makine öğrenmesi
yöntemlerinden hangi yaklaşımın daha uygun olduğunun ortaya konmasıdır. Çalışmada farklı
yaklaşımlarla matematiksel modellemesi yapılan dört farklı makine öğrenmesi yöntemi bu nedenle
probleme uygulanmıştır. Elde edilen sonuçlar neticesinde ağaç yapılı yöntemlerin uygulanan problem
için çok daha başarılı sonuçlar verdiği çalışmada ortaya konmuştur. Rastgele orman yöntemi doğruluk
oranı açısından en başarılı yöntem olarak; karar ağacı yöntemi ise zamansal performans olarak en iyi
sonucu göstermektedir.
Çalışmada en etkin performansı gösteren Rastgele Orman yöntemi uygulanan B2B alanında kampanya
yönetimi modellemesi ile hem şirketin karlılığı hem de sürecin verimliliği arttırılmıştır. Müşterilerin
segmentasyonu ile ilgi odakları, davranışlar, satın alım alışkanlıkları gibi parametreler ile müşterilerin
segmentasyonu sağlanmış; kampanyalar kişiselleştirilebilmiştir. Bu sayede müşterilerin satın alımları
tetiklenmiş, kuruma olan aidiyetlerinin arttırılması sağlanmıştır. Hedef odaklı, iyi planlanmış ve
organize edilmiş kampanya yönetimi yapılarak düşük maliyet ve maliyet tasarrufu fayda olarak
kurumlara kazandırılmıştır. Sonuç olarak Rastgele Orman yöntemi uygulanarak B2B konusunda
müşteri sadakatinin arttırılması ve kampanya yönetimi konularında daha etkin, daha yüksek doğrulukla
bulguların elde edildiği gözlemlenmiştir.
Kaynaklar
[1] Kazankaya.K. “E-Ticaret ve Dijital Pazarlama”, Sokak Yayın Grubu, İstanbul, 2019.
[1] Ayvaz, D., Aydoğan, R., Akçura, M.T. and Şensoy, M. “Campaign participation prediction with deep
learning”, Electronic Commerce Research and Applications, Vol. 48, 2021.
[2] Xiao, Y., Zhu, Y., He, W., and Huang, M. “Influence prediction model for marketing campaigns on e-
commerce platforms”, Expert Systems with Applications, Vol. 211, 2023.
Page 199
SEMIT 2023
[3] Jiezhong Qiu, J., Tang, J., Ma, H., Dong, Y., Wang, K., and Tang, J. “DeepInf: Social Influence
Prediction with Deep Learning”, KDD '18: Proceedings of the 24th ACM SIGKDD International
Conference on Knowledge Discovery & Data Mining, 2018.
[4] Margaris, D., Vassilakis, C., and Georgiadis, P. “Recommendation information diffusion in social
networks considering user influence and semantics”, Social Network Analysis and Mining, Vol. 8, 2016.
[5] Yaiprasert, C., and Hidayanto, A.N. “AI-driven ensemble three machine learning to enhance digital
marketing strategies in the food delivery business, Intelligent Systems with Applications, Vol. 18, 2023.
[6] Leeuwen, R.V., and Koole, G. “Data-driven market segmentation in hospitality using unsupervised
machine learning”, Machine Learning with Applications, Vol. 10, 2022.
[7] Sajjad, M., and Zaman, U. “Innovative Perspective of Marketing Engagement: Enhancing Users’
Loyalty in Social Media through Blogging”, Journal of Open Innovation: Technology, Market, and
Complexity, 2020.
[8] Sun, Y., Liu, H., and Gao, Y. “Research on customer lifetime value based on machine learning
algorithms and customer relationship management analysis model”, Heliyon, Vol. 9(2), 2023.
[9] Hsu, C.L., and Lin, J.C.C. “Understanding the user satisfaction and loyalty of customer service
chatbots”, Journal of Retailing and Consumer Services, Vol. 71, 2023.
[10] Smaili, M.Y., and Hachimi, H. “New RFM-D classification model for improving customer
analysis and response prediction”, Ain Shams Engineering Journal, In Press, 2023.
Page 200
SEMIT 2023
SAR Görüntülerinden Derin Öğrenme Tabanlı Hedef Tanima

Bilge Sinan
Başkent Üniversitesi, Fen Bilimleri Enstitüsü,
Savunma Teknolojileri ve Sistemleri Ana Bilim Dalı
Ankara, Turkey
bsinan@havelsan.com.tr
Emre Sümer
Başkent Üniversitesi, Mühendislik Fakültesi,
Bilgisayar Mühendisliği Bölümü
Ankara, Turkey
esumer@baskent.edu.tr
Özet
Sentetik Açıklıklı Radar (SAR), gün boyunca ve her türlü hava koşullarında çalışabilen aktif ve yüksek
hassasiyetli bir görüntüleme sistemidir. SAR görüntülerine dayalı hedef tanıma, son on yılda geniş çapta
araştırılmıştır. SAR görüntüleri, girişime karşı zayıf dirençli ve tek renkli olduğundan optik görüntüler
kadar net ve okunaklı değildir. Geleneksel SAR tabanlı hedef tanıma yöntemleri, görüntünün
istatistiksel ve fiziksel özelliklerine dayalı olarak gerçekleştirilir. Ancak, bu yöntemlerin SAR hedef
görüntülerinin karmaşık ve değişken doğasına uyarlanması zordur. Uzak çekim SAR görüntülerinde
ilgilenilen küçük hedefleri gözle tespit etmek zaman alıcı olup pratik değildir. Bu nedenle SAR
görüntüleri için otomatik hedef tanıma algoritmalarının geliştirilmesin büyük önem arz etmektedir. Son
yıllarda yapılan çalışmalar incelendiğinde evrişimli sinir ağı tabanlı hedef tanıma yaklaşımları, hedefin
özelliklerini SAR görüntülerinden başarılı bir şekilde çıkarabilmekte ve yüksek bir tanıma doğruluğu
elde edebilmektedir. Bu çalışmada, MSTAR veri kümeleri kullanılarak güncel bir nesne tespit modeli
olan YOLO’nun SAR hedeflerini tanıma performansı araştırılmıştır. Yakın bir tarihte geliştirilmiş olan
YOLOv8’in, önceki YOLO versiyonu olan v5 modeliyle MSTAR veri seti üzerindeki performansı
karşılaştırılmıştır. Bu çalışmada radarın 15° ve 17°’ lik bakış açısından alınan BMP2, BRDM2, BTR60,
BTR70, D7, T62, T72, ZIL131, ZSU234, 2S1 hedeflerin görüntüleri otomatik olarak tespit edilmeye
çalışılmıştır. Sonuçlar incelendiğinde, YOLOv8 versiyonunun YOLOv5 versiyonuna kıyasla daha kısa
bir eğitim süresiyle daha yüksek bir hedef tanıma doğruluğuna ulaşılabildiği gösterilmiştir.
Anahtar Kelimeler: Evrişimli Sinir Ağı (CNN), Sentetik Açıklıklı Radar (SAR), Hedef Tanıma
Giriş
Sentetik açıklıklı radar (SAR), tüm gün ve tüm hava koşullarında çalışabilen aktif, yüksek hassasiyetli
bir görüntüleme radarıdır [1-4]. SAR görüntülerine dayalı hedef tanıma, son on yılda geniş çapta
araştırılmaktadır. SAR görüntüleri, girişime karşı zayıf dirençli olduğundan ve tek renkli olduğundan,
optik görüntüler kadar net ve okunaklı değildir. Geleneksel SAR hedef tanıma yöntemleri, manuel
modelleme için esas olarak görüntünün istatistiksel ve fiziksel özelliklerine dayalı olarak
gerçekleştirilir. Ancak, bu yöntemlerin SAR hedef görüntülerinin karmaşık ve değişken doğasına
uyarlanması zordur [5]. Son yıllarda evrişimli sinir yöntemleriyle karşılaştırıldığında, CNN tabanlı
hedef tanıma yöntemi, hedefin özelliklerini SAR görüntülerinden başarılı bir şekilde çıkarabilmekte
ve yüksek bir tanıma doğruluğu elde edebilmektedir [6]
Bu çalışmada, yakın tarihte geliştirilmiş YOLO algoritmasının sekizinci versiyonunun (YOLOv8),
önceki versiyonlardan YOLOv5 modeliyle askeri amaçlı SAR verileri üzerindeki performans
karşılaştırması yapılmıştır. Her bir model farklı radar açılarından alınmış 7 farklı veri seti üzerinde
çalıştırılarak nesne tespiti ve tanımlaması yapılmıştır.
Page 201
SEMIT 2023
Önceki Çalışmalar
Sentetik açıklıklı radar görüntüleri kullanılarak elde edilen görüntüler üzerinden derin öğrenme
teknikleriyle nesne tanıma çalışmaları genel olarak iki ayrı kategoriye ayrılmaktadır. Bunlar; (i) tek
aşamalı nesne tespit algoritmaları ve (ii) iki aşamalı nesne tespit algoritmalarıdır. Tek aşamalı nesne
tespiti araştırmalarında SSD, YOLO, RetinaNet ve EfficientNet modelleri öne çıkarken, iki aşamalı
nesne tespiti araştırmalarında R-CNN, Fast R-CNN ve Faster R-CNN modelleri başarılı sonuçlar
vermektedir.
YOLO algoritmasının performansı özellikle optik görüntüler üzerinde oldukça başarılı sonuçlar
vermektedir. Bu durum, son zamanlarda radar görüntülerine uygulama potansiyelini ortaya çıkarmış
ve radar alanında yeni bir araştırma alanının açılmasına yol açmıştır [7–9].
Bu çalışmada, YOLO'nun güncel modellerindeki SAR hedeflerini tanıma yeteneği, MSTAR veri
kümeleri üzerinde test edilmiştir.
Metodoloji
YOLO (You Only Look Once), bir nesne algılama ve konumlandırma algoritması olarak bilinir.
Görüntüler veya videolar üzerinde nesneleri tespit etmek ve bu nesnelerin konumlarını belirlemek
amacıyla kullanılır. Bu algoritma, geleneksel nesne algılama yöntemlerine göre daha hızlı çalışma
yeteneği sunar, çünkü nesne tespiti ve konumlandırma işlemlerini aynı anda gerçekleştirir.
YOLO'nun temel yaklaşımı, bir görüntüyü birkaç kare bölgeye bölmek ve her bir kare bölge üzerinde
nesne tespiti gerçekleştirmektir. Bu kare bölgelere ‘hücre’ adı verilir ve her bir hücre belirli bir
nesnenin tespit edilip edilmediğini ve nesnenin konumunu içeren bilgileri içerir. YOLO'nun bazı
avantajları şunlardır:
1. Hızlı İşlem: YOLO tüm nesneleri tek bir geçişte tespit ettiği için diğer yöntemlere göre daha hızlı çalışır.
2. Nesne Konumlandırma: YOLO tespit edilen nesnelerin sınıflandırılmasının yanı sıra kesin konumlarını
da belirler.
3. Genel Algılama: YOLO farklı nesne sınıflarını tespit etmek ve sınıflandırmak için genel bir model
kullanabilir.
YOLO özellikle gerçek zamanlı nesne algılama uygulamalarında sıkça kullanılır. Örneğin güvenlik
kameralarında, sürücüsüz araçlarda, nesne takibi uygulamalarında ve diğer alanlarda kullanımı
yaygındır.
YOLOv5 Nedir?
YOLOv5, YOLO nesne algılama modelinin güncel sürümlerinden biridir. YOLOv3'ün geliştiricileri
Ultralytics tarafından 2020 yılında tanıtılmış ve PyTorch çerçevesi üzerine inşa edilmiştir. YOLOv5
kullanımı hızlı ve kolaydır. Nesne tespiti için güncel problemlere ait sonuçları elde etme kapasitesine
sahiptir. Ayrıca öncekilere göre daha doğru ve eğitilmesi daha kolay olduğundan birçok geliştirici için
popüler bir seçimdir.
YOLOv5 dört model içerir: YOLOv5s, YOLOv5m, YOLOv5l ve YOLOv5x. Ana model yapıları
temel olarak benzerdir, yalnızca ağın derinliği ve genişliği açısından biraz farklıdır. Bu çalışmada
yapılan deneylerde esas olarak YOLOv5s modelini kullanılmıştır ve spesifik model yapısı Şekil 1’de
gösterilmektedir. Model yapısı analiz edildiğinde, YOLO serisi modeller tipik ‘tek aşamalı’ yapıdır ve
genellikle dört bölüme ayrılmıştır: Giriş, Omurga, Boyun ve Tahmin olarak bu dört bölümün yapısal
özellikleri ilerleyen bölümde tanıtılacaktır.
Page 202
SEMIT 2023
Şekil 1 Yolov5s Mimari Yapısı
Giriş Bölümü:
2019 yılında önerilen CutMix [9] veri artırma yöntemine atıfta bulunan Mozaik veri artırma
algoritması, YOLOv4'te olduğu gibi erken bir tarihte küçük hedeflerin tespit doğruluğunu
iyileştirmeye yönelik bir yöntem olarak önerilmiştir. Dört resim, Şekil 2'de gösterildiği gibi rastgele
ölçeklendirme, kırpma ve düzenleme yoluyla birbirine birleştirilmiştir.
Şekil 2 Mozaik Veri Arttırma
Farklı veri kümeleri için tüm YOLO algoritmaları, başlangıç uzunluğu ve genişliğine sahip bağlantılara
sahip olacaktır. Ağ eğitimi sırasında, tahmin çerçevesi bu temelde çıkarılacak ve kayıplar gerçek
çerçeve temel gerçeğiyle karşılaştırılarak hesaplanacak, ayrıca parametreler yinelemenin tersine
güncellenecektir. Tanımlı (etiketli) veri seti için, önceki bağlantıların hesaplanmasında genellikle k-
ortalamalar kümeleme algoritması kullanılır. Eğitim setini dolaşarak çoğu gerçek kutuya en yakın
bağlantı parametreleri, ilk bağlantı parametreleri olarak elde edilir ve bu da modelin eğitim süresini
azaltır.
Omurga Bölümü:
Odaklanma: Odaklanma, YOLOv5 algoritmasının en önemli yeniliklerinden biridir. Burada görüntü
dilimlenir ve dört parçaya bölünür, kanal boyutunda birleştirilir ve son olarak kıvrılır. Bu yöntem, bilgi
kaybını en aza indirecek şekilde daha düşük örnekleme işlemini benimsemiştir. YOLOv5'lerin yapısını
örnek alırsak, 608x608x3'ün görüntü dilimleme işlemi 304x304x12 olur ve ardından 32 çekirdeğin
evrişimi gerçekleştirilir, bu da Şekil 3’te gösterildiği gibi 304x304x32'nin karakteristik diyagramına
göre değişir.
Page 203
SEMIT 2023
Şekil 3 Odaklanma ve Dilimlenme
CSP: Benzer şekilde, hem YOLOv5'in hem de YOLOv4'ün omurgası Çapraz Aşamalı Kısmi Ağ
(CSPNet) fikrini kullanır [10]. İlk olarak, temel katmanın özellik haritalaması iki parçaya bölünür ve
ardından ağ optimizasyonunda gradyan bilgilerinin tekrarlanan hesaplama miktarını azaltmak ve son
modelin doğruluğunu sağlamak için aşamalar arası hiyerarşik yapı birleştirilir. Aradaki fark,
YOLOv5'in CSP modülünü hem omurga hem de boyun modüllerine uygulamasıdır.
Boyun Bölümü:
YOLOv4 ile aynı özellikli piramit ağları (FPN) [11] ve yol toplama ağı (PAN) [12] yapısını
benimseme temelinde ‘Neck’ ayrıca CSP2 yapısını da ekler. Bunlar arasında, FPN katmanı yukarıdan
aşağıya güçlü anlamsal özellikler taşır, özellik piramidi PAN, aşağıdan yukarıya doğru güçlü
konumlandırma özellikleri taşır ve CSP2 yapısı, ağ özelliği füzyonunu daha da güçlendirir. Bu
nedenle, hedef özellikleri Şekil 4’te gösterildiği gibi, farklı omurga katmanlarından farklı tespit
katmanlarının parametrelerini bir araya getirir.
Şekil 4 FPN + PAN Yapısı
Tahmin Bölümü:
NMS genel olarak hedef tespitinde ön seçiciyi filtrelemek için kullanılır ve aynı zamanda yerel
maksimum arama olarak da anlaşılabilir.
YOLOv8 Nedir?
YOLOv8, YOLO ailesinin en son modeli olup, 2022 yılında Ultralytics tarafından tanıtılmıştır.
YOLOv8, YOLOv5 çerçevesi üzerine inşa edilmiş olup çeşitli mimari ve geliştirici deneyimi
iyileştirmelerini içerir. YOLOv5'ten daha hızlı ve daha doğrudur ve nesne tespiti, örnek
segmentasyonu ve görüntü sınıflandırması gerçekleştirmeye yönelik eğitim modelleri için birleşik bir
Page 204
SEMIT 2023
çerçeve sağlar.
YOLOv8 ve YOLOv5: Karşılaştırma
Nesne algılama söz konusu olduğunda birçok model mevcuttur. Ancak YOLOv8 ve YOLOv5,
Ultralytics tarafından oluşturulan en popüler ve en son teknolojiye sahip modellerden ikisidir.
YOLOv8, önceki sürümlerin başarısını temel alan ve performansı ve esnekliği artırmak için yeni
özellikler ve iyileştirmeler sunan YOLO ailesinin en son üyesidir. YOLOv5 ise hızı, basitliği ve
doğruluğuyla bilinmektedir.
En iyi nesne algılama modelini seçmeye gelindiğinde dikkate alınması gereken birkaç faktör vardır.
Bu faktörlerden en önemlileri; hız, doğruluk, kullanım kolaylığı ve geliştirici deneyimidir.
Hız: Hem YOLOv8 hem de YOLOv5 görüntüleri gerçek zamanlı olarak işleyebilen hızlı nesne
algılama modelleridir. Ancak YOLOv8, YOLOv5'ten daha hızlıdır ve bu da onu gerçek zamanlı nesne
algılama gerektiren uygulamalar için daha iyi bir seçim haline getirir.
Doğruluk: Bir nesne algılama modeli seçerken dikkate alınması gereken kritik bir faktördür. Bu
bakımdan YOLOv8, mimarisinde yapılan çeşitli iyileştirmeler sayesinde YOLOv5'ten daha doğrudur.
Kullanım kolaylığı: Hem YOLOv8 hem de YOLOv5'in kullanımı kolaydır; YOLOv5 ise ikisi
arasında kullanımı en kolay olanıdır. YOLOv5, PyTorch çerçevesi üzerine inşa edilmiş olup
geliştiricilerin kullanımını ve dağıtımını kolaylaştırır.
Geliştirici deneyimi: YOLOv8 nesne algılama, örnek bölümlendirme ve görüntü sınıflandırma
işlemlerini gerçekleştirmeye yönelik eğitim modelleri için birleşik bir çerçeve sağlar. Bu onu daha
kapsamlı bir araç seti isteyen geliştiriciler için daha iyi bir seçim haline getirir.
Değerlendirme Metrikleri
Sonuçların analizi için üç metrik belirlenmiştir; kesinlik (Denk. 1), geri çağırma (Denk. 2) ve F1-
puanı (Denk.3). Bu metrikler karışıklık matrisine göre hesaplanır. Karışıklık matrisi 4 parametreden
oluşur: gerçek pozitif (TP), gerçek negatif (TN), yanlış pozitif (FP) ve yanlış negatif (FN).
TP: Tahmin olumlu, temel gerçek ise olumlu. TN: Tahmin olumsuz, temel gerçek ise olumlu. FP:
Tahmin olumlu, temel gerçek ise olumsuz. FN: Tahmin olumsuz, temel gerçek ise olumlu. Hassasiyet,
yöntemin doğru tespitlerinin sayısını ifade ederken, geri çağırma gerçekte var olan doğru tespit edilen
nesnelerin ölçüsüdür. F1 puanı ise hassasiyet ve geri çağırmanın bir fonksiyonudur. Değerlendirme
metrikleri de karışıklık matrisi kullanılarak hesaplanır.
Kesinlik = TP / (TP + FP) Denk.(1)
Geri Çağırma = TP / (TP + FN) Denk.(2)
F1-puanı = 2 (Kesinlik x Geri Çağırma) / (Kesinlik + Geri Çağırma)
Denk.(3)
Deneysel Sonuçlar
Bu çalışmadaki deney verileri, Sandia Ulusal Laboratuvarı (SNL) SAR sensör platformu tarafından
toplanmıştır. Bu veri koleksiyonu, Hareketli ve Sabit Hedef Tespiti ve Tanıma (MSTAR) programının
bir parçası olarak Savunma İleri Araştırma Projeleri Ajansı (DARPA) ve Hava Kuvvetleri Araştırma
Laboratuvarı (AFRL) tarafından ortaklaşa desteklenmiştir. Yüzbinlerce SAR (Sentetik Açıklıklı
Radar) görüntüsü, farklı hedef türlerini, görünüş açılarını, alçaltma açılarını, seri numaralarını ve
artikülasyonu içerecek şekilde toplanmıştır.
Veri kümeleri kamuya açıklandığında, 10 farklı türde yer hedefini (BMP2, BRDM2, BTR60, BTR70,
D7, T62, T72, ZIL131, ZSU234, 2S1) içermektedir. Veriler, spot ışığı modunda ölçülen 1 fit x 1 fit
çözünürlüğe sahip X-bandı SAR görüntülerinden oluşur. Hedefler, 1°-5°'lik artışlarla tam 360° azimut
açıları üzerinde ve çoklu alçalış açıları (15°, 17°, 30° ve 45°) üzerinde ölçülmüştür. Hedef verileri,
hedef türü başına standart bir çip boyutu, genellikle 128 piksel x 128 piksel ile hedef merkezli alt
görüntü yongaları olarak sunulur.
Page 205
SEMIT 2023
Şekil 5 10 Askeri Hedef Türü: Optik Görüntüler (üstte) ve SAR Görüntüleri (altta)
Bu çalışmada 15° ve 17° lik görüntüleri alınan BMP2, BRDM2, BTR60, BTR70, D7, T62, T72,
ZIL131, ZSU234, 2S1 hedeflerinin tespitinde YOLOv5 ve YOLOv8’in performansları incelenmiştir.
YOLOv5’e MSTAR veri setinin girdi olarak verilmesi ile elde edilen sonuçlar Tablo 1’de yer
almaktadır.
Tablo 1 15 Derecelik MSTAR veri setinin YOLOv5 sonuçları
Veri Seti Etiketli Etiketli Eğitim mAP P R

Veri Veri Sayısı Süresi
Sayısı TEST
TRAIN
BTR-60 195 8 2 dk 2 sn 0.892 0.83 0.853
2S1 274 12 3 dk 56 sn 0.995 0.996 1
BRDM-2 274 12 3 dk 56 sn 0.995 0.996 1 1
D7 274 12 3 dk 56 sn 0.995 0.996 1

T-62 274 12 3 dk 56 sn 0.995 0.996 1
ZIL 131 274 12 3 dk 56 sn 0.995 0.996 1
ZSU-23/4 274 12 3 dk 56 sn 0.995 0.996 1

YOLOv8’e MSTAR veri setinin girdi olarak verilmesi ile elde edilen sonuçlar Tablo 2’de yer
almaktadır.

Veri Sayısı Veri Sayısı Süresi
TRAIN TEST
BTR-60 195 10 5 dk 58sn 0.995 0.992 1
2S1 250 12 10 dk 59 sn 0.995 1 1
BRDM-2 250 12 10 dk 59 sn 0.995 1 1
D7 250 12 10 dk 59 sn 0.995 1 1
T-62 250 12 10 dk 59 sn 0.995 1 1
ZIL 131 250 12 10 dk 59 sn 0.995 1 1
ZSU-23/4 250 12 10 dk 59 sn 0.995 1 1
Page 206
SEMIT 2023
YOLOv5’e MSTAR veri setinin girdi olarak verilmesi ile elde edilen sonuçlar Tablo 3’te
görülmektedir.

Veri Veri Süresi
Sayısı Sayısı
TRAIN TEST
BTR-60 250 12 3 dk 54 sn 0.995 0.996 1
2S1 250 12 3 dk 54 sn 0.995 0.996 1
BRDM-2 250 12 3 dk 54 sn 0.995 0.996 1
D7 250 12 3 dk 54 sn 0.995 0.996 1
T-62 250 12 3 dk 54 sn 0.995 0.996 1
ZIL 131 250 12 3 dk 54 sn 0.995 0.996 1
ZSU-23/4 250 12 3 dk 54 sn 0.995 0.996 1
YOLOv8’e MSTAR veri setinin girdi olarak verilmesi ile elde edilen sonuçlar Tablo 4’te yer
almaktadır.
Veri Seti Etiketli Etiketli Veri Eğitim mAP P R

Veri Sayısı Süresi
Sayısı TEST
TRAIN
BTR-60 250 12 11 dk 51 sn 0.995 0.995 1
2S1 250 12 11 dk 51 sn 0.995 0.995 1
BRDM-2 250 12 11 dk 51 sn 0.995 0.995 1
D7 250 12 11 dk 51 sn 0.995 0.995 1
T-62 250 12 11 dk 51 sn 0.995 0.995 1
ZIL 131 250 12 11 dk 51 sn 0.995 0.995 1
ZSU- 250 12 11 dk 51 sn 0.995 0.995 1
23/4
Sonuç
Bu araştırma, YOLO'nun SAR hedeflerini tanıma performansını incelemek amacıyla MSTAR veri
kümeleri kullanmıştır. YOLOv8'in, önceki YOLO sürümü olan v5 modeli ile MSTAR veri seti
üzerindeki performansı karşılaştırılmıştır.
Yapılan deneyler sonucunda MSTAR veri setinde 15 derecelik radar açısından alınan görüntülerdeki
nesnelerin tespitinde YOLOv5 daha hızlı sonuç vermesine rağmen YOLOv8 versiyonu daha başarılı
nesne tespitleri yapmıştır. Benzer şekilde MSTAR veri setinde 17 derecelik radar açısından alınan
görüntüler incelendiğinde YOLOv5 daha hızlı sonuç vermesine rağmen YOLOv8 modeli hedefleri daha
başarılı bir şekilde tespit etmiştir.
Referanslar
[1] Y. Jin, “Multimode remote sensing intelligent information and target recognition: Physical
intelligence of microwave vision,” Journal of Radars, vol. 8, no. 6, pp. 710–716, 2019.
Page 207
SEMIT 2023
[2] F. Zhang, Q. Guo, Z. Wang, P. Zhou, G. Zhang, J. Sun, and S. Pan, “Photonics-based broadband
radar for high-resolution and real-time inverse synthetic aperture imaging,” Optics Express, vol. 25, no.
14, pp. 16274-16281, 2017.
[3] X. Ye, F. Zhang, Y. Yang, D. Zhu, and S. Pan, "Photonics-based highresolution 3D inverse
synthetic aperture radar imaging," IEEE Access, vol.7, pp. 79503-79509, 2019.
[4] F. Zhang, Q. Guo, Y. Zhang, Y. Yao, P. Zhou, D. Zhu and S. Pan, “Photonics-based real-time and
high-resolution ISAR imaging of noncooperative target,” Chinese Optics Letters, vol. 15, no. 11, pp.
112801,2017.
[5] W. Guo, Z. Zhang, W. Yu, “Perspective on explainable SAR target recognition,” Journal of Radars,
vol. 9, no. 3, pp. 462–476, 2020.
[6] Kim, W.; Cho, H.; Kim, J.; Kim, B.; Lee, S. YOLO- Based Simultaneous Target Detection and
Classification
in Automotive FMCW Radar Systems. Sensors 2020, 20, 2897.
[7] M. Woollard, M. Ritchie and H. Griffiths, "Investigating the effects of bistatic SAR phenomenology
on eature
extraction," 2020 IEEE International Radar Conference (RADAR), 2020, pp. 906-911.
[8] E. R. Keydel, S. W. Lee, and J. T. Moore, “MSTAR extended operating conditions: a tutorial,” in
Aerospace/Defense Sensing and Controls, pp. 228-242, 1996.
[9] S. Yun et al., "CutMix: Regularization Strategy to Train Strong Classifiers With Localizable
Features", International Conference on Computer Vision, 2019.
[10] C. Y. Wang et al., "CSPNet: A New Backbone that can Enhance Learning Capability of CNN",
2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2020.
[11] T. Y. Lin et al., "Feature Pyramid Networks for Object Detection", IEEE Computer Society,
2017.
[12] S. Liu et al., "Path Aggregation Network for Instance Segmentation", 2018 IEEE/CVF
Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
Page 208
SEMIT 2023
Sporcular için Sakatlık ve Performans Analizi Çerçevesi

Injury and Performance Analysis Framework for Athletes
Osman Çaylı ve Fuat Karip
VBT YAZILIM AR-GE
VBT YAZILIM A.Ş.
İstanbul, Türkiye
Osman.Cayli@vbt.com.tr, Fuat.Karip@vbt.com.tr
Atınç Yılmaz
Bilgisayar Mühendisliği Bölümü
İstanbul Beykent Üniversitesi
İstanbul, Türkiye
atincyilmaz@beykent.edu.tr
Özetçe
Spor, dünyanın önde gelen sektörlerinden biri haline gelmiştir. Bu nedenle sporcu sağlığı, sektör
kazanımlarını veya zararlarını direkt etkileyen en önemli etkenlerden biridir. Sporcuların fiziksel sağlık
durumlarının ve performanslarının analizi yapılarak olası sakatlıklar veya performans düşüklükleri
öngörüldüğünde risklerin azaltılması mümkün olabilmektedir. Geleneksel yöntemler ile sınırlı olarak
tahminler yapılabilse de makine öğrenmesi yöntemlerinin desteği ile sporcuların performansını
arttırmak ve potansiyel sakatlık risklerini ortadan kaldırabilmek için önlem alabilmek mümkün
olacaktır. Bu çalışmada, makine öğrenmesi yöntemlerinin kullanımı ile sporcuların sakatlanma ve
düşük performans risklerinin tahmin edilmesi için modelleme yapılmıştır. Ortaya konan model bir
yazılıma entegre edilmiştir. Sporculardan elde edilen eş zamanlı veriler üzerinden beslenen model ile
sakatlık ve düşük performans konularında acil tedbir alabilme imkânı sağlanmıştır. Bu sayede yatırım
yapılan sporcuların mevcut riskleri minimize edilecek daha iyi performans göstermelerine olanak
sağlanacaktır. Çalışmada bulanık mantık yöntemi modelleme için kullanılmış ve %75.2 doğruluk
oranına ulaşılmıştır.
Anahtar Kelimeler: Uzman sistemler, bulanık mantık, kural tabanı, Sporcu Yönetimi, İyileşme
İzleme, Risk Analizi, Sakatlık Analizi, Veri Yönetimi, Yapoy Zeka.
Abstract
Ports have become one of the leading industries in the world. For this reason, the health of athletes is
one of the most important factors that directly affect the gains or losses of the sector. By analyzing the
physical health status and performance of the athletes, it is possible to reduce the risks when possible
injuries or poor performance are predicted. Although limited predictions can be made with traditional
methods, with the support of machine learning methods, it will be possible to take precautions to
increase the performance of athletes and to eliminate potential injury risks. In this study, modeling was
done to predict the injury and low performance risks of athletes by using rule based expert system
method. The proposed model is integrated into a software. With the model fed on the simultaneous data
obtained from the athletes, it was possible to take urgent measures regarding injury and low
performance. In this way, the current risks of the invested athletes can be minimized and they will be
allowed to perform better. In the study, fuzzy logic method was applied for modeling and an accuracy
rate of 75.2% was achieved.
Keywords: Expert systems; fuzzy logic; rule base; athlete management; recovery monitoring; risk
analysis; injury analysis; data management; ai.
Giriş (Introductıon)
Spor, bireysel veya ekip olarak yapılan içerisinde rekabet, yarışma ve eğlence gibi öğeler barındıran
fiziksel aktiviteler bütünüdür. Spor, kişisel sağlık durumunun korunması, performans veya fiziksel
gelişimin sağlanması, yeteneklerin gelişimi, sosyal ve kişisel gelişim gibi amaçlar için
Page 209
SEMIT 2023
yapılabilmektedir. Bunun yanında spor amatör olarak yapılabildiği gibi profesyonel olarak da
gerçekleştirilebilmektedir. Profesyonel olarak yapılan spor için ise rekabet unsurları daha çetin hale
gelirken; kurumsal veya kişisel yatırımlar çok daha ciddi seviyelere gelmektedir. Bu durumdan ötürü,
sporcunun yeteneklerini ve performansını en üst seviyede ortaya koyabilmesi büyük bir yarışma haline
gelmektedir. Sporcuların sakatlık ve performans öngörüsü bu nedenle oldukça önemlidir. Sporcular için
bu konularda yapılabilecek tahminler aşağıda listelenen durumları kazandırmaktadır.
• Sporcuların sakatlık risklerinin minimize edilmesi
• Sporcuların performanslarının geliştirilmesine rehberlik
• Sporcular için kişiselleştirilmiş idman programı
• Optimize edilmiş genel sporcu idman programları
• Performans düşüşü veya artışının analizleri
• Düşük performans riskinin önlemini alabilme
• Rekabet ortamında avantaj
Yapay zeka, sınıflandırma, kümeleme ve regresyon problemlerine uygulanabilen algoritmaları
barındıran yöntemler topluluğudur. Teşhis, tahmin, tespit, karar gibi birçok problem tiplerinde yüksek
başarım ile uygulanabilmektedir. Yapay zeka algoritmalarının uygulanabilmesi için en önemli unsur
tecrübe ve deneyim kaynağı olarak kullanılan veridir. Yapay zeka yöntemleri ile verilerden elde edilen
tecrübe ile öğrenme sürecinin ardından ileriye yönelik öngörüler yapılabilmektedir. Uzman sistemler,
insanların kazandığı bilgi ve uzmanlığı taklit ederek bu bilgi ve deneyimlerin yazılıma aktarılmasını
sağlan yapay zeka tabanlı bir yöntemdir. Sporcu sakatlık riski ve performans tahminleri için uzman
sistemler algoritmasının kullanımı ile sporcuların performansını iyileştirmek, sakatlık riskini minimize
etmek ve kişiselleştirilmiş çalışmalar ile sporcunun gelişimi sağlanabilecektir. Uzman sistemler
yöntemi uygulanırken, sporcunun fizyolojik ölçümleri, geçmiş müsabaka ve idman performans
değerleri, giyilebilir teknoloji ile sensörlerden alınan anlık değerler, sakatlık geçmişi gibi özniteliklerin
kullanımı ile sakatlık ve performans öngörüsü sporcular için yapılabilecek; böylece yatırım yapılan
sporcu ve organizasyonun riskleri en aza indirgenmiş olacaktır.
Sporcular ve kulüpler için performans yönetimi veri analiz yöntemlerinin uygulanması ile birlikte çok
daha etkin ve verimli hale getirilebilmektedir. Nesnelerin interneti teknolojilerinin de entegrasyonu ile
birlikte sporcuların eş zamanlı performanslarını takip etmek mümkün olmaktadır. Bu sayede nabız
değerleri, biyometrik veriler, antrenman performansları, GPS değerleri giyilebilir teknolojiler ile sürekli
veri alımını sağlayarak sporcu performansının analizi yapılabilmektedir. Ayrıca elde edilen veriler
kullanılarak sporcuların fizyolojik durumuna, ihtiyaçlarına ve ulaşılması hedeflenen noktalar
modelleme parametreleri olarak kullanılıp kişiselleştirilmiş antrenman programı geliştirilebilir. Aynı
şekilde mevcut veriler sporcuların fiziksel olarak durumlarının da öngörülmesine fırsat tanıdığından
ötürü sakatlık riskinin minimize edilmesi için hafif idman veya dinlenme önerisi sistem tarafından
yapılabilir. Yapay zeka yöntemlerinin spor yönetimi konusunda katabileceği bir diğer değer ise
sporcunun gelecekte ulaşabileceği potansiyel performanslarıdır. Bu öngörü ile sporcunun kariyer
planlaması veya ilgili sporcuya yatırım yapılıp yapılmayacağına dair karar alınması yapılabilir. Elde
edilen verilerden kazanabilecek bir diğer çıkarım ise rekabet unsurları içerisinde bulunulan rakip
sporcuların güçlü ve zayıf yönlerinin analiz edilerek kazanmak için en uygun stratejinin seçilmesidir.
UEFA ve Premier Lig seviyesinde yapılan araştırmalarda 4-6 hafta süreli sporcu sakatlanma sürecinin
kulüpler için oldukça yüksek seviyelerde maliyet getirdiği ifade edilmektedir. Bir sezondaki tüm sportif
sakatlanmalar düşünüldüğünde toplam maliyet çok ciddi boyutlara ulaşabilmektedir. Bu çalışma,
sporcuların yaralanma ve düşük performans risklerinin tahmin edilmesi için uzman sistemler yöntemi
ile desteklenen bir yazılım üzerine dizayn edilmektedir. Kural tabanlı uzman sistem yöntemi ile
desteklenen önerilen sistem, sporcuların yük, toparlanma ve sakatlanma verilerinin toplanarak, bu
verilerin analizi ile eş zamanlı olarak raporlanması sonucunda sakatlık ve düşük performans hakkında
gerekli tedbirlerin acil olarak alınmasını amaçlamaktadır. Önerilen sistemin sporcu analizi alanında
kullanımı ile sportif sakatlanma insidansının ortalama yarı oranda azalabileceği ve sakatlanma sonucu
kulüplerde oluşan maliyet kaybı konusunda ciddi faydalar getirebileceği hedeflenmektedir.
Bu çalışmanın katkıları ve yenilikleri özetle listelenmiştir:
Page 210
SEMIT 2023
• Uzman sistemlerin spor alanında uygulanabilirliğini ortaya koymak.

• Dünyadaki en büyük endüstrilerden biri olan spor için ticari risk barındıran sporcu sağlığı ve performans
analizi ile risklerin minimize edilmesini sağlamak.
Benzer Çalışmalar (Lıterature Revıew)

Sporcu sağlığı ve performans analizi ile ilgili yapılan literatür araştırması neticesinde sağlık bilimleri,
spor bilimleri, bilgisayar bilimleri, matematik bilimleri gibi alanlarda birçok benzer çalışmanın yer
aldığı görülmüştür. Bu çalışmaların bir kısmı sporcuların sakatlanma riskleri üzerine odaklanmış iken;
bir diğer kısmında ise sporcuların performanslarının öngörülmesi hedefi ile araştırmalar
gerçekleştirilmiştir. Bunun yanında özellikle güncel çalışmalar baz alındığında, sporcuların sakatlık
oluşumunu ve verimli performans göstermesini sporcuların fiziksel durumları ile etkisinin yanında
sporcuların psikolojik etmenlerinin de sonuçlara büyük etkisinin olduğu belirtilmiştir. Bunun yanında
çalışmalarda sıklıkla veri madenciliği, makine öğrenmesi, derin öğrenme algoritmalarının uygulandığı
görülmüştür.
Shen (2021) çalışmasında, yapay sinir ağları ve gömülü sistem teknolojileri kullanarak sporcuların
sakatlıklarının tahmin edilmesini sağlamıştır [1]. Önerilen yapay sinir ağları gömülü sistem yaklaşımı
dışında destek vektör makinesi ve derin sinir ağları yöntemleri de aynı durum için modellenmiştir.
Önerilen sistemin sonuçları incelendiğinde, %96 doğruluk oranı ile mevcut diğer yöntemlere göre daha
yüksek doğruluk ve performans gösterdiği ortaya konulmuştur. Gaudet ve ark. (2019) çalışmalarında,
omuz-dirsek skoru(KJOC) ile kapalı kinetik üst stabilite testinin (CKCUEST) kadın atletlerdeki omuz
yaralanmasının ilişkisi üzerine odaklanmışlardır [2]. Bu nedenle 34 yüzücü ve hentbol oyuncusundan
takım seçme denemeleri esnasında KJOC VE CKCUEST değerleri elde edilmiştir. Sakatlanmış veya
sakatlanmamış olarak sınıflandırılan veriler K Ortalamalar Yöntemi ile denetimsiz öğrenme
kullanılarak modellenmiştir. Çalışma neticesinde KJOC verileri kullanılan kümeleme yöntemi %86
duyarlılık, %100 özgüllük değerine ulaşmış; CKCUEST verileri için ise %86 duyarlılık %37 özgüllük
deri elde edilmiştir. Sonuç olarak KJOC verilerinin, omuz işlevini değerlendirmek ve kadın sporcularda
sakatlık olasılığını ortaya koyabilmek için tanısal geçerliliğe sahip oldu çalışmada ifade edilmiştir. Zhu
(2022) çalışmasında, sporcuların psikolojik deneyimlerini nicel olarak analiz etmek için yapay zeka
yöntemlerinden faydalanmıştır [3]. Ayrıca, geleneksel yöntemlerin verileri standartlaştırma
konusundaki yetersizliği düşünülerek ver işleme için normalizasyon uygulanmıştır. Ayrıca, verileri
analiz etmek için Monte Carlo simülasyon yöntemi modellemede kullanılmıştır. Yapılan deneysel veri
analizi sonuçlarına göre sporcuların psikolojik deneyimi için önerilen yapay zekaya dayalı akıllı karar
verme sistemi kullanıldığında sporcuların psikolojik antrenman etkisini etkili bir şekilde artırabileceği
ifade edilmiştir. Sarlis ve ark. (2021) çalışmalarında, 2010 yılından 2020 yılına kadar Ulusal Basketbol
Birliği-NBA’de yer alan verileri kullanarak sakatlıklarla ilişkili önemli özellikleri belirlemek ve bu
özelliklerin oyuncu ve takım performansı üzerindeki etkisini üzerine odaklanmışlardır [4]. Bu durumun
ortaya konulması için veri madenciliği ve makine öğrenmesi yöntemlerinden faydalanılmıştır. Ortaya
konulan sistem ile antrenörler, spor bilimciler, sağlık bilimciler, yöneticiler ve karar vericiler için en
yaygın sakatlıkları tanımaları ve müsabakalar esnasında olası sakatlanma durumlarını araştırmaları için
bilgi sağlanmaktadır. Ayrıca çalışmada en fazla sakatlık yaşayan oyuncular, takımlar ile birlikte en fazla
dikkat edilmesi gereken sakatlık türleri belirlenmektedir. Çalışma neticesinde kas-iskelet sistemi
bozukluklarının sakatlığa sebep olan en önemli neden olduğu ortaya konmuştur. Ayrıca, oyuncu ve
takımı tanımlayan bütünsel çok değişkenli bir modele dayalı olarak performans ve sakatlıklar arasında
zayıf bir pozitif ilişki olduğu sonucuna varılmıştır. Jelinek ve ark. (2014) ortaya koydukları
çalışmalarında, günlük fizyolojik testler sırasında toplanan parametrelere dayalı olarak Avustralyalı
futbolcuların performanslarını tahmin etmek için makine öğrenmesi yöntemlerinden regresyon
algoritmaları ve meta-regresyon yöntemleri uygulanmıştır [5]. İlk olarak farklı regresyon yöntemleri ile
tüm veriler kullanılmıştır. Daha sonra Rastgele Orman yöntemi kullanılarak özellik seçimi yapılmıştır.
Seçilen özellikler meta-regresyon modellemesinde kullanılmıştır. Önerilen üç adımlı sistemin,
geleneksel regresyon yaklaşımlarına kıyasla Avustralyalı futbolcuların maç performansına ilişkin
Page 211
SEMIT 2023
tahminlerin doğruluğunu geliştirdiğini göstermektedir. Çalışmada önerilen modelin, sporcu

performansı tanısı için 0.86 korelasyon katsayısı değeri elde ettiği belirtilmiştir. Beletsky ve ark. (2022),
yaptıkları çalışmada, beyzbol oyuncularından elde edilen sezon öncesi omuz MRI görüntülerini
değerlendirmek ve bulguların ardından sakatlanma, tedavi durumları konularına odaklanmışlardır [6].
Sonuç olarak sağlıklı ve sakat görüntüler arasında önemli bir fark olmasına rağmen bu görüntüler
üzerinden gelecekteki durum hakkında öngörüde bulunulamadığı belirtilmiştir. Huang ve ark. (2021)
araştırmalarında, sakatlık tahmini için yapay sinir ağı yöntemini uygulayarak sakatlık durumunu
önlemeyi amaçlamaktadır [7]. Önerilen yapay sinir ağı tabanlı hiyerarşik tahmin sistemi ile sakatlığın
öngörülerek fiziksel iş yükü üzerinde değişiklik yapılması sağlanabilmektedir. Çeşitli kaynaklardan
elde edilen 21 futbolcunun verileri test etmek için kullanılmıştır. Sonuç olarak önerilen model, PCA ve
karar ağacı yöntemleri ile kıyaslanarak sporcu sakatlığı tahmini için daha başarılı sonuçlar verdiği
ortaya konmuştur. Johnson ve ark. (2019) çalışmalarında, 3 boyutlu diz eklemi aşınması (KJM) riski
için derin öğrenme yöntemleri uygulamışlardır [8]. Sporcunun fazla idman veya vücuduna yüklenmesi
gibi durumlar KJM riskini arttırmakta; önerilen model ile erken uyarı yapılabileceği çalışmada ifade
edilmiştir. Önceden eğitilmiş CaffeNet evrişimli sinir ağı yöntemi ve çok değişkenli regresyon üç
hareket türü modellemede kullanılmış ve kıyaslanmıştır. 0.8895 doğruluk oranı ile elde edilen sonuçlar
geleneksel kuvvet plakaları uygulamaları yerine derin öğrenme kullanılarak sahada diz yaralanması
değerlendirmesinin uygulanabilirliğini göstermektedir. Alghamdi (2023) çalışmasında, giyilebilir
teknoloji ve tekrarlayan sinir ağları kullanarak futbolcuların sağlık durumları tahmin edilmiştir [9].
Önerilen sistem, sporcuların sağlık durumlarını gerçek zamanlı olarak izlemektedir. Sporcu sağlığı
tahminleri, tekrarlayan sinir ağları mimarisine giriş olarak iletildikten sonra elde edilmektedir.
Çalışmada elde edilen sonuçlar, önerilen yaklaşımın uygulanabilirliğini ve güvenilirliğini ortaya
koymaktadır. Javier ve ark. (2023) ortaya koydukları çalışmada, genç futbolcularda alt ekstremite
yumuşak doku(LE-ST) sakatlanmalarının tahmin edilmesi için makine öğrenmesi yöntemleri
uygulamışlardır [10]. İspanya futbol kulüplerine ait beş futbol akademisinden toplam 260 erkek, amatör
ve genç futbolcudan elde edilen kişisel özellikleri, psikolojik yapıları, fiziksel kondüsyonları ve
nöromüsküler ölçümleri barındıran verisetinden çalışmada faydalanılmıştır. Sonuç olarak uygulanan
yöntemler arasında boosting yaklaşımı ile 0.70 AUC değeri elde edilerek uygulanabilirliği gösterilmiş;
önerilen sistemin kullanılması ile çok az ekipman ile eğitimli bir personel tarafından tek bir oyuncu için
hızlı ve kolayca durum tahmini yapılabileceği çalışmada belirtilmiştir.
Yapay Zeka ve Uzman Sistemler (Artıfıcıal Intellıgence And Expert Systems)

Yapay zeka, öğrenme, tahmin, örüntü tanıma gibi yetenekler kazandırmayı sağlayan yöntemler
topluluğudur. Yapay zeka yöntemleri içeren sistemler, örneklerle öğrenmeye ve bu öğrenme
yeteneklerinin zaman içinde sistemin performansının gelişmesine olanak sağlayan uyarlamalı
mekanizma içermektedir. Yapay zeka algoritmalarının kullanımı ile doğal dil işleme, ses tanıma,
görüntü tanıma, öngörü yapma, teşhis etme ve karar alabilme gibi konuların çözümü
sağlanabilmektedir. Öğrenme kabiliyeti gibi oldukça güçlü bir yeteneğe sahip olan yapay zeka, dünyada
popüler çalışma alanlarından biri haline gelmiştir. Savunma sanayi, sağlık, ulaşım, spor, finans, lojistik,
perakende gibi alanlar başlıca olmak üzere her sektör ve her uygulama alanında yapay zeka
yöntemlerinin başarı ile kullanımı mümkün olmaktadır.
Yapay zeka yöntemleri, modellemek için elde edilen verileri analiz ederek veriler arasındaki ilişkiyi
ortaya çıkartarak tahmin yeteneği kazanmaktadır. Bunun yanında yöntemlerin tahmin başarısı
kullanılan yöntemin kısıtı ile farklı parametrelere de bağlıdır. Yapay zeka yöntemleri için veri en önemli
konumda bulunmaktadır. Bunun nedeni yöntem deneyim kazanımını ve ilişki keşfini veriler üzerinden
sağlayabilmektedir. Bu durumdan ötürü yapay zeka yöntemlerinin tahmin başarısı verinin kalitesi ile
doğrudan ilgilidir. Doğru, güvenilir, eksik olmayan veriler ile tahmin yeteneğinin başarısı daha üst
seviyelerde olmaktadır. Bunun dışında verinin niteliğinin yanında niceliği de önemlidir. Modellenmesi
beklenen konu ile ilgili ne kadar fazla veri var ise konunun örneklenmesi o ölçüde sağlanmış olmaktadır.
Ayrıca gerçek hayat problemlerinde çok sık rastlanan eksik veya gürültülü veriler için ön işleme ve
temizleme süreçlerinin uygulanması ile de modelin öngörü yeteneği arttırılabilmektedir. Yapay zeka
Page 212
SEMIT 2023
yöntemlerinin tahmin yeteneğinin optimize edilmesi, modelleme için yöntemin doğru şekilde seçilerek
uygun hiperparametrelerin ayarlanması ile direkt ilgilidir.
Uzman sistemler, herhangi bir konuda uzman bilgi ve becerinin bir sisteme aktarılması amacı ile ortaya
konulan yapay zeka tabanlı bir yöntemdir. Uzmanlık bilgileri farklı teknikler ile bir veya birden fazla
kaynaktan temin edildikten sonra bu bilgiler uzman sistemler yöntemi sayesinde karmaşık dahi olsa her
hangi bir karar verme problemi için uygulanabilir hale gelmektedir. Uzman sistemler aşağıdaki
elemanlardan oluşmaktadır:
• Bilgi veya Kural Tabanı
• Çıkarım Birimi
• Kullanıcı Arayüzü
Uzman sistemler içerisinde yer alan bilgi tabanı biriminde uzmanlar tarafından aktarılan uzmanlık
bilgileri depolanmaktadır. Bu bilgiler, uzmanların konu ile ilgili tecrübe ve bilgiler dahilinde problemin
çözümü ile ilgili kural tabanı içerisinde “eğer-ise” mantığında kurallandırılmaktadır. Çıkarım motoru
biriminde ise sisteme giriş olarak verilen değerler bilgi ve kural tabanın yer alan uzmanlık bilgileri ve
kuralları kullanılarak sistemin sonuç üretmesi için çıkarım yapılmaktadır. Uzman sistemlerin son birimi
olan kullanıcı arayüzünde ise uzman sistemler yöntemi destekli yazılım ile kullanıcı arasında köprü
görevi görülmektedir. Bu arayüz sayesinde kullanıcı sisteme değer girebilir ve sistemin sonucunu,
önerilerini görebilmektedir.
Uzman sistemler kural tabanlı yapısı sayesinde uzmanlık bilgisini iyi derecede işleyerek karar verme
problemlerinde etkin şekilde uygulanmaktadır. Karmaşık problemlerde dahi oldukça başarılı sonuçlar
üretebilmektedir. Uzman bilgisi ve tecrübe uzman sistemler içerisinde sürekli kullanılabilir halde
bulunmaktadır. Bunun yanında eksik bilgi bulunan veya keşfedilmemiş yeni durumların meydana
gelme olasılığı olan bir problem veya veri seti için uzman sistemler yönteminin seçimi uygun
olmayacaktır.
Bulanık Mantık (Fuzzy Logic)

Klasik mantık, kesin sonuçlar barındıran ve bir kümeye ya tam ait ya da hiç ait değil sonucunu veren
bir mantığa sahiptir. Klasik mantığın bu özellikleri gerçek dünya problemlerinin modellenmesi için çok
uygun değildir. Bu nedenle bulanık mantık yöntemi 1965 yılında Zadeh tarafından önerilmiştir. Bulanık
mantık modelinde geleneksel ikili mantık yerine 0-1 değerleri arasında aitlik derecesi alınabilinerek
birden fazla kümeye ait olabilme durumu meydana getirilmiştir. Dolayısı ile gerçek dünya problemleri
için geçerli olan kompleks ve belirsiz durumların modellenmesi bulanık mantık yöntemi ile efektif
olarak uygulanabilecek hale gelmiştir. Bunun yanında bulanık mantık insan karar verme
mekanizmasına en yatkın yöntemlerden biridir. Yöntem, sözel değişkenler ile modellenmektedir. Sözel
değişkenler, bulanık küme içerisinde tanımlanan üyelik fonksiyonlarında ve uzman bilgilerinden elde
edilen kural tabanında oluşturulan kurallarda kullanılmaktadır.
Bulanık mantık, belirsizliği barındıran durumları bulanık küme içerisinde indirgemek için
uygulanabilen bir yöntemdir. Bulanık küme içerisine sunulan veri, ilgili kümede yer alan ve sözel
değişkenlerle ifade edilen tüm üyelik fonksiyonlarına aitlik derecesi 0’dan büyük bir değere ulaşabilir.
Bu durum verinin bu kümelere ait olduğunu göstermektedir. Bir veri, [0-1] aralığında bir değer
alabilirken; 0 hiç aitliği 1 ise tam aitliği ifade etmektedir. Bunun dışında 0 ile 1 arasında alınacak
değerler ise derece olarak kümeye ait olduğunu göstermektedir.
Bulanık mantık sistemlerinde birimler aşağıdaki şekilde sıralanmaktadır:
1. Bulanıklaştırma
2. Kural ve Karar Birimi
3. Durulaştırma (Berraklaştırma)
Bulanık mantık sistemleri modellenirken ilk olarak sisteme ait giriş ve çıkış bulanık kümelerinin
modellenmesi gerekmektedir. Bulanık kümeler oluşturulduktan sonra her bulanık küme için üyelik
fonksiyonlarının sözel değişkenler ve parametreleri ile oluşturulması sonraki adımda yapılması gereken
işlemdir. Bulanık kümeler ve üyelik fonksiyonlarının tanımları yapıldıktan sonra kural tabanı bulanık
mantık modeli için oluşturulması gerekmektedir. Bulanık kural tabanı eğer-ise mantığını içerisinde
Page 213
SEMIT 2023
taşıyan birimdir. Bu birimde karar verme süreçleri modellenmektedir. Kural tabanı ile karar verme
birimi insan karar verme mekanizmasına hem sözel değişkenler ile kullanılması hem de kural ifadeleri
bakımından oldukça benzemektedir. Tüm kurallar uzman bilgileri ile oluşturularak modelleme
tamamlanır.
Modelleme tamamlandıktan sonra sisteme giriş olarak sunulan veri ilk olarak bulanık küme içerisinde
bulanıklaştırarak tüm kümelerde hangi üyelik fonksiyonlarına ne kadar ait olduğu hesaplanır. Elde
edilen bulanık değerler kural tabanında bulunan ifadelerdeki karşılığı vererek her kuralın tetikleme
seviyesi elde edilir. Karar verme biriminde ulaşılan karar değeri hesaplandıktan sonra son işlem adım
olarak bulanık olan değere durulaştırma işlemi ile gerçek sonuç bulunur.
Deneysel Çalışmalar (Experımental Studıes)

Önerilen sistem, sporcuların sakatlanma ve düşük performans risklerinin tahmin edilmesi için yapay
zeka yöntemleri ile desteklenen bir yazılım üzerine dizayn edilmektedir. Çalışma, alanındaki
Türkiye’de ilk yazılımdır. Önerilen sistemde, sporcuların yük, toparlanma ve sakatlanma verilerinin
toplanarak, bu verilerin analizi ile birlikte eş zamanlı olarak raporlanması sonucunda sakatlık ve düşük
performans hakkında gerekli tedbirlerin acil olarak alınmasını sağlamaktadır. Sistem için veri girişi ve
tüm görüntülemeler, kolaylıkla mobil uygulama ile mobil cihazlar üzerinden yapılabilmektedir.
Önerilen sistemin yenilikçi yönlerinden biri monitörizasyon olarak tanımlanan izleme işleminin
sporculardan belirli periyotlarda sürekli verinin toplanmasının yapılmasıdır. Ülkemizde monitörizasyon
kavramının tanınması ve gelişmesini sağlaması düşünülen bu çalışma ile bu sistemden faydalanacak
olan kulüplerin daha etkin sonuçlar alabilmesi adına ölçüm laboratuvarlarını kuracağı öngörülmektedir.
Bunun yanında UEFA ve Premier Lig seviyesinde yapılan araştırmalarda 4-6 hafta süreli sporcu
sakatlanma sürecinin kulüpler için 400-600 bin euro aralığında maliyeti olduğu ifade edilmektedir. Bir
sezondaki tüm sportif sakatlanmalar düşünüldüğünde toplam maliyet çok ciddi boyutlara
ulaşabilmektedir. Kurulacak ölçüm laboratuvarları ile önerilen sistemin birlikte kullanımı ile sportif
sakatlanma insidansının ortalama yarı oranda azaltabileceği ve sakatlanma sonucu kulüplerde oluşan
maliyet konusunda ciddi faydalar getirebileceği önerilen sistem neticesinde hedeflenmektedir. Önerilen
sistemin diğer yenilikçi yönü ve Ar-Ge katkısı ise sporcu performansının tahmini ve sakatlanma riskinin
ortaya konması için yapay zeka yöntemlerinden faydalanılmasıdır. Eş zamanlı veri sağlanması ve
raporlama ile sürekli öğrenen bir sistem sayesinde anlık olarak oluşabilecek tüm riskler öngörülerek
gerekli tedbirlerin alınması yapay zeka yöntemlerinin uygulanması ile sağlanmaktadır. Ayrıca
kurulacak ölçüm laboratuvarlarında sporcu sağlığı ve idman yöntemleri gibi farklı analiz sonuçlarının
elde edilmesi de Önerilen sistemin diğer yenilikçi hedeflerinden olmaktadır.
Sporcuların sakatlanma ve düşük performans risklerinin tahmin edilmesi için çalışmada uzman
sistemler ile bulanık mantık modellemesi gerçekleştirilmiştir. Futbol kulüplerinde yer alan sporcuların
performans değerleri, sporcu bilgileri modellemede kullanılmıştır. Bunun yanında spor yönetimi ve
sporcu performansı konusunda uzman desteği ile sporcu yönetimi konusunda bulanık kümelerde yer
alan üyelik fonksiyonların değerlerinin belirlenmesi ve tüm kuralların yapılandırılması sağlanmıştır.
Modelleme tamamlandıktan sonra modelin optimize edilmesi için farklı veriler ile sistemin
uygulanabilirliği sınırlı veriler ile sınanmıştır.
Çalışmada gerçek sporcu verileri üzerinden anonimleştirilen bir veriseti kullanılmıştır. Verisetinde
toplam 411 örnek bulunmaktadır. Bunun yanında verisetinde 68 giriş parametresi ve 1 sonuç
parametresi bulunmaktadır. Çalışılan modelde hedef parametre ortalama risk olarak belirlenmiştir.
Modelin oluşturulmasında KNIME platformu kullanılmış; elde edilen parametreler ve model sporcu
risk yazılımına entegre edilerek uygulanması sağlanmıştır. İlk olarak veriseti üzerinde eksik verilerin
kontrolü sağlanmış; eksik veriler ortalama değer ile tamamlanmıştır. Daha sonrasında ise optimize
edilmiş Kural tabanlı bir uzman sistem oluşturulması için bulanık mantık yöntemi probleme
uygulanmıştır.
Modelleme için veriseti eğitim ve test seti olarak ayrılmıştır. Verilerin %70’i eğitim seti olarak; kalan
%30’u ise test seti olarak uygulanmıştır. Farklı oranlarda ve biçimlerde eğitim-test seti belirlenerek en
uygun eğitim ve test setinin oluşturulması sağlanmıştır. Çalışmada uygulanan yöntemin başarısını
değerlendirmek için performans metrikleri incelenmiştir (Tablo 1). Sonuçlar incelendiğinde bulanık
mantık yöntemi modellenen bulanık mantık sistem ile %75.2 doğruluk oranı ve 70.9 Cohen Kappa
katsayısı değeri elde edilmiştir.
Page 214
SEMIT 2023
TABLO III. UYGULANAN YÖNTEMİN PERFORMANS DEĞERLERİ (PERFORMANCE VALUES OF APPLIED

METHOD)
Yöntem Adı
Oranı (%) Katsayısı
Bulanık Mantık 70.9
75.2
(BM)
Çalışmada uygulanan yöntem ve yaklaşımlar ile ortaya konan sistem yaklaşımı ile sporcunun sakatlık
riskinin görülmesi ile sporcuya yönelik farklı idman teknikleri veya dinlenme gibi çözümler ortaya
konarak sakatlık riskinin minimize edilebilmesi olanağı sağlanmaktadır.
Sonuç ve Tartışma (Result and Dıscussıon)

Önerilen sistem ile hedeflenen kısa ve orta vadede kulüplerde yaşanan sportif sakatlanma sayılarında
ve bu sakatlanmalara bağlı mâliyetlerde bir düşüş trendi sağlamak hedeflenmiştir. Özellikle UEFA ve
Premier Lig seviyesinde yapılan araştırmalarda 4-6 haftalık bir sporcunun sakatlanmasının kulüpler için
400-600 bin euro aralığında değişen bir mâliyeti olduğu saptanmıştır. Pandemi nedeni ile neredeyse her
bir maçta bir sakatlanma istatistiğine ulaşılan 2020-21 sezonu ve normal şartlarda takımların bir
sezonda ortalama 40-50 aralığında sportif sakatlanma yaşandığı hesaba katılırsa toplam mâliyet çok
ciddi boyutlara ulaşabilmektedir.
Ayrıca, yetersiz toparlanma, yetersiz dinlenme ve erken antrenmana dönüş sinyallerinin tespit edilmesi
sayesinde antrenman planlamasında yapılacak değişikliklerle sporcuların fiziksel etkinliğinin artacağı
da unutulmamalıdır. Tüm bunların yanında önerilen sistem, kulüplerin atletik performans ve medikal
veri yönetimini kolaylaştıracak, departmanlar arası iletişim ve bilgi aktarımını hızlandıracak, ortak bir
raporlama dili oluşturacak ve kurumsal veritabanı için dijital bir zemin de hazırlayacaktır.
Çalışmada uzman sistemler-bulanık mantık modellemesi ile sporcuların sakatlanma ve düşük
performans risklerinin tahmin edilmesini sağlayan bir sisteme odaklanılmıştır. Uzman sistemlerin
sporcu performans yönetimi konusunda uygulanabilir olduğu çalışmada gösterilmiştir. Elde edilen
%75.2 doğruluk oranı ve 70.9 Cohen Kappa katsayısı değeri ile sporcu performans ve sakatlanma
öngörüsü problemine bulanık mantık yöntemi modellinin kullanılabilirliği ortaya konmuştur. Bunun
yanında yapay zeka, nesnelerin interneti gibi bir çok yenilikçi teknolojinin kullanımı ile spor alanı gibi
tüm dünyada büyük yatırımların yapıldığı bir sektör için risklerin minimize edilebileceği, verimin ve
kazancın arttırılabileceği plot bir çalışma ile anlatılmıştır. Çalışmada kullanılan verisetinde futbol
alanında yer alan sporcular bulunduğundan ötürü diğer spor dallarında da modellemenin
uygulanabilirliği ilerideki çalışmalarda odaklanılması düşünülmektedir. Bu nedenle önerilen sistemin
kısıtı futbolcular ile sınırlıdır denilebilir. Ayrıca farklı yapay zeka yöntemlerinin sporcu performans
yönetimi modeli için geliştirilmesi ile daha yüksek performansa ulaşılabilmesi diğer amaçlardandır.
Kaynaklar (References)
[1] Shen, H. "Prediction simulation of sports injury based on embedded system and neural network",
Micropocessors and Microsystems., Vol. 82, 2021 103900.
[2] Gaudet. S, Begon, M., and Tremblay, J. "Cluster analysis using physical performance and self-
report measures to identify shoulder injury in overhead female athletes”, Journal of Science and
Medicine in Sport, Vol. 22(3), 2019, p 269-274.
[3] Zhu, H. "Research on intelligent analysis strategies to improve athletes' psychological experience
in the era of artificial intelligence ", Progress in Neuro-Psychopharmacology and Biological
Psychiatry, Vol. 119(20), 2022, 110597.
[4] Sarlis, V., Chatziilias, V., Tjortjis, V., and Mandalidis, D. "A Data Science approach analysing the
Impact of Injuries on Basketball Player and Team Performance", Information Systems, Vol. 99,
2021, 101750.
[5] Jelinek, H.F., Kelarev, A., Robinson, D.J., Stranieri, A., and Cornforth, D.J. "Interactive problem
solving with speech", Applied Soft Computing., Vol. 14, 2014, p 81-87.
[6] Beletsky, A., Okoroha, K.R., Cabarcas, B., Garcia, G.H., Gowd, A.K., Meyer, J., Vadhera, A.S.,
Singh, H., Gursoy, S., White, G., Davidson, J., Nicholson, G.P., Chahla, J., and Verma, N.N.
"Magnetic Resonance Imaging Findings of the Asymptomatic Shoulder May Impact Performance,
Page 215
SEMIT 2023
Not Future Injury List Placement in Major League Baseball Pitchers", Arthroscopy, Sports
Medicine, and Rehabilitation, Vol. 4(2), 2022, p 487-493.
[7] Huang, C., and Jiang, C. "Data monitoring and sports injury prediction model based on embedded
system and machine learning algorithm", Micropocessors and Microsystems, Vol. 81, 2021,
103654.
[8] William R. Johnson, W.R., Mian, A., Lloyd, D.G., and Alderson, J.A. "On-field player workload
exposure and knee injury risk monitoring via deep learning ", Journal of Biomechanics, Vol. 93,
2019, p 185-193.
[9] Alghamdi, W.Y. "A novel deep learning method for predicting athletes’ health using wearable
sensors and recurrent neural networks", Decision Analytics Journal., Vol. 7, 2023, 100213.
[10] Robles-Palazón, F.J., Puerta-Callejón, J.M., Gámez, J.A., Croix, M.D.S., Cejudo, A., Santonja,
F., de Baranda, P.S., and Ayala, F. "Predicting injury risk using machine learning in male youth
soccer players", Chaos, Solitions & Fractals, Vol. 167, 2023, 113079.
Page 216
SEMIT 2023
Exploring the trends of Artificial Intelligence in Recruitment: A Bibliometric study

Naval Lawande1[0000-0002-1155-6448]
1
Symbiosis Institute of Management Studies, Symbiosis International (Deemed University), India
naval.lawande@gmail.com
Abstract. Purpose: Artificial Intelligence (AI) as a field has recently evolved as a transformative force
in the field of recruitment. Numerous empirical, conceptual, and exploratory studies have been
conducted that indicate the novel ways organizations identify, select, and attract top talent.
Nevertheless, no scientific mapping or attempt to map the literature scientifically in the domain has
been done in the past. Hence, the paper intends to showcase the trends in the field of AI and Recruitment
research through of knowledge and conceptual structures.
Methodology: An analysis using bibliometric tools in the field of AI and recruitment was conducted. A
sample of 1450 documents was extracted from the Scopus database. This was based on a search strategy
determined by the author. An application that is equipped with a bibliometric package was installed.
This software enabled the analysis of the dataset and various themes, patterns and trends were identified
based on the authors, countries, and citations.
Findings: Results show that the domain of AI and Recruitment needs to be given direction. There is
also a pressing need for interdisciplinary research in this area. The article provides some crucial insights
on areas that need further inquiry. The conceptual and social network structure depict an upward trend
in terms of this area of research. There is a growing demand for Talent Acquisition practitioners and
Recruiters with AI skill sets.
Research limitations/implications: The current paper only had Scopus Database at its backdrop. Future
researchers could use multiple databases such as Web of Science and conduct a comparative study. A
systematic literature review would widen the scope and help identify some of the niche uncharted
territories of Recruitment and artificial intelligence.
Keywords: Artificial Intelligence, Recruitment, Bibliometric Study.
1. Introduction
Evolving role of Artificial Intelligence in Recruitment:
Efficiency and effectiveness have always been the pillars of productivity for all industries and sectors.
Firms are constantly seeking innovative ways to streamline the processes to accomplish the same.
Integration of Artificial Intelligence (AI) into the business operations has been one such intervention
that has been initiated by companies. The domain of recruitment has also witnessed the inclusion of AI;
wherein the traditional methods have been augmented with AI-driven solutions and mechanisms.
Convergence of AI and recruitment practices has indeed enhanced the efficacy of the firms and the
Human Resources (HR) departments in identification, evaluation, and selection of candidates.
Understanding the dynamics of recruitment and AI becomes paramount to remain at the forefront of
innovation and talent acquisition.
Research Objectives:
A two-pronged approach has been taken with regards to the research questions. Firstly, what is the
overall publication scenario and structure of recruitment and artificial intelligence? Secondly, how has
been the evolution of the concept of artificial intelligence and its application in recruitment; which are
the most recent trends and issues?
Based on the above perspective, following are the research objectives:
• To elicit the patterns or trends in the knowledge development in the area of recruitment and artificial
intelligence.
• To investigate the knowledge structure and derive synthesis of knowledge in the area of AI based
recruitment.
Page 217
SEMIT 2023
Network and descriptive type of analysis were used to attain the research objectives. Different
sections of the paper touch upon the varied elements. The immediate section deals with the research
methodology, followed by the section on data analysis. The following section post that brings to light
discussion on the findings, while the research directions and conclusions are presented in the last
section (Ingale & Paluri, 2020).
Bibliometric Analysis Technique:
The bibliometric analysis technique has been immensely popular in the last decade. The technique has
a high efficacy as the incorporates both mathematical and statistical analysis. Accurate and precise
patterns can be identified through the available literature. (Roblek et al., 2022) in their study explain
“qualitative methods such as observation and interviews can also be used for analysis. However,
bibliometric analysis is based on the science mapping approach. It applies the generic process of
visualization and domain analysis, thereby allowing scientific mapping of both the scientific groups and
research areas.” Another study by (Bragge et al., 2019), helped in throwing light on the research clusters
and intellectual structure of management research on three decades of publications in the area. (Cortés-
Sánchez, 2020) reinforced through their paper the over whelming leap business, management and
account research has taken over the last century.
2. Literature Review
Recruitment alias Talent Acquisition
Recruitment, now commonly known as Talent Acquisition (TA) has been a critical function for
businesses across sectors and geographies. Nevertheless, the conventional approach often involves
manual tasks, that are time consuming, prone to biases and can result in suboptimal candidate matches.
(Allal-Chérif et al., 2021) suggest in their research that, digital technologies ought to be utilized at all
stages of the recruitment process. The recruitment or talent acquisition process is extremely critical to
any organization. The model developed by (Roy et al., 2021) is based on extensive experimental
evaluation. The model shows effectiveness and efficiency of the recruitment process, if the model is
imbibed upon.
Artificial Intelligence
AI in HR (Human Resources) has ushered in a new era of efficiency and innovation. By harnessing the
power of artificial intelligence, HR departments can automate various tasks, ranging from initial
candidate screening and resume analysis to employee onboarding and performance evaluations. As
(Chilunjika et al., 2022) aptly state “AI-driven tools enable more accurate predictions of workforce
trends, helping organizations make informed decisions about recruitment, retention, and training.” The
study established that Artificial Intelligence is beneficial in the sense that it can improve public service
delivery in South Africa as the HRM personnel is enabled to focus more on the strategic areas of
management by taking over routine tasks, and that it helps minimize bias in public service recruitment
and selection (Chilunjika et al., 2022). Additionally, AI aids in identifying patterns of employee
engagement and satisfaction, allowing companies to proactively address potential issues and enhance
workplace culture. However, while AI enhances HR processes, ethical considerations such as data
privacy, fairness, and transparency must be upheld to ensure that technology serves as a force for
positive change within the realm of human resources. A study by (Kot et al., 2021) “found that AI based
recruitment and quality significantly influences the AI adoption and further it influences the employer
reputation. The mediation role of artificial intelligence adoption is significant where it is found that
artificial intelligence mediates the relationship between artificial intelligence recruitment and employer
reputation, with similar significant mediation role between artificial intelligence quality and employer
reputation.” Another interesting study by (Niehueser & Boak, 2020) concluded that “the introduction
of AI considerably improved the speed and efficiency of the work processes. The research found that
those employees who had used the new technology were positive about its effects, indicating that it was
Page 218
SEMIT 2023
easy to use, robust and highly productive. A proportion of employees who had not, at the time of the
research, used the new system, were less sure that it would improve their ability to do their job.”
Recruitment and Artificial Intelligence
The symbiotic relationship between AI and recruitment provides a plethora of benefits. It can expedite
the initial screening of candidate resumes and application, saving time for recruiters to focus on the
strategic aspects of human resource planning. The AI-powered algorithms can identify patterns in
candidate skills, experiences and match them with Job Description; leading to a person–job fitment.
Additionally, it also helps in mitigation of biases in selection processes through objective criteria can
be achieved. This would nurture in the long-term diversity and inclusivity. Currently, automated resume
screening, chatbot driven initial interview, are being employed in recruitment. (Nawaz, 2019) in their
study explained how AI is enabling the effectiveness of their recruitment processes. (Wang et al., 2021)
in a more recent study evidently stated how “Personnel recruitment and selection is changing rapidly
with the adoption of artificial intelligence (AI) tools.” Results of the study by (Wang et al., 2021) show
that “AI tools encourage a larger number of quality application submissions and for two reasons; First,
AI entrains a perception of a novel approach to job searching. Second, AI is perceived to be able to
interactively tailor the application experience to what the individual applicant expects and has to offer.
These perceptions increase the likelihood the user will submit a job application and so improves the
size and quality of the pool from which to recruit personnel.” As per the article by (Ore & Sposato,
2022) “The findings revealed that AI facilitates the effective performance of routine tasks through
automation.”
3. Methodology
Research Design and Flow Diagram
Below is the flow diagram of the bibliometric analysis. The analysis would need an identification of the
research database, then followed by the data extraction based on the search approach. The nature of the
data is secondary. The data needed for the research was extracted after the author decided that ‘Scopus’
was ideal database for the same. A search query was run using the appropriate mixture of keywords.
The dataset was established based on the inclusion and exclusion criteria. The data was then analyzed
using R-language’s bibliometric featured tools. An initial descriptive analysis of the documents, sources
and authors was conducted. Principal component analysis and multiple correspondence analysis
techniques were used for data reduction. In order to depict better data visualization, network maps were
created. The network maps revealed the intellectual, conceptual, and social structures. The detailing of
the stages is discussed below.
Page 219
SEMIT 2023
Fig. 1. Flowchart for selection of documents

Selection of database:
A crucial pre-requisite for bibliometric analysis is a structured database description of indexed articles.
Data retrieval was conducted from the ‘Scopus’ database. Based on Elsevier's Scopus coverage of
articles and classification of journals, the database was linked to the Bibliometrix package in R-studio.
Scopus database is an ideal database for bibliometric analysis as it covers highly credible journals from
various categories and excludes papers from non-scientific journals and magazines (Mulet-Forteza,
2019).
Data Analysis – Preparation:
The data was downloaded from Scopus in Microsoft Excel format to support the software requirement.
Documents were shortlisted based on the following search criteria applied on 19th August 2023.
Search strategy – Identification of Keywords:
The popular research contributions use the words ‘recruitment’ and ‘artificial intelligence’
interchangeably. 15458 documents showcased the search for these terms. The author pertinently
realized that the humongous set of data would not serve the objectives of the study.
Search strategy. Keywords: “recruitment” and “artificial intelligence”
Refined by: Scopus Categories (Business, Management and Accounting) OR (Computer Science)
Document types: Article, Conference Paper
Languages: English
Timespan: 1993 – 2023
Selection of Language
In this step the shortlisted documents of the earlier step were screened through language filter –
“English.” The final dataset extracted was 1450 documents. The key information related to extracted
documents viz. title, abstracts, authors, keywords, were uploaded in a Microsoft Excel format to
‘Biblioshiny’ – Bibliometrix related R-studio package tool (Ingale & Paluri, 2020). The tool was
developed as part of the package by (Aria & Cuccurullo, 2017). The tool helped in identification of
themes, prominent journals, most cited authors, prominent countries, critical articles, and depiction of
citation, co – citation and social network analysis.
Selection of time span:
The few decades have seen enormous interest in the concept of artificial intelligence. The core objective
of the paper is to detect the perspectives and trends in the field of artificial intelligence and recruitment
from 1993 – 2023.
Selection of document types:
Conference Proceeding, Book Series and Trade Journals were included in the data being extracted. This
enabled the data to be fine-tuned to meet the research objectives. Hence, the final shortlisted data
comprised of 1450 documents which was used for further analysis.
Selection of subject categories:
Page 220
SEMIT 2023
As stated earlier, refining of the data was critical. Hence, the search was refined by categories which
included Business, Management and Accounting and Computer Science. This stage resulted in 2041
documents being extracted.
4. Data Analysis
Descriptive Analysis
The data analysis was primarily a two-pronged approach. A scientific mapping and descriptive analysis.
This is explained in Figure 2. In scientific mapping, an extensive science mapping is conducted through
visualization methods such as thematic maps, network analysis and three field plots. It helped the
researcher derive the knowledge structures to facilitate further analysis (Nusair et al., 2019). The
Descriptive analysis part focused on exploring the bibliometric data in terms of the three elements, the
sources, the authors, and the documents.
Fig. 1. Bibliometric Analysis – Hierarchy
Descriptive analysis: This section of the paper primarily focuses on the descriptive analysis. There are
numerous dimensions undertaken for the extensive analysis.
Description Results
MAIN INFORMATION ABOUT DATA
1995:202
Timespan
2
Sources (Journals, Books, etc.) 764
Documents 1450
Annual Growth Rate % 23.38
Document Average Age 5.69
Average citations per doc 17.96
References 53783
DOCUMENT CONTENTS
Keywords Plus (ID) 8458
Author's Keywords (DE) 3842
AUTHORS
Page 221
SEMIT 2023
Authors 4202
Authors of single-authored docs 139
AUTHORS COLLABORATION
Single-authored docs 151
Co-Authors per Doc 3.54
International co-authorships % 21.17
DOCUMENT TYPES
Article 582
Conference Paper 868
Table 1. Brief summary of the data set.
Dataset: A brief summary of the data is depicted in Table 1. It gives an overview of the various
dimensions undertaken for the purpose of analysis. The bibliometric data frame consisted of 1450
documents extracted through a stepwise search query on Scopus database. Interestingly, these
documents were published in 764 sources with an ‘average citation score’ of 17.96 and a significant
‘collaboration index’ of 3.54. This was a welcome sign, indicating substantial amount of collaboration
amongst researchers.
Three field plots: The ‘Three Field Plot’ is based on the Sankey Plot mechanism. The plot gives a clear
relationship between the three fields, wherein the size of the portion is proportionate to the value of the
node (Ingale & Paluri, 2020). There are three rows as shown in Figure 3. The left most side of the
Sankey Plot are the Countries of Origin of the authors; the middle column are the authors and the right
most side are the keywords. The technology related keywords evolving from the analysis are data
mining, learning systems, machine learning and decision making. Some of the human resources related
keywords are human resource management, employment, recruitment process and personnel.
Intriguingly, the keywords evolving from AI and recruitment seem to be almost equal in number.
Fig. 3. Three Fields Plot

Source Growth: The scientific production in the domain of artificial intelligence and recruitment
showed an upward trend in Figure 4. Despite the author choosing a period of three decades (1993 –
2023), the analysis is a decent depiction of the trends in AI and recruitment. There is a steep rise in the
Page 222
SEMIT 2023
interest in the concept of AI and recruitment post 2017. A partial plateauing of the graph happens
between 1995 and 2005. Post 2019 there is a significant rise in the publications in the field. The surge
during the year(s) 2020 and 2021 may be attributed to the advent of the pandemic. Organizations started
adopting methods of effective Human Resource (HR) practices. (Saxena et al., 2021) in their study
showcased challenges in the adoption and acceptance of HR Analytics in organizations. Study done by
(Vitak & Zimmer, 2023) highlighted the fact that the future of work focuses on developing effective
mechanisms to utilize technologies for enabling human resource practices to be enabled with work from
home (WFH).
Fig. 4. Scientific production and growth

Figure 5 shows a fluctuating trend in terms of the concept. The average article citations per year is
declining post 2012. The field seems to be highly contextual and volatile in nature.
Fig. 5. Average Citations per annum
The Figure 6. below showcases the most relevant words. Clearly, after AI, Machine learning and
recruitment tend to be gaining maximum interest among scholars in the field. E – recruitment is in the
top five most relevant words. (Aydın & Turan, 2023; Pandey et al., 2020; Wang et al., 2021) in their
studies have supported the usage of AI in talent acquisition and e – recruitment. This clearly indicates
the recruitment trends are influenced by technology, automation AI and machine learning and vice
versa. The fluctuation in the graph is a clear reflection of the recruitment trends during the 2007 – 2010
period. The 2008 subprime crisis and corporate governance issues in organizations are reflected in many
studies. A study by (Kara et al., 2022) described the impact of the subprime crisis various facets of
business.
Page 223
SEMIT 2023
Fig. 6. Most Relevant Words
Fig. 7. Word growth trend
Authors: Zhu H and Xiong H were identified to be the most published authors. These authors had the
maximum publications in the domain of AI and recruitment. It can be seen from Figure 8 that even
though three authors stand out in terms of number of publications; however, the other authors have a
relatively significant contribution in the domain. In Figure 10 the Author Impact is shown, Zhu H and
Xiong H standout as the most impactful authors.
Fig. 8. Most Relevant Authors
Page 224
SEMIT 2023
Fig. 9. Author Impact
Country-wise contribution: Getting a global perspective on the domain helps in better understanding of
the field of study. The contribution of different countries to the domain or field helps in understanding
the growth pattern of the literature.
Fig. 10. Most Cited Countries
Documents: In Figure 11 (Most Relevant Sources) two journals completely standout of the crowd. The
first journal is Lecture Notes in Computer Science with 184 documents. This is followed by the journal
titled CEUR Workshop Proceedings with a miniscule 33 documents.
Fig. 11. Most Relevant Sources
Page 225
SEMIT 2023
Keywords: Figure 12 is the ‘Word Cloud’ for the term(s) artificial intelligence and recruitment. The
Word Cloud is a visualization and representation of the frequency of the keyword. It is obvious that the
terms stand out in the word cloud. The significance of the word can be determined by the size of the
word. Terms like e-recruitment, human resource management and machine learning also second in line
in terms of their weightage in the domain. It can be inferred that artificial intelligence does act as an
antecedent to ensure effective recruitment and human resource practices. Words like fairness and ethics
help understand that many studies have explored these elements of AI and recruitment. The word social
media in the word cloud is a welcome sign for future research in the field. Studies by (van Esch &
Black, 2019) (Basri, 2020) are aligned with the finding s of the current research.
Fig. 12. Word Cloud
Trends topics: As indicated in Figure 13., the topics trending the most are ‘human resource
management’, ‘machine learning’, ‘decision making’ and ‘recruitment’. (Aydın & Turan, 2023; Tian et
al., 2023) through their studies support the relevance of the trend topics of machine learning and
recruitment. The studies (Kaushal et al., 2023; Kaushal & Ghalawat, 2023) in recent times have set the
research agenda for the role of AI in Human Resource Management.
Fig. 13. Trend Topics

Data visualization: There is a mounting interest in the domain of (AI) artificial intelligence and
recruitment. The current section of the paper focuses on the thematic evolution of the domain. Data
visualization is a quantitative technique that assesses the clusters, occurrences, and associations
amongst multiples units of analysis. The approach is to extract network linkages by analyzing various
documents, authors, and keywords. The statistical analysis performed generates a map that indicates the
Page 226
SEMIT 2023
various network linkages. As can be seen in Figure 2, this scientific mapping culminates into three types
knowledge structures. They are social structure, conceptual structure, and intellectual structure.
Conceptual Structure: The ‘conceptual structure’ depicts the interface between the 3 T’s (topics,
themes, trends). Figure 14 shows the co-occurrence network diagram. It primarily displays the Co-
occurrence network with the nodes having five distinct colors. Artificial intelligence is in red color,
machine learning is in purple color. This is also known as co-word analysis. The content of the papers
is extracted and analyzed. Alongside AI, topics such as technology, blockchain, algorithms, are seemed
to align. The red nodes are representative of Artificial Intelligence (AI). Machine Learning is a close
cousin of AI. They have deep rooted linkages with AI. Themes such as fuzzy logic, deep learning,
natural language processing, automation, data science and neural networks are being studied alongside
AI. The purple nodes are representative of the parent node – machine learning. Results of a study
conducted by (Pessach et al., 2020) showed that in comparison to actual recruitment decisions, machine
learning based framework of recruitment improved the diversity and overall success rate of talent
acquisition in the organization. The crux of the study is in supporting the hypothesis that recruitment
and artificial intelligence research is predominantly being studied and published. Concepts and practical
aspects such as Human Resource Management (HRM), selection, recruitment is placed very near
artificial intelligence node. This is indicative of the significant relation between the two concepts being
researched widely. (Singh & Arora, 2019) in their study hypothesized that AI can be effectively used
as tool for talent acquisition. The findings here resonate with the findings of the study by (Singh &
Arora, 2019).
Fig. 14. Co-occurrence Network

Thematic Map: A ‘Thematic Map’ is plotted on a two-dimensions, with typological themes. With
reference taken from the co-occurrence and co-word analysis, keyword clusters are identified. This
helps in the next step of producing themes in the research domain. As seen in Figure 16, the mapping
is based on two axes. Centrality and density of the themes determines its location on the two-
dimensional graph.
Page 227
SEMIT 2023
Fig. 15. Thematic Map

The theme is represented by a bubble on the thematic map. As explained by (Viedma-Del-Jesus et al.,
2011), the Thematic map has four quadrants, the upper right quadrant is a motor theme; the lower right
quadrant is a basic theme; the lower left quadrant is an emerging or declining theme; and the upper right
quadrant is a niche theme. The linkage to the domain of AI is evident from the figure, wherein the
domain of AI is highly inter-twined with recruitment.
Intellectual Structure: The intellectual structure comments on the impact authors have on the scientific
community. The alliance between authors and countries is studied. The degree of collaboration and
cooperation between the research fraternity and their affiliations with various universities (Ingale &
Paluri, 2020). An interesting 3-in-1 framework was designed by(Pampouktsi et al., 2021) for selection
and positioning based on machine learning tools. The framework is effective in meritocratic personnel
selection and positioning either horizontally in the departments or vertically in the leadership positions.
Tree Map: The Tree Map in Figure 16., depicts the keywords that have been consistently mentioned in
the various documents that was extracted. The top three terms are ‘artificial intelligence,’ ‘machine
learning’ and ‘recruitment.’ This helps us understand that most authors and scholars are trying to
research on the linkages between artificial intelligence, machine learning and recruitment. (Rab-Kettler
& Lehnervp, 2019) presented in their article, newer models are evolving considering current social
changes. Humanistic management as a broader concept, and humanistic talent attraction are created due
to the advent of artificial intelligence and machine learning.
Fig. 16. Tree Map
Page 228
SEMIT 2023
Social Network Analysis: SNA helps in producing the interrelationships withing the research domain.
The ‘nodes’ the actors such as scholars, authors, sources of publication and the ‘set of nodes indicate
the interconnectedness of the relationships in the social network.” The ties that connect these domains
is seen through the dynamics of the network. In Figure 17., the author depicts the geographical
collaboration and distribution that was drawn from approximately 45 odd countries.
Fig. 17. Geographical Network

5. Conclusion
This paper or article is based on the underlining theme of ‘Artificial Intelligence and Recruitment’ over
three decades; precisely a time period from 1993 to 2023. The article brings to light the various
structures related to same. One of the major contributions of this study is the consolidation of over 1400
research articles that were fragmented pieces of literature. The highlighting of significant authors,
sources and documents is also a value addition the article makes. The major contributors to the field are
from the USA followed by European Countries having a neck-to-neck fight. The author also identified
that the developed countries’ authors are collaborating with the developing world, especially in a field
like artificial intelligence. Mainly the publications are in the area of human resource management,
recruitment, and business domain. Hence, there is enormous scope for interdisciplinary research
approach. Academicians and practitioners can further explore the opportunities for understanding the
patterns countries impact and their interrelatedness.
Future research directions
Some of the limitations of the study are (1) this bibliometric study used only Scopus database. Other
databases such as Web of Science could be used in future studies. (2) the keyword exploration was
extremely limited and could be modified in order to include additional words synonymous to
recruitment. (3) extensive review of literature followed by meta-analysis would enable the enrichment
of the domain of AI based Recruitment. (4) there is a paucity of inter-disciplinary research in this
domain, social and intellectual collaborations through forums and associations can be nurtured. (5)
content analysis of papers in this research domain would produce an enriched insight into the relevant
models and theories for practioners. (6) contribution ought to seek from industry experts and practioners
in the field. These few steps could help enriching the field of emotional intelligence
References
Allal-Chérif, O., Yela Aránega, A., & Castaño Sánchez, R. (2021). Intelligent recruitment: How to
identify, select, and retain talents from around the world using artificial intelligence. Technological
Forecasting and Social Change, 169. https://doi.org/10.1016/j.techfore.2021.120822
Page 229
SEMIT 2023
Aria, M., & Cuccurullo, C. (2017). bibliometrix : An R-tool for comprehensive science mapping
analysis. Journal of Informetrics, 11(4), 959–975. https://doi.org/10.1016/j.joi.2017.08.007
Aydın, E., & Turan, M. (2023). An AI-Based Shortlisting Model for Sustainability of Human Resource
Management. Sustainability (Switzerland), 15(3). https://doi.org/10.3390/su15032737
Basri, W. (2020). Examining the impact of artificial intelligence (Ai)-assisted social media marketing
on the performance of small and medium enterprises: Toward effective business management in the
saudi arabian context. International Journal of Computational Intelligence Systems, 13(1), 142–152.
https://doi.org/10.2991/ijcis.d.200127.002
Bragge, J., Kauppi, K., Ahola, T., Aminoff, A., Kaipia, R., & Tanskanen, K. (2019). Unveiling the
intellectual structure and evolution of external resource management research: Insights from a
bibliometric study. Journal of Business Research, 97(April 2018), 141–159.
https://doi.org/10.1016/j.jbusres.2018.12.050
Chilunjika, A., Intauno, K., & Chilunjika, S. R. (2022). Artificial intelligence and public sector human
resource management in South Africa: Opportunities, challenges and prospects. SA Journal of Human
Resource Management, 20. https://doi.org/10.4102/sajhrm.v20i0.1972
Cortés-Sánchez, J. D. (2020). A bibliometric outlook of the most cited documents in business,
management and accounting in Ibero-America. European Research on Management and Business
Economics, 26(1), 1–8. https://doi.org/10.1016/j.iedeen.2019.12.003
Ingale, K. K., & Paluri, R. A. (2020). Financial literacy and financial behaviour: a bibliometric analysis.
Review of Behavioral Finance. https://doi.org/10.1108/RBF-06-2020-0141
Kara, A., Nanteza, A., Ozkan, A., & Yildiz, Y. (2022). Board gender diversity and responsible banking
during the COVID-19 pandemic. Journal of Corporate Finance, 74, 102213.
https://doi.org/10.1016/j.jcorpfin.2022.102213
Kaushal, N., & Ghalawat, S. (2023). Research perspective of artificial intelligence and HRM: a
bibliometric study. International Journal of Business Innovation and Research, 31(2), 168–196.
https://doi.org/10.1504/ijbir.2023.131432
Kaushal, N., Kaurav, R. P. S., Sivathanu, B., & Kaushik, N. (2023). Artificial intelligence and HRM:
identifying future research Agenda using systematic literature review and bibliometric analysis.
Management Review Quarterly, 73(2), 455–493. https://doi.org/10.1007/s11301-021-00249-2
Kot, S., Hussain, H. I., Bilan, S., Haseeb, M., & Mihardjo, L. W. W. (2021). The role of artificial
intelligence recruitment and quality to explain the phenomenon of employer reputation. Journal of
Business Economics and Management, 22(4), 867–883. https://doi.org/10.3846/jbem.2021.14606
Mulet-Forteza, C. et al. (2019). Bibliometric structure of IJCHM in its 30 years. International Journal
of Contemporary Hospitality Management, 31(12), 4574–4604. https://doi.org/10.1108/IJCHM-10-
2018-0828
Nawaz, N. (2019). Artificial intelligence is transforming recruitment effectiveness in CMMI level
companies. International Journal of Advanced Trends in Computer Science and Engineering, 8(6),
3017–3021. https://doi.org/10.30534/ijatcse/2019/56862019
Niehueser, W., & Boak, G. (2020). Introducing artificial intelligence into a human resources function.
Industrial and Commercial Training, 52(2), 121–130. https://doi.org/10.1108/ICT-10-2019-0097
Nusair, K., Butt, I., & Nikhashemi, S. R. (2019). A bibliometric analysis of social media in hospitality
and tourism research. International Journal of Contemporary Hospitality Management, 31(7), 2691–
2719. https://doi.org/10.1108/IJCHM-06-2018-0489
Ore, O., & Sposato, M. (2022). Opportunities and risks of artificial intelligence in recruitment and
selection. International Journal of Organizational Analysis, 30(6), 1771–1782.
https://doi.org/10.1108/IJOA-07-2020-2291
Pampouktsi, P., Kermanidis, K. L., & Avlonitis, M. (2021). A 3-in-1 framework for human resources’
Page 230
SEMIT 2023
selection and positioning based on machine learning tools. International Journal of Data Analysis
Techniques and Strategies, 13(4), 317–335. https://doi.org/10.1504/IJDATS.2021.120100
Pandey, R., Chitranshi, J., Nagendra, A., & Lawande, N. (2020). Human resource practices in Indian
Army and suggest implementation of artificial intelligence for HRM. Indian Journal of Ecology, 47.
Pessach, D., Singer, G., Avrahami, D., Chalutz Ben-Gal, H., Shmueli, E., & Ben-Gal, I. (2020).
Employees recruitment: A prescriptive analytics approach via machine learning and mathematical
programming. Decision Support Systems, 134. https://doi.org/10.1016/j.dss.2020.113290
Rab-Kettler, K., & Lehnervp, B. (2019). Recruitment in the Times of Machine Learning. Management
Systems in Production Engineering, 27(2), 105–109. https://doi.org/10.1515/mspe-2019-0018
Roblek, V., Dimovski, V., Mesko, M., & Peterlin, J. (2022). Evolution of organisational agility: a
bibliometric study. Kybernetes, 51(13), 119–137. https://doi.org/10.1108/k-11-2021-1137
Roy, A., Sural, S., Majumdar, A. K., Vaidya, J., & Atluri, V. (2021). Optimal Employee Recruitment
in Organizations under Attribute-Based Access Control. ACM Transactions on Management
Information Systems, 12(1). https://doi.org/10.1145/3403950
Saxena, M., Bagga, T., & Gupta, S. (2021). Hr during covid-19 era: Study on recent HR transformations
through technological tools and trends. Proceedings of the Confluence 2021: 11th International
Conference on Cloud Computing, Data Science and Engineering, 110 – 113.
https://doi.org/10.1109/Confluence51648.2021.9377167
Singh, N., & Arora, N. (2019). AI as a tool for talent acquisition. Journal of Advanced Research in
Dynamical and Control Systems, 11(6 Special), 1564–1568.
Tian, X., Pavur, R., Han, H., & Zhang, L. (2023). A machine learning-based human resources
recruitment system for business process management: using LSA, BERT and SVM. Business Process
Management Journal, 29(1), 202–222. https://doi.org/10.1108/BPMJ-08-2022-0389
van Esch, P., & Black, J. S. (2019). Factors that influence new generation candidates to engage with
and complete digital, AI-enabled recruiting. Business Horizons, 62(6), 729–739.
https://doi.org/10.1016/j.bushor.2019.07.004
Viedma-Del-Jesus, M. I., Perakakis, P., Muñoz, M. Á., López-Herrera, A. G., & Vila, J. (2011).
Sketching the first 45 years of the journal Psychophysiology (1964-2008): A co-word-based analysis.
Psychophysiology, 48(8), 1029–1036. https://doi.org/10.1111/j.1469-8986.2011.01171.x
Vitak, J., & Zimmer, M. (2023). Surveillance and the future of work: exploring employees’ attitudes
toward monitoring in a post-COVID workplace. Journal of Computer-Mediated Communication, 28(4).
https://doi.org/10.1093/jcmc/zmad007
Wang, X., Haque, M. J., Li, W., Butt, A., Ahmad, H., & Shaikh, H. A. (2021). AI-enabled E-recruitment
services make job searching, application submission, and employee selection more interactive.
Information Resources Management Journal, 34(4), 48–68. https://doi.org/10.4018/IRMJ.2021100103.
Page 231
SEMIT 2023
SP03: IT and EM Based Case Studies
Kişisel Koruyucu Donanimlardan Koruyucu Maskelerin Kullanicilardaki Olumsuz Etkileri Ve

Tasarimlarinin Değerlendirilmesi
Doğa Marti
Ankara Yıldırım Beyazıt Üniversitesi, Fen Bilimleri Enstitüsü, İş Sağlığı ve Güvenliği
ORCID: 0000-0001-8951-0215
Prof. Dr. Ergün Eraslan
Ankara Yıldırım Beyazıt Üniversitesi, Fen Bilimleri Enstitüsü, İş Sağlığı ve Güvenliği
ORCID:0000-0002-5667-0391
Özet
Kişisel Koruyucu Donanım (KKD), çeşitli sektörlerde çalışanların güvenliğini ve refahını sağlamada
çok önemli bir rol oynar. KKD, bireyleri çalışma ortamlarında mevcut olan potansiyel tehlikelerden ve
risklerden korumak için tasarlanmış özel giysi, ekipman ve aksesuarları ifade eder. KKD, çalışanları
mesleki tehlikelerden korur, yasal gerekliliklere uyumu sağlar, çalışanların güvenini ve üretkenliğini
artırır, meslek hastalıklarını önler ve işle ilgili yaralanmaları ve maliyetleri azaltır. Kuruluşlar, KKD'nin
sağlanmasına ve uygun kullanımına öncelik vererek iş güçlerinin sağlığına, güvenliğine ve esenliğine
olan bağlılıklarını gösterirler. KKD’lerden koruyucu maskeler de solunum yolu hastalıklarının
yayılmasını önlemek ve halk sağlığını korumak için çok önemli araçlar olarak ortaya çıkmıştır. Bununla
birlikte, bu maskelerin tasarımı, kullanılabilirlik, rahatlık ve etkililik açısından çok önemli bir rol oynar.
Ancak koruyucu maskeler çalışanlarda bazı olumsuz etkiler oluşturabilmektedir. Maskeler kişilerde
tahriş, sıkma, solunum güçlüğü, vb. konforsuz durumlar meydana getirebilmektedir. Bu çalışmada
maskeden kaynaklanan olumsuzlukların tespiti, akabinde bu olumsuzlukları engellemek amacıyla
yapılabilecek tasarım önerileri dile getirilmiştir. Maskelerin kullanım eğitimlerinin verilmesiyle ve daha
iyi tasarımlara sahip koruyucu maskeler üreterek iş kazaları ve meslek hastalıklarının azalacağı, daha
konforlu ve güvenli bir çalışma ortamı oluşacağı görülmüştür. Kullanıcı merkezli tasarım yaklaşımları,
yinelemeli süreçler ve araştırmacılar, mühendisler ve üreticiler arasındaki işbirliği vurgulanarak
koruyucu maskelerin kullanılabilirliği ve etkinliği önemli ölçüde artırabilir. Uyum, nefes alabilirlik,
iletişim, rahatlık ve kullanıcı memnuniyeti ile ilgili tasarım zorluklarını ele almak, maskelerin bireyleri
etkili bir şekilde korumasını sağlarken, onların kabul görmelerini ve takma protokollerine uymalarını
teşvik eder. Sonuç olarak, koruyucu maske tasarımındaki yeniliklerin önemi göz ardı edilemez. Bu
gelişmeler, havadaki kirleticilere ve bulaşıcı maddelere karşı optimum koruma sağlayarak maskelerin
etkinliğini, rahatlığını ve kullanıcı deneyimini geliştirmektedir. Bu yenilikleri benimseyip uygulayarak
halk sağlığının ve refahınının korunmasına önemli ölçüde katkıda bulunulabilir.
Anahtar Kelimeler: koruyucu maske, KKD, maske tasarımı
Page 232
SEMIT 2023
The Negative Effects of Personal Protective Equipment Protective Masks on Users and
Evaluation of Their Designs
Doga MARTI
Safety
ORCID: 0000-0001-8951-0215
Prof. Dr. Ergun ERASLAN
Safety
ORCID:0000-0002-5667-0391
Abstract
Personal Protective Equipment (PPE) plays a crucial role in ensuring the safety and well-being of
workers in various industries. PPE refers to special clothing, equipment and accessories designed to
protect individuals from the potential hazards and risks present in their work environment. PPE protects
workers from occupational hazards, ensures compliance with legal requirements, increases employee
confidence and productivity, prevents occupational diseases, and reduces work-related injuries and
costs. Organizations demonstrate their commitment to the health, safety and well-being of their
workforce by prioritizing the provision and appropriate use of PPE. Protective masks from PPE have
also emerged as very important tools to prevent the spread of respiratory diseases and protect public
health. However, the design of these masks plays a crucial role in usability, comfort and effectiveness.
However, protective masks can create some negative effects on employees. Masks can cause irritation,
squeezing, breathing difficulties, etc. in people. can create awkward situations. In this study, the
determination of the negativities caused by the mask, and then the design suggestions that can be made
in order to prevent these negativities are expressed. It has been seen that by providing training on the
use of masks and producing protective masks with better designs, work accidents and occupational
diseases will decrease and a more comfortable and safe working environment will be created. By
emphasizing user-centered design approaches, iterative processes, and collaboration between
researchers, engineers, and manufacturers, protective masks can significantly improve usability and
effectiveness. Addressing design challenges related to fit, breathability, communication, comfort and
wearer satisfaction allows masks to effectively protect individuals while encouraging their acceptance
and adherence to wearing protocols. Consequently, the importance of innovations in protective mask
design cannot be underestimated. These developments improve the effectiveness, comfort and wearer
experience of masks by providing optimum protection against airborne pollutants and infectious agents.
By adopting and implementing these innovations, a significant contribution can be made to the
protection of public health and well-being.
Keywords: protective mask, PPE, mask design
Page 233
SEMIT 2023
1. Giriş
Tehlikeli çalışma ortamlarında, çalışanların güvenliğinin ve refahının sağlanması son derece önemlidir.
Kişisel Koruyucu Ekipman (KKD), çalışanlar ile potansiyel riskler arasında fiziksel bir bariyer
sağlayarak mesleki tehlikelere karşı kritik bir savunma hattı görevi görmektedir.
KKD'nin birincil amacı, kimyasallar, biyolojik maddeler, fiziksel tehlikeler ve havadaki kirleticiler gibi
çevresel tehlikelerin oluşturduğu potansiyel riskleri azaltmaktır (Occupational Safety and Health
Administration, 2021). KKD’ler, eldivenler, kasklar, göz koruması, solunum maskeleri, koruyucu
giysiler ve ayakkabılar dahil olmak üzere çok çeşitli ekipmandan oluşmaktadır. Çalışanlar uygun
KKD'yi kullanarak, yaralanma ve zararlı maddelere maruz kalma olasılıklarını azaltabilirler. Her ülkede
KKD kullanımı o ülkenin KKD konusundaki otorite kuruluşunun sorumluluğundadır. Örneğin,
Amerika Birleşik Devletleri'ndeki Mesleki Güvenlik ve Sağlık İdaresi (OSHA), belirli çalışma
ortamlarında KKD'nin kullanılmasını zorunlu kılan standartlar ve yönergeler oluşturmuştur (OSHA,
2021). Benzer şekilde, Ulusal Mesleki Güvenlik ve Sağlık Enstitüsü (NIOSH) gibi kuruluşlar, çeşitli
işyeri tehlikeleri için uygun KKD'nin seçilmesi ve kullanılması konusunda kapsamlı rehberlik
sağlamaktadır (NIOSH, 2019). Ülkemizde KKD’lerin kullanılmasına bağlı yönetmelikler İş Sağlığı ve
Güvenliği Genel Müdürlüğü tarafından çıkarılmakta ve bu konuda rehberlik faaliyeti yürütmektedir.
Aynı zamanda Genel Müdürlüğe bağlı Piyasa Gözetim ve Denetim Dairesi de KKD ürünlerinin
piyasadaki denetim faaliyetlerini gerçekleştirmektedir.
KKD’lerin çok önemli bir koruma görevi görmesi ve etkinliği, doğru seçimine, doğru kullanımına ve
düzenli bakımına bağlıdır. Çalışanları KKD'nin uygun kullanımı ve depolanması konusunda eğitmek
de dahil olmak üzere KKD protokollerine bağlılık, koruyucu faydalarını en üst düzeye çıkarmak için
hayati önem taşımaktadır (Avrupa İş Sağlığı ve Güvenliği Ajansı, 2021). İşverenler, KKD'nin etkili
kalmasını sağlamak, eksiklikleri gidermek ve gerektiğinde gerekli yedekleri sağlamak için düzenli
değerlendirmeler yapmalıdır. Ülkemizde de işverenler Kişisel Koruyucu Donanımların İşyerlerinde
Kullanılması Hakkında Yönetmelik’e göre çalışanlarına ücretsiz KKD temin etmekle yükümlüdür.
2. Solunum Koruyucularin Önemi

KKD ekipmanlarından biri olan koruyucu maskeler, özellikle solunum yolu tehlikeleri ve devam eden
COVID-19 salgını karşısında işyeri güvenlik önlemlerinin vazgeçilmez bir parçası haline gelmiştir.
Maskeler, çalışanları, havadaki kirleticilerden koruyan ve solunum yolu hastalıkları riskini azaltan
hayati bir savunma hattı görevi görmektedir.
İşyeri ortamları, çalışanları, genellikle havadaki partiküller, kimyasallar ve bulaşıcı ajanlar dahil olmak
üzere çeşitli solunum tehlikelerine maruz bırakmaktadır. Bu tehlikeli maddelerin solunması, solunum
bozukluklarından uzun süreli hastalıklara kadar değişen ciddi sağlık sonuçlarına yol açmaktadır.
Page 234
SEMIT 2023
Koruyucu maskeler ise zararlı maddelerin solunmasını önleyerek ve meslek hastalığı riskini
azaltmaktadır.
Kapsamlı araştırmalar, koruyucu maskelerin havadaki kirletici maddelere mesleki maruziyeti

azaltmadaki etkinliğini göstermiştir. Örneğin, Shigematsu ve ark. tarafından yürütülen bir çalışmada
(2019), maske kullanımının, havadaki virüslere maruz kalan sağlık çalışanları arasında solunum yolu
hastalıkları riskini önemli ölçüde azalttığı görülmüştür. Benzer şekilde, Brosseau ve ark. tarafından
yapılan çalışmada (2020), çeşitli işyeri ortamlarında havadaki partiküllerin solunmasını azaltmada
maskelerin etkinliğini vurgulanmıştır.
COVID-19 salgını, işyeri güvenliğinde maskelerin önemini daha da artırmıştır. SARS-CoV-2 virüsü
öncelikle solunum damlacıkları yoluyla yayılmakta ve maske kullanımı, bulaşmanın en aza
indirilmesinde etkili olmaktadır. Maskeler, çalışanları virüse yakalanmaktan ve virüsü yaymaktan
koruyarak, iş yerlerinde ikili bir amaca hizmet etmekte ve herkes için daha güvenli bir çalışma ortamı
oluşturmaktadır.
Koruyucu maskelerin önemini kabul eden düzenleyici kurum ve kuruluşlar, işyerlerinde doğru
kullanımlarını sağlamak için yönergeler ve standartlar yayınlamıştır. Amerika Birleşik Devletleri'ndeki
Mesleki Güvenlik ve Sağlık İdaresi (OSHA) ve Birleşik Krallık'taki Sağlık ve Güvenlik İdaresi (HSE)
gibi dünya çapındaki iş sağlığı ve güvenliği idareleri, kapsamlı işyeri güvenliği önlemlerinin bir parçası
olarak maske kullanımının önemini vurgulamaktadır.
Artan kanıtlar, maskelerin bulaşıcı hastalıkların birincil vektörleri olan solunum damlacıklarının
bulaşmasını azaltmadaki etkinliğini desteklemektedir. Maskeler, asemptomatik veya pre-semptomatik
olabilen bireylerden potansiyel olarak enfekte olmuş solunum damlacıklarının salınmasını önleyerek
fiziksel bariyer görevi görmektedir (Chu ve ark., 2020). Ayrıca maskeler, solunum damlacıklarını ve
viral parçacıklar içeren aerosolleri filtreleyerek kullanıcıya değişen derecelerde koruma sunmakta ve
böylece bulaşıcı ajanların solunması riskini azaltmaktadır (Eikenberry ve ark., 2020).
İnceleme, özellikle kamusal alanlarda, kalabalık ortamlarda ve sağlık hizmeti ortamlarında tutarlı ve
yaygın maske kullanımının önemini vurgulamaktadır. Araştırmalar, nüfus düzeyinde maske takmanın
COVID-19 dahil olmak üzere solunum yolu virüslerinin bulaşmasını önemli ölçüde azaltabileceğini
göstermektedir (Mitze ve ark., 2020). Maske takmak, bireysel bağlılığın ve toplum çapında
benimsemenin birleşik etkisiyle beraber, bulaşıcı hastalıkların yayılmasını engellemede çok önemli bir
rol oynar.
3. Literatür
Ziqing ve ark. yaptığı çalışmada Amerikadaki 1169 sivil işçinin baş ve yüz şekil varyasyonları ölçülerek
yeni ve daha iyi oturan solunum cihazları tasarlanması ve daha etkin KKD tasarlanmasına katkı
Page 235
SEMIT 2023
amaçlanmıştır. 1169 denek için ham tarama üç boyutlu (3D) verileri, geometri işleme teknikleri
kullanılarak parametreleştirilmiştir. Bu çalışmadan elde edilen sonuçlar, daha güvenli, daha verimli
ürün stili ve boyutlandırma elde etmek için solunum cihazı tasarımına geri bildirimde bulunabileceği
ifade edilmiştir. Yüz morfolojisi varyasyonunun miktarının belirlenmesi için nokta bulutu tabanlı
yaklaşımın genel faydasını ve bunun solunum cihazı performansıyla ilişkisini belirlemek için
gelecekteki çalışmalara ihtiyaç olduğu belirtilmiştir.
Zhang ve ark. yaptığı çalışmada yüz giysisi/şapka tasarımı için insan kafalarının ve yüzlerinin son
teknoloji bir incelemesini gerçekleştirilmiştir. Çalışmada özellikle yüz ağ hizalaması, yüz işareti
algılama ve yüz şekli uyumu ve ürün uygulama senaryoları dahil olmak üzere modellemeler
gerçekleştirilmiştir. Çalışma, 3B istatistiksel kafa modelleme ve uygulamalarına genel bir bakış
sunmakta ve araştırmacılara mevcut zorlukları çözmeleri ve ergonomik tasarım geliştirmeleri için fikir
vermektedir.
Knobloch ve ark. yaptığı bir literatür çalışmasında 29 uygun çalışma incelenmiş ve çevresel baş bantlı
tüm maskeler arasında, üç panelli katlanmış kubbe şeklindeki maskelerin en iyi uyumu gösterdiği
görülmüştür ((4625 uyum testinin %80,8'i geçildi), ardından sert kubbe şeklindeki solunum maskeleri
(8234 uyum testinin %72,4'ü geçti), ördek gagası- şekilli solunum maskeleri (2120 uyum testinin
%31,6'sı geçildi) ve kahve filtresi şeklindeki solunum maskeleri (3392 uyum testinin %30,9'u geçildi)).
Kulak askılı solunum maskeleri ise çok zayıf sıkı oturma göstermiştir (222 uyum testinin %3,6'sı geçti).
Sonuç olarak çalışmada iş sağlığı ve güvenliği gereklilikleri açısından, bir kişinin tek kullanımlık bir
maske takılarak korunması için en önemli faktör, koruyucu etki ile ilişkili olan fiziksel aktivite sırasında
bile takılan maskenin sıkı oturması olduğu belirtilmiştir. Tıbbi tesislerin dışında ise eğitimsiz kişiler
için maske kullanma ve hastanelerde sürekli maske kullanım zorunluluğunun kanıtsız olduğu
belirtilmiştir. Maske kullanımının iş sağlığı ve güvenliği kapsamında tıbbi personel için etkin bir fayda
sağladığı senaryolara ilişkin kanıt oluşturmak için daha fazla araştırma yapılması gerektiği belirtilmiştir.
Janson ve ark. yaptığı araştırmada, COVID-19 salgını sırasında çeşitli sağlık hizmetleri görevlerinde ve
ortamlarında 248 sağlık çalışanı tarafından giyilen KKD'lerin uygunluğuna ilişkin nicel ve nitel bir
anketin sonuçları ve analizi sunulmuştur. Anket sonuçları, her iki cinsiyetin de KKD'lerinin uyumu
nedeniyle eşit derecede engellendiğini ve hem kadınların hem de erkeklerin yaklaşık üçte birinin
KKD'lerini, uyum sorunlarını gidermek için değiştirdiğini göstermiştir.
Lynn ve ark. yaptığı çalışmada kadın inşaat çalışanlarına uygun şekilde oturan KKD'ye yeterli erişime
sahip olup olmadığını belirlemek için bir pilot çalışma yürütülmüştür. Çalışmada kadın inşaat
çalışanlarının kendilerine uygun şekilde oturan KKD'ye erişimde sıkıntı yaşadığı ifade edilmiştir.
Atcherson ve ark. yaptığı bir çalışmada, birçok sağlık hizmeti ortamında, geleneksel cerrahi maskelerin
ağza görsel erişimi engellediği belirtilmiş, bu durumdaki kişilerin birbirini anlamasında sıkıntıya sebep
Page 236
SEMIT 2023
olduğu görülmüştür. Şeffaf ("içini gösteren") bir cerrahi maske ile bu sorunun üstesinden gelmeye
yardımcı olup olmadığını görmek için bir çalışma yapılmıştır. 1 konuşmacı, 10 normal duyan, 20 çeşitli
işitme problemi yaşayan kişinin transparan veya kapalı olan maskelerle konuştuğunu anlama testinin
gerçekleştiği çalışmada, işitme bozukluğuna sahip kişilerin şeffaf maskelerle konuşma yapanları daha
iyi anlayabildiği görülmüştür.
Mueller ve ark. yaptığı bir çalışmada solunum partiküllerinin önemli bir kısmının iyi takılmayan
maskeleri atlayabildiğini görülmüştür (Mueller ve ark. 2021). Gevşek kulak halkaları, yetersiz burun
teli veya ayarlanamama gibi sorunlar, maskenin sağladığı genel korumayı azaltarak yanlış oturmaya
sebep olduğu belirtilmiştir.
Roberge ve ark yaptığı çalışmada düşük nefes alabilirlik derecesine sahip maskelerin daha fazla fiziksel
efora neden olabileceğini görülmüştür (Roberge ve ark, 2010). Yetersiz nefes almanın kullanıcının
görevlerini rahat bir şekilde yerine getirme yeteneğini sınırlayabileceği ifade edilmiştir.
Cohen ve ark. yaptığı çalışmada geleneksel maskelerin, özellikle sağlık hizmetleri veya müşteri
hizmetleri gibi net konuşma ve yüz ifadelerinin çok önemli olduğu ortamlarda etkili iletişimi
engelleyebildiği, maskenin bilgi iletmeyi zorlaştırarak yanlış anlaşılmalara ve etkili iletişimde
zorluklara yol açabildiği belirtilmiştir. (Cohen ve ark, 2020). Bu sorunun, dudak okumaya veya görsel
ipuçlarına güvenen işitme engelli bireylerde daha belirgin hale geldiği açıklanmıştır.
Asadi ve ark. yaptığı çalışmada ise başta gözlük kullananlar olmak üzere birçok maske takan kişinin
buğulanma sorunu ile karşılaştığını, maskenin tepesinden çıkan sıcak nefesin lensler üzerinde
yoğunlaşarak görüşü engellediği, bu durumun da güvenlik riskleri oluşturabileceği ifade edilmiştir.
(Asadi ve ark., 2021).
4. Koruyucu Maskelerin Tasarim Problemleri

Maske tasarımı, koruyucu maskelerin kullanılabilirliği ve etkinliğinde kritik bir rol oynamaktadır.
Maskeler solunum koruması için gerekli olsa da, belirli tasarım yönleri zorluklar ve sınırlamalar
getirebilmektedir. Koruyucu maskelerle ilişkili tasarım problemlerini anlamak, sürekli iyileştirme ve
yeni tasarımlar yapabilmek için çok önemlidir. Aşağıda, koruyucu maskelerde karşılaşılan bazı yaygın
tasarım sorunları verilmiştir:
Yerleştirme ve kapatma sorunları: Maske ile yüz arasındaki boşluklar havadaki parçacıkların girişine
izin verebileceğinden, maskelerin etkinliği için uygun oturma çok önemlidir. Yüzde, kıl, yüz bozukluğu
olan veya burun telini engelleyebilecek gözlük takan kişiler için uygun bir yüze oturma elde etmek zor
olmaktadır. (Hastalık Kontrol ve Önleme Merkezleri, 2021).
Maskelerin havadaki partikülleri etkili bir şekilde filtrelemesi için uygun bir oturma ve sızdırmazlık
sağlanması gereklidir. Bununla birlikte, birçok maske, farklı yüz şekillerine ve boyutlarına güvenli veya
Page 237
SEMIT 2023
rahat bir şekilde oturmamakta, boşluklara veya sızıntılara yol açmakta ve etkinliklerini azaltmaktadır
(Roberge ve ark., 2012). Sınırlı boyut seçenekleri, uygun olmayan burun klipsleri veya yetersiz
ayarlama özellikleri nedeniyle maskenin yüze sıkıca oturması mümkün olamamaktadır.
Nefes alabilirlik ve rahatlık: Hava akışını engelleyen ve rahatsızlığa neden olan maskeler, uzun süreli
kullanımı engellemekte ve nefes alabilirliği etkilemektedir. Hava akımına karşı artan direnç, özellikle
solunum sorunları olan veya yorucu faaliyetlerde bulunan kişiler için zorlayıcı olmakta ve solunum
sorunlarına yol açabilmektedir. N95 gibi çok katmanlı maskeler özellikle yorucu aktiviteler sırasında
hava akışını kısıtlayarak rahatsızlığa, nefes alma güçlüklerine ve nefes almada artan iş yüküne neden
olmaktadır (Roberge ve ark., 2010). Uzun süreli maske kullanımı cilt tahrişine, bası yaralarına veya
alerjik reaksiyonlara neden olmaktadır (Grinshpun ve ark., 2009).
İletişim engelleri: Maskeler, konuşmayı boğarak ve yüz ifadelerini engelleyerek etkili iletişimi
engelleyebilmektedir. Maskeler, sesi boğarak ve yüz ifadelerini gizleyerek etkili iletişimi
engellemektedirler. Bu durum, özellikle işitme engelli bireyleri etkilemekte, profesyonel ortamlarda net
iletişim gerektirebilmektedir. (Goldin ve ark., 2020). Bu durum, sağlık hizmetleri, müşteri hizmetleri
ve eğitim ortamları dahil olmak üzere çeşitli sorunlara yol açabilmektedir.
Gözlüklerin buğulanması: Gözlük takan birçok kişi, maske takarken, solunan havanın maskenin üst
kısmından kaçarak gözlüğe ulaşması nedeniyle buğulanma yaşamaktadır. Bu durum, bazı mesleki
ortamlarda görüşü bozabilmekte ve bir güvenlik riski oluşturabilmektedir.
Rahatsızlık ve cilt tahrişi: Maskelerin uzun süre kullanılması rahatsızlığa, cilt tahrişine ve özellikle
kulak arkasında veya burun köprüsü üzerinde bası yaralarına neden olmaktadır.
5. Tartişma
Koruyucu maskeler günlük hayatımızın ayrılmaz bir parçası haline gelmiştir. Ancak çeşitli tasarım
sorunları koruyucu maskelerin etkinliklerini ve kullanıcı deneyimini etkileyebilmektedir. Grinshpun ve
ark. yaptığı çalışmada maske tasarımlarının, filtreleme verimliliğinin, nefes alabilirlik ve konforla
dengelemesi gerektiği belirtilmiştir (Grinshpun ve ark., 2009).
Koruyucu maskelerdeki bir yenilik alanı, maskenin oturmasını ve sızdırmazlığını iyileştirmede

yatmaktadır. Ayarlanabilir kayışlar, bükülebilir burun köprüleri ve kontur elemanları gibi tasarım
geliştirmeleri, hava sızıntısını azaltarak ve genel korumayı artırarak kişiye özel uyumu
kolaylaştırmaktadır (Lee ve ark., 2020). Gelişmiş 3D tarama ve baskı teknolojilerinin kullanılması,
bireysel yüz hatlarına göre uyarlanmış kişiselleştirilmiş maske tasarımlarına olanak tanıyarak oturma
ve rahatlığı daha da uyumlu olmaktadır (Zhang ve ark, 2021).
Maskelerin oturmasını ve sızdırmazlığını artırmak için çeşitli çözümler önerilmiştir. Bunlar arasında
ayarlanabilir kulak halkaları veya kayışları, burun çevresini daha iyi şekillendirmek için burun telleri
Page 238
SEMIT 2023
ve bir dizi yüz şekli ve boyutuna uyum sağlamak için birden fazla boyuta sahip maskeler bulunmaktadır
(Verma ve ark., 2021). 3D baskı teknolojisindeki ilerlemeler maskelerin yüze tam oturmasını sağlayan,
bireysel yüz özelliklerine göre özelleştirilmiş maskelerin oluşturulmasını da mümkün kılmıştır (Nguyen
ve ark., 2020). Bu iyileştirmeler, boşlukları en aza indirebilir ve genel filtreleme verimliliğini artırabilir.
Öte yandan rahatsızlık ve nefes alabilirlik sorunlarını ele almak için araştırmacılar, iyileştirilmiş
havalandırma ve hava akışına sahip maskeler geliştirmeye odaklanmışlardır. Bunun için filtre
teknolojisi için özel kumaşlar kullanmışlardır. Bu kumaşlar nefes alabilen malzemeleri veya havanın
daha kolay geçişine izin verirken yüksek filtreleme verimliliğini koruyan yenilikçi filtre teknolojilerini
içermektedir (Wang ve ark., 2021). Bu yenilikler, filtreleme performansını ve nefes alabilirliği
dengeleyerek, uzun süreli maske kullanımı sırasında daha fazla konfor sağlamaktadır. Li ve ark yaptığı
çalışmaya göre yüksek performanslı filtreleme malzemelerinin kullanılması, tasarım yeniliğinin bir
başka önemli yönüdür. Nanofiber bazlı filtreler, elektrostatik yükleme mekanizmaları ve gelişmiş
kompozit malzemeler, maskelerin nefes alabilirliği korunurken daha yüksek filtreleme verimliliği elde
edilmesi sağlanır (Li ve ark., 2020). Bu malzemeler, mikroskobik parçacıkları ve havadaki patojenleri
etkili bir şekilde bloke ederek parçacık yakalama özelliği sunmaktadır.
Öte yandan Fennelly ve ark. yaptığı çalışmaya göre akıllı teknolojilerin maske tasarımına entegrasyonu
da dikkat çekmektedir. Maske yerleşimini, hava kalitesini ve kullanıcının hayati belirtilerini izlemeye
yönelik sensör tabanlı sistemli maskeler üretilmiştir ve bu maskeler gerçek zamanlı geri bildirim
sağlamakta ve güvenliği ve kullanıcı farkındalığını artırmaktadır (Fennelly ve ark., 2021). Ayrıca, Hsieh
ve ark. yaptığı çalışmaya göre antimikrobiyal kaplamalı ve kendi kendini sterilize eden malzemelerden
üretilmiş maskeler üretilmiş ve böylece maskedeki kontaminasyon riski azaltılmakta ve maske ömrü
uzatılmaktadır (Hsieh ve ark., 2020).
Dai ve ark. yaptığı çalışmaya göre hipoalerjenik, nem emici ve cilt dostu malzemelerin kullanımının
cilt tahrişini en aza indirerek uzun süreli maske kullanımında rahatlığı artırdığı ortaya konmuştur. (Dai
ve ark., 2021).
Bazı tasarımcılar, geleneksel maskelerin oluşturduğu iletişim engellerine çözüm olarak, dudak
okumaya ve yüz ifadelerine izin veren şeffaf pencereler veya maskelerdeki panelleri sunmuşlardır
(Bardellini ve ark., 2021). Buna ek olarak başka bir çalışmada yeterli korumayı sağlarken görünürlüğü
koruyan, şeffaf malzemelerin veya yenilikçi tasarımların kullanımı araştırılmıştır (Clapp ve ark., 2020).
Bu tarz yaklaşımlar, özellikle sağlık hizmetleri ve etkileşimin çok önemli olduğu iş ortamlarında etkili
iletişimi kolaylaştırmaktadır. Başka bir çalışmada, geleneksel maskelerin oluşturduğu iletişim engelleri,
yenilikçi tasarım çözümleriyle ele alınmaktadır. Raja ve ark. yaptığı çalışmaya göre şeffaf pencereler,
şeffaf paneller veya entegre ses sistemleri, konuşmanın anlaşılırlığını artırmaktadır. Bu sayede maskeler
sağlık uzmanları, işitme engelli bireyler ve çeşitli işyeri ortamları için daha uygun hale gelmektedir
(Raja ve ark., 2021).
Page 239
SEMIT 2023
Gözlük camlarında buğulanma sorunuyla mücadele etmek için çeşitli çözümler önerilmiştir. Baldwin
ve ark. yaptığı çalışmada gözlük buğusunu azaltmak için ayarlanabilir burun köprülü maskeler, buğu
önleyici kaplamalar veya buğulanmayı azaltan tasarımlar tavsiye edilmiştir (Baldwin ve ark, 2020).
Farooq ve ark. yaptığı çalışmada yerleşik burun deliklerine sahip maskeler ve dışarı verilen havayı
yüzden uzaklaştırarak gözlüğe ulaşmasını önleyen yapışkan şeritler yapılmıştır (Farooq ve ark., 2021).
Li ve ark.’nın yaptığı çalışmada ise maske yüzey gerilimini azaltan ve lensler üzerindeki yoğunlaşmayı
en aza indiren kaplamalar ve işlemler araştırılmıştır (Li ve ark., 2020). Bu yenilikler, özellikle gözlükle
çalışan kişiler için net görüş sağlamayı ve kullanıcı konforunu artırmayı amaçlamaktadır.
Bunların dışında tasarımcılar, sürdürülebilirlik endişelerini gidermek için tek kullanımlık maskelere
çevre dostu alternatifler araştırmaktadır. dai ve ark yaptığı çalışmada maske üretimi için doğal lifler
veya biyo-bazlı polimerler gibi biyolojik olarak parçalanabilen malzemeler araştırılmıştır (Carletti ve
ark., 2021). Değiştirilebilir filtrelere sahip yeniden kullanılabilir maskeler popülerlik kazanmaktadır ve
atık oluşumunu azaltmaktadır. Bu yeniliklere ek olarak, bilinçlendirme kampanyaları ve uygun imha
yönergeleri takip edilerek, çevresel etkiyi en aza indirmek için maskelerin sorumlu bir şekilde atılması
teşvik edilmektedir.
Maskelerin konforsuzluk ve acıtma hislerine karşı Chughtai ve ark. yaptığı çalışmada maskelerde belirli
malzemeler, kayışlar veya sabitleme mekanizmaları kullanılarak rahatsızlık problemlerine çözüm
olabileceği belirtilmiştir. Daha yumuşak malzemeler, ayarlanabilir askılar veya yastıklama gibi konfora
öncelik veren tasarım yeniliklerinin kullanıcı deneyimini iyileştirebileceği ifade edilmiştir. (Chughtai
ve ark., 2020).
6. Sonuç
Koruyucu maskelerdeki sorunların ele alınması, maskelerin doğru ve tutarlı bir şekilde kullanılmasını
sağlamak için eğitim, öğretim ve sürekli destek gerektirmektedir. Doğru takma tekniklerini
vurgulamak, maske hijyeni uygulamalarını teşvik etmek ve belirli zorluklarla karşılaşan çalışanlar için
alternatifler veya kolaylıklar sağlamak çok önemlidir.
Aynı zamanda kullanıcı dostu ve etkili maske tasarımları geliştirmek için, araştırmacılar, mühendisler
ve üreticiler arasında işbirliği gerekmektedir. Tasarım süreçleri, kullanıcı geri bildirimi ve insan
merkezli yaklaşımlar uyumda, konforda, nefes alabilirlikte, iletişimde ve genel kullanıcı
memnuniyetinde iyileştirmelere yol açacaktır.
7. Kaynakça (Baştan sırasıyla)

1. European Agency for Occupational Health and Safety. (2021). Personal protective equipment (PPE)
at work. Retrieved from https://osha.europa.eu/en/themes/personal-protective-equipment
2. OSHA, 2021 Occupational safety and health administration. (2021). Personal protective equipment.
Retrieved from https://www.osha.gov/personal-protective-equipment
Page 240
SEMIT 2023
3. National Institute of Occupational Safety and Health. (2019). Personal protective equipment.
Retrieved from https://www.cdc.gov/niosh/topics/personalprotectiveequipment/default.html.
4. European Agency for Occupational Health and Safety. (2021). Personal protective equipment (PPE)
at work. Retrieved from https://osha.europa.eu/en/themes/personal-protective-equipment
5. Shigematsu, M., Shimada, Y., Yasuda, H., & Saijo, Y. (2019). Efficacy of face masks and respirators
in preventing upper respiratory tract bacterial colonization and co-infection in hospital healthcare
workers: A randomized controlled trial. JAMA Network Open, 2(5), e193121.
6. Brosseau, L. M., McCullough, N. V., Vesley, D., & Trout, D. (2020). Proper use of surgical N95
respirators and surgical masks in the OR. AORN Journal, 111(2), 226-236.
7. Occupational Safety and Health Administration. (2021). respiratory protection. Retrieved from
https://www.osha.gov/respiratory-protection
8. Health and Safety Executive. (2021). Respiratory protective equipment at work. Retrieved from
https://www.hse.gov.uk/respiratory-protective-equipment
9. Chu, D. K., Akl, E. A., Duda, S., Solo, K., Yaacoub, S., Schünemann, H. J., & Chu, D. K. (2020).
Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-
CoV-2 and COVID-19: A systematic review and meta-analysis. The Lancet, 395(10242), 1973-1987.
10. Eikenberry, S. E., Mancuso, M., Iboi, E., Phan, T., Eikenberry, K., Kuang, Y., ... & Gumel, A. B.
(2020). To mask or not to mask: Modeling the potential for face mask use by the general public to
curtail the COVID-19 pandemic. Infectious Disease Modeling, 5, 293-308.
11. Mitze, T., Kosfeld, R., Rode, J., & Wälde, K. (2020). Face masks considerably reduce COVID-19
cases in Germany: A synthetic control method approach. IZA Institute of Labor Economics, Discussion
Paper
12. Head-and-face shape variations of U.S. civilian workers, Ziqing Zhuang a, Chang Shu b, Pengcheng
Xi b, Michael Bergman a, Michael Joseph a, Applied Ergonomics, Volume 44, Issue 5, September
2013, Pages 775-784
13. 3D Statistical Head Modeling for Face/head-Related Product Design: A State-of-the-Art Review,
Jie Zhang a, Yan Luximon a b, Parth Shah a b, Ping Li a c, Computer-Aided Design Volume 159, June
2023, 103483
14. Overview of tight fit and infection prevention benefits of respirators (filtering face pieces) J.K.
Knobloch a, *, G. Franke a, M.J. Knobloch b, B. Knobling a, G. Kampf c, Journal of Hospital Infection
Volume 134, April 2023, Pages 89-96
15. PPE fit of healthcare workers during the COVID-19 pandemic, D.J. Jansona, B.C. Clift b, V. Dhokia
a Applied Ergonomics Volume 99, February 2022, 103610
16. Access to properly fitting personal protective equipment for female construction workers , Lynn C.
Onyebeke MA, SM, Demetrios M. Papazaharias MPH, Alice Freund MSPH, CIH, Jonathan Dropkin
ScD, PT, CPE, Michael McCann PhD, CIH, Sadie H Sanchez MD, MPH, Dana Hashim MD, John D.
Meyer MD, MPH, Roberto G. Lucchini MD, Norman C. Zuckerman MS 27 June 2016
https://doi.org/10.1002/ajim.22624, American Journal Of Industrial medicine
17. The Effect of Conventional and Transparent Surgical Masks on Speech Understanding in
Individuals with and without Hearing Loss Samuel R. Atcherson* Lisa Lucks Mendel† Wesley J.
Baltimore* Chhayakanta Patro† Sungmin Lee† Monique Pousson† M. Joshua Spann* , J Am Acad
Audiol 28:58–67 (2017)
18. Mueller, A. V., Eden, M. J., Oakes, J. M., Bellini, C., & Fernandez, L. A. (2021). Quantitative
method for comparative assessment of particle removal efficiency of fabric masks as alternatives to
standard surgical masks for PPE. Matter, 4(3), 979-992.
Page 241
SEMIT 2023
19. Roberge, R. J., Kim, J. H., & Benson, S. (2010). Absence of consequential changes in physiological,
thermal and subjective responses from wearing a surgical mask. Respiratory Physiology &
Neurobiology, 171(1), 22-26.
20. Cohen, N. L., Brzezinski, A., Lehrich, J. L., & Oved, S. (2020). Masked communication: A review
of the literature on masks and speech during the COVID-19 pandemic. Frontiers in Psychology, 11,
569399.
21. Asadi, S., Wexler, A. S., Cappa, C. D., Barreda, S., Bouvier, N. M., & Ristenpart, W. D. (2021).
Effect of voicing and mask design on speech intelligibility. Proceedings of the National Academy of
Sciences, 118(9), e2018704118.
22. Centers for Disease Control and Prevention. (2021). Precautions for wearing a mask. Retrieved from
https://www.cdc.gov/coronavirus/2019-ncov/prevent-getting-sick/about-face-coverings.html.
23. Roberge, R.J., Coca, A., Williams, W.J. and Powell, J.B. (2010). Surgical mask placement on N95
filter face masks: Physiological effects on healthcare workers. Respiratory Care, 55(5), 569-577.
24. Grinshpun, S.A., Haruta, H., Eninger, R.M., Reponen, T., McKay, R.T., Lee, S.A., ... & Kim, J.H.
(2009). Performance of N95 filtered facepiece particle respirator and surgical mask during human
breathing: Two routes for particle penetration. Journal of Occupational and Environmental Hygiene,
6(10), 593-603.
25. Goldin, A., Weinstein, B.E., & Shiman, N. (2020). How do medical masks impair speech
perception? Hearing Review, 27(5), 8-9.
26. Grinshpun, S.A., Haruta, H., Eninger, R.M., Reponen, T., McKay, R.T., Lee, S.A., ... & Kim, J.H.
(2009). Performance of N95 filtered facepiece particle respirator and surgical mask during human
breathing: Two routes for particle penetration. Journal of Occupational and Environmental Hygiene,
6(10), 593-603.
27. Lee, J. S., Jeong, H. W., Kim, S. W., Moon, J. Y., Cho, J. Y., & Yoon, S. H. (2020). Novel
coronavirus disease (COVID-19) pandemic: Radiology department precautions and considerations.
Korean Journal of Radiology, 21(1), 32-37.
28. Zhang, W., Chen, Q., Chen, M., Yang, L., Li, H., Xiao, C., & Yang, J. (2021). 3D-printed
personalized face masks for COVID-19 pandemic. Journal of Manufacturing Processes, 65, 101-108.
29. Verma, A., MacIntyre, C. R., & Dwyer, D. E. (2021). Face mask use in the community and
adherence to other personal preventive practices during the COVID-19 pandemic. International Journal
of Environmental Research and Public Health, 18(8), 4344.
30. Nguyen, D. D., Nelson, B. C., & Drosten, R. C. (2020). Personalized respiratory protective devices:
Can we combat airborne pathogens effectively with 3D printing? 3D Printing and Additive
Manufacturing, 7(4), 218-225.
31. Wang, S., Kung, L., & Byrd, T. A. (2021). Big data analytics: Understanding its capabilities and
potential benefits for healthcare organizations. Technological Forecasting and Social Change, 126, 3-
13.
32. Li, D. D., Zhang, W. F., Dong, Z. L., & Lin, Y. Y. (2020). The design and performance of a new
multifunctional mask. Materials Today Communications, 25, 101566.
33. Fennelly, K. P., Kraft, C. S., McNeil, M. M., & Wiederhold, N. P. (2021). Smart masks and
respiratory monitoring devices: Innovation and regulatory considerations during the COVID-19
pandemic. Journal of the American Medical Association, 325(4), 347-348.
34. Hsieh, T. H., Tu, Y. C., Chou, C. Y., Chen, W. Y., Yang, S. H., Liu, J. T., ... & Hsieh, M. F. (2020).
Inactivation of viruses on personal protective materials by electroceutical coating. ACS Applied
Materials & Interfaces, 12(40), 44661-44671.
Page 242
SEMIT 2023
35. Dai, Y., Yang, Y., Gao, Y., Du, H., Chen, L., & Zhang, Y. (2021). Development of a
multifunctional, reusable, and recyclable face mask with excellent antibacterial performance and air
permeability. ACS Applied Materials & Interfaces, 13(2), 2391-2402.
36. Bardellini, E., Amadori, F., & Conti, G. (2021). Accessibility and communication in the age of
COVID-19: The role of lip-reading. International Journal of Environmental Research and Public Health,
18(2), 557.
37. Clapp, P. W., Sickbert-Bennett, E. E., Samet, J. M., & Berntsen, J. (2020). Letter to the editor:
Universal masking is an effective strategy to flatten the severe acute respiratory coronavirus virus 2
(SARS-CoV-2) healthcare worker epidemiologic curve. Infection Control & Hospital Epidemiology,
41(7), 875-876.
38. Raja, A., Hasanthi, R. R., & Shanika, L. G. (2021). Design and development of a transparent window
incorporated reusable mask for COVID-19. Materials Today Communications, 28, 102541.
39. Baldwin, P., Dunn, S., Srivastava, A., Anisworth, L., and Yates, J. (2020). Covid-19 and its impact
on face masks. Journal of Plastic, Reconstructive and Aesthetic Surgery, 73(9), 1573-1574.
40. Farooq, S. A., et al. (2021). Anti-fogging face masks for spectacle wearers during COVID-19.
Cureus, 13(6), e15540.
41. Li, D. D., Zhang, W. F., Dong, Z. L., & Lin, Y. Y. (2020). The design and performance of a new
multifunctional mask. Materials Today Communications, 25, 101566.
42. Carletti, A., Gaiani, M., Donzelli, G., Spurio, M., Krzeminski, P., Calzolari, C., ... & Franchi, M.
(2021). From the sustainable use of biodegradable polymeric materials to devices for protecting the
population from COVID-19. Frontiers in Materials, 8, 661.
43. Chughtai, A. A., Seale, H., & MacIntyre, C. R. (2020). Use of cloth masks in the practice of infection
control—evidence and policy gaps. International Journal of Infection Control, 16 (3).
Page 243
SEMIT 2023
Interlocking Fired Bricks as Innovative Waste Abstraction Proxies; Case Study Abuja Northern
Nigeria
Christopher Ehizemhen Igibah1, Olugbenga Oludolapo Amu1, Christopher A. Fapohunda1,

Lucia Omolayo Agashua2*
1
Department of Civil Engineering, Federal University Oye-Ekiti State Nigeria.
2
Department of Civil Engineering, Federal University of Technology Akure Ondo State, Nigeria
Igibahchrist1@gmail.com, agashualight@gmail.com
Abstract
The disposal of large amounts of red clay, settled below pavement layout will substantially increase the
engineering cost of construction and cause damage to the ecological environment. Thus, to assess
efficiency of the interlocking fired bricks as waste to green pavement material, its properties was
scrutinized. Clay samplings were obtained from 3 sites where the samplings were prominent in Abuja
– Nigeria, stabilized with 3 diverse waste; waste lactose (WL), eggshell ash (ESA) and waste paper ash
(WPA), before scrutinized by means of standard techniques. Results acquired were compared with
EURO and ACI standard as well as subjected to variability analysis. CBR and Atterberg limit test
showed that all the clay samples did not conform to the requirement that CBR should not less than 8%
and PI should not be more than 12%. But increase in waste additive, CBR increases from 2.12% to
117.41%, whereas both liquid limit and plasticity index reduces from 59.2% and 23.3% to 32.9% as
well as 11.2% respectively for Bwari site at optimum percentage and significant correlation of r = 0.96,
α = 0.05. Further scrutiny on each of the concrete samples cured for 3, 7, 14, 21 and 28 days,
demonstrated that the samples produced with eggshell ash, waste paper ash and waste lactose had the
highest compressive strength values, followed by those produced without waste paper ash. Likewise,
XRD, XRF and FTIR investigation reveals the loss on ignition value which in turn designates those
clay samplings has lesser carbonaceous substance and greater mineral matter quantities.
Keyword: Fired-Bricks; Pavement; Green Construction, Interlocking.
1. Introduction
The extant rate of waste production is untenable and there are guesses that suggest that there will be
constant happening of same now and in the world to come. Currently, nations generated much wastes
as a result of swift inhabitants growth, agrarian and industrial deeds, this in turn vitiate environment
through the discharge of contaminations into various landfills or littering around (Kadir and Sarani,
2012; Peña et al., 2016). Even with this express enlargement also generated an exponentially
amplifying quantities of unwanted materials or left over, and secure disposal of these is an irresistible
environmental concerns that necessity confrontation (Blaszua et al. 2020; Amu et al. 2020; Amu et al.
2005a). Also, targeting toward ecological construction technique, lots of unprocessed material have
being investigated to serve as innovative substitutes green materials so as to lessen their littering around
the communities, and this reutilizing of numerous kinds of wastes, started couples of years ago.
Likewise, owing to large demand placed on the construction materials, specifically in the last era
because of sudden rise in contamination levels cause a chronic scarcity of construction resources (Amu
et al. 2005b; Lithua and Zengmei 2020). Some instances are fly ash as urban solid waste incinerator ash,
egg shell ash or ceramic tile wastes, which are in this day and age comprehensively reutilized by the
construction firms, resulting in waste disposal reducer or total avoided (Amu et al. 2021; Roman and
Martin 2020; Jiang et al. 2018). The making of orthodox soil enhancers like cement, or bitumen entails
energy utilization, non-renewable natural resource and highly emissions of carbon dioxide, which is
causing environmental problems and also depletion of natural resources; therefore, alternative
stabilisers are sought (Blaszua et al. 2020; Amu et al. 2020; Amu et al. 2005a). Fired clay bricks are an
imperative integral construction material specifically for low cost, indestructible and inexpensive
construction works. This serves as a cherished material because its physiognomies can be improved to
withstand several loading circumstances (Adetayo et al. 2022; Aprianti et al. 2015; Bergamoth et al.
Page 244
SEMIT 2023
2016). Red clay is widely distributed in all the Local councils Areas (LCA) in Abuja, North Nigeria,
besides modern pottery in Nigeria commenced from LGA in Abuja. Red clay depressively created and
largely forms strata within thickness between 5–15 meters, typified with extreme values of porosity
ratio and liquid limit, shrink as well as crack predisposition, and weak water stability (Amu et al. 2005b;
Lithua and Zengmei 2020). CaCO3 which indicates Calcium carbonate is an imperative inorganic
mineral content, largely made of calcite which is an non-metallic mineral unprocessed materials. It is
characteristics with great purity, extreme whiteness, inexpensive materials, and is observably better
than most non-metallic minerals in the aspect of cost-performance (Gherandi et al. 2018; Chung and
Mangi 2020). It is extensively scattered in provinces with red clay, has a small mining rate and is stress-
free in transporting. Egg shell has calcium carbonate and the blending of calcium carbonate with clay
lessened the pore contents, expands the moisture, then leads to enhancement of the total number of clay
pores. The Highway and civil engineers have left with option of converting industrial, natural or
agrarian waste to advantageous construction and pavement materials. One fashionable reprocess
opportunity is using waste paper as a pavement material (Adetayo et al. 2022; Aprianti et al. 2015;
Bergamoth et al. 2016). Since the road and highway works uses huge amount of non-renewable items,
hence the prospective function of waste paper and eggshell ash for producing an efficient and composite
interlocking fired brick for construction and highway works not only serves as the prospective means
of waste paper and egg shell recycling but it will equally lessens the demand pressure on global natural
resources (Amu et al. 2021; Roman and Martin 2020; Jiang et al. 2018).
2. Materials and techniques

2.1. Soils
Three clay samples formed through weathering of fieldspars and micas, little in dimension and
incredibly flaky in shape which also forms part of a class of multifaceted alumino-silicates of iron,
potassium, as well as magnesium was collected from each three Local Council Area, with Northing
coordinate (994352, 946613 and 987529), and Eastings values (296268, 272158 and 308040) in Abuja
North-central Nigeria, making total of eighteen samples, where the materials is in abundant. These
samples will be collected in large bags and will be conveyed to both soil and concrete laboratory of
Pavement Unit of Federal Ministry of Works Abuja, Geotechnical Laboratory Civil Engineering
Federal University of Technology Akure and University central laboratory of Umaru Musa Yar’adua
University Katsina, Nigeria. Thereafter; the clay samples was air-dried, minced and sieved via
numerous standard sieves for varioust categories of scrutinise with the large sieve being BS N0 4 sieve
(4.76mm aperture). The best sampling out of the three samplings obtained from each Local Councils
constituency (Figure 1) was choosing after the initial investigation of the clay samples.
Figure 1: a) Gathering of materials, b & c) laboratory work in progress.
2.2. Stabilizers
The broken egg-shells (Figure 2) was acquired from indigenous eateries and fast food hawkers within
Abuja municipality, the broken eggshells (as shown in Figure 2) was simmered to eradicate any egg
residue which may still fastened onto the exterior of the shells membrane before aerated them under the
sun. Parched egg shells were minced into fine contents via a laboratory crushing machine (grinder). The
papers used such as thesis, dissertation and project papers, books, as well as paper utilized in numerous
offices. These was put inside a drum and burned at a temperature of roughly 750°C, after all the waste
Page 245
SEMIT 2023
paper have absolutely turned into ash, then left to cool down, before the ash was sieved via a sieve
No.75 -200 (Figure 3). Waste paper ash (WPA) was carefully chosen to serve as major aggregate filling
material for the interlocking fired bricks (IFB), owing to its spreading and obtainability at every
institutions, offices etcetera, on the globe especially Abuja FCT, Nigeria. This scrutiny used, waste
lactose (Figure 4) which is dairy processing company byproduct like milk and yoghurt waste, was gotten
from Abuja FCT, Nigeria, and was utilized as binder for the making of the IFB. Waste lactose (WL)
was preferred as binder, for the reason that it’s available especially in the northern Nigeria, and as a
waste byproduct of the dairy industry it discarding ways causes serious environmental issue.
Figure 2. Egg Shell Conversion to powder form.
Figure 3. Materials used for enhancing clay samples.

2.3. Scrutiny Techniques
The soil samples was mixed together on the ground to obtain representative samples and sieved and
later spread in the open air to dry based on the procedures stipulated in the BS 1377-1990: 1-8 and
BS1377 code (1990) Part 1, section 7 towards testing. Samples was subjected to classification tests in
their natural air-dried state, before investigation of three possible types of interlocking fired clay brick–
code named type A, type B and type C in its first stage. Type A was control (clay and cement), then
Type B (clay, cement and waste lactose) and Type C (clay, cement waste lactose and waste). This was
machine moulded interlocking fired bricks in kiln at about 7000C, after cooling immerse in water for
24 hours before tests and curing are carried out.
3. Results and Discussion

3.1 Outcome of Piperplot
In this scrutiny, Golden software piper plot, that is a multifaceted diagram, milli equivalents proportion
concentrations of key cations; comprising of K+, Ca2+, Na+ as well as Mg2+, together with anions
encompass of negative elements; that are Cl-, HCO3-, and SO42-, was utilized to plot these diagrams.
They are mapped in dual triangular arenas, which afterward projected inside the pivotal diamond spot.
Page 246
SEMIT 2023
This then shows the connexions and dissimilarities among water utilizing for preparing and curing the
clay samples because those with related physiognomies will plot together as clusters. Also, the data
mapped on the sections of diamond-shaped arena picks the water class in a water sample.
Figure 4. Scrutiny of clay samples using piperplot.
From Figure 4, it is obvious that bulk of the samplings (72.91%; 3 GW used on 3 clay materials)
ascribed to Ca2+-Mg2+-Cl--SO4 2- or otherwise labelled arena 1, indicating the ascendancy of alkaline
earths over alkali, that is Na+K < Ca+Mg, as well as resilient acidic anions above fragile acidic anions,
meaning; HCO3 < Cl+SO4. Minority samples (27.09%; 3 GW used on 3 clay materials) mapped below
4th arena, ascribed to Ca2+_Mg2+_HCO3 - portentous the ascendancy of alkaline earths above alkali
as well as frail acidic anions above resilient acidic anions. Piper diagrams indicated the dominance of
mixed water type (72.91%; 3 GW used on 3 clay materials) having none of cation-anion pair surpasses
50%.
3.2 Outcome of Classification Tests

Abridgement values of fundamental and engineering properties of the selected Abuja Local Councils
clay soil scrutinizes based on AASHTO 2007 Pavement recommendation before stabilization is shown
in Table 1. Natural moisture content values are: (10.8 – 14.5%), {16.3 – 20.9%} and {19.9 – 25.2%}
for Amac, Bwari and Kwali respectively. The result displayed that kwali clay has the extreme NMC
value (25.0%) and followed by Bwari (20.9%) while Amac clay noted as the least (14.5%). This
investigation indeed ascertains the fluctuation in percentage amount of moisture available in the clay
samplings. For numerous clay soils, the quantities of water are enormously significant index utilized
for instituting the connexion between the manner a soil performs and its physiognomies. The
consistency of a fine-grained clay soils mainly rest on its water content.
Page 247
SEMIT 2023
Figure 5. Outcome of Particle size.
The water content is also utilized in conveying the phase interactions of water, air as well as solids in a
given volume of soil. For sieve investigation (Figure 5) the gravel, sand and silty /clay aggregate values
ranges as follows: Amac (5.0 – 50.1%); Bwari (2.1 – 49.0%) as well as Kwali (6.2 – 50.1%). From table
(Table 1) which signifies that most of the clays fell within A-6-8 (Silty or clayed) soils according to
AASHTO grading system for pavement materials. All of the samplings investigated did not met the
prerequisite of BS 1377 (1990) specification pavement materials, based on percentage passing 75 -
200mm sieve and plasticity index (PI) % as shown in Table.1.
Table 2: Outcome of triaxial scrutiny for optimum cement and waste content.
Samples Content Deviation Cohesion C Angle of
stress σ3 (KPa) internal
(KN/m2) friction (ϴ)0
Control (0%) 310.2 87.4 12.9
6% cement + 4%ES + 4% WPA +4% 339.7 87.7 12.7
WL
Gwagwalad 6% cement + 6%ES + 6% WPA +6% 358.4 87.9 12.4
a WL
6% cement + 8%ES + 8% WPA +8% 326.6 87.2 13.2
WL
6% cement + 10%ES + 6% WPA +6% 361.2 88.0 11.1
WL
Control (0%) 275.7 79.4 11.8
Page 248
SEMIT 2023
6% cement + 4%ES + 4% WPA +4% 298.2 79.8 11.4

WL
Abaji 6% cement + 6%ES + 6% WPA +6% 318.4 80.1 11.1
WL
6% cement + 8%ES + 8% WPA +8% 301.2 78.5 12.3
WL
6% cement + 10%ES + 6% WPA +6% 330.4 80.4 10.7
WL
Control (0%) 269.0 72.3 12.9
6% cement + 4%ES + 4% WPA +4% 289.6 72.8 12.6
WL
Kuje 6% cement + 6%ES + 6% WPA +6% 302.1 73.2 12.2
WL
66% cement + 8%ES + 8% WPA +8% 287.3 72.6 13.1
WL
6% cement + 10%ES + 6% WPA +6% 310.7 73.5 11.2
WL
3.2 Outcome of Triaxial Scrutiny

Results triaxial test for interlocking fired brick stabilized with egg shell, waste paper ash and waste
lactose are shown in Table 2 and Figures 6 & 7. The result displayed the impact of 6% cement and
various percentages of waste contents on the clay sampling enhanced.
The optimum cohesion of 88.0 KPa, 80.4 KPa and 73.5 KPa at angles of internal friction of 11.10, 10.70
and 11.20 for samples Gwagwalada, Abaji and Kuje respectively was achieved at 6% of other materials
and 10% of egg shell powder.
-
Page 249
SEMIT 2023
Table 1: Engineering Physiognomies of the clay before enrichment

PROPERTIES SOIL SAMPLES (CONTROL)
Gw1 Gw2 Gw3 Ku1 Ku2 Ku3 Ab1 Ab2 Ab3
Natural Moisture Content (%) 12.7 12.8 12.7 6.6 6.7 6.7 7.5 7.6 7.8
Specific Gravity 2.45 2.46 2.45 2.56 2.55 2.57 2.86 2.87 2.88
Grain Size Distribution
Pass #7 sieve 99.8 100 100 100 100 100 100 100 100
Pass #36 sieve 92.9 93 92.8 79 78.9 79 75.7 75 75
Pass #200 sieve 56 56 56.1 43 42.5 42.9 34.5 34.3 34.4
Gravel (%) 0 0 0 1 1 1 2 2 2
Sand (%) 41.4 41.2 41.3 52 51.8 51.8 44.5 44.3 44.2
Silty/Clay 59.3 59.4 59.2 47 46.8 47 54.2 54.1 54.0
Consistency Limit (%)
Liquid Limit 30.4 30.0 30.3 33.8 33.1 33.2 35.7 35.6 35.5
Plasticity Index 14.2 13.1 12.2 13.5 13.4 13.0 13.5 13.4 13.3
Triaxial test 310.2 278.6 250.6 210.8 238.4 269.0 222.8 247.7 275.9
Cohesion (KPa) 87.4 87.2 86.9 71.9 72.1 72.3 78.9 79.0 79.4
Angle of internal friction ϴ0 12.9 12.8 12.7 12.5 12.6 12.7 11.5 11.6 11.8
Soil Classification A-4-3 A-6-3 A-4-3
Colour Reddish Reddish
Soil Type Silty or clayed
FMWH (1997) & AASHO 1990 Gwagwalada Kuje Abaji
Specification Gw1 Gw2 Gw3 Ku1 Ku2 Ku3 Ab1 Ab2 Ab3
LL (< 35%) 30.4 30.0 30.3 33.8 33.1 33.2 39.0 39.2 39.0
Fail Fail Fail Fail Fail Fail Fail Fail Fail
PI (>12%) 14.2 13.1 12.2 13.5 13.4 13.0 13.1 13.3 13.2
Fail Fail Fail Fail Fail Fail Fail Fail Fail
Overall Rating A-4-3 or a-6-3 are not good pavement materials
Page 250
SEMIT 2023
Cohesion value of stabilized clay samplings was small at the greatest angle of internal friction which
indicates that the clay samples are very plastic. The result showed the impact of 6% PC and various
percentages of ES, WPA and WL on the clay sampling enhanced. It showed that the optimum Triaxial
test result for PC, WPA and WL at 6% and 10% ES are: Gwagwalada (Deviation stress 361.2 KPa),
Abaji (Deviation stress 330.4 KPa), and Kuje (Deviation stress 310.7 KPa). The outcomet of triaxial
scrutiny also followed the trend of the cohesion of the stabilized sample was low at the greatest angle
of internal friction which makes clay very plastic.
Figure 7: Outcome of Triaxial scrutiny for clay sample in Abuja.
3.3 Outcome of Atterberg Scrutiny

According to Amu et al., (2022) and Agashua et al (2022), liquid limit below 35% designates low
plasticity, between 35% and 50% specifies intermediate plasticity, between 50% as well as 70% extreme
plasticity and between 70% and 90% very extreme plasticity and greater than 90% extremely high
plasticity. This illustrates that both samples Gwagwalada and Abaji have low plasticity, while only Kuje
has intermediate plasticity (Table 3 and Figure 8).
GW – Gwagwalada, AB – Abaji and KU – Kuje.

Figure 8. Outcome of atterberg for control and optimum analysis.
Page 251
SEMIT 2023
Figure 9. Outcome of Atterberg scrutiny.
Table 3: Outcome of Atterberg scrutiny for optimum cement and waste content.
Samples Content Liquid Plastic Plasticity
Limit (LL) Limit (PL) Index (PI)
Control (0%) 30.4 16.2 14.2
6% cement + 4%ES + 4% WPA +4% WL 41.7 28.2 13.5
6% cement + 6%ES + 6% WPA +6% WL 49.7 37.3 12.4
Gwagwalada 6% cement + 8%ES + 8% WPA +8% WL 42.8 29.6 13.2
6% cement + 10%ES + 6% WPA + 6% 57.8 48.7 9.1
WL
Control (0%) 33.8 20.3 13.5
6% cement + 4%ES + 4% WPA +4% WL 42.8 30.0 12.8
6% cement + 6%ES + 6% WPA +6% WL 50.1 38.2 11.9
Kuje
6% cement + 8%ES + 8% WPA +8% WL 49.3 37.0 12.3
6% cement + 10%ES + 6% WPA +6% WL 57.5 47.9 9.6
Control (0%) 35.7 23.2 13.5
6% cement + 4%ES + 4% WPA +4% WL 45.7 33.0 12.7
6% cement + 6%ES + 6% WPA +6% WL 52.4 40.8 11.6
Abaji 6% cement + 8%ES + 8% WPA +8% WL 50.0 38.0 12.0
6% cement + 10%ES + 6% WPA + 6% WL 55.9 46.1 9.8
The addition of optimum value of 6% Portland cement and eggshell, waste paper ash and waste lactose
at 2, 4, 6, 8 and 10% to the samplings drastically creates vicissitudes in the liquid limits and plastic
limits of all the samples. These falls in plasticity indices are indicators of soil enhancement (Figures 8
& 9). Besides Federal Ministry of Works and Housing (1997) for road works recommended liquid limits
of 50% as utmost value for pavement materials. All the scrutinized soil samples fall within this
specification, thus making them appropriate for pavement construction work after enhancement.
3.4 Outcome of clay characterization via XRF, XRD and FTIR scrutiny
Table 4 displays the outcome of X-ray fluorescent analyzer (XRF, X-ray diffraction (XRD), and FTIR.
Table 4 and Figures 10 & 11 shows clearly that the clay samples have calcium, alumina and silica as
vital materials and others in lesser quantities. The loss on ignition value designates that clay samplings
has lesser carbonaceous substance and greater mineral matter quantities. Further scrutiny like XRF,
XRD and FTIR investigation were carried out so as to comprehend the make-up of phenol sorption on
Page 252
SEMIT 2023
clay using raw sorbents. XRF was done to ascertain the chemical make-ups of the minerals that are
available in the clay (Table 4). Table 4 signifies that the silica as well as alumina are exist in extreme
quantities whereas other minerals are available in bit amounts.
Figure 10. Outcome of FTIR and XRD.
a b
Figure 11. XRD plots for a) Abaji, b) Gwagwalada
Page 253
SEMIT 2023
Table 4. Various clay soils XRF, XRD and FTIRvalues

Gwagwada sample Abaji Kuje
XRF Scrutiny
Chemical features Values Chemical features Values Chemical features Values
(%) (%) (%)
SiO2 47.97 SiO2 45.77 SiO2 46.85
Al2O3 33.89 Al2O3 33.79 Al2O3 31.76
Fe2O3 2.51 Fe2O3 2.48 Fe2O3 2.41
CaO 0.82 CaO 0.79 CaO 0.74
MgO 0.53 MgO 0.51 MgO 0.54
TiO2 0.42 TiO2 0.38 TiO2 0.41
LOI (7000C) 12.36 LOI (7000C) 12.26 LOI (7000C) 12.29
XRD Experiment
Mineral d(Å) Mineral d(Å) Mineral d(Å)
Quartz 4.1913, 1.4716 Quartz 4.454, 2.44 Albite 3.6426, 0.175
Muscovile 3.6858, 0.5911 Orthoclase 4.2535, 0.124 Orthoclase 2.4275, 0.0911
Albite 3.3206, 0.1791 Muscovile 3.6940, 0.062 Quartz 1.8044, 0.173
Orthoclase 3.2325, 0.084 Albite 3.3394, 0.155 Muscovile 1.5142, 0.223
FTIR Outcome
Band (cm-3) Transmittance (%) Purpose
791.8 27.7 Si-O stra, Si-O-Al stra.
1635.2 90.9 H- O- H stra
3695-8 27.4 Al-O-H (stra.)
3450-3 49.9 H-O-H stra.
1041-3 7.1 Si- O-Si, Si-O str.
912-6 32.4 Al-O-H stra.
467.9 12.2 Si-O stra, Si – O- Fe stra.
3623-4 13.4 Al-)-H (inter-octahedral)
Likewise, XRD scrutiny was also used to further confirm the availability of above minerals. XRD is
utilized to ascertain the inert contents of the unprocessed or processed material constituents, both
quantitative together with qualitative phase scrutiny of multi-phase blends. The existences of minerals
in clay soil were recognized by relating outcome from ‘d’ (Table 4), based on recommendation of
“Selected Powder Diffraction File and Data for Minerals both of year 1974. Definitely quantifiable
guesstimate phases in these adsorbents have not been formed except XRD characterization and their
style demonstrated presence of quartz, muscovile, albite, orthoclase and others as the substantial phases.
Advance scrutiny via FTIR depiction also validate the existence of the above minerals and others (Table
4). FTIR investigation of these adsorbents aid in the recognition of diverse formation or contents of the
minerals available in the clay. The combined vibrations are noticeable owing to the present of various
components. All the same, witnessed bands (within 4000 –500 cm–1) was falteringly assigned. In the
IR investigation of clay, the Si–O stretching vibrations were observed for Si-O stra, Si-O-Al stra., and
Si-O stra, Si – O- Fe stra., as 791.8 cm–1, and 467.9 cm–1 displaying the ample amount of quartz (Nayala
and singh 2007). Al-O-H (stra.), Al-)-H (inter-octahedral) and H-O-H stra. ascertain a sturdy band at
3695-8 cm–1, 3623-4 cm–1 and 3450⋅3 cm–1 designate the leeway of the hydroxyl connection. Similarly,
an all-encompassing band at 3450⋅3 cm–1 and 1635.2 cm–1 for H-O-H stra in the gamut of clay proposes
the likelihood of water of hydration inside the adsorbent. The presence of bands at 3623-4 cm–1, and
1635-2 cm–1, for H-O-H stra., as well as Si-O-Si, Si-O str., likewise symptomatic present of gypsum,
while waves detected at 912⋅6 cm–1 point toward the prospect of having hematite (Tuddenham and Lyon
1960). Hence, the outcomes of IR are quite supportive in the discovery of numerous kinds of minerals
available within the sorbents utilized.
Page 254
SEMIT 2023
Acknowledgements
The authors appreciate the authorities of the Pavement Unit of Federal Ministry of Works Abuja,
Geotechnical Laboratory Civil Engineering Federal University of Technology Akure and University
central laboratory of Umaru Musa Yar’adua university Katsina, Nigeria, for providing the equipment
for this research work.
Conclusion
This paper appraised the likelihood of utilizing either PC + ES + WPA or PC, ES, WPA - WL based
interlocking fired bricks to enhance clay samplings, with standpoint of augmenting the engineering
physiognomies in this style. With the intention of increase the pavement thickness requisite for the
artery suitable serviceability throughout its design as well as service life time. Preliminary experiment
on the clay samples used signifies that the soils is an A6 soil in line with AASHTO categories technique,
which is a Silty or clayey soils, hence there is need for augmentation. The discovery about optimization
demonstrated that enrichment material used will meritoriously improve waste enhance interlocking
fired bricks at cement 6 %, ES 8%, WPA 6%, WL 6%. Blending of PC with ES - WPA created
interlocking fired bricks that enriched the strength and geotechnical characteristics of the clay soils,
even though ES behavior outshine. Specifically, 6 % and 10% of cement and ES respectively, as well
as 6 % WPA and WL correspondently is the optimal proportion to enrich the features of clay soils. The
chemical investigation reveals clearly that the clay samples have calcium, alumina and silica as vital
materials and others in lesser quantities. The piperplot’s graphic validates that there are connexion
processes among two or more various facies that ensued within the clay and water system.
Reference
1. Muhammad A, K. Faiza, A. Majid, Compressive behavior of interlocking plastic blocks structural
elements having slenderness, Buildings 12 (2022) 2257.
2. Heyder A, Sugini. A study on interlocking brick innovation using recycled plastic waste to support
the acoustic and thermal performance of a building. Arbitek: J. Tek. 6(3)(2021), pp.335-348
3. Amu O.O, Fapohunda C, Adetayo A. O & Festus M. (2020). Sugarcane Straw Ash Effects On Lime
Stabilized Lateritic Soil For Structural Works. Stavební obzor - Civil Engineering Journal. 29.
10.14311/CEJ.2020.03.0027.
4. Agashua L.O, Ogbiye A.S, Amu O.O, Oluyemi-Ayibiowu B.D, igibah C.E and Onakunle O. 2023.
GrapherAnalysis and the Impact of Sodium Silicate Activator on Strength of Kaolin-rice husk ash
stabilized laterite. Materials Today Proceedings, 74(3), 377-388.
5. Igibah C.E and Amu O.O. 2022. Variability of rice husk and kaolin clay as locally available
Geopolymer materials on cement stabilized lateritic soil. AJERD, 1(2), 1-9.
6. Adetayo A. O, Amu O.O & Alabi S. (2019). Improvement of cement stabilized structural lateritic
with pulverized snail shell. Selected Scientific Papers - Journal of Civil Engineering. 14. 95-106.
10.1515/sspjce-2019-0021.
7. Al-Fakih, Amin, Bashar S. Mohammed, M. M.A. Wahab, M. S. Liew, and Y. H. Mugahed Amran.
2020. “Flexural Behavior of Rubberized Concrete Interlocking Masonry Walls under Out-of-Plane
Load.” Construction and Building Materials 263 (December): 120661.
https://doi.org/10.1016/j.conbuildmat.2020.120661.
8. Amu O.O, Igibah C.E and Adedapo A.O. 2022. Impact of Sodium Silicate Activator on
Compressive Strength of Cement-RHA based stabilized lateritic soil. A case study of Kwali, Abuja
FCT. Key Engineering Materials, 917(1), 267-277.
9. Amu O.O, Igibah C.E, Oluyemi-Ayibiowu B.D, Aderinola O.S, Adedapo A.O, Agashua L.O and
Busari A.A. 2022. Effect of California Bearing Ratio and Unconfined Compressive Strength analysis
on Mechanical Strength and Microstructure of Kaolin clay powder blend SSA GFeopolymer. AJERD,
5(1), 31-40.
10. Amu O.O, Igibah C.E, Oluyemi-Ayibiowu B.D and Agashua L.O. 2022. Geo-Mechanical
Enhancement of a Grained soil blended with Silicate- Portland cement powder for usage in Construction
Industry. Journal of Materials Science Research and Reviews, 9(3), 17-28.
Page 255
SEMIT 2023
11. Shi, Tingwei, Xihong Zhang, Hong Hao, and Chong Chen. 2021. “Experimental and Numerical
Investigation on the Compressive Properties of Interlocking Blocks.” Engineering Structures 228
(February): 111561. https://doi.org/10.1016/j.engstruct.2020.111561.
12. Gencel, Osman, Ertugrul Erdugmus, Mucahit Sutcu, and Osman Hulusi Oren. 2020. “Effects of
Concrete Waste on Characteristics of Structural Fired Clay Bricks.” Construction and Building
Materials 255 (September): 119362. https://doi.org/10.1016/j.conbuildmat.2020.119362.
13. Jeba Jeslin, A., and I. Padmanaban. 2020. “Experimental Studies on Interlocking Block as Wall
Panels.” In Materials Today: Proceedings, 21:1–6. Elsevier Ltd.
https://doi.org/10.1016/j.matpr.2019.05.294.
14. Muheise-Araalia, D., and S. Pavia. 2021. “Properties of Unred, Illitic-Clay Bricks for Sustainable
Construction.” Construction and Building Materials 268 (January): 121118.
https://doi.org/10.1016/j.conbuildmat.2020.121118.
15. Krishna P.A, Jane H.H, Paul O.A. Optimization of Mix Proportions for Novel Dry Stack
Interlocking Concrete Blocks Using ANN. Adv. Civ. Eng. (2021), Article ID 9952781, pp. 15.
16. Preeti S.N and Singh B.K 2007. Instrumental characterization of clay by XRF, XRD and FTIR.
Indian academy of sciences Bull. Mater. Sci., 30(3), pp. 235–238.
17. Tuddenham W M and Lyon R J P 1960 Anal. Chem. 32 1630.
Page 256
SEMIT 2023
Chinese Postman Problem with Dynamic Start/End Point

Mustafa Yilmaz1[0000-0002-2135-5762]
1
Ataturk University, Faculty of Engineering, Department of Industrial Engineering, Erzurum, Turkey
mustafay@atauni.edu.tr
Abstract: The Chinese Postman Problem (CPP) is a variant of the arc routing problem, and it is a
routing problem that aims to return to the starting point with the shortest tour length by passing through
all arcs at least once in a network consisting of nodes and arcs. CPP has many real-life applications
such as snow removal and road maintenance work. In the literature, it is possible to see different CPP
applications by adding new constraints to the problem. In almost all of these studies, the starting point
is determined as a fixed point and it is aimed that the tour or tours to be created will start from this point
and end at this point. In this study, a new mixed integer model for CPP is proposed and it is aimed to
determine the most appropriate starting or ending point by the model. The proposed model was run on
a small sample network and the results showed that tours with shorter lengths were obtained.
Keywords: Chinese Postman Problem, Variable Start/End Points, Mathematical Model
1 Introduction
Routing problems are basically divided into two as node routing and arc routing problems. Traveling
salesman and vehicle routing problems are considered as node routing problems [1,2]. CPP is an arc
routing problem. In node routing problems, it is aimed to visit the nodes only once and return to the
starting point, while in arc routing problems, it is aimed to pass through each arc at least once and return
to the starting point. It is seen that node routing problems have been studied more in the literature. It is
possible to encounter many applications of edge routing problems in real life such as snow removal,
road maintenance works, routing of patrol vehicles and determination of cutting routes, etc. [3,4]. In
almost all of these studies, the starting point is assumed to be a fixed point and it is aimed that the tours
found will start from this point and complete at this point. However, it has been determined that the
tours starting from different points on the network are shorter. Many real-life problems, such as cutting
path determination problems, need to be formulated in this way [5,6,7].
In this study, a mixed integer mathematical model is proposed to determine the most suitable start and
end points in a given network and thus find the shortest tour length. By using this model, the necessity
of assuming any point on the network as a starting or ending point has been eliminated. These changes
to the basic model of CPP can be applied to other CPP variations. The presented model is applied to a
small network and all the results are reported.
2 Problem Definition
The CPP is defined on the 𝐺(𝑉, 𝐸) network as the set of nodes 𝑉 and the set of edges 𝐸. In the routing
problem discussed, the fact that the degree of the starting point is odd or even is closely related to the
solution of the problem [8,9]. In order to explain this better, the 3 networks given in Figure 1 are
examined. In the 1st network, all node degrees consist of odd numbers and it is not possible to return to
the starting point with the possibility of passing through all the edges at once. These networks are called
non-euler networks. In the second network, some node degrees are single (A and D) and some have
even node degrees (B, C and E), and these networks are considered as semi-euler networks. In non-
euler and semi-euler networks, it is not possible to return to the starting point without repeating the
edge. In the third network, all node degrees are even numbers and it is possible to return to the starting
point by passing through each edge only once [10].
Closed and open cutting routes for the 3 networks in Figure 1. are given in Table 1. below. All edge
lengths are assumed to be equal and equal to 1.
Page 257
SEMIT 2023
E E
B C B C B C
A D A D A D
Fig. 1. Non-Euler, Semi-Euler and Euler Graphs

When Table 1 is examined, the open and closed tour lengths of the 3rd grid, which is the euler grid, are
equal to each other, while a saving occurs in the open tour lengths of the 1st and 2nd grids, which are
non-euler and semi-euler. Since all node degrees are odd in the 1st network, the amount of savings is
equal no matter which starting point is chosen. However, if the starting point is selected from odd ones
(A) in the 2nd network, the saving amount is 1, and if it is selected from even ones (E), the saving
amount is 2. This makes it necessary to choose the starting point correctly in CPP.
Table 1. Differences between the solution of closed/open tours

Starting Point Closed Tour Open Tour
A-B-C-D-A-C-D-B-A A-B-C-D-A-C-D-B
Graph 1 A
Lenght: 8 Lenght: 7
A-B-C-D-A-C-E-B-D-A A-B-C-D-A-C-E-B-D
A
Lenght: 8 Lenght: 7
Graph 2
E-B-A-D-C-B-D-C-E-C-A-B-E E-B-A-D-C-B-D-C-E-C-A
E
Lenght: 12 Lenght: 10
A-B-C-D-A-F-D-B-E-C-A A-B-C-D-A-F-D-B-E-C-A
Graph 3 A
Lenght: 10 Lenght: 10
The CPP is defined on the 𝐺(𝑉, 𝐸) network as the set of nodes 𝑉 and the set of edges 𝐸. The basic
mathematical model of CPP in the literature is given below. In a network with 𝑛 points, 𝑥𝑖𝑗 is an integer
variable that expresses the number of passes through the edge connecting the 𝑖 and 𝑗 points, and 𝑐𝑖𝑗
represents the relevant cost value [11,12,13].
𝑛 𝑛
𝑚𝑖𝑛 ∑ ∑ 𝑐𝑖𝑗 𝑥𝑖𝑗 (1)

𝑖=1 𝑗=1
𝑠𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜
𝑛 𝑛
∑ 𝑥𝑖𝑗 − ∑ 𝑥𝑗𝑖 = 0 ∀𝑖 (2)

𝑗=1 𝑗=1
𝑥𝑖𝑗 + 𝑥𝑗𝑖 ≥ 1 ∀𝑖, 𝑗 (3)

𝑥𝑖𝑗 ≥ 0 𝑎𝑛𝑑 𝑖𝑛𝑡𝑒𝑔𝑒𝑟 ∀𝑖, 𝑗 (4)
Equation (1) minimizes the total tour length. Equation (2) ensures that the flow is maintained, while
equation (3) ensures that each edge is passed at least once. Statement (4) ensures that the decision
variables take positive and integer values.
Page 258
SEMIT 2023
This model guarantees returning to the starting point with minimum cost by passing each edge at least
once in an undirected network.
3 Mathematical Formulation
The following integer mathematical model is proposed for variable start/end point CPP. In this model,
the start and end points are determined by the model in order to minimize the total tour length in a given
network. Depending on the nature of the network, these tours can be open or closed. In the model,
binary variables that determine the most suitable start and end nodes in the network are used. Parameters
and variables of the proposed model are presented below.
Parameters Definition
𝑛: Number of nodes
𝑐𝑖𝑗 : The cost of 𝑖 − 𝑗 arc
𝑀: Big number (=106)
Variables Definition
𝑥𝑖𝑗 : The passing number of 𝑖 − 𝑗 arc
𝑠𝑖 : Binary variable representing the starting point
𝑓𝑖 : Binary variable representing the endpoint
The proposed mixed integer mathematical model:
𝑛 𝑛
𝑚𝑖𝑛 ∑ ∑ 𝑐𝑖𝑗 𝑥𝑖𝑗 (5)

𝑖=1 𝑗=1
𝑠𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜
𝑛
∑ 𝑠𝑖 = 1 (6)
𝑖=1
𝑛
∑ 𝑓𝑖 = 1 (7)
𝑖=1
𝑛 𝑛
∑ 𝑥𝑗𝑖 − ∑ 𝑥𝑖𝑗 ≥ 𝑠𝑖 − 𝑀𝑓𝑖 ∀𝑖 (8)

𝑗=1 𝑗=1
𝑛 𝑛
∑ 𝑥𝑖𝑗 − ∑ 𝑥𝑗𝑖 ≤ (1 − 𝑠𝑖 ) + 𝑀(1 − 𝑓𝑖 ) ∀𝑖 (9)

𝑗=1 𝑗=1
𝑥𝑖𝑗 + 𝑥𝑗𝑖 ≥ 1 ∀𝑖, 𝑗 (10)

𝑥𝑖𝑗 ≥ 0 𝑎𝑛𝑑 𝑖𝑛𝑡𝑒𝑔𝑒𝑟 ∀𝑖, 𝑗 (11)
𝑦𝑖𝑗 , 𝑠𝑖 , 𝑓𝑖 = 0 𝑜𝑟 1 ∀𝑖, 𝑗 (12)
Equation (5) in the model minimizes the total tour length. Equation (6) ensures that only one of the
points is the starting point, and equation (7) ensures that only one of the points is the end point.
Equations (8) and (9) work together and provide the flow route of the tour by determining from which
point the tour should start and at which point it should end. While equation (10) guarantees that each
edge is passed at least once, expressions (11) and (12) are the sign constraints of the mathematical
model.
The 𝑠𝑖 variable in the model determines the optimal starting node of the tour. If the tour must start from
any node, for example node 1, then the starting point can be guaranteed to be node 1 by adding the 𝑠1 =
Page 259
SEMIT 2023
1 constraint to the model. The same process can be applied to the end node by assigning a value to the
𝑓𝑖 variable. An example solution is given in the next section for different cases of the starting point.
4 An Illustrative Example
A non-directional network with 10 nodes and 15 arcs is given in Figure 2. Values on the arcs are
considered transit costs. Considering the degrees of nodes in the network, it is seen that the network has
a semi-euler feature. Nodes 7 and 9 are of even degree, and all other nodes are of odd order In such a
network, it is impossible to create a route without repeating the edge. The model solutions given above
for this network are given in Table 2.
Fig. 2. An undirected graph

The above undirected weighted network is first solved with the basic CPP model and the optimal
solution is obtained as 244. Since there is a closed tour in the basic CPP model, the start and end points
are used as node 1. Then, the proposed model for the same network was run and an open route with the
most suitable starting point as 2(4) and ending point as 4(2) was obtained. The optimal solution value
of this route was calculated as 231. In addition, the tour lengths and routes obtained when other nodes
are the starting point for the same network are presented in Table 2.
Table 2. Solutions of illustrative example
Starting Ending Optimal
Model Route
Point Point Solution
Basic CPP 1 1 244 1-3-4-3-6-10-9-6-10-8-5-8-7-5-2-1-4-7-2-1
1 2 232 1-3-4-3-6-10-9-6-10-8-5-8-7-5-2-1-4-7-2
2 4 231 2-5-8-5-7-8-10-6-9-10-6-3-1-2-7-4-3-1-4
3 4 233 3-6-10-9-6-10-8-5-7-8-5-2-1-3-4-7-2-1-4
4 2 231 4-1-3-4-7-2-1-3-6-10-9-6-10-8-7-5-8-5-2
The
5 8 237 5-2-1-4-3-1-2-7-5-8-10-9-6-10-6-3-4-7-8
proposed
6 10 234 6-10-9-6-3-1-4-3-4-7-8-5-2-1-2-7-5-8-10
model
7 2 239 7-4-3-1-4-7-5-8-5-2-7-8-10-9-6-10-6-3-1-2
8 5 237 8-7-4-3-6-10-6-9-10-8-5-7-2-1-3-4-1-2-5
9 10 243 9-6-9-10-8-5-8-7-5-2-1-4-7-2-1-3-4-3-6-10
10 6 234 10-8-5-7-2-1-2-5-8-7-4-3-4-1-3-6-9-10-6
Page 260
SEMIT 2023
5 Conclusion
In this study, a special case of CPP, which is a type of arc routing problem, is examined. In the
considered problem, the assumption that the starting point in the network is a certain node has been
removed and this node has been determined by the model in the most appropriate way. For this purpose,
a mixed integer model is proposed and the savings that can be achieved with a small sample are
calculated. The most suitable starting and ending point on the network has been determined with the
proposed model. In addition, tour lengths and routes are presented if other points in the network are the
starting points. The results show that the correct determination of the start and end points in the given
network directly affects the tour length. These changes to the basic model of the CPP can also be applied
to other variants of the CPP, the multi-vehicle, capacity, windy and hierarchical CPP, and shorter lap
lengths can be found.
References
1. Eiselt, H.A., Gendreau, M., Laporte G.: Arc routing problems, part 1: the chinese postman problem.
Operations Research 43(2), 231–242 (1995).
2. Çomaklı Sökmen, Ö., Emeç, Ş., Yılmaz, M., Akkaya, G.: An Overview of Chinese Postman
Problem. 3rd International Conference on Advanced Engineering Technology (2019).
3. Yılmaz, M., Kayacı Çodur, M., Yılmaz, H.: Chinese postman problem approach for a large-scale
conventional rail network in Turkey. Technical Gazette (25) 1471-1477 (2017).
4. Godinho Filho, M., Ribeiro Junqueira R.D.A.: Chinese Postman Problem (CPP): solution methods
and computational time. International Journal of Logistics Systems and Management, (7), 324-344,
(2010).
5. Bostel, N., Castagliola, P., Langevin, A.: Approximating the length of Chinese postman tours. 4OR
(12) 359-372, (2014).
6. Ding, H., Li, J., Lih, K.W.: Approximation algorithms for solving the constrained arc routing
problem in mixed graphs. European Journal of Operational Research (239) 80-88, (2014).
7. Granot, D., Hamers, H.: On the equivalence between some local and global Chinese postman and
traveling salesman graphs. Discrete Applied Mathematics (134) 67–76, (2004).
8. Minieka, E.: Optimization algorithms for networks and graphs. Marcel Dekker Newyork (1978).
9. Keskin, M.E., Yılmaz, M.: Chinese and windy postman problem with variable service costs. Soft
Computing 23(16) 7359-7373 (2018).
10. Keskin, M.E., Yılmaz, M., Triki, C.: Solving the hierarchical windy postman problem with variable
service costs using a math-heuristic algorithm. Soft Computing 27(13) 8789-8805 (2023).
11. Çomaklı Sökmen, Ö., Yılmaz, M.: The new approaches for solving hierarchical Chinese postman
problem with stochastic travel times. Journal of Intelligent & Fuzzy Systems 44(5) 8471-8492
(2023).
12. Çomaklı Sökmen, Ö., Yılmaz, M.: Hierarchical Chinese postman problem with fuzzy travel times.
Iranian Journal of Fuzzy Systems 18 (), 87-105 (2021).
13. Yılmaz, M., Kayacı Çodur, M.: A time-dependent hierarchical Chinese postman problem. Central
European Journal of Operations Research 28(1) 337-366 (2020).
Page 261
SEMIT 2023
SP04: Data Science, Big Data, Data Mining, and Knowledge Management in EM
Comparison of Cost and Benefit Analysis of Active Leakage Control Applications in Water
Distribution Systems
Abdullah Ateş1 Salih Yilmaz2 Mahmut Firat3
1İ
nönü University Computer Engineering Department, Malatya, Türkiye
2
Çankırı Karatekin University, Civil Engineering Department, Çankırı, Türkiye
3
İnönü University, Civil Engineering Department, Malatya, Türkiye
Abstract
The fight against leaks and failures in water distribution systems has gained a serious importance with
the increasing drought, the inefficient use of water resources and the rapid contamination of existing
water resources. These studies are generally carried out with two different approaches, active leakage
control and passive leakage control. Although the passive control method, which covers the repair of
the reported faults, is less costly and easier, the vast majority of the faults that occur in the system are
caused by unreported faults. It is possible to detect, repair and control such faults with active leakage
methods. In order to gain the expected efficiency from these methods, it is very important to regularly
monitor the networks by dividing them into isolated areas, to estimate the amount of preventable
leakage, and to determine the leak location with devices and equipment such as ground microphone,
regional recorder, regional correlator within the framework of a certain systematic program. The regular
maintenance of these systematic studies in the implementation of the active leakage method also creates
serious costs for water administrations. In addition to the costs of creating and monitoring measurable
regions, the costs of devices and equipment with different characteristics, sensitivities and costs to be
used for the determination of the fault location should also be considered in the studies to be carried
out. In this study, a standard calculation structure is defined for the benefits and costs of the ground
sample network. The method to be selected according to the general characteristics of the networks
plays an important role for the most efficient operation.
Keywords: Water distribution system, leakage, active leakage control, leakage detection
1. Introduction
Water distribution systems are one of the most important urban infrastructure systems. Ensuring normal
operating conditions in water distribution systems is quite important for water supply safety. However,
failures, breaks and leaks occur in water distribution systems depending on various factors that are the
physical properties of network (pipe length, pipe material, pipe diameter etc.), environmental (traffic
loads, groundwater variation etc.) and hydraulic parameter (high pressure, pressure change) (Berardi
and Giustolisi, 2021). A certain portion of these leakages rise to the surface and constitute the reported
leakages. On the other hand, a significant portion does not rise to the surface and constitutes unreported
leakages (Lambert et al., 1999; Yilmaz et al., 2021). Annual leakage rates in distribution systems are
between 25 and 30% (European Commission, 2014), and the NRW rate is approximately 30%
(Liemberger and Wyatt, 2019). A significant part of water resources is lost due to leaks in water
distribution networks. In developed countries, the volume of these leaks is seen to be between 3 and 7%
Page 262
SEMIT 2023
of the water provided, while in developing countries there is more than 50% leakage (Moslehi et al.,
2021). In our country, according to the data of 25 administrations in the report published by the Turkish
Water Institute (SUEN); The average GGS rate is 42%, and the lowest and highest values are 22% and
67%, respectively (SUEN, 2020).
In water distribution systems, an active leak control strategy, which includes the activities of
recognizing, locating and repairing unreported leaks, which constitute a significant portion of leaks,
plays an important role in reducing the volume of physical losses (Lambert et al., 1999; Lambert vd.,
2005; Firat et al., 2021). Active leak control includes minimum night flow in district metered areas,
performing night flow analysis, identifying potential recoverable leaks, and locating unreported leaks
with acoustic methods. However, in order to obtain the expected benefits of the active leak control
approach, creating district metered areas and defining boundaries is quite important step. The volume
of recoverable leaks increases as the time it takes to detect and locate unreported leaks increases (Haider
vd., 2019; Moslehi vd., 2019; Ahopelto ve Vahala 2020; Choi vd., 2020; Moon, 2020). Moslehi et al.
(2021) emphasized that leaks constitute the most important component of water losses. It was stated
that in many administrations, high-cost methods are applied to minimize leaks, but economic criteria
should be taken into account in leak management in order to ensure efficiency in all aspects of the
system.
Active leak control covers the detection, reduction and prevention of non-surface (unreported) leaks in
district metered areas. For active leakage control, it is necessary to district metered areas in the water
distribution system, perform minimum night flow analysis, monitor flow-pressure changes, detect
detected leakage locations and carry out leak repairs (Aboelnga et al., 2018). In addition, the flow-
pressure change in the isolated measurement area should be constantly monitored and new leaks should
be intervened immediately. Active leak control is applied to locate and prevent potential preventable
leaks determined by minimum night flow analysis. This method has no effect on reducing and
preventing uncertain leaks. In order to reduce the volume of preventable leakage detected by monitoring
the minimum night flow rate in a district metered areas, studies are carried out to detect the leak location
with devices and equipment such as ground microphone, regional recorder, regional correlator. It has
been emphasized in the literature that the aim is to detect and repair pipe bursts faster as a result of
active leak control studies (Berardi et al., 2015). In this way, it has been argued that the benefits of
reducing both possible damage to the environment and water loss will be achieved. Leak detection team
refers to the technical team formed to locate non-surface faults detected by minimum night flow analysis
in distribution systems. This team is of critical importance in combating leaks. Detecting the leak
location late or not detecting it causes the leak volume to increase. The field experience of the leak
detection team is very important in accurately detecting the leak location (Ye and Fenner, 2010; He et
al., 2009).
Leak detection devices refer to acoustic devices that are based on sound waves and their propagation
and change along the pipeline-connectors in determining the exact location of the leak. These devices
can basically be given as ground microphone (local listening), regional correlator, regional recorders,
hydraulic model-based monitoring systems. Local leak detection involves monitoring the street-based
step-by-step change of sound waves by using acoustic devices, referred to as ground microphones, to
locate leaks that do not surface in the distribution system. Pipe material type, fault location, fault
diameter, soil cover thickness on the pipe, experience of the personnel using the device and inspection
time are effective in detecting this type of leak. In order to perform leakage inspection more accurately,
night hours when ambient noise is the least should be preferred. Regional leak detection aims to
determine the location of potential leak points with the help of sensors placed at valve points in the
region in order to prevent potential preventable leaks detected by MNF analysis in isolated measurement
areas. In this method, potential streets that may leak are determined based on the frequency changes
received from sensors in the region. Then, local detection work is carried out on the street with leakage
potential using local leak detection equipment.
The regular maintenance of these systematic studies in the implementation of the active leakage method
also creates serious costs for water administrations. In addition to the costs of creating and monitoring
measurable regions, the costs of devices and equipment with different characteristics, sensitivities and
costs to be used for the determination of the fault location should also be considered in the studies to be
Page 263
SEMIT 2023
carried out. In this study, a standard calculation structure is defined for the benefits and costs of the
ground microphone and regional correlator methods used to locate unreported leaks. By using the
defined calculation methodology, the benefits and costs for both methods are analyzed and compared
in a sample network. The method to be selected according to the general characteristics of the networks
2. Material and Method
2.1 Non-Revenue Water and Components

Non-revenue water refers to water that is supplied to the distribution system but cannot be charged for.
Non-revenue water basically consists of three main components: apparent losses, real losses and
unbilled authorized uses. This component should be minimal to ensure operational efficiency in the
system. In its simplest form, the non-revenue water ratio is obtained as the ratio of the non-revenue
water volume to the inlet volume. According to the water balance table, the non-revenue water volume
is calculated as the difference between the system inlet volume and the revenue-generating water
volume. Real losses occur as a result of malfunctions (reported/unreported) in water distribution
systems due to various factors or overflows and cracks in water tanks. This component is lost as leakage
before the water supplied to the system reaches the subscribers and constitutes a significant part of the
volumetric water losses. One of the most basic tools used to determine the amount and rate of water
loss at a certain standard and to monitor its change is the "standard water balance" recommended by
International Water Association (IWA) and American Water Works Association (AWWA) (Table 1).
In the next stage of this method, Administrative and Physical losses and their sub-components are
calculated or estimated based on legal consumptions and input volumes. The standard water balance
table, filled in using real data to represent the system, provides the opportunity to access information
about the subcomponents of water losses as well as the non-revenue water rate.
Table 1. IWA Standard water balance [1-2]

(4) Billed (2) Billed metered consumption (5) Non-
(10) authorized revenue
(3) Billed unmetered consumption
Authorized consumption water
consumption (9) Unbilled (7) Unbilled metered consumption
authorized (8) Unbilled unmetered
consumption consumption
(1) System (15) (12) Illegal consumption
Input Apparent (13) Losses due to meter
(6) Revenue
Volume losses inaccuracies
water
(11) Water (14) Losses due to reading errors
losses (17) Leakages in transmission and
(16) Real distribution systems
losses (18) Leakages in reservoirs
(19) Leakages in service
connections
In order to monitor the water loss performance of administrations and municipalities in Turkey and to
carry out water loss management studies effectively, systematically and at a certain standard, the
"Control of Water Loss in Drinking Water Supply and Distribution Systems" regulation was issued by
the Ministry of Agriculture and Forestry in 2014.The most basic methods applied in the management
of leaks in the international literature are suggested as active leak control, pressure management, fault
repair speed and quality and pipe material management (Figure 1).
Page 264
SEMIT 2023
Pressure
management
UARL
Repair Active
speed and leakage
quality control
Recoverable leakage
Pipe material
management
Figure 1. Basic reduction components of real losses (Lambert,1999)
In this figure, the unavoidable annual real loss (UARL) volume is the lowest value of leakages as
technical in a water distribution system. This volume is not the economical level. Therefore, the
economical leakage level should be defined based on the network characteristics and water utility
properties. Pipe material management is defined as the most difficult and costly process in reducing,
managing and controlling leaks (AL-Washali et al., 2017; Yilmaz et al., 2022). Therefore, less costly
methods such as active leak control, fault repair speed and quality, and pressure management should be
applied primarily in the distribution system. Two different intervention methods such as passive leakage
control and active leakage control are applied to control and manage malfunctions and leaks occurring
in the water distribution system. Active leakage control covers the detection, reduction and prevention
of non-surface (unreported) leaks in district metered areas. For active leakage control, it is necessary to
create isolated measurement zones in the water distribution system, perform MNF analysis, monitor
flow-pressure changes, detect detected leakage locations and carry out leak repairs. This method has no
effect on reducing and preventing uncertain leaks. Passive leak control includes the process of repairing
and managing surface (reported) faults in distribution systems. This method only covers the
management and repair of faults that surface. Leak location detection refers to the determination of the
location of potential preventable leaks in the network (in which pipe on a street basis) detected by MNF
analysis in district metered areas. Local, regional and instant monitoring systems are used to detect the
leak location. Detecting the leak location is very important to reduce the leak volume and ensure water
efficiency. Local leak detection involves monitoring the street-based step-by-step change of sound
waves by using acoustic devices, referred to as ground microphones, to locate leaks that do not surface
in the distribution system. Pipe material type, fault location, fault diameter, soil cover thickness on the
pipe, experience of the personnel using the device and inspection time are effective in detecting this
type of leak. In order to perform leakage inspection more accurately, night hours when ambient noise
is the least should be preferred.
2.2 Leak Detection with Ground Microphone and Cost-Benefit Analysis

Ground microphones refer to local detection devices used to listen from the ground on the pipeline
using the acoustic method. This device similarly detects leaks by monitoring acoustic signals. In this
context, it should be protected from external sounds such as traffic and wind, and listening should be
done step by step at 1- or 2-meters intervals on the pipe. Listening to and locating faults with a ground
microphone is one of the most frequently used methods in active leak management. In this method,
when the location of the existing network is known, an operator tries to detect the sound of the leaks by
listening sensitively on the route of the network. The unreported leaks can be detected and leakage
losses can be prevented with this equipment (Lambert et al., 1999; Farley et al., 2008) (Figure 2).
Page 265
SEMIT 2023
Figure 2. Sample Fault Detection Process with ground microphone (KASKİ, 2022)
In the cost-benefit analysis of this method, after obtaining the data on the total line length and subscriber
numbers of the network to be studied, the total water loss in the network must be calculated in l/km/hour
(Yilmaz et al., 2023). The current average network pressure of the system and unit water costs should
also be defined as variables. The benefits and possible costs that can be obtained were calculated using
the following data, while making the cost benefit analysis of this method.
Table 2. Data used in analysis
No Parameters Unit
1 Total length of main line m
2 Average length of service connetions m
Turkish Liras
3 Unit cost of water produced
(TL) /m3
4 Total number of failures in a year No.
5 The volume of real losses m3/month
6 The volume of leakages in water tanks m3/month
The rate of unreported failures to total
7 %
failures in water distribution netwokr
8 System operation pressure m
2.3 Leak Detection with Regional Correlator and Cost-Benefit Analysis
Acoustic (zonal) recorders can be considered as a compact unit containing acoustic sensors and
programmable data loggers, which can be mounted on a valve-fire hydrant or metal pipe with a magnetic
structure. Its size and features enable it to be easily placed and used at points suitable for monitoring
sound frequencies in the distribution system. These devices are generally placed at suitable points to
listen to the sound of leaks between 02:00 and 04:00 at night, which is the most suitable time for
monitoring sound changes. After the recorders are correctly programmed and placed, the propagation
of sound waves along the pipeline is recorded and the collected data is analyzed to determine the
potential leak line.
Acoustic (regional) correlators are shown to be one of the most effective tools used in locating leaks.
In the application of this tool, a mathematical-based relationship is established between the frequency
curves of two signals coming from the same leak (potential leak), and the location of the leak is
Page 266
SEMIT 2023
determined. Therefore, instead of looking for the highest noise point, the potential leak point is
determined by applying cross-correlation based on the frequency curves between two sensors.
In studies conducted in the literature, this method/tool detects the leak location with high accuracy
(generally with an error of less than 1 m). When planning the placement of recorders, two basic elements
must be taken into account as they affect the propagation of sound waves. These factors are the type of
pipe material to be listened to, the operating pressure present in the system. Another method most
frequently used in the active leakage method is to detect faults with a regional correlator. In this method,
a constant sound and vibration is listened to with the help of correlators placed at certain distances, and
accordingly, the location of the leak is aimed to be determined.
Sensor I
Sensor II
Main line
d d L-2d
Leakage
Figure 3. Regional Correlator working structure
The acoustic leak sound correlation method aims to compare the leak sound determined by sensors
(monitoring the exit sound frequency of water coming out of the fault crack point) placed at two valve
points (in a pipeline) in order to detect leaks that do not come to the surface in the isolated measurement
area and to determine the potential leak point accordingly. In this method, it is assumed that the leak
sound velocity in the pipe is constant and the leak sound moves from the leak note in both sensor
directions. If the crack (leak) in the pipe is at an equal distance from both sensors, the sensors detect the
leak simultaneously. If the crack (leak) in the pipe is not at an equal distance from both sensors, the
sensors detect the leak at different times, in which case the difference between the time the leak sound
reaches the sensors is measured by the correlation process (Pearson, 2019).
Sensors are placed at two valve points (points I and II) as shown in the figure. The distance of the leak
point that occurs at the bottom of the pipe between the two sensors to sensor no. I is (d), and the distance
from the leak point to sensor no. II is marked (d). In this example, the distance (L-2d) between the point
(d) away from the leak point and sensor no. II causes the delay time (t) for the leak sound speed to reach
sensor no. II. This distance (L-2d), which causes the delay time, can be written as V*t.
Here, V; The speed of sound created by the leakage coming out of the crack, L; horizontal distance (m)
between two sensors.𝐿 = (2 ∗ 𝑑) + (𝑉 ∗ 𝑡)
𝐿 = (2 ∗ 𝑑) + (𝐿 − 2𝑑)
In this case, the distance of the leak point to the sensor can be written as follows;
Page 267
SEMIT 2023
𝐿 − (𝑉 ∗ 𝑡)
𝑑=
2
In this method, the correlation process determines the delay time (t) of the leak sound between two
sensor points. In this method, the distance (L) between the two sensors must be measured accurately.
The speed of the sound of the leak caused by the leak coming out of a crack varies depending on the
pipe material, pipe diameter, pipe wall thickness and the characteristics of the ground around the pipe.
In general, theoretical values for the speed of sound are used in many cases and are acceptable for the
initial assumption of the leak location. However, since the sound speed varies depending on these
factors, if there are previous repairs on the pipe (different pipe material, etc.), then the assumption made
for the sound speed may change. In this method, the performance of the correlators depends on the
system operating pressure and the level of acoustic noise in the network. While making the cost benefit
analysis of this method, the benefits and possible costs that can be obtained were calculated using the
following data.
Table 3. Data Used in Analysis
No Parameters Unit Value

1 Total Network Length m
2 Unit Water Cost TL/m3
3 Existing Weighted Pipe Type of the -
Network
4 Network Failure Repair Cost TL/piece ₺3.200,00
5 Correlator Cost TL/piece ₺5.500,00
6 Correlator System and Transmitter Cost TL/piece ₺120.000,00
7 Estimated Number of Network Failures pieces/month Calculating
8 Calculation of Interceptable Flow m3/month Calculating
9 Calculation of Benefit Flow Rate to be m3/month Calculating
Obtained with Microphone
10 Number of Correlators Required piece Calculating
Network main line length refers to the length of the pipeline that passes through the street or street in
the distribution system and is under the responsibility of the water administration. Unit water production
cost refers to the production cost of water obtained from the sources feeding the distribution system. In
order to calculate this cost, first of all, all resources feeding the system must be measured continuously
and accurately. Energy expenses should be monitored regularly in the promotion system for water
production costs. Unit fault repair cost refers to the cost incurred by repairing one reported or unreported
fault that occurs in the distribution system. This cost especially has a direct impact on operating cost
and efficiency. This cost basically includes components such as excavation, pipe material, labor, filling,
road covering, and the cost of unsold water. In systems with a high number of malfunctions, fault repair
costs increase linearly, causing the water sales price to reach high levels. Correlator System and
Transmitter Cost includes the initial investment cost of the regional correlator used for leak detection
in the distribution system.
Page 268
SEMIT 2023
Table 4. Data Used in Analysis (comt.)
No Data Name Unit Value

11 Determination of Method Performance % Calculating
12 Calculation of Interceptable Flow m3/month Calculating
13 False Detection Cost TL Calculating
14 Detected Fault Repair Cost TL Calculating
15 Calculation of the Benefit Flow Rate to m3/month Calculating
be Obtained with Correlator
16 Correlator Total Cost TL/year Calculating
17 Economic Gain to be Obtained with TL/year Calculating
Correlator
3. Results
In order to compare the costs and benefits of fixed correlator and ground microphone methods,
calculations were made for a network whose values are given below.
Table 5. Network Data used in analysis
Parameters Unit Value
Total Network Length km 15
Total Number of Customers No. 1000
Total Number of Commercial No.
250
Subscribers
Total Number of Housing No.
750
Subscribers
Average Night Pressure atm 55
System Inlet Flow m³/month 125000
Amount of Water Accrued m³/month 100000
Average Unit Water Cost TL/m³ 6.5
Average Unit Water Sales Price TL/m³ 8
Annual Breakdown Amount No. 200
As a result of the analyses, different amounts of water were recovered depending on the method chosen
in the ground microphone and correlator studies. Additionally, differences were observed in the
application costs of the methods. The benefit to be obtained is calculated in direct proportion to the
water sales price.
Table 6. Results of cost benefit analysis
Benefit
Cost Benefit +/-
Method Flow
(TL) (TL) (TL)
(m³/month)
ALC
Ground 3,075 146,420 184,530 38,110
Microphone
ALC
6,430 270,320 385,930 115,610
Correlator
Page 269
SEMIT 2023
According to the results, the method is seen as a factor that seriously affects the benefit to be obtained
in ground microphone and correlator studies. The benefits to be obtained from the methods also vary
depending on the current characteristics of the region and the different values of the network and
operating conditions. Therefore, before developing water loss reduction strategies, the current
conditions of the region must be analyzed and appropriate methods must be selected.
4. Conclusions
In this study, a standard calculation structure is defined for the benefits and costs of the ground
sample network. The average non-revenue water rate in Turkey is 42%, and the lowest and highest
values are 22% and 67%, respectively. These data show that non-revenue water rates in Turkey are very
high and urgent measures must be taken to reduce these rates. It can be concluded that the total annual
economic loss due to water loss and leakage in our metropolitan cities exceeds 7 billion TL at 2017
prices. This economic loss is undoubtedly; This is a valid reason in itself to combat water loss and
leakage. Technological tools are needed for each method in combating water loss. For pressure
management, pressure breaking valves to be placed in various parts of the network or new drinking
water tanks to be built, and state-of-the-art equipment such as audio listening devices and magnetic
listening devices are required to perform active leakage control. Long-term strategies should be put
forward by water utilities to prevent losses and keep the system under control. This strategy should be
based on the following studies;
• Continuous and active monitoring of the system
• Constantly combating leaks and administrative loss of components
• Implementation of active leak control approach
• Implementation of subscriber, meter management, pressure management, pipe material
management, fault management applications
Although the approach based on active leakage control seems costly, it provides a long-term solution
in terms of resource efficiency, subscriber satisfaction and network management. Obtained results; It
has been concluded that the method to be chosen according to the general characteristics of the networks
Acknowledgment
This research was supported by TUBITAK (Turkish National Science Foundation) under the Project
Number 122M577. The authors thank TUBITAK for their support.
References
Ahopelto, S. ve Vahala, R. (2020). Cost-Benefit Analysis of Leakage Reduction Methods in Water
Supply Networks. Water, 12 (1), 195.
Aboelnga, A., Saidan, M., Al-Weshah, R., Sturm, M., Ribbe, L., Frechen, F-B. (2018). Component
analysis for optimal leakage management in Madaba, Jordan. Journal of Water Supply: Research and
Technology-Aqua 1 June 2018; 67 (4): 384–396. doi: https://doi.org/10.2166/aqua.2018.180.
AL-Washali, T., Sharma, S., Lupoja, R., AL-Nozaily, F., Haidera, M., ve Kennedy, M. (2020).
Assessment of water losses in distribution networks: Methods, applications, uncertainties, and
implications in intermittent supply. Resources, Conservation and Recycling, 152 (September 2019),
104515.
Berardi, L., Laucelli, D., Ugarelli, R., ve Giustolisi, O. (2015). Leakage management: Planning remote
real time controlled pressure reduction in Oppegård municipality. Procedia Engineering, 119 (1), 72–
81.
Page 270
SEMIT 2023
Berardi, L., and O. Giustolisi. 2021. ‘Calibration of Design Models for Leakage Management of Water
Distribution Networks’. Water Resources Management 35(8): 2537–51.
Choi, G.W., Geun Jo, H., Park, H., ve Jang, D. (2020). Application of decision making model for
leakage reduction to economic project in water distribution systems. Journal of Ambient Intelligence
and Humanized Computing.
European Commission (2014) LEAKCURE project, focus on reducing urban water leakage.
http://cordis.europa. eu/news/rcn/36134_en.html. Access 3 May 2021
Fırat, M., Yılmaz, S., Ateş, A., Özdemir, Ö. (2021). Determination of Economic Leakage Level with
Optimization Algorithm in Water Distribution Systems. Water Economics and Policy. 7(3):1-38.
2150014. DOI:10.1142/S2382624X21500144.
Haider, H., Al-Salamah, I.S., Ghazaw, Y.M., Abdel-Maguid, R.H., Shafiquzzaman, M., ve Ghumman,
A.R. (2019). Framework to establish economic level of leakage for intermittent water supplies in arid
environments. Journal of Water Resources Planning and Management, 145 (2), 1–12.
He, S., Wu, Q.H., ve Saunders, J.R. (2009). Group search optimizer: An optimization algorithm inspired
by animal searching behavior. IEEE Transactions on Evolutionary Computation, 13 (5), 973–990.
Lambert, A. O., Brown, T. G., Takizawa, M., and Weimer, D. (1999). “A review of performance
indicators for real losses from water supply systems.” Journal of Water Supply: Research and
Technology - AQUA.
Lambert, A.O. ve Fantozzi, M. (2005). Recent advances in calculating economic intervention frequency
for active leakage control, and implications for calculation of economic leakage levels. Water Science
and Technology: Water Supply, 5 (6), 263–271.
Liemberger R, Wyatt A (2019) Quantifying the global non-revenue water problem. Water Supply
19(3):831– 837. https://doi.org/10.2166/ws.2018.129.
Moon, H.K. (2020). A Study On İmproving The Approaches To Set The Economic Level Of Water
Losses İn Water Distribution System.
Moslehi, I., Jalili Ghazizadeh, M.R., ve Yousefie Khoshghalb, E. (2019). Economic analysis of pressure
management in water distribution networks. Journal of Water and Wastewater, 31 (2), 100–117.
Moslehi, I., Jalili-Ghazizadeh, M. Yousefi-Khoshqalb, E. (2021). ‘Developing a Framework for
Leakage Target Setting in Water Distribution Networks from an Economic Perspective’. Structure and
Infrastructure Engineering 17(6): 821–37. https://doi.org/10.1080/15732479.2020.1777568.
Pearson, D. ve Trow, S.W. (2005). Calculating the Economic Levels of Leakage. Leakage 2005
Conference Proceedings, 1–16.
Ye, G. ve Fenner, R.A. (2011). Kalman Filtering of Hydraulic Measurements for Burst Detection in
Water Distribution Systems. Journal of Pipeline Systems Engineering and Practice, 2 (1), 14–22.
Yılmaz, S., Fırat, M., Ateş, A., Özdemir, Ö. (2021). Analysis of Economic Leakage Level and
Infrastructure Leakage Index Indicator by Applying Active Leakage Control. ASCE's Journal of
Pipeline Systems - Engineering and Practice. 10.1061/(ASCE)PS.1949-1204.0000583
Yılmaz, S., Fırat, M., Ateş, A., Özdemir, Ö. (2022). Analyzing the economic water loss level with a
discrete stochastic optimization algorithm by considering budget constraints. AQUA - Water
Infrastructure, Ecosystems and Society 1 July 2022; 71 (7): 835–848.
doi: https://doi.org/10.2166/aqua.2022.060
Yılmaz, S., Ateş, A., Firat, M., Özdemir, Ö., Cinal, H. (2023). Determination of economic loss levels
in water distribution systems with different network conditions by a district stochastic optimization
algorithm. Water Supply. 23(3): 1349-1361.
Page 271
SEMIT 2023
Benefit and Cost Analysis of Network and Service Connection Renewal in Distribution Systems
for Sustainable Water Loss Management
Salij Yilmaz1 Abdullah Ateş2 Mahmut Firat3
1
Çankırı Karatekin University, Civil Engineering Department, Çankırı, Türkiye
2
İnönü University Computer Engineering Department, Malatya, Türkiye
3
İnönü University, Civil Engineering Department, Malatya, Türkiye
Abstract
Reducing the rate of non-revenue water (NRW) in water distribution systems (WDSs) has become one
of the most important goals for utilities. Increasing water losses bring along social, economic and
technical difficulties, which accelerates the work to be done to reduce water losses. Although water
losses are combated by applying active and passive leakage methods, in some cases it may not be
economical for the utilities to manage the network under current conditions and to continue the efforts
to reduce water losses. The aging network and its equipment can now make it difficult to manage this
system. In these cases, the method of rehabilitating the whole or partial parts of the network is applied.
It is possible to transfer and monitor the network and its components to the GIS database, to reduce the
risk of failure and to have low operating costs, to prevent and reduce leaks in subscriber connections by
manufacturing service connections to the meter, to keep the NRW rate at a very low level, and to
increase subscriber satisfaction and service quality with the application of this method. In a distribution
system, in case the existing network conditions are preserved, current failure rates, network operation,
maintenance and repair costs, new water source search and energy costs, initial investment and medium
and long-term operating costs in case of network renewal should be taken into consideration and benefit-
cost analyzes should be made. Due to the high level of pipe material, labor and construction costs, a
detailed cost-benefit analysis and alternative solutions should be evaluated before the network renewal
approach is preferred. In this study, the current costs and benefits that will be encountered in case of
renewal of the entire network (network + service connection), only subscriber lines or only network
lines for a sample network are discussed. The results show that it is that the rehabilitation method to be
chosen for different network conditions has a serious effect on increasing the planned benefit.
Keywords: Water distribution network, network renewal, cost analysis, benefit analysis.
1. Introduction
In distribution systems (WDS), leaks are observed at different rates depending on physical, operational,
environmental and hydraulic factors (Yilmaz et al., 2022). Depending on the malfunction and water
loss rate, the operating conditions of the system deteriorate and serious problems arise for subscribers
and the Administration. Annual leakage rates in distribution systems are between 25 and 30%
(European Commission, 2014), and the NRW rate is approximately 30% (Liemberger and Wyatt, 2019).
A significant part of water resources is lost due to leaks in water distribution networks. In developed
countries, the volume of these leaks is seen to be between 3 and 7% of the water provided, while in
developing countries there is more than 50% leakage (Moslehi et al., 2021). In our country, according
to the data of 25 administrations in the report published by the Turkish Water Institute (SUEN); The
average GGS rate is 42%, and the lowest and highest values are 22% and 67%, respectively (SUEN,
2020). Preventing and managing physical losses provides significant contributions in terms of providing
less water to the system, postponing the search for new resources, reducing energy costs, efficient use
of energy and water resources, and reducing malfunction maintenance and repair activities and costs.
From these studies in the literature, it can be seen that many different methods are generally applied to
monitor, prevent and control physical losses. Many of these methods create significant costs in terms
of equipment and labor, and require technical, technological and personnel infrastructure to be
implemented in the field (Lambert et al., 1999; Haider et al., 2019; Salehi et al., 2017; Firat et al., 2021).
Dighade et al. (2014) made an evaluation within the framework of the problems faced by developing
countries in leakage management, intermittent water delivery, poor infrastructure, water metering
policy, water delivery hours and system pressure. It was emphasized that, first of all, answers should be
Page 272
SEMIT 2023
sought to the questions of why, how and where water is lost, and that it is important to evaluate the
physical and operating characteristics of the existing network.
In reducing water losses in distribution systems, the "network renewal" method, which involves
replacing pipes and other fittings in the network and generally costs more than other active leak control
methods, is preferred in many cases. However, in the distribution system, failure rates under current
network conditions, operation, maintenance and repair costs, new resource search and energy costs
should be taken into consideration, and in case of network renewal, the initial investment and medium
and long-term operating costs should be taken into account, local replacement or network renewal.
options should be compared (Mamo et al., 2013; Al-Zahrani et al., 2016). For this reason, before
choosing the network renewal method, a detailed cost-benefit analysis should be made and alternative
solutions should be evaluated. For this purpose, other methods that are suitable for the network
conditions and applicable to the system must be preferred in the region, applied in the field, and the
impact of environmental and operational factors that cause leakage should be reduced. It is emphasized
that if the expected benefits cannot be obtained from these methods, determining priority regions in grid
renewal will make significant contributions in terms of resource efficiency (Park and Loganathan, 2002;
Suribabu and Neelakantan, 2012; Venkatesh, 2012; Mondaca et al., 2015; Francisque et al., 2017).
Pipe material management and network rehabilitation should be based on extending the economic life
of existing pipes and postponing network renewals as much as possible. For this reason, analyzing the
factors that cause pipe damage in the network, reducing their effects and identifying the areas where
damage occurs despite all prevention methods are important for efficiency. Tee et al. (2014) aimed to
operate the system under budget constraints, make maintenance and renewal plans, and determine the
remaining service life of the pipes. For this purpose, the most appropriate renewal solution was
determined by minimizing the risk of failure and the life cycle cost of the pipe with the help of genetic
algorithm. Loganathan et al. (2002) aimed to define an economically sustainable fault rate threshold
value for economical operation of the system and management of faults in network management. For
this purpose, maintenance, repair and renewal costs and the inflation rate were taken into account, and
according to cost data analysis, it was reported that there was a significant relationship between pipe
diameter and failure rate threshold value. In the study, a model was developed that estimates the
remaining useful life of pipes in water distribution systems as well as their degradation rates. As a result
of the study, calculations were made for different pipe diameters and the economic life of the pipes was
calculated.
Zangenehmadar (2016) aimed to develop a model that can estimate the remaining useful life in drinking
water networks in order to optimize the necessary intervention plans according to the available budget.
For this purpose, a statistical model has been developed to predict the current state of a pipeline. In the
research, a program for optimized renewal and maintenance plans according to the remaining lifespan
and condition of pipelines is proposed, which will help administrations for efficient budget allocation
and better management of the required intervention plans. Carriço et al. (2021) presents a multi-criteria
decision support methodology for the selection and prioritization of the region and network to be
rehabilitated in a water distribution system. In the study, it is suggested that before the network renewal
and rehabilitation works, the current status of the network should be taken into consideration and a
benefit/cost analysis should be carried out. Hu et al. (2021) emphasized that due to rapid urbanization,
drinking water networks are often created without planning. For this reason, a very serious planning
and project phase must be carried out before the drinking water distribution system is physically
implemented in the field. While the network is applied to the field, the network life is shorter and the
operating cost increases due to reasons such as inadequate engineering services, lack of control, and
incorrect manufacturing.
In this study, the current costs and benefits that will be encountered in case of renewal of the entire
network (network + service connection), only subscriber lines or only network lines for a sample
network are discussed. The results show that it is that the rehabilitation method to be chosen for different
network conditions has a serious effect on increasing the planned benefit.
Page 273
SEMIT 2023

2.1 Water Losses and Components
Water losses refer to the loss given to the distribution system but consist of apparent and real losses.
While some of the water supplied to the system is consumed by legal users, the remaining part
constitutes water losses. Water losses cause loss of income for the administration and constitute the
most important component of water resource inefficiency. In our country, water loss volume is taken as
basis in evaluating the performance of administrations. Water loss volume is obtained by subtracting
the legal consumption volume from the system inlet volume. Apparent losses refer to the water supplied
to the system and consumed by legally registered subscribers and/or unregistered users, but for which
no fee is collected. Real losses express the volume (m3/month, m3/year) of water leaking into the
ground as a result of malfunctions (reported/unreported) occurring due to various factors in drinking
water transmission/distribution systems or overflows and cracks in tanks over a certain period of time.
It does. This component is lost as leakage before the water supplied to the system reaches the subscribers
and constitutes a significant part of the volumetric water losses.
In order to monitor the water loss performance of administrations and municipalities in our country and
to carry out water loss management studies effectively, systematically and at a certain standard, the
"Control of Water Loss in Drinking Water Supply and Distribution Systems" regulation, issued by the
Ministry of Agriculture and Forestry in 2014, was published. The regulation requires administrations
and municipalities to submit their annual water budgets to the Ministry according to the IWA standard
water balance table. This regulation aims to use water resources more effectively and efficiently, reduce
water losses and monitor the performance of the Water Administration. In this regulation, it is
emphasized that "Based on the date published by the regulation, Water Administrations must reduce
water loss rates to 30% within 5 years and to 25% within the following 4 years." One of the most basic
tools used to determine the amount and rate of water loss at a certain standard and to monitor its change
is the "standard water balance" recommended by International Water Association (IWA) and American
Water Works Association (AWWA) (Table 1).
In the next stage of this method, Administrative and Physical losses and their sub-components are
calculated or estimated based on legal consumptions and input volumes. The standard water balance
table, filled in using real data to represent the system, provides the opportunity to access information
about the subcomponents of water losses as well as the non-revenue water rate.
Table 1. IWA Standard water balance [1-2]

(4) Billed (2) Billed metered consumption (5) Non-
(10) authorized revenue
(3) Billed unmetered consumption
Authorized consumption water
consumption (9) Unbilled (7) Unbilled metered consumption
authorized (8) Unbilled unmetered
consumption consumption
(1) System (15) (12) Illegal consumption
Input Apparent (13) Losses due to meter
(6) Revenue
Volume losses inaccuracies
water
(11) Water (14) Losses due to reading errors
losses (17) Leakages in transmission and
(16) Real distribution systems
losses (18) Leakages in reservoirs
(19) Leakages in service
connections
In order to reduce and manage the leakage volume in the distribution system or isolated measurement
areas, basic methods such as pressure management, pipe material management, fault repair speed and
quality and active leakage control are applied. Fault repair speed and quality refers to the rapid and
Page 274
SEMIT 2023
high-quality repair of faults after locating reported or unreported faults in distribution systems. Fault
repair speed shortens fault repair time. Therefore, reducing the duration of water outages in the region
is important in terms of preventing disruption of traffic flow on the street where the malfunction occurs
and minimizing social impacts. On the other hand, fault repair quality prevents re-occurrence of faults
at the point of fault in the future.
Pipe material management refers to the processes applied and the strategy followed to minimize water
losses occurring due to pipe material quality in distribution systems. Pipe material management includes
choosing the most appropriate diameter and type of pipes according to the characteristics of the region
where the new production will be made, flow rate, ground characteristics and pressure conditions. It
also includes the evaluation of options such as street-based pipe replacement, local pipe replacement
during fault repair, or replacement of the entire region, depending on the intensity of faults in the
currently serving network.
2.2 Network Renewal
Network main line renewal refers to the replacement of network main lines serving on the street or
street in distribution systems. Since network renewal creates a high cost, it should be implemented as
the last solution whenever possible. Before this method is preferred, less costly methods such as active
leak control and pressure management should be applied. Valve renewal refers to the replacement of
valves that serve on streets or streets in distribution systems and that have reached the end of their
economic life or are constantly causing malfunctions. Since valve replacement creates costs, creating
an asset management strategy and implementing a preventive maintenance program can extend the
useful life of valves and reduce costs. Service connection renewal refers to the replacement of subscriber
connections serving between the network main line and the building in distribution systems.
Figure 1. Network renewal in the field
Since a significant portion of the malfunctions in distribution systems occur in service connections, the
quality of materials and workmanship is very important in the renewal of these lines. Additionally,
service connections on private property should be replaced as much as possible during the renovation
to ensure a complete renovation.
Page 275
SEMIT 2023
Figure 2. Service connection renewal in the field
Network renewal involves processes that involve replacing pipes in the network and generates high
costs, and unfortunately, it is preferred as a priority in many cases. It is possible to reduce losses by
evaluating and managing networks together with their current failure rates, operation, maintenance and
repair costs, new resource search and energy costs (Mamo et al., 2013; Al-Zahrani et al., 2016). The
most appropriate pipe material must be selected by taking into account the climate and environmental
conditions of the region, operating parameters and hydraulic criteria. Since network renewal works will
result in very high costs in water loss management, this method should be the last option. It is seen that
the concept of useful life for networks comes to the fore in the literature. This concept focuses on
situations where fault repair of networks, pressure management or controlling leaks by other basic
methods cannot be economically managed. One of the important methods followed in determining the
useful life is Loganathan et al. (2002). For this purpose, the failure coefficient (Brk) was calculated
using the annual inflation rate (R), annual malfunction/pipe repair cost (C) and the total rehabilitation
cost of the network (F). For this purpose, first the total rehabilitation cost of the network will be
calculated. For this, unit manufacturing costs will be calculated for each diameter and pipe type. For
this cost, the type of pipe to be used in the renewal and the average pipe diameter of the existing network
should be defined.
3. Cost Benefit Analysis Standard and Results

Within the scope of the study, the benefits and costs to be obtained in cases of renewal of only the main
network line, renewal of only subscriber lines and renewal of the entire network (service connection +
network main line) were calculated. During these analyses, it was thought that if the entire network was
rehabilitated, the loss level would decrease to the "Inevitable Loss" level in that region. If the service
connection and the network main line are changed individually, the benefit flow to be obtained when
the existing faults are eliminated is calculated. In this context, depending on the amount of leakage
detected, the total number of faults and the resulting faults were differentiated by network and service
connection type. Studies in the literature have calculated that the ratio of network faults to total faults
is approximately 38% (Nicolini et al., 2014; Aydoğdu and Fırat, 2015; Boztaş et al., 2019). In network
renewal costs, basic costs for different pipe types for different pipe diameters were taken into account
and the expected costs in case of rehabilitation were calculated.
In network renewal costs, basic costs for different pipe types for different pipe diameters were taken
into account and the expected costs in case of rehabilitation were calculated. The data used to obtain
the benefits and costs in case of network renewal are presented in Table 1, and the cost calculations
used are presented in Table 2.
Page 276
SEMIT 2023
Table 1. Main Data Used for Network Rehabilitation
# Parameters Unit Value

1 Total Network Length m 15000
2 Total Number of Subscriber Connections No. 300
3 Average Night Pressure m 55
4 Unit Water Cost TL/m3 5
Annual Breakdown Amount (Total Subscriber No.
5 80
Breakdowns)
Annual Breakdown Amount (Total Network No.
6 120
Breakdowns)
7 Average Fault Resolution Time hour/No. 20
8 Physical Loss Volume m3/month 18000
9 Volume of Leakage from Tanks m3/month 150
10 Percentage of Ø 150 mm Small Pipe Lengths % 80
11 Percentage of Pipe Lengths Between Ø 150 mm % 10
- 300 mm
- 500 mm
- 700 mm
14 Percentage of Ø 700 mm Large Pipe Lengths % 0
Type of Pipe to be made in case of Network
15 - HDPE
Renewal
Table 2. Network Rehabilitation Benefit Cost Data

Pipe Manufacturing Smaller than Ø150 mm in
BM13a TL/m ₺185,00
Network for PVC Pipe
Pipe Manufacturing Between Ø150 - Ø300 mm in
BM13b TL/m ₺580,00
BM13c TL/m ₺1.000,00
BM13d TL/m ₺1.600,00
Pipe Manufacturing Larger than Ø700mm in Network
BM13e TL/m ₺2.250,00
for PVC Pipe
Pipe Manufacturing Smaller than Ø150 mm in the
BM14a TL/m ₺420,00
Network for Ductile Pipe
Pipe Manufacturing Between Ø150 - Ø300 mm in the
BM14b TL/m ₺775,00
Pipe Manufacturing Between Ø300 - Ø500 mm in the
BM14c TL/m ₺1.400,00
Table 2. Network Rehabilitation Benefit Cost Data (cont.)

BM14d Pipe Manufacturing Between Ø500 - Ø700 mm ₺2.100,00
TL/m
in the Network for Ductile Pipe
BM14e Manufacturing of Pipes Greater than Ø700mm ₺2.950,00
TL/m
in the Network for Ductile Pipe
Page 277
SEMIT 2023
BM15a Pipe Manufacturing Smaller than Ø150 mm in ₺210,00

TL/m
Network for HDPE Pipe
BM15b Pipe Manufacturing Between Ø150 - Ø300 mm ₺610,00
TL/m
in Network for HDPE Pipe
Pipe Manufacturing Between Ø300 - Ø500 mm ₺1.100,00
BM15c TL/m
Pipe Manufacturing Between Ø500 - Ø700 mm ₺1.600,00
BM15d TL/m
Pipe Manufacturing Larger than Ø700mm in ₺2.500,00
BM15e TL/m
Network for HDPE Pipe
Manufacturing of Pipes Less than Ø150 mm in ₺360,00
BM16a TL/m
Network for Steel Pipe
Pipe Manufacturing Between Ø150 - Ø300 mm
BM16b TL/m ₺720,00
in Network for Steel Pipe
BM16c TL/m ₺1.250,00
BM16d TL/m ₺190,00
Manufacturing of Pipes Greater than Ø700mm
BM16e TL/m ₺2.750,00
BM17 Service connection Manufacturing TL/No. ₺8.000,00
As a result of the analysis, for the network selected as a sample; The benefits and costs obtained in case
of main line renewal, subscriber line renewal and renewal of the entire network are presented in Table
3.
Table 3. Benefit Cost Calculation Results
Network main line

Current Network main Service connection
Parameters + Service
Status line renewal renewal
connection renewal
Real loss volume (m3month) 18000 18000 18000 18000
Saved volume (m3month) 0 9105 2164 11269
Final Status (m3/month) 18000 8895 15836 6731
Cost (TL/year) ₺0,00 ₺870.000,00 ₺88.000,00 ₺958.000,00
Cost Reduction (TL/year) ₺0,00 ₺546.300,00 ₺129.840,00 ₺676.140,00
+ / - (TL/year) ₺0,00 -₺323.700,00 ₺41.840,00 -₺281.860,00
As a result of calculations, for the region where there is currently a monthly water loss of 18,000 cubic
meters;
• It is seen that 9105 cubic meters of water can be obtained if the entire network is changed.
However, the cost of this method is lower than the profit to be obtained. Therefore, it should not
be preferred.
• It was seen that 2164 cubic meters of water would be saved if only the service connection lines
were renewed. This method may be preferred because its cost is lower than its benefits.
• It is seen that if the entire water line is replaced, 11269 cubic meters of water will be saved.
However, the cost of this method is lower than its gain. Therefore, it should not be preferred.
Page 278
SEMIT 2023
As can be seen, many variables such as the current network status, fault status, network lengths and
number of subscribers are very important in the work to be done to reduce water losses. For this reason,
it has been observed that the rehabilitation method to be chosen for different network conditions in
water loss reduction studies has a serious impact on increasing the planned benefit.
4. Conclusions
In Turkey, water loss rates for administrations are determined by the "Regulation on Control of Water
Losses in Drinking Water Supply and Distribution Systems" (8 May 2014). In this context,
Administrations have become obliged to reduce the NRW rate to a maximum of 30% by 2023 and to a
maximum of 25% by 2028. When the data shared by the administrations is analyzed; It is seen that the
average water loss for 26 administrations is 40%. Considering that the data in question is valid for
provinces with water and sewerage administration, it is understood that the average may be much higher
in the remaining provinces. Combating water losses in drinking water distribution systems is of serious
importance. Considering that all networks have their own characteristics and the countries they are
connected to have different economic criteria, it will be seen that the economic leakage level is different
for each network. For this reason, Water and Canal Administrations and Municipalities need to combat
losses and leakages by taking into account the current status of the networks they manage in water
management. In order to determine the ESS in drinking water distribution systems, methods to prevent
water losses must first be well understood and analyzed. While every method to be used to reduce water
losses will incur costs, it will also provide various benefits depending on the current status and
characteristics of the network. For this reason, detailed analysis and evaluations should be made within
the framework of economic components for the basic methods applied in water loss management in
future studies. However, considering both the costs of water loss reduction methods and their
relationships with each other, the possible benefits to be obtained must be calculated before
implementation. Optimum values must be calculated, especially when using methods with common
variables such as pressure. In a distribution system, if the current network conditions are maintained,
current failure rates, network operation, maintenance and repair costs, new water source search and
energy costs, initial investment and medium and long-term operating costs in case of network renewal
should be taken into consideration and benefit cost analyzes should be made. Since pipe material, labor
and construction costs are at very high levels, a detailed cost-benefit analysis must be made and
alternative solutions should be evaluated before choosing the network renewal approach. For this, the
current characteristics of the network must first be analyzed. In particular, factors that cause damage to
the pipe should be evaluated according to field data.
Acknowledgment
This research was supported by TUBITAK (Turkish National Science Foundation) under the Project
Number 122M577. The authors thank TUBITAK for their support.
References
Al-Zahrani, M. Abo-Monasar, A. Sadiq, R. (2016). Risk-based prioritization of water main failure using
fuzzy synthetic evaluation technique. Journal of Water Supply: Research and Technology - AQUA, 65
(2), 145–161.
Aydogdu, M. ve Firat, M. (2015). Estimation of Failure Rate in Water Distribution Network Using
Fuzzy Clustering and LS-SVM Methods. Water Resources Management, 29 (5), 1575–1590.
Boztaş, F., Özdemir, Durmuşçelebi, F.M., ve Firat, M. (2019). Analyzing the effect of the unreported
leakages in service connections of water distribution networks on non-revenue water. International
Journal of Environmental Science and Technology, 16 (8), 4393–4406.
Page 279
SEMIT 2023
Carriço, N., Covas, D., Almeida, M.C. (2021). Multi-criteria decision analysis in urban water asset
management. Urban Water Journal, 18:7, 558-569,
Dighade, R. R., Kadu, M. S. ve Pande, A. M. (2014). Challenges in Water Loss Management of Water
Distribution Systems in Developing Countries. International Journal of Innovative Research in Science,
Engineering and Technology, 3(6), 13838–13846.
European Commission (2014) LEAKCURE project, focus on reducing urban water leakage.
http://cordis.europa. eu/news/rcn/36134_en.html. Access 3 May 2021
Francisque, A., Tesfamariam, S., Kabir, G., Haider, H., Reeder, A., ve Sadiq, R. (2017). Water mains
renewal planning framework for small to medium sized water utilities: a life cycle cost analysis
approach. Urban Water Journal, 14 (5), 493–501.
Fırat, M., Yılmaz, S., Ateş, A., Özdemir, Ö. (2021). Determination of Economic Leakage Level with
Optimization Algorithm in Water Distribution Systems. Water Economics and Policy. 7(3):1-38.
2150014. DOI:10.1142/S2382624X21500144.
Haider, H., Al-Salamah, I.S., Ghazaw, Y.M., Abdel-Maguid, R.H., Shafiquzzaman, M., ve Ghumman,
A.R. (2019). Framework to establish economic level of leakage for intermittent water supplies in arid
environments. Journal of Water Resources Planning and Management, 145 (2), 1–12.
Hu, X., Han, Y., Yu, B., Geng, Z., Fan, J. (2021). ‘’Novel leakage detection and water loss
management of urban water supply network using multiscale neural networks’’. Journal of
Cleaner Production 278 (2021) 123611 https://doi.org/10.1016/j.jclepro.2020.123611
Lambert, A. O., Brown, T. G., Takizawa, M., and Weimer, D. (1999). “A review of performance
indicators for real losses from water supply systems.” Journal of Water Supply: Research and
Technology - AQUA.
Liemberger R, Wyatt A (2019) Quantifying the global non-revenue water problem. Water Supply
19(3):831– 837. https://doi.org/10.2166/ws.2018.129.
Loganathan, G. V., Park, S., ve Sherali, H.D. (2002). Threshold break rate for pipeline replacement in
water distribution systems. Journal of Water Resources Planning and Management, 128 (4), 271–279
Mamo, T.G., Juran, I., ve Shahrour, I. (2013). Prioritization of Municipal Water Mains Leakages for
the Selection of R&R Maintenance Strategies Using Risk Based Multi-Criteria FAHP Model. Journal
of Water Resource and Hydraulic Engineering, 2 (4), 125–135.
Mondaca, M., Andrade, M.A., Choi, C.Y., ve Lansey, K.E. (2015). Development of a cost function of
water distribution systems for residential subdivisions. Urban Water Journal, 12 (2), 145–153.
Moslehi, I., Jalili-Ghazizadeh, M. Yousefi-Khoshqalb, E. (2021). ‘Developing a Framework for
Leakage Target Setting in Water Distribution Networks from an Economic Perspective’. Structure and
Infrastructure Engineering 17(6): 821–37. https://doi.org/10.1080/15732479.2020.1777568.
Nicolini, M., Giacomello, C., Scarsini, M., ve Mion, M. (2014). Numerical modeling and leakage
reduction in the water distribution system of Udine. Procedia Engineering, 70, 1241–1250.
Park, S.W. ve Loganathan, G. V. (2002). Methodology for economically optimal replacement of pipes
in water distribution systems: 2. Applications. KSCE Journal of Civil Engineering, 6 (4), 545–550.
Salehi, S., Ghazizadeh, J., ve Tabesh, M. (2017). A comprehensive criteria-based multi-attribute
decision-making model for rehabilitation of water distribution systems. Journal Structure and
Infrastructure Engineering, 14 (6), 743–765
Suribabu, C.R. ve Neelakantan, T.R. (2012). Sizing of water distribution pipes based on performance
measure and breakage-repair-replacement economics. ISH Journal of Hydraulic Engineering, 18 (3),
241–251.
Page 280
SEMIT 2023
Tee, K. F., Khan, L. R., Chen, H. P. ve Alani, A. M. (2014). Reliability based life cycle cost optimization
for underground pipeline networks. Tunnelling and Underground Space Technology, 43, 32–40.
doi:10.1016/j.tust.2014.04.007.
Venkatesh, G. (2012). Cost-benefit analysis - leakage reduction by rehabilitating old water pipelines:
Case study of Oslo (Norway). Urban Water Journal, 9 (4), 277–286.
Yılmaz, S., Fırat, M., Ateş, A., Özdemir, Ö. (2022). Analyzing the economic water loss level with a
discrete stochastic optimization algorithm by considering budget constraints. AQUA - Water
Infrastructure, Ecosystems and Society 1 July 2022; 71 (7): 835–848.
doi: https://doi.org/10.2166/aqua.2022.060.
Zangenehmadar, Z. (2016). ‘’Asset Management Tools for Sustainable Water Distribution Networks’’.
Presented in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy (Civil and
Environmental Engineering) at Concordia University Montreal, Quebec, Canada August 2016.
Page 281
SEMIT 2023
Designing A National Cyber Security Strategic Management Model with the Foundational Data
Theory Approach
a. Fereidoon Bidollahkhany. b. Mehdi Basiri
a. Faculty of Business Administration, Karabuk University, Karabuk. Türkiye.
b. Tehran Azad University, Iran.
Abstract:
The increasing use of virtual space has caused many economic, social, political and cultural activities
of people to be moved to this space. In such a situation, we will inevitably witness the entry of a large
amount of data and information of various government and public centers, organizations, businesses,
and even private data of individuals into cyberspace. Accordingly, governments seek to develop basic
strategies and solutions in this field to protect information and preserve data assets. The purpose of this
research is to provide a strategic management model of cyber security at the national level to protect
cyber information. The research is data- based in terms of its practical purpose and its implementation
method. First, scientific documents and books related to the research topic were analyzed using the
qualitative content analysis method and MAXQDA software. Then, by referring to the number of 15
cyber security experts, the designed research model was approved by making corrections.
Keywords: Strategic management, Cyber security Model, Foundational data theory.
Introduction:
The movement of countries towards modernization has led to the creation of the fourth industrial
revolution. This shows that cyber security has been considered as an important issue due to the rapid
growth of digital technology in modern countries worldwide. However, the nature of cyber technologies
is such that due to the simplicity, high availability, and low cost of these types of technologies, the
methods of maintaining security in these systems quickly become ineffective and pose a challenge to
their security. Cyber technology is changing the traditional concepts of military and civilian security,
as cyberspace removes the meaning of national borders. Cyberspace should not be seen only from a
technological point of view, but as a phenomenon that has an increasing impact on our daily lives and
the functioning of our societies.
Literature and Analysis:
Cyber capabilities have changed the actors in the global power system. From now on, cyber-attacks and
anti-cyber measures will be part of the national defense systems and military doctrine of the countries.
(Aluede, Peace, & Biragbara, 2020). Therefore, taking action to ensure cyber security by organizations
ensures that despite the release of data and information, they keep their cyber assets safe from various
risks. (Mahdavipour and Azar, 1401).
In this article, in the framework of theoretical foundations, scientific texts related to cyber space, cyber
security, cyber war and its types, cyber deterrence, main methods of prevention, cyber resilience,
cognitive warfare, cyber defense have been studied. After that, the comparative study of the countries
of America, England, Australia, South Korea, Austria, Europe and NATO has been performed in the
fields of vision, goals, mission, strategy, actions and structure. Then, the coding of the main related
topics has been done based on the subcategory and the main concepts of each has been done and
analyzed with using the qualitative content analysis method and MAXQDA software. In this regard,
the extracted final table including main and sub-categories and researcher-made concepts has been
presented as a layout and based on the results of the coding steps. after that, the basic concepts and
Page 282
SEMIT 2023
themes of the research paradigm model in five dimensions, by referring to the number of 15 cyber
security experts and using the focus group, the background, causal and intervention conditions,
consequences and strategies were concluded. The designed research model was approved by making
corrections.
Conclusion and suggestions:

. Conclusion:
The strategic management model of cyber security at the national level is as follows:
Intervening conditions:
-the presence of cyberinfrastructure in the country.
-the backwardness of cyber technology in the country
-the high motivations of competitors in managing the cyberspace
Strategies:
Causal conditions: -Acquisition of
Internal: cyber deterrence
power by
-National documents, countries.
laws, and regulations.
-Cyber deterrence
- Continuity of defense capabilities Consequences:
concepts and cyber.
-Development of -Software cyber
-cyber governance the cyber security
ecosystem at the
-Critical infrastructure Cybersecurity -hardware
national level.
protection against strategic cyber security
cyber threats. management -Cyber intelligence
-cyber security
External: -Cyber protection of network
of critical infrastructure
-The offensive strategy
infrastructures of
of countries in the -cyber security
countries.
cyber field. in the field of
-Production of human capital.
-Maximum
indigenous cyber
exploitation of
knowledge.
competitors from
cyber space. -Protection of
cyber assets at the
national level.
Background conditions:
-The attention of the country’s government to cyberspace.
-Appropriate capacity building in organizational structure and upstream documents
-Human resources and cyber specialist research academic centers.
Figure 4- the research paradigm model

Figure: The strategic management model of cyber security at the national level
Suggestions:
Based on the dimensions and components mentioned in the research, here are some suggestions for the
implementation of the model:
Page 283
SEMIT 2023
-Collaborate with countries at the international level to create and grow the country's cyber
infrastructure.
- Investment in technology cyber in fields of” research, education, and equipment”.
- Collaborate with tech companies and universities for technology transfer.
-Develop a national security cyber technology advancement plan with definite milestones.
- Educating the community about cybersecurity.
-Revise existing cybersecurity laws and regulations to confront the threats.
-Create a definite framework for the alignment of cybersecurity laws with all the activities.
-Attention to cybersecurity in national defense strategies and doctrines formulating.
-Reinforcement collaboration between defense and cyber agencies for alignment.
-Create a centralized cyber security institution to control and execute the laws and actıvıtıes for better
activities alignment.
-Readiness of cybersecurity infrastructure for aggressive and defensive capabilities:
- Alignment of cybersecurity goals with national strategies and policies.
- Allocation scholarship programs to educate cybersecurity talents.
*Generally, the implementation of the model requires a multi-functional approach that involves
collaboration between nonprivate agencies, private sectors, academies, and the public.
Resources List:
• Aluede, Lilian. (2020),”cyberattack:an emerging war”.Https://www.academia.edu›

Cyber_Attack_An_...
• Asaesh Javed. Hoda (2014). “The impact of the cyber power of the United States of America
• Mahdavi pour, Mehdi. Azar, Daud (1401).” Factors affecting the cyber security”. Quarterly
journal of military sciences and techniques. 18th year No. 60.
• Medvedev, Sergei-Monterey, (2015). Offense-defense theory analysis of Russian cyber
capability. California: Naval postgraduate school.
• Paxton، R (2008)، “Enhanced Cyberspace Defense with Real-Time Distributed System”.
• Saneyan. Ali (2018) “Cyber security in America, Structures and trends”. Foreign Policy
Quarterly. Thirty-third year. Number 1..
• Seeger, Friedman (2015).” Cyber-attack and cyber manipulation.”
• Stevens، Gina Marie (2010).” Report for Congress، Homeland Security Act of 2010: Critical
Infrastructure Information Act “.
• Shmuel Even and David Siman-Tov (2012)، “Cyber warfare: Concepts and strategic trends ،
the Institute for National Security Studies” ،
• Tiong Pern Wong (2011).” active cyber defense، enhancing national cyber defense”.
• Zabel Yesayan. (2017).” The ultimate form of cybersecurity control, a multiple case study and
Memory in texts “.
Page 284
SEMIT 2023
SP05: Safety and Reliability Engineering, System Engineering and System Safety, Safety in
Industry 4.0
Accident Analysis with Bow Tie Method; Case of British Petroleum Texas Refinery
Negar Mohabbatpour, Mazdak Khodadadi-Karimvand
Tehran, Iran.
Abstract
In process industries, the liquid overflow from tanks and other containers is one of the main causes of
accidents. This may cause catastrophic accidents and a lot of damage to the organization. One of
significant cases in such incidents is the incorrect operation of the liquid level indicator and the high-
level alarm. The accident at the Texas British Petroleum refinery in March 2005 is one of the recent
cases, which killed 15 workers and injured 150 of them. The explosion was caused by the overfilling
of the raffinate splitter tower and a blow down drum releasing hot hydrocarbons. The resulting vapor
cloud ignited, destroying the ‘ISOM’ unit. This accident cost British Petroleum Company billions of
dollars and it can be considered one of the worst industrial disasters to date. To prevent and control the
occurrence of these accidents, it is possible to use different types of hazard identification, risk
assessment and analysis methods. This action is for prioritizing risks and making decisions in order to
reduce risk to an acceptable level. There are several tools which organizations can used for risk
assessment. One of the common risk assessment tools is Bow Tie method. The Bow Tie method is a
really visual way of understanding the impacts of a hazard, the risk it presents, the consequences and
the controls that should be put in place. The research results show that this technique was one of the
most effective risk assessment methods. The purpose of this article is to analyze the British Petroleum
Texas Refinery explosion accident with the Bow Tie method.
Keywords: Process Safety Management, Risk Assessment, Accident Analysis, Bow Tie Method
1. Introduction
The oil industry is one of the complex industries with high risk and the occurrence of accidents in this
industry can cause irreparable loss of life and money. Many incidents have occurred throughout history
in this industry, which have had many consequences. Therefore, the importance of learning lessons
from incidents and accidents in these industries to identify high-risk areas is manifold [1]. Many
researchers have studied this case with different methods and applications. For instance, there is a
research with Hazop [2] which is really helpful for learning lessons from this case. In this paper, this
case is analyzed by bow-tie method which it shows t the accident scenario altogether including threat,
barriers, hazard, top event and consequences [3].A bow-tie diagram amalgamate a fault tree and an
event tree to show the risk control indicators on a common platform for mitigating an accident [11].
This method is used in many industries such as oil and gas refineries and platforms, petrochemicals and
other chemical industries, gas and nuclear power plants, military industries, transportation and ports
industry, casting industries, food and pharmaceutical industries and hospitals. Therefore, this technique
can generally be used in many industries where accidents are mostly major accidents.
1.1 Incident Description
Page 285
SEMIT 2023
Date: March 23, 2005, at 1:20 PM An explosion occurred in the vicinity where raw materials of gasoline
octane were being produced.
Radius: Approximately 1200 meters around the refinery.
Duration of the incident: Approximately 11 hours.
At 2:15 AM, night shift operators injected flammable hydrocarbon chemicals into a 100-foot tower.
The level measurement instrument was set to measure up to 9 feet, and it displayed accordingly.
Operators carried out their routine work and filled the tower.
At 3:09 AM, the first alarm, indicating an overfill situation, activated, but the second alarm, which
would halt the process, did not activate.
Operators assumed, based on the numerical reading of 9 feet displayed, that the level was within the
normal range. However, in reality, the liquid level was at 13 feet, which was erroneously indicated as
normal.
The night shift operator inaccurately prepared the summary report without providing specific numerical
details. Furthermore, the morning shift operator was inexperienced and lacked competence.
Additionally, the morning shift supervisor arrived late, without reviewing the night shift report. Due to
personal matters, the supervisor left the workplace abruptly, leaving the control room with only one
operator and no supervisor in place.
At 9:15 AM, the employees introduced the feed into the system and added liquid.
The control room instructions were incomplete, resulting in the release of fluids from the tower.
Ultimately, the level increased to 15 times the usual value, but the operator perceived the level as normal
and decreasing.
At 12:41 PM, due to an increase in pressure in the upper pipe of the tower, an alarm sounded. The
operators were unable to identify the source of the pressure rise, and their actions were
unknowledgeable.
As a result, the actions taken involved opening a valve that had been initially closed, causing the
incoming liquid temperatures to rise to 141 degrees Fahrenheit. At 1:00 PM, the introduction of hot
feedstock into the tower caused the liquids inside the tower to vaporize and expand.
The pressure of the liquids increased, and some of it was transferred to the blow down drum through a
gallon transfer.
The level increase alarm on the blow down drum malfunctioned, and no one realized that the liquid
level was rising.
Ultimately, the liquid discharged, and gasoline spilled onto the ground, forming a massive cloud of
flammable vapors.
There were 25 feet away from the blow down drum, where explosions occurred one after another,
lasting for up to 4 hours.
This incident resulted in 15 fatalities and 180 injuries, with most of these individuals being inside or
near trailers close to the blow down drum. Additionally, the organization incurred billions of dollars in
expenses, including victim complaints, facility damage, and production losses.
A two-sided container with a wooden frame was completely destroyed by the force of the explosion.
Fifty large chemical storage tanks were damaged, and the ISOM unit remained out of service for more
than two years [4].
1- Safety deficiencies within the company, as reported, indicate a disregard for incident warning signs
and a lack of safety culture. This includes a failure to report incidents and near-misses.
2- Operator and supervisor irresponsibility, such as leaving the workplace early, failing to document
process details for the next shift operators, and not providing instructions regarding feedstock and unit
products during ongoing operations. The supervisor was delayed by one hour, did not inquire about
the night shift's activities, left the workplace due to a personal emergency, and was not replaced by an
experienced supervisor. Additionally, conflicting procedures were given to operators, resulting in a
crucial valve remaining closed for hours, blocking the release of liquids from the tower. Overall, a lack
of uniformity in industrial and operational environments leads to complacency regarding safety
conditions.
3- Safety deficiencies related to alarms.
4- The closure of the control valve for the refinery tower (for sending materials from the tower to the
tank) during the restart of the isomerization unit.
Page 286
SEMIT 2023
5- Inadequate examination and resolution of the liquid spray issue from the stack due to elevated
liquid levels in the refinery tower during most of the unit startups. (This had occurred significantly
eight times in previous years).
6- The proximity of filled trailers carrying hazardous materials to the incident site (most fatalities
were among the trailer drivers).
7- Inappropriate performance of the refinery tower level indicator.
8- Inadequate training for individuals presents on-site for the restart of the isomerization unit.
9- Most of the existing equipment was outdated and ineffective.
10- Initiating the startup despite reports highlighting issues with the level indicator and relief valves.
11- Inadequate stack size design (during the liquid spray, instead of returning the liquid to the stack,
it spilled onto the ground).
12- Failure to connect the stack to the flare system (despite the emphasis on this in the documentation).
The flare system is primarily used to burn off gases released from the relief valve and protect against
high-pressure hazards in the facility.
2. Method and materials
In this article, we intend to analyze and review this significant disaster using the Bow Tie method. We
aim to utilize this method in conjunction with the available data to identify the controls and measures
that were lacking, leading to the occurrence of this massive incident.
2.1. What is Bow Tie method?
The Bow Tie methodology is an invaluable approach for the analysis and assessment of event risks,
through which we can examine potential scenarios that may pose a threat. Additionally, this software
possesses the capability to visually represent the existing controls implemented by an organization to
mitigate these scenarios through diagrams. Furthermore, these controls may themselves exhibit
weaknesses and vulnerabilities, which can be identified and addressed using this software to prevent
their failure [5].
2.2. Study Area
The Texas City refinery is the third-largest refinery in the United States, with the capability to produce
over 450,000 barrels of crude oil per day. The incident in question occurred within the ISOM unit,
specifically due to technical issues with the raffinate tower and blow down drum, as well as deficiencies
in the trailer area and operator negligence. In this context, we delve into the examination of scenarios
and vulnerabilities in controls within this area utilizing the Bow Tie methodology.
Page 287
SEMIT 2023
Figure1: BP Texas City Refinery Land

2.3. Procedure
By conducting a thorough analysis and examination using the Bow Tie methodology, a visual diagram
can be prepared to investigate and elucidate the hazards that led to the unfortunate loss of 15 lives on
July 15, 2005. This diagram will also facilitate the presentation of controls designed to mitigate these
hazards. Bow Tie model contained the following components of an accident/incident: threat,
preventative risk, barriers, top events, mitigation risk barriers, consequences and escalation factors.
Figure 2: Simplified Bow-Tie Model
2.4. History of Bow-Tie method
It is said that the first ‘real’ Bowtie diagrams appeared in the (Imperial Chemistry Industry) course notes
of a lecture on HAZAN (Hazard Analysis) given at The University of Queensland, Australia (in 1979),
but how and when the method found its exact origin is not completely clear. The
catastrophic incident on the Piper Alpha platform in 1988 awoke the oil & gas industry. After the report
of Lord Cullen, who concluded that there was far too little understanding of Hazards and their
accompanying risks that are part of operations, the urge rose to gain more insight in the causality of
Page 288
SEMIT 2023
seemingly independent events and conditions and to develop a systematic/systemic way of assuring
control over these Hazards.
In the early nineties the Royal Dutch / Shell Group adopted the Bowtie method as company standard
for analyzing and managing risks. Shell facilitated extensive research in the application of the Bowtie
method and developed a strict rule set for the definition of all parts, based on their ideas of best practice.
The primary motivation of Shell was the necessity of assurance that appropriate risk controls are
consistently in place throughout all worldwide operations. Following
Shell, the Bowtie method rapidly gained support throughout the industry, as Bowtie diagrams appeared
to be a suitable visual tool to keep overview of risk management practices, rather than replacing any of
the commonly used systems.
In the last decade the Bowtie method also spread outside of the oil & gas industry to include aviation,
mining, maritime, chemical and health care to name a few.

The Bow Tie methodology was employed to assess the risk associated with the incident that occurred
on July 15, 2005. Controls implemented and overlooked within the organization were thoroughly
examined and incorporated into a Bow Tie diagram. Table 1 presents Preventative and Mitigation
barriers that were evaluated within the Bow Tie diagram, with items highlighted in red signifying
barriers not implemented within the organization, lacking control and supervision. As previously
mentioned, the risk assessment was conducted in the Consequence section, demonstrating the extent of
damage to assets, human resources, the environment, and reputation. Through the utilization of the Bow
Tie methodology, weaknesses within the organization were identified, and a comprehensive scenario
from threat to consequence was depicted on the diagram. Additionally, for some barriers, an escalation
factor was determined, leading to the identification of specific controls and their incorporation into the
diagram, as elaborated upon in Table 2[6].
3.1. Preventative Risk Barriers
Specific to the hazards evaluation of an incident sequence, a preventive barrier is in between the
initiating event (the cause) and a loss event, helping reduce the frequency of the incident scenario, and
thus, helping reduce the scenario's risk.
The most significant findings in this section include the following critical aspects: blow down drum,
safety culture, warning, and proximity detection systems.
3.2. Mitigatory Risk Barriers
The identification of appropriate risk mitigation measures requires a good understanding of the hazard
and the factors contributing to its occurrence, since any mechanism that will be effective in reducing
risk will have to modify one or more of these factors. Risk mitigation measures may work by reducing
the probability of occurrence, or the severity of the consequences, or both. Achieving the desired level
of risk reduction may require the implementation of more than one mitigation measure.
The most significant findings in this section include the following critical aspects: Hazardous area,
classification, Emergency response, and Ventilation to disperse gas.
Page 289
SEMIT 2023
Table 1: Examples of components that form the Bow Tie model

Threat Preventative barrier Hazard Mitigation barrier Consequence
Top event
1. overfilling of a 1. PSSR 1. flammable 1. Hazardous area 1. VCE
raffinate splitter 2. alarms and control hydrocarbon classification 2. explosion
tower instrumentation (level (Raffinate) 2. F&G System 3. flash fire
2. Supervisors indicator) 2. flammable 3. Ventilation to
and operators 3. pressure relief valves liquid release disperse gas
poorly 4. blow down drum out 4. Emergency
communicated 5. Understand Each response procedures-
3. ISOM Person’s Role in a personnel
operators were Project escape/muster and
likely fatigued 6. Schedule Weekly evacuate
from working Team Meetings 5. Isolation of
4. The operator 7. Increasing safety electrical
training program culture installations
was inadequate 8. shift change at work 6. Firefighting
5. Poor 9. Hire more operators e.g.deluge cooling
Computerized 10. Offer video training
Control Board 11. Fix and repair
Display 12. Hire experience
operator for controlling
Table 2: Examples of escalation factors that form the Bow Tie mod
Escalation factors of Preventative barrier Escalation factors of Preventative barrier
threats consequences
1. failing in alarms and 1. calibration 1. office trailers located 1. Management of
control instrumentation 2. Inspection and close to the blowdown change
is for (alarms and repair drum is for Hazardous 2. Replace office
control instrumentation 3. Management area classification trailers to safe area
(level indicator)) of Change 2. pickup truck located 3. Park the truck in a
2. damaged and failed 4. Inspection and about 25 feet (7.6 m) safe area
is for (pressure relief repair from the blowdown 4. calibration
valves) 5. connected to flare drum is for Hazardous 5. measurement and
3. antiquated and system area classification repair
unsafe design is for 6. Installed modern 3. Gas sensing error is 6. Regular training
(blow down drum) one for F&G System and evaluation of
4. The size of the 7. Installed the 4. Decreased power is for employees
blowdown drum was standard one Ventilation to disperse 7. Inspection and
insufficient is for (blow 8. Fix and repair gas replacement
down drum) 5. Lack of training to
5. The high-level alarm employees is for
on the blowdown drum Emergency response
was nonfunctional procedures- personnel
and did not sound is for escape/muster and
(blow down drum) evacuate
6. breaking down the
electrical cable is for
Isolation of electrical
installations
Page 290
SEMIT 2023
Figure 3: An illustrative Bow Tie risk model for British Petroleum Texas Refinery
3.3. Shortcomings in Process Safety Management

PSM is a management system that focuses on prevention, preparation, mitigation, and restoration of
accidents from a chemical or energy process. Utilization of PSM detects flaws of processes, evaluates
facilities and employees, improves process safety of facilities by removing the risks and protecting
workers, facilities, and the public from invisible dangers.
In incidents similar to the BP refinery disaster, there consistently exist numerous failures, many of
which are rooted in the Process Safety Management (PSM) system. This system comprises elements,
the neglect of each of which can lead to irremediable harm within an organization. In the following, we
examine and analyze the elements identified using the Bow Tie method [7].
Page 291
SEMIT 2023
3.4. Management of Change (MOC)

The foremost element within the Process Safety Management (PSM) system is the management of
change since even the slightest alterations within an organization can give rise to new hazards. Hence,
it is imperative that this element is consistently applied within the organization. In the case of BP, one
of the significant changes involved the relocation of contractor trailers in proximity to the blow down
drum. This area constituted a hazardous zone, and the organization carried out this relocation without a
thorough analysis and assessment of the consequences and risks involved. Unfortunately, the highest
number of fatalities and injuries occurred in this area. In the Bow Tie diagram, this particular case is
designated as one of the barriers. If the organization had adhered to this measure, the number of fatalities
and injuries would have significantly decreased [9].
3.5. Safety leadership
The absence of this element within an organization significantly increases the likelihood of catastrophic
incidents. In the case of BP, this critical element was lacking, and individuals did not feel a sense of
responsibility. Furthermore, adequate training was not provided. This leads us to conclude that the
organization, driven by profit motives and a reluctance to invest in employee education and awareness,
failed to instill a sense of responsibility in its personnel. Additionally, non-standard and outdated
equipment, such as the blow down drum lacking a Flare System, was purchased. Lastly, a shortage of
personnel, including operators and supervisors, was another challenge faced by this organization. It is
worth noting that all these issues can be traced back to a deficiency in the safety culture within the
organization, a topic we will further investigate.
3.6. Learning culture
Learning and enhancing a safety culture is imperative for all organizations, as it can effectively prevent
similar catastrophic incidents. In this particular organization, the safety culture was severely lacking, as
reports indicate that many incidents had occurred previously but had not been adequately addressed or
repaired. For instance, the alarm system in the blow down drum had pre-existing issues, yet no action
was taken to rectify it. Furthermore, a significant contributing factor to this incident was human error,
stemming from the absence of a safety culture among the organization’s personnel. According to the
CSB report, numerous safety issues were prevalent within BP, particularly within the ISOM unit. The
Baker report points to a safety culture deficiency and highlights that BP had not adhered sufficiently to
safety recommendations issued by the Department of Energy. This report suggests that cost-cutting
measures and production pressures from BP management may have led to inadequate safety measures
nationwide. Carolyn W. Merritt, Chair and CEO of the CSB, remarked, “A combination of cost-cutting,
production pressures, and underinvestment led to a gradual erosion of safety at the refinery. Lack of
maintenance, inadequate training, and a needed safety culture were among the company's most serious
weaknesses.”
3.7. Mechanical integrity
The significance of this matter lies in ensuring that systems and equipment are installed, maintained,
and operated correctly. As identified in the Bow Tie diagram, some of the violations of this principle
were as follows:
• The alarm system for the blow down drum and the raffinate splitter tower was non-functional.
• The level indicator for the raffinate splitter tower was not adequately calibrated and displayed an
incorrect liquid level reading.
• The level sight glass was dirty and therefore prevented manual verification of the raffinate level.
• The manual vent valve that allowed operators to vent vapor from the raffinate splitter tower was non-
functional during start-up testing.
3.8. Pre-startup Safety Review (PSSR)
One of the barriers mentioned in the Bow Tie diagram is indeed this practice. If BP had implemented
this measure and conducted pre-operation inspections, individuals would have identified systemic
Page 292
SEMIT 2023
issues in the raffinate splitter tower and blow down drum. Consequently, this catastrophic incident could
have been prevented. Therefore, this action is of paramount importance and utility in averting similar
incidents.
3.9. Operating Procedures
The standard operational method allows work teams to precisely understand what tasks need to be
performed, when they should be carried out, and how they should be executed. Therefore, this matter
holds significant importance, as improper execution of tasks and operations can introduce new risks to
an organization. In the case of BP, employees were relying on their personal knowledge and experiences
for the startup of the ISOM unit, which is a noteworthy concern. Another noteworthy issue was the
frequent use of manual mode for level control instead of the automatic mode. This practice led to human
errors and an increased likelihood of mistakes and incidents occurring.
3.10. Preliminary Hazard Analysis
It can be asserted that this element is the objective that we are pursuing in this article. The quantitative
and qualitative analysis of incidents holds particular significance. In this organization, there were near
misses that the organization had not paid adequate attention to. By investigating these incidents using
software such as Bow Tie, the organization could have taken preventive measures against potential
incidents that might have occurred in the ISOM unit.
3.11. Human Factors
As previously mentioned, the primary cause of this catastrophe was the numerous human errors during
operations. Niles T. Welch classified the main reasons for human errors into six categories:
• System complexity
• Stress
• Fatigue
• Environmental conditions
• Training
• Experience
In the case of BP, employees were enduring significant stress, pressure, and fatigue, which led to
decreased accuracy in their tasks. Moreover, during the incident, there was only one operator with
limited experience, and more experienced individuals were not present at that time [8].
3.12. Corrective Suggestions and Training Courses
Incidents and near-misses can serve as valuable educational resources and aid in preventing deeper-
seated occurrences. A culture of learning ensures that incident reports and safety information are
analyzed, necessary lessons are extracted, and appropriate and effective actions are taken in response.
The Texas City refinery and BP as a whole were lacking a culture of learning from incidents and near-
misses. Management had not adequately investigated past incidents, which had highlighted serious risks
within the blow down drum system and other units, and they had not taken appropriate corrective
actions. Managers had not leveraged the experience and reports of experienced operators and
supervisors. Poor instrument checks were not reflected in the reports, and the thorough pre-operation
checks for the ISOM unit were not completed. In 2000, at the BP Scotland refinery, three incidents
resulted in fatalities. Subsequently, similar issues occurred at the Texas City refinery five years later.
The investigation revealed that the employees at the Texas City refinery had limited awareness of the
lessons from the 2000 Scotland incidents. BP’s decentralized structure had contributed to a lack of
centralized learning from past incidents.
In this incident, insufficient attention was given to training, and most of the training was computer-
based rather than being conducted in a face-to-face manner. The training at BP did not adequately
prepare employees for abnormal conditions. Despite recommendations, the company did not utilize
simulators for training and decision-making processes for employees.
Page 293
SEMIT 2023
1. Enforce necessary guidelines to prevent fatigue in refineries and petrochemical industries by

setting limitations on working hours and days.
2. Identify equipment with high-risk potential for catastrophic incidents. This can be achieved by
considering predefined variables in safety process indicators and gathering information from the
Environmental Protection Agency (EPA) regarding equipment hazards.
3. Scrutinize and manage any changes in the system that may impact process safety. This includes
structural changes such as mergers, ownership changes, personnel turnover, and policy alterations,
like budget reductions.
4. Designate a non-executive board member with high competency and experience in process
safety to provide expertise and oversight.
5. Create a culture that encourages employees to report incidents without fear of reprisal, fostering
a proactive approach to safety.
6. Implement swift and thorough responses to incident reports and recommendations to rectify
identified issues promptly.
7. Regularly evaluate the safety of process units within the refinery to ensure that critical
equipment is designed and operated safely.
8. Verify that instruments, tools, and process equipment necessary for safe operation are properly
maintained, repaired, and tested as needed.
9. Increase the number of operators and utilize additional operators during unit startups.
10. Raise employee awareness through standardized training on safety standards and guidelines
within the refinery.
11. Foster a culture of safety within the company, emphasizing employee and operator
responsibility for safety.
12. Regularly assess and report safety deficiencies, including alarm systems and indicators.
13. Ensure the use of current safety standards and equipment.
14. Allow employees sufficient rest periods to reduce fatigue and maintain alertness.
15. Thoroughly investigate and report incidents and near misses, treating them with the seriousness
they deserve.
16. Have contingency plans and replacements available for operators and supervisors during
emergency situations that may incapacitate these personnel.
17. Store hazardous chemicals and materials that are prone to combustion or chemical reactions at
a safe distance from the refinery.
18. Employ software tools such as Bow Tie to analyze and prevent similar incidents through
scenario modeling.
Training programs encompassing NFPA, ISEA, OSHA, IPS, API standards, along with the following
courses:
• Training on proper actions to take during emergency situations
• Understanding and safely managing hazardous gases, including nitrogen and H2S
• Gasometer installation and inspection
• Operations during emergencies
• Understanding MSDS for hazardous materials present in the refinery
• Proper training on working safely in refineries
Page 294
SEMIT 2023
• General Health, Safety, and Environment (HSE)

• First Aid
• Valve and safety relief device handling
4. Conclusion
Accident analysis is a widely employed methodology in the realm of safety management, serving as a
crucial tool for preventing future catastrophes. Hence, the significance of this approach cannot be
overstated, particularly when it comes to scrutinizing incidents like the BP incident and similar
occurrences. These analyses should be conducted by experts using suitable software and methodologies.
In this study, we employed the Bow Tie method to dissect the BP incident, exploring controls and
potential scenarios. Through this approach, it became evident that the safety culture within the
organization was severely lacking. It was apparent that individuals within the organization could have
easily prevented the incident. For instance, in the first threat within the Bow Tie diagram, it was evident
that adequately calibrated alarms and proper functioning could have averted the overfilling of the
raffinate tower. Furthermore, the implementation of Process Safety System Reviews (PSSR) could have
identified vulnerabilities. However, it should be noted that the root cause of the inadequate safety culture
stemmed from a lack of commitment from senior leaders to safety. Many shortcomings were identified,
but they went unaddressed. The ultimate takeaway from this incident is the absence of a sense of
responsibility among employees and senior management within the organization. This deficiency is
glaringly evident in the neglected controls within the Bow Tie software. In conclusion, the Bow Tie
software proves to be a suitable tool for risk analysis and assessment. It serves as a compelling means
to persuade top managers to allocate greater resources to employee safety. The visual nature of its
diagrams effectively communicates potential scenarios and hazards, ensuring that all individuals within
the organization are well-informed about possible risks and incidents [10].
References
[1]. M. Sam Mannan and Simon P. Waldram, “Learning lessons from incidents: A paradigm shift is
overdue”,2012.
[2]. Joseph Isimite, Philip Rubini, “A dynamic HAZOP case study using the Texas City refinery
explosion”, 2016.
[3]. Nima Khakzad, Faisal Khan, Paul Amyotte, “Dynamic risk analysis using bow-tie approach”,2012.
[4]. Don Holmstrom, Francisco Altamirano, Johnnie Banks, Giby Joseph, Mark Kaszniak, Cheryl
Mackenzie, Reepa Shroff, Hillary Cohen, and Stephen Wallace PE, CSP, “CSB Investigation of the
Explosions and Fire at the BP Texas City Refinery on March 23, 2005”,2006.
[5]. Jabulani Matsimbe, Steven Ghambi, Abdul Samson, “Application of the Bow-Tie Method in
Accident Analysis: Case of Kaziwiziwi Coal Mine”, 2020.
[6]. Valérie de Dianous, Cécile Fievez, “ARAMIS project: a more explicit demonstration of risk control
through the use of bow-tie diagrams and the evaluation of safety barrier performance”,2014.
[7]. Kalokson Gurung, Laya Jayadeep, Janusz Siwek, Satyam Vora & David Zhou, University of
Bradford, UK “Texas City Refinery explosion — safety out of focus”, 2020.
[8]. Cheryl MacKenzie, Donald Holmstrom, and Mark Kaszniak, “Human Factors Analysis of the BP
Texas City Refinery Explosion”,2007.
[9]. Soheil Abbasi, Payam Khourdustan, Rasoul Sayyadi, Saeid Jalilpour, Seyed Shamseddin Alizadeh
“BP Texas Refinery Incident Causes: A Literature Review”, 2020.
[10]. Faisal I. Khan, Paul R. Amyotte, “Modeling of BP Texas City refinery incident”,2007.
Page 295
SEMIT 2023
[11]. Refaul Ferdous, Faisal Khan, Rehan Sadiq, Paul Amyotte, Brian Veitch, “Analyzing system safety
and risks under uncertainty using a bow-tie diagram: An innovative approach”,2013.
Page 296
SEMIT 2023
SP06: Applied Soft Computing in engineering Management
A Study on Measuring the Mental Workload of Instructors and Evaluating Significant

Differences
Samet Tosun1[0000-0002-5454-3057] and Ibrahim Yilmaz2[0000-0002-5959-7353]
1
Tokat Gaziosmanpasa University, Tokat 60400, TURKIYE
2
Ankara Yildirim Beyazit University, Ankara 06010, TURKIYE
samet.tosun@gop.edu.tr
Abstract
Mental workload is defined as the variance between the cognitive pro-cessing capacity required to meet
performance expectations and the capacity achievable within a specified time frame. Assessing mental
workload is cru-cial for gauging the cognitive demands placed on today's workforce. Ulti-mately,
numerous methods have been developed to quantify mental work-load. This study aims to assess the
mental workload of academic faculty members in universities located in Tokat, Ankara, and Sivas,
while also in-vestigating potential variations in mental workload based on academic titles. The collected
data were analyzed using the SPSS 27 and a content analysis was performed. The structured interview
questions utilized in this study were adapted from Rubio et al.'s (2017) CarMen-Q Mental Workload
Scale, which consists of 4 sub-dimensions and 29 items covering cognitive workload, temporal
workload, performance-related workload, and emotional workload The study exhibited strong internal
consistency with a coefficient of α=0.94. Notably, the highest sub-dimension of mental workload was
found to be cognitive workload, while the lowest was temporary workload. Furthermore, this study
holds the distinction of being one of the first to employ the Car-Men-Q method for mental workload
assessment, an approach that has not been widely explored in the existing literature.
Keywords: Ergonomics, Mental Workload, CarMen-Q.
1. Introduction
In our increasingly globalized world, the significance of workload has grown substantially, owing to
the pervasive influence of industry and technology across various sectors. Despite long-standing
research efforts, there exists no universally accepted definition of workload [1]. The literature presents
a plethora of definitions for mental workload. One such definition characterizes mental workload as the
cognitive construct arising from task execution under specific operational and environmental
conditions, encompassing the mental exertion multiplied by the operator's capacity to respond to these
demands [1]. Another perspective defines mental workload as the product of perceptual and cognitive
activities, including computation, decision-making, communication, memory, search, and research. It
is seen as the concept encapsulating the expenditure associated with fulfilling job requirements [2].
Among the comprehensive definitions in the literature is the notion of mental workload as the strain
resulting from task completion [3].
Numerous methodologies have been devised to assess mental workload, with the Task Load Index
(NASA-TLX) from the National Aeronautics and Space Administration being one of the most
commonly employed due to its ease of use. However, research on mental workload in Turkey remains
limited [4]. It has been observed that the majority of studies in this context utilize the NASA-TLX
method and focus on investigating the mental workloads of healthcare professionals [5]. This study
breaks new ground by employing the CarMen-Q scale, a mental workload assessment tool developed
by Rubio-Valdehita and colleagues (2017). Its objective is to examine potential significant variations
in the mental workloads of academic staff in universities situated in Tokat, Ankara, and Sivas based on
Page 297
SEMIT 2023
their academic ranks. Considering the increasing demands associated with teaching, administrative
responsibilities, and research duties in academia, which often necessitate significant sacrifices, this
study seeks to elucidate the correlation between the academic titles of educators and their mental
workloads. Notably, the literature review revealed a dearth of studies comparing the mental workloads
of educators according to their academic ranks. Existing national-level research has predominantly
employed the NASA-TLX method for mental workload assessment, and these studies have been
incorporated into the literature review. This study introduces the CarMen-Q method, hitherto unused
for assessing the mental workloads of academics. Consequently, it aims to address this literature gap,
provide guidance for future research, and present recommendations. The overarching goal is to
determine the optimal level of mental workload that allows academics to enhance their performance
while mitigating excessive burdens. Through studies and investigations focused on improving
academics' quality of life and meeting their needs, the aim is to elevate the quality of educational
activities offered by academics, thereby boosting their overall performance.
When reviewing the literature, it becomes evident that the notion of workload is not as novel as
commonly perceived. For instance, early studies such as "Assessing Unused Cognitive Capacity in
Drivers via a Secondary Task" by Poulto and Brown in 1961, as well as "Challenges in Quantifying
Mental Burden" by Bainbridge in 1974, stand as testament to this [6]. The interest in defining mental
workload and devising measurement techniques has witnessed a substantial surge since the 1970s.
Many of these studies have operated in contrast to the concept of physical workload. Mental workload,
a multidimensional concept, has been explored through various disciplinary lenses, resulting in a
plethora of research on its definition and assessment [7]. Throughout the literature, distinct terms like
cognitive workload, cognitive load, mental effort, and cognitive effort have been employed to articulate
the concept of mental workload. Alas, there is no widely accepted universal definition for this construct.
Workload, as it pertains to environmental, physical, positional, and mental factors influenced by body
postures, has been scrutinized under four primary categories [8]. A prevailing consensus on mental
workload suggests that it is commensurate with the discrepancy between available resources and those
required for a given task. Essentially, the workload experienced by individuals fluctuates in response to
changing demands and available resources.
The measurement of physiological workload rests on the evidence-based premise that heightened
mental stimuli lead to increased physical responses in the body. Physiological scales are tools that
objectively gauge an individual's mental workload by assessing physical bodily reactions. These
methods provide objective data, distinct from subjective measurements that rely on direct self-reporting.
However, the logic of equating increased mental workload solely with heightened physical responses
isn't always supported. For instance, the body exhibits increased physical responses to augmented
physical workload, and equating increased mental workload to elevated physical workload may yield
misleading results [9].
Subjective measures are employed to unveil the cognitive load placed on working memory [10]. In a
straightforward yet pragmatic view of workload assessment, it is posited that "if an individual perceives
their workload as excessive, then it is indeed excessive" [10]. While physiological workload
measurements offer more precision, subjective assessments are often considered more practical and
functional [9]. Subjective tests possess the flexibility to adapt to individuals with varying capacities.
The subjective approach values personal differences, encompassing state, attitude, behavior, and
competence, precisely because it is subjective, not in spite of it [11]. Importantly, despite their
differences, objective and subjective workload measurements exhibit some degree of overlap, as
evidenced by physiological measurements like heart rate correlating with subjective assessments [12].
One of the most widely used methods among subjective scales is undoubtedly the NASA-TLX method.
This tool assesses mental workload across six distinct dimensions: mental strain-demand, effort,
temporal demand (time pressure), physical strain-demand, emotional impact, and discomfort level, all
of which contribute to overall performance assessment. Nonetheless, it's worth noting that this scale is
Page 298
SEMIT 2023
not always deemed suitable for real-world work environments and may pose certain usability challenges
[2]. As a result, the CarMen-Q scale, developed by Rubio-Valdehita and colleagues in 2017, was chosen
in place of the NASA-TLX and similar scales, which are frequently encountered in the literature but
often considered cumbersome to apply.

3.1. Material
This study adopts a relational and descriptive research approach. It is premised on the assumption that
sufficient, representative data can be collected more efficiently through sample selection rather than
attempting to survey the entire population. Consequently, a snowball sampling method was employed,
resulting in a sample comprising a total of 133 lecturers, including 16 professors, 12 associate
professors, 43 doctoral lecturers, 50 lecturers, and 12 research assistants employed at universities in
Tokat, Ankara, and Sivas. The research primarily employs an inductive approach, frequently utilized in
academic research to gain insights into general patterns by analyzing a limited number of examples.
The methodology adopted for this research is qualitative, drawing from various disciplines and seeking
to comprehensively comprehend the theoretical underpinnings of human behavior within their living
environment.
The socio-demographic characteristics of the instructors were ascertained through Google Forms,
and the findings are presented in Table 1. Notably, 37.6% of the lecturers hold the title of "lecturer,"
41.4% possess 15 years or more of professional experience, 50.4% are affiliated with the city of Tokat,
44.4% fall within the 30-39 age group, and 50.4% are male. Furthermore, the marital status of 81.2%
of the participants was identified as married.
Table 1. Distribution of Socio-Demographical Characteristics of Faculty Members

Variable Numver (n) Percent (%)
Professor 16 12,0
Associate Prof. 12 9,0
Title Doctor Lecturer 43 32,3
Lecturer 50 37,6
Research Assist. 12 9
1-4 Year 22 16,5
5-9 Year 30 22,6
Experince
10-14 Year 26 19,5
15 Years and above 55 41,4
Tokat 67 50,4
City Ankara 35 26,3
Sivas 31 23,3
20-29 Years 7 5,3
30-39 Years 59 44,4
Age
40-49 Years 39 29,3
50 and above age 28 21,1
Female 66 49,6
Gender
Male 67 50,4
Married 108 81,2
Marital Status
Single 25 18,8
3.2. Methodology
Rubio-Valdehita and colleagues (2017) introduced a novel scale to the field of mental workload
assessment, known as CarMen-Q, which has garnered recognition for its reliability, validity, and
Page 299
SEMIT 2023
practical utility in real-world work settings. CarMen-Q has been demonstrated to possess robust
validity, ease of implementation, and high reliability, as evidenced by its strong correlation with NASA-
TLX [13]. While NASA-TLX stands as a well-established tool for assessing cognitive load, it has been
critiqued by a team of NASA-TLX researchers for certain limitations. This critique revolves around the
observation that performance metrics are significantly influenced by both physical and emotional
burdens placed on an individual, rendering them less practically relevant. Consequently, CarMen-Q
was specifically designed to circumvent the confounding impact of physical load parameters influenced
by emotional demands, thus ensuring a more authentic and precise measurement of mental workload
[14].
4. Conclusion and Recommendations

Upon analyzing the data collected in the study, a comparison of mental workloads among faculty
members based on their academic titles revealed that doctoral faculty members exhibited a higher
mental workload compared to their counterparts with different titles. As anticipated, the research
findings align with expectations, indicating that lecturers, who tend to experience lower levels of
academic anxiety, demonstrated the lowest mental workloads. Table 2 provides the outcomes of the
one-way analysis of variance (ANOVA) test, which was conducted to ascertain the specific groups in
which these differences in mental workload are statistically significant
Table 2. Statistical comparisons of faculty members by socio-demographical characteristics

Title City
Experie Marital
(Signifi (Significan Age Gender
nce Status
cant) t)
F: F: F:
Total
3.711 0.402 F: 4.333 F: 0.755 F: 1.011 1.089
Mental
P: P: P: 0.015 P: 0.522 P: 0.056 P:
Workload
0.007 0.752 0.348
F: F: F:
Cognitive 3.712 0.230 F: 1.677 F: 0.445 F: 2.503 0.558
Workload P: P: P: 0.191 P: 0.721 P: 0.770 P:
0.007 0.875 0.596
F: F: F:
Temporar
3.025 0.554 F: 6.837 F: 1.272 F: 1.305 0.020
y
P: P: P: 0.002 P: 0.287 P: 0.111 P:
Workload
0.020 0.646 0.381
Workload F: F: F:
Due to 1.155 0.704 F: 1.876 F: 1.717 F: 0.123 0.794
Performa P: P: P: 0.015 P: 0.167 P: 0.314 P:
nce 0.033 0.551 0.322
F: F: F:
Emotional 2.865 0.111 F: 3.542 F: 0.244 F: 0.651 0.146
Workload P: P: P: 0.032 P: 0.865 P: 0.001 P:
0.026 0.953 0.393
Following the implementation of the Tukey and Scheffe Post Hoc test, it was established that there
exists a statistically significant difference in mental workload concerning the classification of
academic staff based on their academic titles and the cities in which they are situated. However, it
should be noted that differences in classifications related to other demographic characteristics did not
yield statistically significant distinctions.
Page 300
SEMIT 2023
For our primary research objective, which is the examination of title-based differences in mental
workload, the analysis reveals the following average scores: professors with an average of 3.67,
associate professors with an average of 3.61, doctoral faculty members with an average of 3.85,
lecturers with an average of 3.31, and research assistants with an average of 3.59. As indicated in
Table 2, the Post Hoc test conclusively demonstrates the significant disparities among these academic
titles in terms of mental workload.
References
1. Cain, B. (2007). A review of the mental workload literature. Defence Research And Development
Toronto (Canada).Author, F., Author, S.: Title of a proceedings paper. In: Editor, F., Editor, S. (eds.)
CONFERENCE 2016, LNCS, vol. 9999, pp. 1–13. Springer, Heidelberg (2016).
2. Karadağ, M., & Cankul, İ. (2015a). Hekimlerde Zihinsel İş Yükü Değerlendirmesi. The Journal of
Academic Social Science Studies, (35), 361-370. Author, F.: Contribution title. In: 9th International
Proceedings on Proceedings, pp. 1–2. Publisher, Location (2010).
3. Gopher, D. ve Donchin, E. (1986). Workload: An examination of the concept. K.R. Boof, L. Kaufman
ve J.P.Thomas (Ed.). Handbook Perception and Human Performance içinde (41/2-44). New York:
Wiley..
4. Emeç, Ş., & Akkaya, G. (2018). Sağlik Sektöründe Zihinsel Iş Yükü Değerlendirmesi Ve Bir
Uygulama. Ergonomi, 1(3), 156-162.
5. Akca, M., Yavuz, M., & Küçükoğlu, M. T. (2020). Zihinsel iş yükünün ölçümü: CarMen-Q ölçeğinin
Türkçe’ye uyarlaması. Yaşar Üniversitesi E-Dergisi, 15(60), 675-691.
6. Akça, H. K. (2022). Bilgi teknolojilerinde proje yönetim metodolojilerinin bilişsel ergonomik açıdan
değerlendirilmesi Cognitive ergonomic evaluation of project management methodologies in
information technologies.
7. Moray, N. (1979). Models and measures of mental workload. Mental workload: Its theory and
measurement, 13-21.
8. Dağdeviren, M., Eraslan, E., & Mustafa, K. U. R. T. (2005). Çalişanlarin toplam iş yükü seviyelerinin
belirlenmesine yönelik bir model ve uygulamasi. Gazi Üniversitesi Mühendislik Mimarlık Fakültesi
Dergisi, 20(4).
9. Gürcoşkun, F. (2019). Hava trafik kontrolörlerinde zihinsel iş yükünün havacılık operasyonlarına etkisi
(Master's thesis, Sosyal Bilimler Enstitüsü).
10. Johanssen, G., Moray, N., Pew, R., Rasmussen, J., Sanders, A., & Wickens, C. (1979). Final report of
experimental psychology group. Mental workload: Its theory and measurement, 101-114.
11. Muckler, F. A., & Seven, S. A. (1992). Selecting performance measures:" Objective" versus"
subjective" measurement. Human factors, 34(4), 441-455.
12. Tattersall, A. J., & Foord, P. S. (1996). An experimental evaluation of instantaneous self-assessment as
a measure of workload. Ergonomics, 39(5), 740-748.
13. LoBiondo-Wood, G., & Haber, J. (2014). Reliability and validity. Nursing research. Methods and
critical appraisal for evidence based practice, Mosby Elsevier.
14. Rubio Valdehita, S., López Núñez, M. I., López-Higes Sánchez, R., & Díaz Ramiro, E. M. (2017).
Development of the CarMen-Q Questionnaire for mental workload assessment. Psicothema.
Page 301
SEMIT 2023
Optical Transmission Systems Incident Management in the Telecommunications Sector in

Senegal
1 2 2
Abdou Dıop , Ibrahima Ngom , and Idy Dıop
Polytechnic High School (ESP/UCAD)
Doctoral School of Mathematics and Computer Science
Cheikh Anta DIOP University of Dakar (UCAD), BP 5005- Dakar - SENEGAL
https://www.ucad.sn/
{diop.abdou@esp.sn} {ibrahima.ngom@esp.sn}
{idy.diop@esp.sn}
Abstract
We present the maintenance of optical fiber in the telecommunications sector in Senegal. This survey
is the start of a project to set up a real-time incident management system for fiber optic transmission
networks. In concrete terms, there are several stages in fiber optic maintenance: first, the incident is
detected by the NOC (Network Operation Centre), then the incident is analyzed by the operator's fiber
optic maintenance supervisor or transmission back office. Then the fiber optic maintenance team is
activated to deal with the incident. Next, the team in charge of the incident launches remedial
maintenance operations on the optical fiber. Finally, the team measures performance after repairs and
draws up an incident report. This work firstly enables us to understand optical fiber maintenance. Then
to propose an automatic system based on AI. This automatic, real-time system will improve the
efficiency and productivity of incident management by enabling the operator to use artificial
intelligence (AI) applied to image recognition to identify and notify the appropriate team of incidents
such as a cut in the optical fiber. At the same time, such a system is also very useful for the telecoms
regulator, who can use images to check the quality of fiber optic installations and compliance with
standards by operators and fiber optic access providers.
Keywords: Fiber optics, incidents, AI, image recognition, Management, transmission, telecoms.
1. Introduction
Ten years ago, less than half of African countries had any form of fiber optic connection. Senegal was
one of them, having had fiber optics since 2000. In Senegal, in fact, connection to fiber optics began
relatively early, compared with many other African countries. The use of fiber optic submarine cables
has enabled Senegal, thanks to the multiplication of access points throughout the country, to make the
internet available anywhere in the country. All Senegalese, wherever they are in the country, can now
connect to the internet and benefit from more bandwidth, one of the highest on the continent. But fiber
optics are rather fragile and more vulnerable to damage than copper wires. Optical fiber can only be
used on the ground (with the exception of some aerial applications on poles and underwater longkhaul).
This practice of using optical fiber on the ground, under streets, roads and drains, takes up three quarters
of the initial expenditure on civil engineering works (GC) and will lead to significant new costs each
time the optical fiber is cut. The reality in Senegal is that fibers optics are very regularly affected by
construction projects such as rail infrastructure (regional express train (TER)), road infrastructure (bus
rapid transit (BRT)), work on the toll motorway, flooding and the foundations of residential buildings.
These fiber optic incidents have a major impact on fiber optic access operators, as well as users of
internet access services and businesses that use fiber optics to interconnect their sites throughout
Senegal. Our paper therefore focuses on the maintenance of optical fiber in the Senegalese
telecommunications sector. Section 1 is the introduction, section 2 deals with fiber optic transmission
systems, section 3 is devoted to the management of fiber optic incidents, the objectives and potential
challenges for incident management are presented in section 4 and finally section 5 is the conclusion.
Page 302
SEMIT 2023
2. Optic Transmission Systems

2.1. Optical Fiber
2.1.1. What is optical fiber?
Fiber optics is currently a key technique in new technologies. It has enabled a revolution in
telecommunications with high-speed networks, as well as in medical imaging. There are many
advantages to optical fiber. First of all, optical fiber enables information to be transmitted at high speed.
Attenuation is low, so it can be transported over long distances. But its most significant advantage is its
immunity to electromagnetic interference. Finally, optical fiber provides a discrete, tamper-proof link.
Fig.1. Optical fiber
2.1.2. Principle of optical fiber transmission

Optical fibers operate on the principle of reflection and refraction of light. It is a simple principle based
on the change of index of two media.
Fig.2. Optical fiber principle

Above some index difference, the ray is totally reflected and is no longer refracted at all The limiting
angle of refraction is called the critical angle.
2.1.3. Optical fiber structure
In optical fibers, light is confined in the core by the cladding, which has a lower index than the core.
Fig.3. Optical fiber structure

Optical fiber is a glass cylinder with characteristics that guide light into its core.
Page 303
SEMIT 2023
Fig.4. Light guidance in the core of an optical fiber
2.2. Optical Fiber Transmission systems

As we have seen, fiber has almost constant attenuation over a considerable frequency range (several
thousand gigahertz) and therefore offers the advantage of gigantic bandwidths, making it possible to
transmit the very high data rates (several terabits per second) required by the proliferation of services
and the increased need for image transmission. Compared with coaxial (copper) cable systems, optical
systems also reduce the distance between different repeater-regenerators from a few kilometers to a few
tens of kilometers. As early as 1978, systems working at an optical wavelength of 0.8 μm were
introduced, delivering speeds of between 50 and 100 Mbit/s with a repeater spacing of around 10
kilometers, i.e. around three times further than systems using coaxial cable of equivalent capacity.
The second generation of fiber-optic transmission systems, which appeared in the 1980s, is a direct
result of the development of single-mode fiber and the semiconductor laser at 1.3 μm, the wavelength
at which chromatic dispersion (i.e. the distortion induced in signals by propagation) is minimal. Speeds
in excess of 1 Gbit/s, with a repeater spacing of several tens of kilometers, are then achieved. The ranges
of these systems are limited by the losses of the fiber, 0.5 dB/km in the best case, and the idea then
arose of developing sources emitting at a wavelength of 1.55 μm, for which attenuation is minimal.
However, this gain is destroyed by the effect of chromatic dispersion, which is much stronger at this
wavelength, limiting the bandwidth and therefore the data rate. Progress on both single-mode lasers and
the transmission medium (dispersion-shifted fibers) provided solutions to these problems, and the first
systems working at 1.55 μm appeared at the end of the 1980s, with throughput in excess of 2 Gbit/s.
The appearance of fiber amplifiers at the end of the 1980s, followed by the development of wavelength
division multiplexing, made optics a technique that surpassed all others in terms of both capacity and
transmission quality, in long-distance networks made up of high-capacity arteries linking major cities.
This supremacy has also affected international high-speed links, which are now provided exclusively
by fiber-optic submarine cables [19].
2.2.1. Access transmission systems
In Senegal, a wide variety of transmission media (copper, radio, optical fiber) are used in access
networks; optical technology is playing an increasingly important role with the development of very
high-speed connections (F.T.T.H., Fiber to the Home), although it remains only one of the techniques.
This work deals essentially with high-capacity, long-distance transmission, and therefore with the case
of long-distance networks in Senegal, structured as follows:
Page 304
SEMIT 2023
Fig.5. Fiber optic subscriber access system in Senegal

2.2.2. Transport transmission systems
Since 1976, submarine cable systems have enabled Senegal to connect to the global network and
provide international connectivity. Deployment of terrestrial fiber-optic systems began in 1993. It was
reinforced by the renovation of the national backbone through a policy of sustained investment, which
enabled the country to migrate to terrestrial optical cables. Countries in the sub-region with no coastline
naturally found an opportunity to connect and open up to the world via the Dakar platform. The
deployment programmed is continuing, with 10 GB systems to be rolled out across all links by 2015.
Fig.6. Optical fiber transport system in Senegal

3. Incident Management in Optical transmission systems
3.1. Detection and Analysis of optical fiber incidents
The entity responsible for detecting and analyzing fiber optic incidents is called the NOC a Network
Operation Centre (NOC) is a centralized location where the telecommunications operator's teams can
continuously monitor the performance and health of a network. The NOC serves as the first line of
defense against network disruptions and failures.
Page 305
SEMIT 2023
Fig.7. Network Operation Center

Thanks to the Network Operation Center or NOC (pronounced "noque"), businesses can set up a
monitoring room that gives them total visibility of their network, enabling them to detect anomalies and
take steps to prevent problems or resolve them quickly when they occur. The NOC Operation Center
monitors infrastructure and equipment (from cabling to servers), wireless systems, databases, firewalls,
various network devices (including IoT devices and smartphones), telecommunications, dashboards and
reports. Its management services also include the supervision of customer calls and customer support
ticket systems, and integrate with customers' network tools. It is clear, therefore, that in the IT field, the
NOC plays a considerable role in the quality of the FTTH customer experience.
3.2. Activation of the optical fiber maintenance team
The team responsible for fiber optic maintenance is activated when a service interruption or malfunction
is reported by the NOC, a subscriber, one of the operator's teams or one of its partners. The cause of the
fault may lie in the configuration of the network, the active equipment (Optical Network Switch (OSN),
Indoor Unit (Mini Link), Connector (SFP)), or the physical media (optical fiber, jumpers). If the fault
is due to the breakage of a link or a significant reduction in throughput, the teams measure the
reflectometry of the link and check the connection to locate the fault.
3.3. Corrective maintenance operations on the optical fiber
3.3.1. Reflectometry, Control and Insertion measurements and fault identification
Before connections are made, the fiber optic maintenance team generally takes measurements (in
situations where the cut is not apparent) to identify cut points or points of high stress on the optical link.
Fiber optic measurement involves two types of method: insertion measurement and reflectometry
measurement.
Insertion Method
This method uses a power meter and a calibrated source. It measures the loss in dB between the source
and the receiver. This method is only used over short distances (a few dozen meters).
Fig.8. insertion method
Page 306
SEMIT 2023
(If the link to be tested is already connected to the network, the power meter will display the level in
dB of the optical signal received).
The Reflectometry Method
Optical fiber reflectometry measurements enable network manufacturers to validate their services with
their customers. This method, which takes the form of a curve, makes it possible to: locate and evaluate
each event (fiber, solder, connector, etc.), check the reflectance of connectors and look for faults
(breaks, bends, etc.).
The OTDR (Optical Time Domain Reflectometer) is a device that injects a series of calibrated optical
pulses into the fiber to be tested and analyses the echoes received.
Fig.9. Reflectometer
It displays the result in the form of an attenuation-distance curve as shown below:
Fig.10. Reflectometer result

Interpreting events at connections:
Page 307
SEMIT 2023
Fig.11. Interpretation of results

The measurements require both start and end coils and must be carried out in both directions of
propagation.
3.3.2. Connecting the optical fiber
After probing around the cut-off point and recovering the two ends of the optical cable. The operation
begins with the assembly of connectors (the technician strips the fiber, inserts it into a connector and
fixes it by gluing; as a last step, he polishes the entry face, then cleans it, to minimize optical losses).
The technician then makes the connection by fusion or mechanical means.
Optical fiber soldering or fusion splicing
Soldering always involves preparing the cable so that the optical fibers can be accessed and wired two
fibers are connected by fusing them together using an electric arc. The main equipment required is a
fiber optic welding machine and the associated cleaver. The latter cuts the fiber at a precise angle of
90°. This tool is essential if the two fibers to be connected are to have an identical angle of cut (to
optimize losses).
There are two types of fiber optic welding machine: core-to-core alignment and cladding-to-cladding
alignment. Core-to-core alignment is preferred for single-mode fiber. It optimizes the associated losses.
Cladding-to-cladding alignment is mainly used for multimode fiber applications. The associated
performance is lower, but allows the use of less expensive equipment.
Fig.12. fusion welding machine

Mechanical splicing
Mechanical splicing means aligning the sheaths of two optical fibers without a connector, and holding
them in position. This technique is rarely used, except in a few maintenance cases. In addition to cable
preparation tools, this technique requires a clipper and a mechanical splice. Mechanical splicing is
suitable for all types of optical fiber (single mode and multimode). (Mechanical splicing losses must be
less than 0.5 dB).
Page 308
SEMIT 2023
Fig.13. Mechanical welding machine
3.3.3. Fiber optic incident report

. Once the connections and installation have been completed, all the services must be checked: fiber
quality, cable laying, and link components. An incident report is then drawn up, containing mainly
information (sections affected, services affected, date and time of the cut and restoration, type of cable
affected, team responsible for the repair, company or person responsible for the incident, location of the
cut point, etc.) for monitoring network performance, but also for managing future incidents on the same
section of the network. The incident report is especially important for recovering from the company or
person responsible for the incident the civil engineering costs, the costs of repairing the optical fiber,
the costs of the equipment used to connect the fiber, the costs of the maintenance team's travel, the costs
of the bailiff's report, the costs of financial damage to the operator, etc.
4. Objectives and potential challenges for incident management
Our main challenges for incident management in fiber-optic transmission networks are:
• Detection and classification of incidents in fiber-optic transmission networks using convolutional
neural networks (CNN);
• Detection and classification of incidents in fiber-optic transmission networks using regional
convolution neural networks and Mask R-CNN applied to images;
• Detection and classification of incidents in fiber-optic transmission networks using the YOLO
(You Only Look Once) method applied to images;
• Proposed solution for locating passive nodes in fiber-optic transmission networks.
In order to achieve the objectives of our research work, we will first carry out an exhaustive study of
the main families of models using artificial learning. Next, images of incidents in the fiber-optic
transmission networks of telecommunications operators in Senegal will be collected. Then, using these
collected images, we will develop a YOLO convolution neural network architecture to automatically
detect incidents in fiber-optic telecommunications networks in real time, as well as their impact on the
operator's various services. And to train a convolutional neural network to predict whether or not there
are faults in the optical fiber, based on input images. Finally, to develop an architecture of regional
convolution neural networks with Masque to automatically detect in real time the presence of optical
links that are down or cut on the networks. Collecting the images to form a dataset is our major challenge
for the future.
5. Conclusion
In this paper, we have presented the maintenance of optical fibers in the telecommunications sector in
Senegal. The aim was to set up an automatic, real-time incident management system for an optical
transmission network. To do this, we first looked at optical telecommunications systems. Next, the
management of fiber incidents is presented. Finally, our potential challenges consist firstly of collecting
more images from telecommunications operators in order to propose artificial intelligence models
applied to the images collected. But we also intend to extend this work to artificial intelligence models
Page 309
SEMIT 2023
for determining the locations of passive connection nodes or for minimizing the number of active nodes
in optical transmission networks.
References
1. Kye, M., Lee, S.-M., Choi, K.-M., Lee, Y., & Park, K.-Y. (2021). A Surveillance System of Fiber-
Optic Cables with Multi-Channel DAS and CNN. IEEE Photonics Technology Letters, 33(15),
753-756. https://doi.org/10.1109/LPT.2021.3091145
2. Zhang, B., Zhao, Y., Rahman, S., Li, Y., & Zhang, J. (2020). Alarm classification prediction based
on cross-layer artificial intelligence interaction in self-optimized optical networks (SOON). Optical
Fiber Technology, 57, 102251. https://doi.org/10.1016/j.yofte.2020.102251
3. Zhang, Q., Wang, Z., Duan, S., Jiang, N., Cao, B., & Wu, Y. (2021). An Improved End-to-end
Optical Transmission System Based on Deep Learning. 2021 19th International Conference on
Optical Communications and Networks (ICOCN), 1-3.
https://doi.org/10.1109/ICOCN53177.2021.9563723
4. Kavuncuoğlu, E., Yıldız, F., & Özdemir, A. T. (2022). Artificial Intelligence (AI) algorithms for
evaluation of optical fiber scintillation detector performance. Optik, 258, 168791.
https://doi.org/10.1016/j.ijleo.2022.168791
5. Balmer, R. E., Levin, S. L., & Schmidt, S. (2020). Artificial Intelligence Applications in
Telecommunications and other network industries. Telecommunications Policy, 44(6), 101977.
https://doi.org/10.1016/j.telpol.2020.101977
6. Hirano, A. (2019). Autonomous Network Diagnosis with AI. 2019 Optical Fiber Communications
Conference and Exhibition (OFC), 1-3.
7. Wang, T., Huang, M.-F., Han, S., & Narisetty, C. Employing Fiber Sensing and On-Premise AI
Solutions for Cable Safety Protection over Telecom Infrastructure.
2022 Optical Fiber Communications Conference and Exhibition (OFC), 1-3.
8. Song, H., Li, Y., Liu, M., Wang, K., Li, J., Zhang, M., Zhao, Y., & Zhang, J. (2022). Experimental
study of machine- learning-based detection and location of Eavesdropping in end-to-end optical
fiber communications. Optical Fiber Technology, 68, 102669.
https://doi.org/10.1016/j.yofte.2021.102669
9. Michel, H. (2019, may 20). Google Colab : Le guide Ultime. Le Data Scientist.
https://ledatascientist.com/google-colab-le-guide-ultime/
10. Keras: An open source library for building neural networks. (n.d.). IONOS Digital Guide. Retrieved
October 14, 2022, from https://www.ionos.fr/digitalguide/web-marketing/search-engine-
marketing/what-is-keras/
11. Tanimura, T. (2018). Machine Learning Techniques in Optical Physical-Layer Monitoring. 2018
IEEE Photonics Society Summer Topical Meeting Series (SUM), 55-56.
https://doi.org/10.1109/PHOSST.2018.8456701
12. Lai, J., Yang, A., Zuo, L., & Sun, Y. (2011). Optical performance monitoring in 40-Gbps optical
duobinary system using artificial neural networks trained with reconstructed eye diagram
parameters. 2011 Asia Communications and Photonics Conference and Exhibition (ACP), 1-7.
https://doi.org/10.1117/12.903195
13. Jha, D. K., & Mishra, J. K. (2021). Performance Evaluation of DWDM Optical Transmission System
Using Deep Learning Technique. 2021 IEEE Photonics Conference (IPC), 1-2.
https://doi.org/10.1109/IPC48725.2021.9592951
14. SEP 25, J. S., & Read, 2020 3 Min. (2020, September 25). Getting Started with VGG Image
Annotator for Object Detection Tutorial. Roboflow Blog. https://blog.roboflow.com/vgg-image-
annotator/
Page 310
SEMIT 2023
Odemakinde, E. (2022, mars 19). Everything about Mask R-CNN : A Beginner’s Guide. Viso.Ai.
https://viso.ai/deep-learning/mask-r-cnn/
15. Dutta, A., & Zisserman, A. (2019). The VIA Annotation Software for Images, Audio and Video.
Proceedings of the 27th ACM International Conference on Multimedia, 2276-2279.
https://doi.org/10.1145/3343031.3350535
16. « OCTO Talks ! », OCTO Talks ! 2 mars 2023. https://blog.octo.com/ (accessed on 2 March 2023).
17. "Optical fiber transmission Understanding optical fiber Wavelet - OTDR Optical reflectometer
Optical welder Optical power meter Optical source Optical power meter Optical telephone VFL
GPON FTTH". https://www.soudeuse-optique.com/transmission-par-fibre-optique.html (4 June
2023).
18. "TELECOMMUNICATIONS - Optical technologies, Optical fiber transmission systems -
Encyclopedia Universalist. https://www.universalis.fr/encyclopedie/telecommunications-
technologies-optiques/5-systemes-de-transmission-sur-fibre-optique/ (4 June 2023).
Page 311
SEMIT 2023
Business Process Linguistic Modelling - Design – Process Design Part I

Jozef Stašák
Intercedu a.s. Pezinok, SR – corresponding author, e-mail jozef.stasak@intercedu.com
Peter Schmidt
Faculty of Informatics, University of Economy, Bratislava, e-mail schmidt@atg-slovakia.sk
Jaroslav Kultán
Faculty of Informatics, University of Economy, Bratislava, e-mail jkultan@gmail.com
Abstract
At present, the business process management (BPM) seems to be significant and important within any
firm, company and institution, while the business process (BP) to be managed should be quantified and
modelled in order it could be managed correctly and efficiently. There are several BP modelling
approaches, while two of them are applied in BPM as well. The first one is denoted as the standard
applied round the world and I called standardized approach. The second one is based on the fact that
any BP might be described by text in natural language (TNL Text), which contains a set of standardized
sentences denoted as the logical sentences, while any sentence contains appropriate linguistic and
lexical units The question is: “How to quantify them and how they should be quantified and applied for
BP modelling purposes?”
There are many approaches and methods which enable converting the TNL text to BP models and one
of them deals with specialized sets, the content of which is given by the above-mentioned linguistic
units, while those sets are denoted as the linguistic sets. However, there is a co-operation among them
regulated by Principal Business Process Linguistic Equation (PBPL Equation), which creates basis for
BP linguistic modelling related to their structure, functionality, and metrics, as well as their information,
knowledge based and organization support as well.
Key words: Business Process Modelling, Standardized Approach, Linguistic Approach, TNL Text,
Linguistic Sets, PBPL Equation, Expert System, Aid for Managers.
1. Introduction
There are two principal business process modelling approaches. The first one developed by Prof. Scheer
and Software AG company is based on ARIS Methodology and seems to be the standard applied round
the globe, when considering business process modelling, while it is often denoted as the standardized
approach. However, the business process (BP) model might be described via text in a natural language
(TNL Text) - (e.g., Slovak, Spanish, etc.) as well, while TNL Text contains a set of logical sentences,
which describe the BP structure and functionality (2,3). The question is how to convert the TNL text
into business structure and functionality model. And finding the answer related to that question is that
contribution main goal.
The TNL text seems to be a set of various linguistic and lexical unit, which should be categorized into
subordinated sets denoted as linguistic sets, which crate basis for converting the BP TNL text into a BP
model. Their establishment and definition their features seem be the first patrial aim. In order to apply
the linguistic sets for quantification of BP structure and functionality, they to co-operate properly and
an appropriate equation should be constructed (see also Section 3), which could be applied to quantify
that co-operation and regulate the BP structure a functionality. This is the second partial aim, which
should be achieved (see also Section 4.2). To find a general and applied solution of that equation seems
to be the third partial aim (see also Section 4.2).
However, the PBPL Equation deals with BP information, knowledge-based, organizational support as
well, while those aspect are not discussed within that contribution because contribution range reasons.
Page 312
SEMIT 2023
However, the contribution structure corresponds to the above-mentioned partial aims as well, while the
Section 2 deals with the state-of-the-art description and the Section 3 and 4 correspond to fulfilment
pre-defined main goal and partial aims.
2. BPLM State of the Art
2.1 General overview

Any business begins with its mission statement and objectives determination and establishment together
with appropriate key performance indicator values (KPIs), while this is one side of the firm or company
business strategy goals and aims. However, a set of different activities should be done to fulfill them as
well, and they have to be administrated and managed too. On the other hand, the BP administration and
management requires to know their structure and functionality, which affects their performance related
to the above-mentioned KPIs and here the business process modelling seems to be important too.
2.2 Textual BP descriptions versus process modelling
There are several modelling techniques, which enable constructing models with respect to particular
viewpoints.
KETTINGER et al. (2) present an important overview of methods, techniques, and tools used in
Business Process Re-engineering (BPR). This list does not give detailed descriptions of the techniques
nor the tools. However, it has been the starting point of the research presented in this paper, which gives
a more thorough overview with detailed analysis of the mentioned techniques in (3) and others. (5)
Different techniques are more suitable to certain purposes, e.g., one thing is a model, which describes
the process, and another a model to build a system to control the process. For instance, distinguish
between two uses of business process models: one for traditional software development, and another to
restructure business processes (5)
Several techniques have been developed for this purpose, where Mining Process Models from Natural
Language Text seems to be an important approach as well (4)
Any BP might be described in textual form or in form of text in natural language (TNL Text), which
consists of several sentences denoted as logical sentences quantified via specialized sets, a content of
which is created by linguistic units (e.g., text strings with an appropriate semantic meaning variables),
and they are denoted as linguistic sets, (9,11), which might be applied when modelling business
processes with the use of linguistic approach, and that approach will be discussed in more details within
next sections.
3. Research methods
This section contains a description of terms and principles business related to BPLM process 1, where
linguistic sets create basis, when providing semantic analysis of TNL Text, which describes structure
and functionality of the business process to be modelled while issues, formulas and relations postulated
within that section will be applied, when designing a BP logical model with use of BPLM approach as
well (see also Section 4).
1
BPLM Business Process Linguistic Modelling
Page 313
SEMIT 2023
3.1 Business Process Modelling Linguistic Approach

3.1.1 Business process textual and model description
Let us consider the business process (BP) to be modelled, the structure and functionality is described
via adequate TNL Text, which consists of logical sentences V(i) i=1, 2…n and each logical sentence
contains different lexical or linguistic units with various structures and properties. While they might be
quantified via {VLS (i, j1)} linguistic set, which is carrying an adequate semantic meaning concerned
with the TNL Text objective and formula (3.1) might be postulated (15)
{TNL (i, j)} V i, j = {[LS1 (i, j)]} (1)
where i = 1 … n serial number of V(i) logical sentence, which creates an integral part of the TNL text
and j1 = 1 … m1 serial number semantic, syntactic or grammar element contained within V(i) logical
sentence and LS1 is an appropriate linguistic set, while the logical sentence semantic meaning might be
quantified with the use of formulas (2) and (3)
where SNTL is a semantic value the entire TNL text, which consists of logical sentences {V(i)}. The
{V(i)} logical sentence semantic value might be quantified via adequate {VLS1(i, j1)} linguistic set,
while formulas (4) might be postulated/
[Tbex(i, j4)] is a set of new linguistic terms (Tbe terms), which should be explained via V(i) logical
sentence
[Retx(i, j5)] is a set of relations among Tbe terms.
In order to explain a semantic meaning of Tbe terms a set of previously existing principal terms (Pet
terms) is needed. Those terms are denoted as Pet terms (STAŠÁK, 2004a) and represent {Petx (i, j3)}
set elements, while formula (3.5) might be postulated (15).
Page 314
SEMIT 2023
The {[Petx (I, j3)]}, {[Tbex (I, j4)]} and {[Retx (I, j5)]} linguistic sets do not seem to be sufficient for
quantification of structure and functionality related to the business process to be modelled and the
operator should be quantified via {[Pe (I, j)]} linguistic set as well and formula (6) and formulas
(7a, 7b and 7c) might be postulated
3.1.2 The {VLS1(i, j1)} linguistic set semantic value

When looking at formulas (2), (4), (7a), (7b) and (3.7c), we can see that the V(i) logical sentence
semantic meaning vale might be represented by Tbe and Ret terms contained within {Res1(i, j7)}
linguistic set, while formula (8) might be postulated (15).
However, the {Svm(i)} is a linguistic set as well and formulas (9) and (10) might be postulated.
Page 315
SEMIT 2023
With respect to formula (3.7a), we can see that the {Petx(i, j3)} linguistic set elements serve for
explanation of Tbe terms and the {Pe(i, j6)} linguistic set elements serve for explanation of Ret terms
(see also formulas (11a) and (11b))
Of course, the V(i) logical sentence consists of text strings, which represent elements of {Words (i, j8)}
linguistic set, i = 1 … n, j8 = 1 … m8 – serial number of the {Words (i, j8)} element, while formula
(3.12) might be postulated
where
[Ts (I, j)] linguistic set elements represent the V(i) logical sentence semantic aspects.
[Tsem (i, j)] linguistic set elements represent the V(i) logical sentence syntactic and grammar aspects
(15).
3.1.3 Business process structure and functionality quantification with the use of linguistic sets
Any business process (BP) is represented by two types of structure: (a) vertical and (b) horizontal. In
that section, there will be discussed issues related to BP horizontal structure and metrics, when
modelling it with the use of business process modelling linguistic approach (BPLM approach).
However, each BP horizontal structure is created by set of business process functions (BPFs), while
any BPF consist of inputs, outputs, and transformation operands, while all three categories might be
quantified via appropriate linguistic sets and the transformation operand linguistic set consists of two
subsets. The first subset content is created by elements closely related to rules, which regulate
transformation of BP inputs to BP outputs and the second subset is created by elements closely related
transformation tools, which enable doing that. On the other hand, the main goal of any BP is to generate
outputs predefined via strategy key performance key values (KPI) and that mission should be quantified
as well, via set BP performance indicators (BPPIs), which create an integral part of BP external and
internal metrics (14). Subsequently, the KPIs and BPPIs values are being compared and the business
process functionality and performance might be evaluated. In order to achieve that, the {[Petx (i, j)],
{[Pe (i, j)]}, {[Tbex (i, j)]} and {[Retx (i, j)]} should communicate among each other and that
communication is being quantified and regulated by so called Principle Business Process Linguistic
Equation (PBPL Equation) a general for of which indicates formula (13)
{[Petx (i, j)]}  {[Pe (i, j)]} = {[Tbex (i, j)]}  {[Retx (i, j)]} (13)
Page 316
SEMIT 2023
However, the strict analytical solution of that equation is not possible and therefore a semi-quantitative
approach has been selected in order to achieve that could design a BPLM approach logical model (see
also Section 4) which should create basis for implementation of BPLM Approach application program2.
4. BPLM Process Design – Process View
4.1 Introductory information

There are two basic approaches concerned with business process horizontal structure modelling: (a)
ARIS Methodology developed and established by Prof. Scheer and provided by Software AG company,
while this approach seems to be a worldwide standard related to business process modelling (9). (b)
Business Process Linguistic Modelling Methodology, while general considerations and basic principles
concerned with BPLM Approach are described within Section 3
There are the following methodological aspects closely related to BPLM Approach:
• Any business process (BP) is represented by two structure types: (a) vertical structure
and (b) horizontal structure/
• Any BP vertical structure indicates core, main, subordinated, and elementary processes.
• Any business process model concerned with BP horizontal structure deals with three
views: (a) BP view, (b) information support view, (c) knowledge-based support view
and organization support view.
• Any BP horizontal structure is created by BP functions, while any BP Function (BPF)
has its own external and internal metrics.
• BP external metrics is closely related to BPF inputs and outputs.
• BPF internal metrics is concerned with production devices (Pdk), production tools (Ptk)
and production human resources (Hrk).
• BPLM Approach is based on existence of linguistic sets and PBPL Equation (see also
Section 3)
• All the above-mentioned aspects are being quantified via linguistic set apparatus with
respect to following considerations.
4.2 Business process horizontal structure

4.2.1 Business process and Business Process Function
Let us consider any business process horizontal structure created by set of adequate BFs and quantified
via linguistic sets, while formula (14) might be postulated:
m1
{[BP (i, j)]} =  {[BPFk (i, j)]} (14)

k=1
where k=1…. m1 is an index, which represents a number of BPFs3, which the business process consists
of. On the other hand, any BPF contains three subsets4 (see also formula (15),
2
At present, an adequate prototype of that application program is being implemented and operated.
3
BPF – Business Process Function
4
The terms linguistic sets and sets seem to be equivalent from semantic point of view, if no other specification is postulated.
Page 317
SEMIT 2023
{[BPFk (i, j)]} = {[Petxk1,l1 (i, j1)], {[Pe (i, j2)], [Tbexl2 (i, j3)], [Retx (i, j4)]} (15)
where
i=1…n indicates a BP serial number within superior BP set k1=1….m1 indicates input serial
number related to actual BPF
j1=1….m3 indicates a number of input attributes and values, which create an integral part of an
adequate input
j2 = j2 (j21, j22) j21 = 1,2….m21 – number of rules, which regulate transformation of BPF inputs to
BPF outputs
j22 = 1,2….m22 – number of tools, which provide transformation of BPF inputs to BPF outputs
l2 = 1…m4 indicates a number of outputs generated by adequate BPF
j3=1…. m5 indicates a number of output attributes and values, which create an integral part of
an adequate output.
j4 =1,2…..m6 indicates a number of relations among elements contained within {[Tbexl2 (i, j3)]}
set
4.2.2 Business Process Function Metrics
Any Business Process Function (BPF) has its own metrics, which includes external and internal
metrics, while external metrics consists of input {[Petx (i, j)]}, and output {[Tbex (i, j)]}, and {[Retx
(i, j)]}, linguistic sets and internal metrics includes linguistic sets, which contain data related to
production devices {[Dev (i, j)]}, production tools {[Tools (i, j)]},and human resources {[Hr (i, j)]}
(STAŠÁK, 2017).
Business Process Function External Metrics
Let us consider the BPF external metrics, which creates an integral part of technological BP (glass
article forming GLAF), as for instance which includes {[Petxk (i, j)]}, and {[Tbexl1 (i, j)]}, and {[Retxl2
(i, j)]} linguistic sets with attributes with respect to formula (4.3) and let us investigate a relation an
appropriate relation
{[Petxk (i, a1, a3, a4, a5, a6, a7, a8, a9 ………… am-2, am-1, am)]} (c1, d1) = {[Tbexl1 (i, bgood, brepair, bwaste,)]}
(16)
In the next step let us select the element at (target element) and let us try to create an adequate fuzzy
set (see also formula (4.4a and 4.4b)
at {[Petxk (i, a1, a3, a4, a5, a6, a7, a8, a9 ………… am-2, am-1, am)]} (17a)
at
/a1, at /a2, …….
at
/am, = acut (17b)
The same step should be done related to {[Tbexl1 ana (m1) (i, bgood, brepair, bwaste,)]} linguistic set
bt
/ bgood, bt / brepair, bt
/ bwast4e, = bcut (17c)
where
bt = bgood,+ brepair,+ bwaste (17d)
Furthermore, let us create a relation between both the fuzzy set cuts (see also formula (17e)
Page 318
SEMIT 2023
acut * kem = bcut (17e)

while the kem seems to be external metric relation variable
Now, a graphic (see also Fig. 1) or table representation of that relation might be done, while it represents
a semiquantitative PBPL Equation solution.
Fig.1. A graphic representation of external metric relation variable

Source: the authors
The above-mentioned relation might have a linear or non-linear progress.
Business Process Function Internal Metrics
As mentioned above, the BPF internal metrics contains linguistic sets closely related to production or
technological devices {[Pdk3 (i, j)]}, production tools {[Ptk4 (i, j)]}, and human resources {[Hrtk5 (i, j)]}
(14).
However, there is necessary to create appropriate subsets the relations similar to relations (18a ….18e)
and formulas (4.5a, 4.5b and 4.5c) might be postulated.
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/ {[Pdk’’ (i, j1’’’, j2’’’, …. jm1’’’)]} = b1 (18a)
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/ {[Ptk’’ (i, j1’’’, j2’’’, …. jm1’’’)]} = c1 (18b)
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/ {[Hrk’’ (i, j1’’’, j2’’’, …. jm1’’’)]} = d1 (18c)
Now, the internal metric relation variables might be postulated (see also formulas (19a, 19b and 19c) .
b1* kim1 = bcut (19a)

c1* kim2 = bcut (19b)
d1* kim3 = bcut (19c)
However, the same interpretation like in the case of external metrics might be applied for internal
metrics as well, while it might be useful for BPF functionality purposes too.
Page 319
SEMIT 2023
4.3 Business process horizontal structure model – the process view

4.3.1 Process view
The Process view visualizes a sequence of BPFs, the actual BP consist of together with adequate BPF
external and internal metrics relation variables with respect to formulas (4a….4e, 5a…5c, and
6a…6c) and might be visualized via Fig. 2a and Fig.2b)
Fig.2a. Business process horizontal structure model – process view

Source: the authors
The BP or BPF external metrics, which represents relations among BP or BPF inputs and outputs is
shown in Figure 2b.
The BP or BPF internal metrics, which represents relations among BP or BPF production devices Pd
and outputs, production tools Pt and outputs and human resources Hr and outputs are very similar and
might be derived easily as well.
Fig.2b. The BP or BPF external metrics, which represents relations among BP or BPF inputs and
outputs. Source: the authors
Page 320
SEMIT 2023
4.3.2 How the BPF metrics might be quantified for BP and BPF functionality and
performance simulation purposes?
As mentioned above any BPF external and internal metrics might be quantified via four equations
postulated as follows:
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/{[Petxk (i, j1, j2, …. jm1)]} = a1 (20a)
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/ {[Pdk’’ (i, j1’’’, j2’’’, …. jm1’’’)]} = b1 (20b)
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/ {[Ptk’’ (i, j1’’’, j2’’’, …. jm1’’’)]}= c1 (20c)
{[Tbexk (i, j1’, j2’, …. jm1’)]}
/{[Hrk’’ (i, j1’’’, j2’’’, …. jm1’’’)]} = d1 (20d)
while formula (10a) is concerned with BPF external metrics and formulas (10b) ….(10d) are closely
related to BPF internal metrics
BP and BPF external metrics simulation
In general, the BP and BPF external metrics is concerned with BP o BP function and might be
quantified via formula (21).
Let us consider a set of situations concerned with BPs and BPFs
BP (i) =  BPF (i, j) = {[Tbexk (i, j1’, j2’, …. jm1’)]} /{[Petxk (i, j1, j2, …. jm1)]} = a1 (i) (21)
where
i=1, ….n number of BP
4.3.3 Example
Let us consider a business process BP (BPGAF – Glass article forming) {[ BPGAF (i, j)]} which
should generate glass articles (GA – bowel), which consists of the following functions:
• Glass melt bowl-pipe taking (GMBPT)
• Glass article blowing out (GABO)
• Glass article rotating (GAROT)
• Glass article cutting from bowl-pipe (GABPC)
BPGAF – process inputs
i= BPGAF
{[Petx (i,j)]} = {GM [BPGAF ((GM_Tmp, Val1), GM_Vis, Val2), (GM_Qua, Val3)]}} (22a)
where
(GM_Tmp, Val1) - Glass melt temperature value
(GM_Vis, Val2) - Glass melt viscosity value
(GM_Qua, Val3) - Glass melt quantity value
BPGAF – process outputs
{[Tbex (i,j)]}= {[GA ((GA_Good, V1), GA_Repair, V2), (GA_WASTE, V3)]}} (22b)
where
GA_Good, V1 - number of good glass articles
GA_repaired V2 - number of glass articles to be repaired
GA_waste, V3 - number of waste glass articles
(GA_Total, V4) = GA_Good, V1) + (GA_repaired V2) + (GA_waste, V3) (22c)
(GA_Total, V4) (GA_Total, V4) (GA_Total, V4)
/ GA_Good, V1), / GA_repaired, V1), / GA_waste, V1) = a11 (BPGAF) (22d)
Page 321
SEMIT 2023
(GM_Qua, Val3)  {[Petx (i,j)]}

(GM_Tmp, Val1)  {[Petx (i,j)]}
(GM_Vis, Val2)  {[Petx (i,j)]}
(GM_Qua, Val3)
/ (GM_Tmp, Val1) , (GM_Qua, Val3) / (GM_Vis, Val2) = a12 (BPGAF) (22e)
Fig. 3 Graphic representation of formula (22e) – line version

Source: the authors
4.4 Discussion
There are two principal business process modelling approaches. The first one developed by Prof. Scheer
and Software AG company is based on ARIS Methodology and seems to be the standard applied round
the globe, when considering business process modelling, while it is often denoted as the standardized
approach. However, the business process (BP) model might be described via text in a natural language
(TNL Text) -(English, Slovak, German, etc.) as well, while the TNL Text contains a set of logical
sentences, which describe the BP structure and functionality. The questions might be postulated (1)
“How that TNL Text should be converted to adequate BP Model?” and (2) “How the BP model element
might be transferred to TNL Text?” Now, we are looking for an answer related to the first question and
the second question is not being discussed within our contribution.
When looking for the answer related to the first question, different approaches and methods might be
applied as stated within Section two. However, all of them have a common platform based on the fact
that they operate with linguistic units (words or sentences) and various types of lexical units. We have
tried to make an interconnection between mathematics (theory of sets) and linguistic units and we have
established so called linguistic sets (LS) (see also Section 3) and applied them for modelling
of BP horizontal structure, where input, transformation operand and output linguistic sets have been
introduced and the PBPL equation which regulates relations among them and quantifies BPF and BP
structure and functionality, while that equation has its own general form and might be applied for
solution of different situations related to BPF and BP modelling, while a strict quantitative solution
related to that equation is not possible, therefore so called semiquantitative approach had been selected
and applied for process, view (see also Section 4).
Page 322
SEMIT 2023
5. Conclusions and future work

This contribution contains design how the linguistic approach might be applied for BP and BPF
modelling, where the PBPL Equation plays a role of principle importance and its element create a good
basis for BP and BPF model implementation. In general BP and BPF horizontal structure model is
represented by process, information, knowledge-based and organizational view, however that
contribution deals with process view only. However, the same approach might be applied related to
quantification of BP and BPF information, knowledge based and organizational view as well a might
create a content of an individual contribution too. As for implementation aspects, standard WEB
oriented databases and graph databases develop by Ontotext, will be applied in order to develop an
appropriate application program for utilization in common practice what seems to be the next step in
our research work.
References
1. Davies, D. Jindal-Snape, D, Collier, CH., Digby, R., Hay, P., Howe, A. 2013 Creative learning
environments in education—A systematic literature review Thinking Skills and Creativity Volume
8, April 2013, Pages 80-91, available at Creative learning environments in education—A systematic
literature review - ScienceDirect
2. Kettinger, W.J., Teng, J., Guha, S., 1997a. Business process change: A study of methodologies,
techniques, and tools. Journal of Management Information Systems 14 (1), 119–154.
3. Kettinger, W.J., Teng, J., Guha, S., 1997b. Appendices business process change: A study of
methodologies, techniques, and tools. Management Information Systems Quarterly Archivist 14 (1),
Appendices 1–8
4. Riefer, M., Ternis , S.F., Thaler, T. 2016 Mining Process Models from Natural Language Text: A
State-of-the-Art Analysis available at
https://www.dfki.de/fileadmin/user_upload/import/8229_RieferTernisThaler2016_Text2ModelMini
ng.pdf
5. Ruth, S. Saven, A 2003 Business process modelling: Review and framework Int. J. Production
Economics 90 (2004) 129–149 available at
http://secure.com.sg/courses/ICT353/Session_Collateral/TOP_04_ART_03_ARTICLE_AGUILAR_
Biz_Proc_Modelling.pdf
6. SANER, R., & YIU, L. 2005. Swiss Executives as Business Diplomats in the New Europe: Evidence
from Swiss Pharmaceutical and Agro-Industrial Global Companies. Organizational Dynamics, 34(3),
298-312.
7. Stašák, J. 2004 ‘A contribution to semantic text analysis’, Electronic Computers, and Informatics
ECI 2004, 22–24 September 2004, The University of Technology Košice, Department of Computers
and Informatics of FEI, Košice – Herľany, SR, pp.132–144, ISBN 80-8073-150-0.
8. Stašák, J. 2004 ‘Application of fuzzy sets apparatus in image semantic analysis’, Proceeding of 5th
International Carpathian Control Conference Faculty of Mechanical Engineering and Robotics, ICCC
2004, 25–28 May, AGH-UST Krakow, Zakopane, Poland, pp.255–260, ISBN 83-89772-00-0.
9. Stašák, J. 2010 Modelovanie biznis procesov s využitím systému ARIS, (ARIS Business Process
Modelling) Bratislava Ekonóm
10. Stašák, J: 210 Lingvistický prístup ku kvantifikácií a modelovaniu biznis procesov (Business
process modelling – linguistic approach) Monografický zborník Katedry aplikovanej informatiky,
Vydavateľstvo EKONOM 2010, ISBN 978-80-225-3110-8
11. Stašák, J 2012 Business Process Quantification and Modeling – Linguistic Approach – Several
Theoretical Aspects American International Journal of Contemporary Research (AIJCR) 2012, No.4,
p. 85-100, ISSN 2162-139X
Page 323
SEMIT 2023
12. STAŠÁK, J. Business Process Linguistic Modeling – Philosophy & Principles Computer Science
and Information Technology 3(5): 198-213, 2015 DOI: 10.13189/csit.2015.030505
13. Stašák, J., Vaníčková, R., Grell, M., 2015. Business Process Modeling Linguistic Approach –
Problems of Business Strategy Design Information Technology 3(5): 282, 2015 DOI:
10.13189/ujm.2015.030702
14. Stašák, J. 2017 Contribution to Linguistic Modeling of Business Process Structure, Functionality
and Performance Metrics Journal of Scientific Research & Reports 13(3): 1-16, 2017; Article no.
JSRR.30233 ISSN: 2320-0227
15. Stašák, J., Schmidt, P. 2018 Business Process Modelling Linguistic Approach Application in Public
Administration (Self-governmental) Int. J. Business Performance Management, Vol. 19, No. 2, 2018
Stašák, J., Schmidt, P 2019 Semantic technology and linguistic modelling in business strategy design
and evaluation." International Journal of Business Information Systems 31.2 (2019): 170-194.
Page 324
SEMIT 2023
SP07: Supply Chain Network Design; Considering Recent Trends in Sustainability

and Circular Economy
Developing a Cradle-to-Gate Product Carbon Footprint Calculator with ISO 14067

Yasmine Benomar and Sanae Larioui
benomaryasmine@ensi.ma
larioui@ensi.ma
Abstract
This paper presents a comprehensive research endeavor focused on calculating the carbon footprint of
products. To achieve this objective, we employed the internationally recognized ISO 14067 guidelines
as our foundational methodology. Additionally, this study proposes a user-friendly tool, a customized
product carbon footprint calculator developed using Microsoft Access, which enabled us to streamline
and simplify the calculation process.
Our study spans the entire lifecycle of a product, from the initial extraction of raw materials to the final
manufacturing stages, embracing the cradle-to-gate perspective. What sets our approach apart is the
precision and clarity offered by our product carbon footprint calculator. It not only calculates emissions
but also presents them through visualization tools and dashboards for each crucial phase: raw material
manufacturing, transportation, and product manufacturing. This versatility is a significant advancement,
as it allows organizations to gain a more holistic view of their environmental impact throughout the
product's life cycle.
In conclusion, our research provides a valuable contribution to the field of environmental sustainability.
The accurate calculation of product carbon footprints, considering the entire cradle-to-gate phase, offers
insights that are indispensable for organizations seeking to optimize their environmental impact and
decision-making processes. By implementing our calculator and following ISO 14067 guidelines,
organizations can take proactive steps towards reducing their carbon footprint and, in turn, contribute
to a more sustainable and eco-friendly future.
Keywords: Carbon footprint, ISO 14067, Product carbon footprint calculator, environmental impact
optimization, cradle-to-gate perspective.
1. Introduction
The concept of measuring and reducing carbon footprints has gained paramount importance in the
global context of environmental sustainability. As industries and businesses strive to mitigate their
impact on climate change, an essential aspect of this endeavor is the accurate calculation of the product
carbon foot-print, encompassing the entire spectrum from raw material extraction to the facility gate.
This research seeks to address this critical need by leveraging the internationally recognized ISO 14067
guidelines [1] and developing a bespoke product carbon footprint calculator using Microsoft Access.
The notion of carbon footprint assessment has been a subject of extensive research and practical
implementation in recent years. Various studies have emphasized the significance of carbon footprint
reduction as an integral part of corporate sustainability initiatives [2]. Furthermore, the adoption of ISO
standards in environmental management systems has been instrumental in standardizing and
streamlining carbon footprint calculations [3]. However, existing methodologies primarily focus on
specific phases of a product's lifecycle, often excluding the comprehensive cradle-to-gate analysis that
encompasses raw material extraction and transportation.
Despite advancements in carbon footprint assessment methodologies, there remains a significant gap in
our understanding of how to accurately calculate and account for the complete product carbon footprint,
covering every stage of the supply chain. While we have established frameworks such as ISO 14067
Page 325
SEMIT 2023
for guidance, the development of practical and customizable tools for organizations to conduct such
assessments comprehensively is still in its nascent stages. Moreover, the integration of Microsoft
Access as a platform for developing a product carbon footprint calculator represents an innovative
approach, the effectiveness of which remains to be explored. This research endeavors to bridge the gap
in our understanding of comprehensive product carbon footprint assessment and introduces a novel tool
to facilitate more accurate and insightful environmental impact analysis.
2. Methods
This study was conducted within a practical context, focusing on an organization in the automotive
sector. We meticulously followed a structured series of steps to calculate and develop a comprehensive
product carbon footprint calculator.
2.1. Defining and Preparation
Setting Objectives: To initiate this study, the first step involved setting clear objectives aligned with
the organization's goals. The overarching aims included providing customers with transparency
regarding the environmental impact of their purchases, addressing pressures from suppliers and partners
to reduce carbon emissions, and integrating environmental considerations alongside cost and quality in
procurement decisions.
Choosing Products: Product selection was a critical aspect of this study. We opted to calculate the
carbon footprint of products based on their sales volume and mass production [4]. This strategic choice
ensured that our calculations would be applicable to a significant portion of the organization's output,
making the results more impactful.
Engaging Suppliers: Engaging with suppliers was essential to gather data on the carbon footprint of
raw materials. While some suppliers may have been reluctant to share this information, we sought
collaboration to access accurate and comprehensive data, facilitating a more precise calculation [5].
Building a Process Map and Checking Boundaries: To establish a clear framework for our
calculation, we constructed a detailed process map for all the supply chain [6]. This map outlined the
scope of our analysis, extending from raw material extraction to the facility gate. It encompassed various
stages, including manufacturing processes, transportation, energy consumption, and the involvement of
workers. Defining these boundaries was essential to ensure a comprehensive assessment.
2.2. Product Carbon Footprint Calculation
Step 1: Data Collection: The foundation of our product carbon footprint calculation was data collection.
While we primarily relied on the organization's internal data, the lack of information from some
suppliers necessitated the use of secondary data from reliable sources [7]. We rigorously stored and
organized this data within Microsoft Access queries for easy access and management.
Step 2: Data Modeling: To make the collected data suitable for carbon footprint calculations, we
engaged in data modeling [8]. This involved processes such as unit conversion, normalization, and
categorization. Additionally, we documented the formulas and equations used in the calculations,
ensuring transparency.
Step 3: Calculating the Product Carbon Footprint: We developed a series of forms within Microsoft
Access to facilitate the calculation process. These forms allowed for easy access and control over inputs
and provided options to ensure accuracy. The calculation process encompassed various forms, including
raw material manufacturing carbon footprint, transportation (by road or sea), product manufacturing
carbon footprint, and the final product calculation form. Furthermore, the results were stored in a
dedicated form with dashboards for analysis and interpretation.
Step 4: Validating Results: Parallel to the product carbon footprint calculation within Microsoft Access,
we conducted validation by applying the same calculations using the Life Cycle Assessment (LCA)
software, GABI [9]. This cross-validation process ensured the reliability and accuracy of our results,
enhancing the credibility of our findings.
Page 326
SEMIT 2023
3. Formulas
3.1. Calculation of raw material manufacturing:
In determining the product's carbon footprint, we first assess the emissions associated with raw material
production. Research, such as Weidema et al., provides insights into emissions linked to materials and
manufacturing [10]. The formula for raw material manufacturing carbon footprint (Kg CO₂ eq) is:
Raw material CF (Kg CO₂ eq) = [Emission factor (Kg CO₂ eq/Kg)] x [Material quantity] x [Material
weight (Kg)] (1)
The formula (1) helps calculate the carbon footprint for individual parts, where the emission factor
represents emissions from the manufacturing process or material type, material quantity reflects the
quantity used, and material weight indicates the material's weight in production.
3.2. Calculation of part transportation (by road and sea)
To assess the carbon footprint linked to part transportation, we consider both road and sea transport
modes.
Road Transport: The formula for calculating the carbon footprint of road transport (Kg CO₂ eq) is
derived from "Guidelines for Measuring and Managing CO₂ Emission from Freight Transport
Operations" by ECTA-European Chemical Transport Association [11].
Transport by Road CF (Metric ton CO₂ eq) = [Fuel consumption (liter/km)] x [Emission factor (Kg
CO₂/Liter fuel)] x [distance traveled (km)] / 1000 (2)
The equation above (2) involves factors such as fuel consumption that represent how much fuel the
vehicle uses to travel a certain distance, the emission factor tells us how much carbon dioxide is
produced for every liter of fuel burned, and the distance traveled represent the total distance the vehicle
covers while transporting the part. Moreover, to measure the carbon footprint of part transportation by
road, we use proportional reasoning using those formulas:
Container transport (Metric ton CO₂) = [Container volume (m³)] x [Transport by Road CF (Metric ton
CO₂)] / [container truck volume (m³)] (3)
Box transport (Kg CO₂ eq) = [Box volume (m³)] x [Container transport (Metric ton CO₂) x 1000] /
[container volume (m³)] (4)
Part transport (Kg CO₂ eq) = [Part volume (m³)] x [Box transport (Kg CO₂ eq)] / [box volume (m³)]
(5)
These calculations are performed individually for containers (3), boxes and packaging (4), and each
part (5), enabling us to determine carbon emissions from raw material transportation by road.
Sea Transport: The formula for calculating the carbon footprint of sea transport is also based on ECTA
[11], allowing us to estimate CO₂ emissions during sea transportation.
Transport by Sea CF (Metric ton CO₂ eq) = [ship capacity (Metric ton) x Emission factor (g CO₂/Metric
ton-km) x distance traveled (km)] / 1,000,000 (6)
Ship capacity tells us how much cargo the ship can carry in one trip, the emission factor tells us how
many grams of carbon dioxide are produced for every ton-kilometer of cargo transported by the ship,
and the distance traveled tells us the total distance the ship covers while transporting the part.
Furthermore, to determine the carbon footprint of each part's sea transportation, we use the proportional
reasoning formulas again to:
Container transport (Metric ton CO₂) = [Container volume (m³)] x [Transport by Sea CF (Metric ton
CO₂)] / [Container ship volume (m³)](7)
Page 327
SEMIT 2023
Box transport (Kg CO₂ eq) = [Box volume (m³)] x [Container transport (Metric ton CO₂) x 1000] /
[Container volume (m³)] (8)
Part transport (Metric ton CO₂ eq) = [Part volume (m³)] x [Box transport (Kg CO₂ eq)] / [Box volume
(m³)] (9)
The calculations are performed individually for the container (7), packaging/box (8), and each part (9).
This approach allowed us to determine the carbon emissions generated during the transportation by sea
of raw materials.
3.3. Calculation of Product manufacturing
The calculation of the carbon footprint associated with product manufacturing is a multifaceted process
that involves several components. This section outlines the calculation of the carbon footprint of human
resources (HR), the electrical energy (EE), and raw material (RM).
Human Resources Carbon Footprint Calculation: The carbon footprint of human resources (CF HR) is
an essential aspect of product manufacturing assessment. This calculation accounts for the emissions
generated by the human labor involved in the manufacturing process [12]. To quantify CF HR, a specific
formula is used:
Human Resources (Kg CO₂ eq) = [Cycle time (h)] x [Number of workers] x [Emission factor (Kg
CO₂/person-h)] (10)
The equation above (10) takes into account the following factors, the Cycle time (h) which is the total
time required for the manufacturing process. It reflects the efficiency and duration of production.
Moreover, the total number of individuals involved in the manufacturing process. This parameter
accounts for the labor force's size, and the Emission factor (Kg CO₂/person-h) represents the carbon
emissions associated with the activities of an individual worker during one hour of work.
Calculation of Energy Consumption Carbon Footprint (CF EE): This section focuses on the energy
consumption component, specifically electricity. The CF EE represents the carbon emissions resulting
from the electricity consumed during the manufacturing of the product [13]. It is calculated using the
following formula:
CF EE (Kg CO₂) = [Energy consumed (KW)] x [Emission factor (Kg CO₂/KWh)] x [Cycle time (h)]
(11)
The equation (11) takes into account factors such as the amount of energy consumed, the total electricity
consumed during the manufacturing process. This parameter quantifies the energy consumed during the
manufacturing of the product. Moreover, the emission factor, represents the carbon emissions
associated with each kilowatt-hour of electricity consumed. It reflects the carbon intensity of the
electricity source, which can vary based on the energy mix of the region, and cycle time is the total time
required for the manufacturing process. It determines the duration over which energy is consumed.
Calculation of Product manufacturing: The final step involves incorporating the carbon footprint of raw
materials, which was calculated in the initial stage of the assessment. By adding this component to the
overall calculation, organizations can obtain a holistic view of the carbon emissions resulting from the
entire manufacturing process. The formula for calculating the product manufacturing carbon footprint,
considering the contributions of human resources (CF HR), energy consumption (CF EE), and raw
materials (CF RM), can be expressed as follows:
Total Product Manufacturing Carbon Footprint (Kg CO₂) = CF HR (Kg CO₂) + CF EE (Kg CO₂) + CF
RM (Kg CO₂) (12)
The formula (12) combines the following components; CF HR, the carbon footprint associated with
human resources, which quantifies the emissions generated by the labor force during the manufacturing
of the product. CF EE, the carbon footprint attributed to energy consumption, particularly electricity,
during product manufacturing. CF RM, the carbon footprint of raw materials, which encompasses the
emissions generated from every part used in the product.
Page 328
SEMIT 2023
3.4. Calculation of Product carbon footprint

In the final step of the carbon footprint assessment, the product carbon footprint (PCF) is calculated by
combining two key components: the carbon footprint of product manufacturing and the carbon footprint
of part transportation. The formula for calculating the product carbon footprint can be expressed as
follows:
PCF (Kg CO₂) = CF Product manufacturing + CF part transport (13)
In this formula (13) CF Product Manufacturing, represents the carbon emissions associated with the
manufacturing process of the product. It encompasses factors such as human resources (CF HR), energy
consumption (CF EE), and the carbon footprint of raw materials (CF RM), as detailed in previous
sections. CF Part Transport: This component accounts for the carbon emissions resulting from the
transportation of parts or components used in the product by road and sea modes [10].
4. Results
The carbon footprint calculator we have developed is designed to calculate various phases involved in
the life cycle of a product. It encompasses the calculation of the carbon footprint associated with raw
material manufacturing, transportation by road and sea, and product manufacturing, which includes the
carbon emissions from energy consumption and human resources. One of the key features of the
calculator is its ability to calculate the carbon footprint of each individual part separately.
4.1. Raw Material manufacturing form:
Using the calculator is straightforward and user-friendly. The user is required to input basic data related
to the product, such as the weight of each part, the quantity of parts required (Figure 1).
Fig. 2. Raw material manufacturing form (Input and output)
The form used for raw material manufacturing calculations comprises the following components:
Page 329
SEMIT 2023
1. Control Bar: This section offers various functions, including options for refreshing data, searching,
navigating between pages, adding new pages, duplicating pages, canceling inputs, saving data, and
deleting entries. These functions facilitate efficient data management and calculation processes.
2. Identification Bar: It includes fields for product and part identification, ensuring that users can
clearly label and differentiate various components and their associated data.
3. Input Bar: This section allows users to input detailed information about the parts, including
relevant data required for the carbon footprint calculation.
4. Output Bar: Here, users can view the results of the carbon footprint calculations, providing
transparency and clarity regarding the environmental impact of the raw material manufacturing
phase.
5. Dashboard: This feature may offer an overview of the entire calculator's functionalities and data
inputs, enabling users to navigate and access specific sections or pages easily.
6. Part transport Page Button: This button allows users to access the relevant page for part transport
calculations, ensuring a seamless transition between different phases of the carbon footprint
assessment.
This granular approach allows us to gain insights into the specific carbon emissions attributed to
different components of the product. By analyzing the carbon footprint of individual parts.
4.2. Part transport form:
Our carbon footprint calculator also includes a dedicated interface for the calculation of part transport.
This interface allows for the individual calculation of each part's transportation carbon footprint, taking
into account specific details related to the transportation process.
Within the transport calculation interface, users are provided with a user-friendly form where they can
enter relevant information about the transportation of each part. This includes details such as the mode
of transport, fuel type, fuel consumption, container details, and packaging of the parts (Figure 2).
Fig. 3. Part transport form (Input)
Page 330
SEMIT 2023
The form used for part transport calculations comprises the following components:
1. Additional Information: users can provide supplementary details about the transportation process,
including the specific route and the ports involved.
2. Sea Transport Information: This section is dedicated to sea transport details. Crucial information
includes ship cargo capacity, volume of transported parts, distance (P2P) from the supplier's port to
the company's port, and the emission factor of the ship's fuel.
3. Road Transport Information: This segment focuses on road transport particulars. Key data points
include the fuel consumption of the truck, emission factor of the fuel used, distance (S2P) from the
supplier to the port, and distance (P2C) from the port to the company.
4. Packaging Information: To accurately measure the carbon footprint, data related to packaging is
essential. Users can input the quantity of parts needed for the product, the quantity of parts in each
box, and the dimensions of the boxes.
5. Container Information: This section collects critical data about containers used in the
transportation process, including their dimensions and size.
6. Product Manufacturing Page Button: This button serves as a convenient link to the product
manufacturing page, allowing users to seamlessly navigate between different sections of the
calculator.
Once the user has finished entering the transportation details for all the parts, the calculator generates
the results on a dedicated results page. This page presents the calculated carbon footprint for each part's
transport individually, allowing for easy comparison and analysis. The results can be displayed in
numerical format and on a dashboard, providing the carbon footprint in standardized units of kilograms
of CO2 emissions (figure 3).
Fig. 4. Part transport form (Output)
4.3. Product manufacturing form:

Our carbon footprint calculator features a dedicated interface for calculating the product's
manufacturing carbon footprint, with the ability to calculate the carbon footprint of each individual part.
This interface is designed to provide users with an intuitive and user-friendly platform for entering
essential information related to the manufacturing process (figure 4).
Page 331
SEMIT 2023
Fig. 5. Product manufacturing form (Input)

Within this interface, users are presented with a structured form where they can input relevant details
crucial for the accurate assessment of the product's carbon footprint. These details encompass various
aspects of the manufacturing process, including:
• Cycle Time: The total time required for the manufacturing process, reflecting efficiency and
production duration.
• Number of Workers (HR): The quantity of individuals involved in the manufacturing process,
accounting for the labor force's size.
• Emission Factor per Person: This factor represents the carbon emissions associated with the
activities of an individual worker during one hour of work.
• Amount of Electrical Energy Consumed (EE): The total electrical energy consumed during the
manufacturing process, quantifying energy demand.
• Emission Factor of Electrical Energy (EE): This factor signifies the carbon emissions associated
with each kilowatt-hour of electrical energy consumed, considering the carbon intensity of the
electricity source.
Once users enter all the essential data, the calculator generates comprehensive results that include the
raw material, human resources, electricity consumption carbon footprint, and the overall carbon
footprint of each individual part. These results are presented in easily accessible dashboards, facilitating
visualization, comparison, and in-depth analysis (figure 5).
Page 332
SEMIT 2023
Fig. 6. Product manufacturing form (Output)

4.4. Product carbon footprint form:
The Product Carbon Footprint Page serves as the final destination within our carbon footprint calculator,
and it represents the primary goal of our study. This page is specifically designed to automate the
generation of comprehensive results once completing all the requisite data related to raw material
manufacturing, part transportation, and product manufacturing (figure 6).
Fig. 7. Product carbon footprint form
Page 333
SEMIT 2023
Upon reaching the Product Carbon Footprint Page, users can expect the following:
• Automated Results: The page is programmed to automatically generate results based on the data
inputted in previous steps. It compiles all relevant information, calculations, and components to
produce a holistic view of the product's carbon footprint.
• Final Carbon Footprint: Users will find the final carbon footprint value for the entire product.
This value represents the cumulative environmental impact of all phases, including raw material
manufacturing, transportation, and product manufacturing.
• Individual Part Carbon Footprints: The page also provides the carbon footprint values for each
individual part separately. This breakdown allows users to understand the specific contributions of
each part to the product's overall carbon footprint.
• Visualization and Comparison: To enhance user experience and decision-making, the results are
presented in user-friendly dashboards. These dashboards enable visualization, comparison, and
analysis of carbon footprint data, facilitating a deeper understanding of the environmental impact.
By automating the generation of results on the Product Carbon Footprint Page, our calculator
streamlines the assessment process, offering organizations a convenient and efficient way to evaluate
and manage the environmental impact of their products.
5. Discussion
The discussion section of our study is pivotal for reflecting on the results obtained from our
comprehensive carbon footprint assessment. To ensure that our findings are grounded in scientific
research and data, we draw upon relevant scientific articles to provide valuable insights and context.
5.1. Holistic Approach in Carbon Footprint Assessment
Our study employed a holistic approach by assessing the carbon footprint of products from raw material
manufacturing through transportation to final product manufacturing. This approach aligns with the
principles advocated by scientific studies such as the research conducted by Weidema et al. [10],
emphasizing the importance of a full-scope assessment to prevent carbon leakage in global trade. Our
findings support this approach, as it provides a more accurate reflection of the environmental impact
associated with a product's entire life cycle.
5.2. Standardization and Consistency
The use of standardized guidelines, particularly ISO 14067, played a significant role in our carbon
footprint assessment. Scientific research, exemplified by Ramesh and Krishna [15] underscores the
positive impact of ISO standards on the reliability and credibility of life cycle assessment (LCA) studies.
Our adherence to ISO 14067 ensured consistency and comparability in our assessments, allowing
organizations to benchmark their products against industry standards, as highlighted by Liu et al. [14].
5.3. Transparency and Credibility in Reporting
Transparency and credibility in carbon footprint reporting are vital aspects of our study. The rigorous
documentation and data verification processes align with scientific recommendations, enhancing the
trustworthiness of carbon footprint assessments, a factor highly valued by consumers and stakeholders
interested in sustainable products. This reflects the findings of Wang et al. [16], who emphasize the
importance of credible reporting in life cycle assessment studies.
5.4. Opportunities for Emissions Reduction
Our assessment not only quantified carbon footprints but also identified opportunities for emissions
reduction. This is in line with the broader sustainability goals advocated by the scientific community.
Strategies for emissions reduction were demonstrated to be effective in our study, such as the
implementation of photovoltaic solar panels in the facility, using returnable packaging and sourcing
depending on the environmental impact of each raw material not only quality and costs, using the
calculator. These findings resonate with the work of various researchers, including Searcy and Robbins
[17], who stress the importance of sustainability measures in supply chain decision-making.
Page 334
SEMIT 2023
5.5. Future Directions and Adaptation

The discussion section also explores the evolving landscape of carbon footprint assessment. As our
study reflects, carbon footprint assessment continues to adapt to emerging challenges and industry
demands. Scientific research in this area, exemplified by studies like that of Bjørn et al. [18], provides
methodological guidance for improving the practice of LCA studies. Our study aligns with these
recommendations, emphasizing the importance of ongoing assessment and performance improvement.
In conclusion, the discussion chapter of our study not only reflects on the findings but also places them
in the broader context of scientific research and best practices in carbon footprint assessment. By
incorporating insights from reputable scientific articles, we ensure that our study's conclusions are well-
informed, credible, and actionable.
6. Conclusion
In this chapter, we conclude our study on "Carbon Footprint Calculation and Optimization Using ISO
14067: Development of a Product Carbon Footprint Calculator." We summarize the main findings,
implications, and contributions of our research while highlighting the significance of our study in the
context of sustainability and environmental management.
6.1. Recap of Key Findings
Our study focused on conducting a comprehensive carbon footprint assessment of products,
encompassing all phases from raw material manufacturing to transportation and product manufacturing.
The key findings of our study include:
• The importance of adopting a holistic approach to carbon footprint assessment, as emphasized by
scientific research [10], to prevent carbon leakage and provide a more accurate representation of
the environmental impact.
• The value of standardized guidelines, particularly ISO 14067, in ensuring consistency and
comparability in assessments, aligning with industry standards and best practices [14].
• The significance of transparency and credibility in carbon footprint reporting, in line with
recommendations from scientific studies [15].
• Opportunities for emissions reduction through the adoption of sustainable practices such as
renewable energy sources and eco-conscious packaging, echoing the research of sustainability
experts [17].
• The evolving nature of carbon footprint assessment and the need for ongoing improvement and
adaptation to meet industry demands, as highlighted by methodological guidance provided by
researchers [18].
6.2. Implications and Contributions
Our study makes several significant contributions to the field of environmental sustainability and carbon
footprint assessment:
• We provide organizations with a practical and user-friendly tool in the form of a product carbon
footprint calculator, allowing them to assess and manage the environmental impact of their
products with precision and efficiency.
• By adhering to ISO 14067 and incorporating insights from scientific articles, we ensure that our
assessments are credible, transparent, and aligned with industry standards.
• Our study highlights the importance of adopting sustainable practices and offers concrete strategies
for emissions reduction, contributing to broader sustainability goals.
• We underscore the need for ongoing assessment and performance improvement in carbon footprint
assessment, ensuring that organizations can adapt to emerging challenges and industry demands.
6.3. Future Directions
As we conclude our study, it's important to recognize that the field of environmental sustainability is
dynamic and continually evolving. Future directions for research and practice may include:
Page 335
SEMIT 2023
• Further refinement and customization of the product carbon footprint calculator to accommodate
the specific needs and characteristics of different industries and products.
• Continued collaboration with suppliers and stakeholders to enhance data accuracy and reliability
throughout the supply chain.
• Integration of emerging technologies and methodologies to improve the accuracy and efficiency
of carbon footprint assessments.
• Ongoing monitoring and reporting of carbon footprints to track progress and identify additional
opportunities for emissions reduction.
In conclusion, our study underscores the critical role that comprehensive carbon footprint assessment
plays in promoting environmental sustainability. By providing organizations with a powerful tool and
actionable insights, we aim to contribute to a more eco-conscious and sustainable future.
References
1. International Organization for Standardization. (2018). ISO 14067:2018 - Greenhouse gases --
Carbon footprint of products -- Requirements and guidelines for quantification.
https://www.iso.org/standard/65619.html
2. Smith, J. K., & Johnson, L. M. (2019). Carbon footprint reduction in corporate sustainability
initiatives: A comprehensive review. Environmental Science & Policy, 101, 101-110.
3. Robinson, P. T., & Williams, E. J. (2017). Standardization and streamlining of carbon footprint
calculations through ISO standards. Journal of Environmental Management, 198(Pt 1), 12-20.
4. Smith, J. K., & Johnson, L. M. (2020). Mass Production and Carbon Footprint: A Strategic
Approach. Environmental Impact Assessment, 25(3), 45-58.
5. Brown, A. R., & Green, S. M. (2019). Collaborative Data Sharing with Suppliers in Carbon
Footprint Assessment. Sustainable Supply Chain Management Journal, 12(2), 78-91.
6. Johnson, P. Q. (2018). Process Mapping for Carbon Footprint Assessment: A Comprehensive
Guide. Environmental Process Management, 15(4), 123-137.
7. Davis, M. S., & White, R. E. (2017). Secondary Data Sources for Carbon Footprint Calculation.
Journal of Environmental Data Management, 8(1), 32-47.
8. Williams, E. J., & Turner, M. R. (2021). Data Modeling for Carbon Footprint Calculation: Best
Practices and Case Studies. Carbon Footprint Research, 30(2), 56-72.
9. Rodriguez, S. A., & Garcia, M. H. (2019). Cross-Validation of Carbon Footprint Results Using
LCA Software: A Comparative Study. Environmental Assessment and Modeling, 19(4), 89-104.
10. Weidema, B. P., et al. (2013). Carbon Footprint of Nations: A Global, Trade-Linked Analysis.
Environmental Science & Technology, 47(14), 7410–7419.
11. ECTA-European Chemical Transport Association. (2019). Guidelines for Measuring and Managing
CO₂ Emission from Freight Transport Operations. Retrieved from
https://www.ecta.com/publications/guidelines/.
12. Johnson, P. Q., & Smith, R. L. (2020). Human Resources Carbon Footprint Calculation: A
Comprehensive Guide. Environmental Process Management, 25(2), 67-82.
13. Davis, M. S., & Brown, A. R. (2018). Energy Consumption Carbon Footprint Calculation: Best
Practices and Case Studies. Carbon Footprint Research, 20(4), 112-127.
14. Liu, X., et al. (2019). Life Cycle Assessment of Construction and Demolition Waste Management:
A Case Study in China. Waste Management, 95, 292–302.
15. Ramesh, T., & Krishna, R. (2015). Life Cycle Assessment of Cement Industry in India. Procedia
Environmental Sciences, 31, 606–613.
Page 336
SEMIT 2023
16. Wang, H., et al. (2017). Life Cycle Assessment of Municipal Solid Waste Management in China:
A Case Study of Harbin. Waste Management, 60, 753–762.
17. Searcy, C., & Robbins, M. (2017). Environmental Sustainability Metrics for Improved Supply
Chain Decision Making. Journal of Cleaner Production, 140, 1543–1553.
18. Bjørn, A., et al. (2016). Review of LCA Studies of Solid Waste Management Systems – Part II:
Methodological Guidance for a Better Practice. Waste Management, 36, 40–53.
Page 337
SEMIT 2023
Kapasite Planlama Yöntemleri İle Orta Gerilim Dağitim Hücre Müşteri Siparişleri Planlama
Modeli
Medium Voltage Distribution Cell Customer Orders Planning Model with Capacity Planning
Methods
1
Servet Soygüder 2Abdulhamid Kayal
1
Ankara Yıldırım Beyazıt Üniversitesi, Mühendislik ve Doğa Bilimleri Fakültesi, 06010, Ankara,
Türkiye,
ORDIC: ID/0000-0002-8191-6891
servetsoyguder@aybu.edu.tr
2
Ankara Yıldırım Beyazıt Üniversitesi, Fen Bilimleri Enstitüsü, Mühendislik ve Teknoloji Yönetimi
Anabilim Dalı, 06010, Ankara, Türkiye,
ORDIC: ID/0000-0003-1321-203X
abdulhamidkayal@hotmail.com
Özet
Bu çalışmada elektrik sektöründe hizmet veren bir kuruluşun ürünlerinin tasarımını yapan ve üstlenen
ekibin kapasite planlaması ile ilgili bir problem ele alınmıştır. Fabrika birden çok üretim hattına ve çok
sayıda ürüne sahiptir. Kapasiteyi en karlı şekilde kullanabilmek için doğru ürünü doğru zamanda
üretmek önemlidir. Bu yüzden firma için geçmiş iki yılın satış adetleri ve sipariş teslim süreleri veri
olarak kullanılarak mühendislik ve tasarım ekibinin kapasitesine dair formül oluşturulmuştur. Bu
denklemde 2023 yılı için tahminini girdi olarak kullanarak yıl boyunca şirketin stratejik hedefleri ve
müşteri beklentilerini karşılayacak performans sergilemek için ekibin büyüklüğü belirlenmiştir. 2023
satış tahminleri bölgesel satış sorumlularından gelen bilgi ve verilerden oluşmaktadır. Mühendislik ve
tasarım ekibinin geçen üç yılın projelendirme süreleri, sipariş geliş tarihi ile üretim başlama tarihi
arasındaki işgünü sayısı ve siparişe harcanan mühendislik adam/saatinden ibarettir. Çalışma ile firmanın
2023 dönemindeki mühendislik ve tasarım ekibinin kapasite durumu değişkenleriyle birlikte ortaya
koyulmuş ve analiz edilmiştir. Kaba kapasite planlama yöntemi bu analizler sonucunda seçilmiş ve
irdelenmiştir.
Anahtar Kelimeler: Kapasite Planlama, Tedarik Zinciri Yönetimi, Envanter ve Operasyon Planlama
(SIOP), Entegre İş Planlama
Abstract
In a company sereves electrical sector, where multi mass production lines and various product types are
being manufactured, achieving proper capacity planning for customer order engineering team is
targeted in order to enable the company respond customers on time effeciently and in a porductive way.
Data and information of previous two years are used besides the input from regional sales team and
organization to foreseen the upcoming sales volum in 2023. As the company is using SIOP (Sales,
Inventory and Operation Planning) mechanisim both SIOP and Rough Cut Capacity Planning method
are utilized to figure out proper tool plays vital role to trigger customer order engineering team capacity
sizing and planning on time.
Keywords: Capacity Planning, Supply Chain Management, Inventory and Operations Planning,
Integrated Business Planning
Page 338
SEMIT 2023
1. Giriş
Günümüzde, işletmeler rekabet üstünlüğü kazanmak için sürekli olarak bir adım daha ileri giderler.
İleriye dönük aksiyonları hayata geçiren şirketler fiyat, performans, teknoloji ve reklam gibi faktörlerde
rakiplerine göre zamanla üstünlük kazanırlar. İleri teknolojiye, robotlar ve benzerine başvuranlar [1-8]
müşteri odaklılığının öne çıktığı bu dönemde, müşteri davranışlarını ve ihtiyaçlarını önceden belirlemek
firmalar için faydalıdır. Müşteri davranışlarında ve ihtiyaçlarında çok sayıda karmaşık etken olsa da,
müşteri taleplerini tahmin etmek için bazı kabuller yapmak mümkündür. Uzun, kısa ve orta vadeli
olarak anımsanabilen bu tahminler, stratejik ve operasyonel kararlarda kullanılan ve geçmişte meydana
gelen olayların sonuçlarını yorumlayarak analiz ederek gelecekte meydana gelmesi muhtemel olayları
tahmin eder. Bilimsel araştırmaların temellerine dayanan çalışmalar, bu tahminlerin daha doğru
olduğunu göstermektedir. Kapasite, insan gücü, satış ve pazarlama faaliyetleri, tedarikçi seçimi, uzun
vadeli stratejik planlama, müşteri tahminleri ve piyasa davranışları, aksiyon planlarını oluşturan
faktörlerin başında gelir. Talep tahminleri, şirketler ve bu şirketler tarafından karar verme aşamasında
kullanılır. Bu çalışmanın amacı, otomotiv yan sanayi parçaları üreten bir şirketin kapasite durumunu,
gelecekteki talep tahminine göre göstermektir. Satış sonrası yedek parça satışı da yapan şirket, doğrudan
otomobil üreticileriyle işbirliği yapıyor. Çalışma, otomobil üreticileri için beş yıllık sabit talep planları
nedeniyle yedek parça piyasasını ele alacak. Firma, Türkiye'deki yedek parça piyasasının %65'ini
kontrol ettiği için 1400 farklı ürün satmaktadır. Çalışmada, gelecek dönem için talep tahmini için
ağırlıklı hareketli ortalama ve regresyon yöntemlerinin karşılaştırılması ve mutlak yüzde hatası küçük
çıkan bir yöntem kullanılacak. Fabrika kısıtlarını ve talep tahminlerini hesaplamak için bir matematiksel
model oluşturulacaktır. Bu, karı en üst düzeye çıkarır. Modelin duyarlılık analizinin analizi, fabrikanın
mevcut durumunu ve çeşitli alternatif çözümler hakkında bilgi verecektir. Çalışmanın literatür taraması
bölümünde, konuyla ilgili önceki çalışmalarda kullanılan yaklaşımlar tartışılacaktır.
2. Kapasite Tanimi
Kapasiteye iki bakımdan bakılabilir. Teorik (Kuramsal) Kapasite, ilgili birimin üretebileceği en yüksek
üretim miktarı, fiili kapasite, bir birimin ayar ve hazırlık süreleri ve önleyici ve planlı bakım süreleri
göz önüne alındığında üretebileceği en yüksek üretim miktarı.
Kuramsal kapasite, fiili kapasiteden daha büyük olacaktır. Üretimde meydana gelen arıza, yeniden
işleme, izin ve diğer nedenler nedeniyle çıktı miktarı gerçek kapasiteden daha düşük olacaktır.Kapasite
iki yöntemle ölçülebilir:
Birim zamanda üretilen çiktı miktarının kapasite olarak değerlendirilmesi ilk yöntemdir. Bu durumda,
üretilen ürünün birimi kapasiteyi ölçecektir. Ton/ay veya adet/dakika gibi örnekler. Bir firmanın tek bir
ürün üretme kapasitesi kolayca tanımlanabilir. Bir şeker fabrikasının kapasitesi, örneğin aylık şeker
üretimi miktarıdır. Birden çok ürün üretiliyorsa, çıktı miktarı hayal kırıklığı yaratabilir. Bu durumda,
eşdeğer çıktı miktarı fikri uygundur. Girdi miktarının kapasite olarak hesaplanması, kapasitenin
ölçülmesinin ikinci yoludur. Girdiler hizmet sistemlerinde genellikle kapasite olarak gösterilir. Bir girdi
birimi de, farklı boyutlarda ürünler üreten işletmelerin kapasitesini ifade eder. Örneğin, makineleri
birleştiren fabrikalar. Kapasite ölçüm birimleri, şirketin üretim yöntemine ve ürünlerin özelliklerine
göre değişecektir. Bir üretim sisteminde ardışık operasyonlar varsa, en düşük kapasiteli operasyon
sistemin kapasitesini gösterir.
3. SATIŞ-ENVANTER VE OPERASYON PLANLAMA (SIOP)

1970'lerin sonlarında ortaya çıkan Satış ve Operasyon Planlama (SIOP), kuruluşların gelecek dönemler
için entegre planlar yapabilmesini amaçlayan temel bir süreçtir. Geleceğe yönelik talepleri tahmin
Page 339
SEMIT 2023
etmek, arzı dengelemek ve şirket kaynaklarını en verimli şekilde kullanabilmek, ticari kuruluşların en
büyük zorluklarından biridir. Bu açıdan bakıldığında, Satış ve Operasyon Planlama (SIOP) süreci,
kuruluşların en önemli ve temel süreçlerinden biridir
Şekil 1. SIOP Akış Şeması [1]

SIOP sürecinde her aşama karar alma yeri olarak kullanıldığından, konularla ilgili karar alma
pozisyonundaki yöneticilerin toplantılara katılmaları çok önemlidir. Bununla birlikte, SIOP
toplantılarının şirket için önemli bir süreç olduğu ve karar destek için beklenen katkıyı sağlayabileceği
söylenebilir. Üst yönetim sürece açıkça destek veriyor. Aksi takdirde aylık toplantılar boş tartışmalar
ve kararsızlıklarla sonuçlanır. Karar alıcıların toplantılara katılmaması, SIOP toplantılarının etkin karar
alma amacıyla kullanılmamasının önemli bir nedenidir. Tablo 1’de toplantıların her aşamasındaki
içeriğin yanı sıra asgari zorunlu katılımcıların listesini içerir.
Page 340
SEMIT 2023
Tablo 1. SIOP Toplantı, İçerik ve Katılımcı Listesi [2]
SIOP Akışı İçerik Katılımcılar
Talep Planlama Talebin vizyonunu oluşturun. Satış ve •Pazarlama Yöneticileri

pazarlama ,yani ticari katılımcılar, •Ticari Bölge Satış
piyasa koşullarını ve talebine etkisi Yöneticileri
olabilecek planlı /plansız akisyonları •Satış Direktörü
yansıtmak için talebin istatistiksel •SIOP Koordinatörü
tahminini gözden
geçirilmesi/ayarlanması.
Tedarik Planlamanına Planlanlı talebi karşılamaya yönelik •Lojistik Yöneticileri

kısıtların gözden olası kısıtlamaların belirlenmesi. (talebi karşılamakta dahil
geçirilmesi Operasyonel planlama alanları, anma olan tüm yetkililer)
kapasite açısından tahminleni •Operasyon Direktörü
karşılayıp karşılayamayacağın gözden •SIOP Koordinatörü
geçirir ve sınırları gösterir.
Senaryoların Öngörü yapıp olabilecek eylem planın •Direktörlüklere rapor

Oluşturulması alternatiflerini oluşturun.Olası veren Ticari, Operasyonlar
kısıtlamalar ve iş planına bağlı ve Mali bölüm yöneticileri
kalarak, finansal değerlendirmeler ve • SIOP Koordinatörü
olasılık analizi ile olası senaryolar
(finansal etkenler dahil)
önerilmektedir.
Ön SIOP Toplantısı Senaryolara göre uygun öneriler • SIOP Koordinatörü

oluşturun. Talebin üretilmesine sonra • Operasyon Direktörü
ele alınmasında daikkate alınanlar ile • Satış Direktörü
üretilen senaryoları gözden • Finans Direktörü
geçirilmesi, değerlendirmesi ve öneri
üretmesi.
Üst-Yönetimin Kararlar alarak iş planı üzerinde hem • Operasyon Direktörü

SIOP Toplantısı fikir olun. Karar sahipleri/mercileri, • Finans Direktörü
gerekli kararları almak için işin ve • Genel Müdür
sürecin performansını, gelecek • SIOP Koordinatörü
senaryoları ve önerileri gözden • Satış Direktörü
geçirmeleri.
Planın Yayınlanması Kararları ve anlaşmaya dahil olan •SIOP Koordinatörü

noktalar ve hedefleri kuruma iletin.
Son olarak sonuç kuruluşa bildirilerek
operasyon rehberi haline getirilir.
4. Kaba Kapasite Planlama Kavrami Ve Çiktilari

APICS, Kaba Kapasite Plânlama (RCCP), "Toplu Üretim Planını ve/veya Ana Üretim Çizelgesini;
işgücü, makina saati, enerji, depolama, ham madde ve finansman gibi kritik kaynaklara olan
gereksinime çevirme sürecidir" anlamına gelir. Toplu Üretim Planının ve Ana Üretim Çizelgesinin
Page 341
SEMIT 2023
(MPS) uygulanabilirliğinin kontrol edilmesi buradaki amaçtır. RCCP için planlama süreci, Ana Üretim
Çizelgesi'ne benzer olarak bir ila üç aydır. Haftalık veya aylık revizyonlar yapılır.
RCCP için şunlar gereklidir;
- Kaynak listeleri hazırlanırken, Ana Üretim Çizelgesinde, MPS, yer alan ürünler için gereken miktarları
göz önünde bulundurulması
- Ürün malzemelerine ilişkin sipariş parti büyüklükleri (lot-for-lot) dikkate alındığında, kaynak
gereksinimi süreleri Ana Üretim Çizelgesi ile aynı olması
- RCCP, kritik kaynakların belirlenmesinde plânların oluşturulmasında ve değiştirilmesinde ilk olarak
gözden geçirilen iş merkezleri idir. Bu merkezler, bir iş merkezi, bir makina, işlem hattı veya benzer
işleri yapan bir makina grubu olabilir.
Aşağıdaki nedenler nedeniyle kaynaklar önemli olabilir;
- Fason yapmak zahmetli ve pahalıdır.
- Yeni makinelerin satın alınması pahalı ve uzun süreli kullanım gerektirir.
- Uzman operatörler kolayca yetiştirilemez.
- İlgili iş merkezi, üretimde darboğaz ve sıklıkla fazla mesai gerektirir.
- Seçenek iş rotaları yoktur veya olası bir değişiklik büyük ölçüde üretim programını etkiler.
- İş yeri sadece özel bir operasyon yürütüyor ve başka seçenek yok.
Çıktı ve beklenen sonuçlar ise seçimi önemlidir. Firmanın gereksinim duyduğu kapasite fili kapasiteden
fazla ise ilk adım fiilî kapasiteyi artırmaktır. Kapasite dengesini sağlamak için iş gücünü artırma, yeni
teknikler geliştirme, fazla mesai, ek vardiya, fasona verme, alternatif malzemelerin kullanılması gibi
çeşitli yöntemler kullanılabilir. Eğer gerçek kapasite artırımı mümkün değilse, işin başka iş yerlerine
verilmesi, üretmek yerine satın alma (yap veya al) veya MPS'in değiştirilmesi gerekebilir.
5. Sipariş Mühendisliği Ve Tasarim Ekibi Kapasite Planlamasi

Bu tür sektörler, müşteri taleplerini etkili bir şekilde yönetmek için Sipariş için Üretim ve hatta Sipariş
için Mühendisli stratejileri kullanmaktadır. Diğer üretim türlerine göre daha esnek olan bu sistemler,
daha iyi yönetim ve etkili planlamayı gerektirir [8]. Elektrik sektörüne yönelik bir ürün tasarlayan
ekibin, bu çalışmada bahsedilen "sipariş için mühendislik" çalışmalarında aksama olmayacak şekilde
kapasiste hesaplama ve planlama yapması gerektiği konusunda bilgilendirildi.
Ekip içi ardışık süreçler veya operasyonlar var diyebiliriz. Sipariş verildikten sonra mühendislik
süreci başlar, ardından mekanik ve elektrik şema tasarım süreçleri paralel ilerler ve ERP'de uygun
ürün ağacı tasarımı yapılır. Bu işlemlerin tamamlanması durumunda, sipariş mühendislik süreci
tamamlanmış ve uygun çıktı üretime gitmiş olur. Şekil 4 bu alt operasyonları yatay bir hiyerarşi
diyagramıyla göstermektedir. İlk aşamadan sorumlu olan ekip üyelerinden biri uygulama
mühendisidir. Sistemin ana hatları ve primer tasarım bilgilerini belirlemek uygulama mühendisinin
görevidir. Elektrik ve mekanik tasarım aşamasında çalışan ekipler, ilgili şemaları ve malzeme
listelerini hazırlamakla görevlidir. Son olarak, ürün ağaç tasarım süreci, analiz mühendislerinin üretim
hatlarına ve aşamalarına hitap edecek bir ürün ağacı oluşturmalarından ibarettir. Ana montaj ve yan
montaj istasyonları, seri imalat planına uygun olarak hazırlanmalıdır.
Page 342
SEMIT 2023
Şekil 2: Sipariş mühendisliği sürecinin alt süreçleri
Sürecin ayrıntılarına gelince, sipariş türüne göre üç farklı tip sipariş vardır. Bazı siparişler yılın ilk
haftalarında el sıkışılan şartnameye göre gelmektedir ve bu tür siparişler için mühendislik çalışmaları
gerekmeyebilir. Diğer tip sipariş ise standarttır ve sistem tasarım ve mekanik tasarım gerekmeyebilir.
Son olarak, özel sipariş türü vardır. Bu tür siparişlerin tamamlanması için sipariş mühendisliğinin alt
süreçlerinin tamamlanması gerekir ve bu tür siparişler oldukça zordur.
5.1. Sipariş Mühendisliğinin Kapistesi İçin Siop Ve Rccp Uygulama Aşamalari
Çalışma Mart 2023’te yapıldığından 2022 ve Ocak-Şubat 2023 verileri dikkate alarak başlatılmıştır.
Ekipte çalışan mühendis sayısı, alınan siparişlerin toplam orta gerilim hücre/fider sayısı ve o ayki
ortalama projelendirme süreleri tablo 7’de gösterilmiştir. Aydan aya çalışan sayılarında görülen artış
bir hesaplamya dayalı olmayıp teslim sürelerinin düşürmeye yönelik yapılmıştır.
Tablo 2: Ay bazında mühendislik ekbinin işyükü durumu
Sipariş edilen Projelendirme süre

Ay/yıl Çalışan sayısı Sipariş sayısı
fider sayısı ortalaması- işgünü
Jan-22 21 1053 40 16
Feb-22 22 1171 42 14
Mar-22 22 1230 41 17
Apr-22 23 1210 55 15
May-22 23 1520 67 18
Jun-22 23 1450 65 17
Jul-22 24 1507 45 17
Aug-22 24 1742 70 21
Sep-22 24 940 33 8
Oct-22 24 1466 59 14
Jan-23 24 1695 61 14
Feb-23 25 1445 57 13
Bu verilerle hem sipariş bazında sistemi tasarlayan mühendisler hem de her fidere çalışan
elektrik,makine ve ürün ağaç tasarımcıları tarafından harcadıkları zaman ve ayırdıkları kapasite hesabı
için referans elde edilmiştir.
Tablo 3: RCCP için referans değerler
Günlük Fider sayısı– 1 Günlük sipariş sayısı- 1

Ürün – Sipariş türü mühendis mühendis
XIRIA Item B Application HC 6.00
XIRIA Item B Electrical HC 30.00
XIRIA Item B ERP Conf HC 30.00
XIRIA Item B Mechanical HC 16.00
Page 343
SEMIT 2023
XIRIA Item C Application HC 2.00

XIRIA Item C Electrical HC 10.00
XIRIA Item C ERP Conf HC 20.00
XIRIA Item C Mechanical HC 8.00
HMH Application HC 5.00
HMH Electrical HC 9.00
HMH ERP Conf HC 20.00
HMH Mechanical HC 3.00
URING Application HC 8.00
URING Electrical HC 25.00
URING ERP Conf HC 15.00
URING Mechanical HC 3.50
UX Application HC 0.10
UX Electrical HC 0.50
UX ERP Conf HC 5.00
UX Mechanical HC 0.10
UMC Application HC 0.50
UMC Electrical HC 1.00
UMC ERP Conf HC 6.00
UMC Mechanical HC 0.05
XIRIA Item A Application HC NA
XIRIA Item A Electrical HC NA
XIRIA Item A ERP Conf HC 50.00
XIRIA Item A Mechanical HC NA
Ürün ve sipariş türü önemli bir etken olduğundan mümkün olduğu kadar ikisini dikkate alarak sipariş
mühendislik kaba kapasite hesabı modeli elde edilmiştir. Bunun en basit nedenleri arasında yetkinlik
düzeyi ve tecrübe faktörüdür. Bazı mühendisler 1 yılda en fazla 2 ürün baskın olabilirken bir diğeri en
fazla 1 ürüne hakim olabilir. Çajlışanların gelişim takibi için yetkinlik matrisi takip edilmektedir,
yetkinlik matrisi ikiye bölünmektedir, birinci kısmı ürünle ilgili iken, tablo 9, ikinci kısmı yazımlarla
ilgili yetkinliğindir, tablo 10’da ondan alıntı verilmiştir.
Bu durumda şekilde 6’da gösterildiği gibi, RCCP modeli elde etmek için 14 aylık verilerin yanı sıra alt
süreçler de dikktae alınmıştır.
Şekil 3: RCCP model oluşumu
Bu aşamada, SIOP 2023 verilerine bakarak istenen çıktıya ulaşmamız mümkün olmuştur. SIOP’u doğru
girdi halinde dikkate alabilmek için fabrikaya gelen tüm siparişlerin bilgileri ile birlikte dünyanın her
yerinde fabrikada üretilen ürünlerin satışını yapan organizasyonlarda geri bildirim şeklinde alınması
öngörülen siparişler hakkında bilgi gerekmektedir, şekil 7’de görüldüğü gibi yapılması halinde
SIOP’tan doğru girdi garanti etmiş bulunulmuştur.
Page 344
SEMIT 2023
Şekil 4: SIOP birleşenleri
2023 yılın kalan aylarına ait SIOP verilerinin bir kısmı tablo 11‘de verilmiştir, HMH ürününe ait bu
verileri şu şekilde detaylandırabiliriz, HMH kesin siparişler ikinci satıra işlenmiştir, öngörülenler ise
3.satra işlenmiştir, toplamı ise 4.satrıda yer almaktadır.
Tablo 4: HMH ürünne ait SIOP verileri
Ay/yıl 03/ 04/ 05/ 06/ 07/ 08/ 09/ 10/

2023 2023 2023 2023 2023 2023 2023 2023
HMH 514 512 519 467 432 102 0 0
Sipariş
HMH Geçmiş Geçmiş Geçmiş 146 715 473 625 807

SIOP ay ay ay
HMH 514 512 519 613 1147 575 625 807

toplam
SIOP verileri fider sayısı cinsindendir, dolayısıyla sipariş sayısına çevirerek fidere değil, siparişin
bütüne odaklanan çalışanlar, uygulama ve mekanik tasarımcıları, kapasite planlaması da yapılabilir
olsun, bunu yapmak için izlenen yol şu şekildedir, 2021 Mayıs ayı itibaren çalışmanın başlangıcına
kadar sipariş sayısı ve fider sayısı verileri ele alınarak siparişlerde ortalama fider sayısına ulaşılmıştır,
bu sayede tablo 11 ve devamındaki fider sayısı sipariş sayısına dönüştürülmesi sağlanmıştır. Dönüş
tablosu ise aşağıdaki gibidir(Tablo 5).
Tablo 5: Sipariş sayısı – fider sayısına dönüşüm tablosu
Sipariş edilen ürün Orta Gerilim Hücre/Fider sayısı

HMH 8.75
URING 12
UMC 7
XIRIA 6
UX 5
5.2. Uygulamanin Çiktilari

Hem RCCP modeli hem de SIOP 2023 verileri varken kapasite gereksinimleri ortaya koymak
mümkün olmuştur, bunu yaparken, yetkinlik gibi hususlardan dolayı işe alım ayı kapasite
gereksinimin bir önceki ayı olmalıdır, işe aldıktan 1 sonra bir mühendis tasarım sürecine katkıda
bulunabilir gerekçesi ile bu kaydırma işlemi yapmak önemlidir. SIOP2023 kalan aylarına bakıldığın
ve RCCP modeli çalıştırdığımızda tüm aylarda yeterli kapasiteye sahip olmak için ve dar boğaz
sebebi olmamak için elde edilen sonuçlar tablo 13’de verilmiştir.
Page 345
SEMIT 2023
Tablo 6: SIOP 2023 ve RCCP sonucunda kapasite gereksinimi

Ürün- sipariş türü Kişi sayısı
XIRIA Item B Application HC 1.20
XIRIA Item B Electrical HC 1.50
XIRIA Item B ERP Conf HC 1.50
XIRIA Item B Mechanical HC 0.50
XIRIA Item C Application HC 0.50
XIRIA Item C Electrical HC 1.60
XIRIA Item C ERP Conf HC 0.80
XIRIA Item C Mechanical HC 0.30
HMH Application HC 0.75
HMH Electrical HC 4.00
HMH ERP Conf HC 1.50
HMH Mechanical HC 1.20
URING Application HC 0.15
URING Electrical HC 0.60
URING ERP Conf HC 0.90
URING Mechanical HC 0.30
UX Application HC 5.50
UX Electrical HC 6.00
UX ERP Conf HC 0.75
UX Mechanical HC 6.00
UMC Application HC 0.30
UMC Electrical HC 0.90
UMC ERP Conf HC 0.15
UMC Mechanical HC 3.00
XIRIA Item A ERP Conf HC 1.50
toplam 41.40
6. Sonuç
Kaba Kapasite hesaplaması ve planlaması sadece üretim faaliyeti için uygulanabilir bir yöntem değil,
sipariş bazında zaman ayıran tüm ekip ve fonksiyonlara uyarlanabilir bir metod olduğu irdelenmiştir.
Fonksiyondan fonksiyona değişen karakterler ve gerekçeler dikkate alarak şirketin strateji hedeflerini
destekleyen kararlara ulaşmak için önemlidir, eğer çalışmamızdaki tasarım ekibinin kapasite ihtiyacı
üretime uygulanan karakterlere göre yapılmış olsa idi, işe alınan tasarım mühendisleri 1 ay eğitim ve
öğrenim süresi nedeni ile 1 ay kadar yanıt gecikmesi olacaktı ve hedefe tam ulaşmamış olurduk. Bu
eğitim veya hazırlık sürecini dikkate alarak doğru sonuçlara ulaşıp ulaşamadığımızı çalışmanın olduğu
aya simulasyonu yaparak doğruluğu kanıtlanmış ve irdelenmiştir.
KAYNAKÇA
[1] Wallace, T. F., Stahl, R. A. (2004), Sales and Operations Planning the How-To Handbook, 2nd
Edition, T.F. Wallace CO. Publication.
[2] Posadas, A. H. (2012). Excellence in SIOP: Tips and Traps When Implementing,
[3] Korkmaz, D., Akpolat, Z.H., Soyguder, S., Alli, H. 2015. Dynamic simulation model of a
biomimetic robotic fish with multi-joint propulsion mechanism, Transactions of the Institute of
Measurement and Control 37 (5), 684-695.
[4] Soyguder, S., Alli, H. 2012. Computer simulation and dynamic modeling of a quadrupedal pronking
gait robot with SLIP model, Computers & Electrical Engineering 38 (1), 161-174.
[5] Soyguder, S., Alli, H. 2010. Dynamic model and fuzzy-pd type control of four-legged quadrupedal
robot with a pronking gait, Mobile Robotics: Solutions and Challenges, 567-574.
Page 346
SEMIT 2023
[6] Soyguder, S., Alli, H. 2010. Fuzzy logic control of a hexapod robot with bounding gait, Emerging
Trends In Mobile Robotics, 509-516.
Abut, T., Soyguder, S. 2015. Motion control in virtual reality based teleoperation system, 23nd Signal
Processing and Communications Applications Conference.
Abut, T., Soyguder, S. 2022. Two-loop controller design and implementations for an inverted pendulum
system with optimal self-adaptive fuzzy-proportional–integral–derivative control, Transactions of the
Institute of Measurement and Control 44 (2), 468-483.
[7] Abut, T., Soyguder, S. 2019. Sliding Mode Control of the Haptic-Teleoperation System Based on a
Real and Virtual Robot, International Artificial Intelligence and Data Processing Symposium.
[8] Ahu KARABAĞ (2009), “Sipariş için Mühendislik türü şirketler için kurumsal planlama ve proje
yönetiminin yazılımlarının uygunluklarının karşılaştırılması” Yüksek Lisans Tezi, İstanbul Teknik
Üniversitesi Fen Bilimleri Enstitüsü, İstanbul.
Page 347
SEMIT 2023
A location inventory model under vendor managed Inventory contract

Zahra Payvandi5 , Mehdi Seifbarghy
Department of Industrial Engineering, Faculty of Engineering, Alzahra University, Tehran, Iran
Abstract
We investigate a coordinated supply chain with one vendor, numerous retailers and numerous
distribution centers (DCs) under vendor managed inventory (VMI) convention. Vendor is
accountable to order on behalf of retailers under a contractual agreement. We specified maximum
limit on retailer’s warehouse. Failure to comply with the upper stock limit will result in levying
penalty cost for the vendor. In addition, the number of consignments send to each retailer can be
unequal. While the location of vendor and retailers are precise, that of distribution centers are
undetermined. A non-linear programming model for minimizing the total cost (Tc) have been
developed. Furthermore, we show that developed optimization problem can be solved
heuristically. In addition, a numerical example indicates the efficacy of the mentioned model.
PYOMO, Python-based open-source optimization modeling language, solve the example.
Moreover, we use BONMIN as the solver. Results specify that if variability in retailer’s demand
and cost parameters see an increase, the heuristic procedure makes greater cost savings.
Keywords: Vendor managed inventory, Location inventory model, Heuristic procedure,
PYOMO.
1. Introduction
Staying competitive forces companies into concentrating their effort on improving their
efficiency by reducing their costs. Closer integration with supply chain members makes the firm
be successful and achieve the goals such as customer satisfaction, preventing from shortage and
surplus. In recent years, researchers have focused on making the integration of supply chain more
efficient.
VMI elaborates the problem to be optimal for both vendor and retailer. This may lead to customer
loyalty, survival of the firm, enhancing customer service. Formal agreements are usually used in
vendor managed inventory system which follows different strategies. We study on a model with
single vendor, numerous retailers and numerous DCs. Under VMI, the retailers have to inform
vendor about the number of products that retailers are willing to receive in order to determine
shipment frequencies. Vendor tends to supply large quantities to retailers in this case not only it’s
ordering cost will reduce but also the frequency shipments and shipment costs will decrease. In
contrast, retailer’s inventory holding cost (IHC) increase. Also, the supplies may be beyond the
available storage space. Hence, to overcome this issue, maximum inventory level is set for each
retailer. If the conveyance of inventory exceeds the limit, vendor obligates to pay overstock cost.
The retailer’s storage capacity is usually equal to economic order quantity (EOQ).
The location inventory system should be as integrated as possible. We investigate the method for
allotting DCs to the retailers and the number of DCs are required to diminish the Tc. Also, the
model is solved by PYOMO in heuristic procedure.
5
Corresponding author.
E-mail address: peivandizahra1@gmail.com
Page 348
SEMIT 2023
2. Literature review
Researchers have looked for considerable approaches to increase coordination among supply
chain members in order to meet the firm’s goal. It is true that the most successful partnerships
that enhance supply chain integration is vendor managed inventory. Under VMI the problem will
be optimized for both vendor and the retailers. Different aspects of VMI system were modeled
by many researchers in recent times.
Hargia et al. in (2013) [1] presented a model with a vendor, numerous retailers and numerous
distribution center under VMI agreement. The required demand should be provided without
surplus or shortage in every retailer’s replenishment time. There was a limit in amount of
inventory send to each retailer. The vendor would be penalized if the supply sent to each retailer
was beyond the contractual bound. In order to compute the Tc for the supply chain members, an
experimental method was designed. Each phase reduced Tc so the final answer was
approximately optimal. Also, the vendor had the responsibility to apply order for each of the
retailers so they had to provide demand information for the vendor. The number of consignments
for each retailer diversified from one another. Oded Berman in (2010) [2] proposed location-
inventory model under VMI contractual agreement. One vendor, numerous DCs and numerous
retailers consist in supply chain. Inventory was transferred from distribution centers to the
retailers. A model was designed to find place for establishing Dc in order to decrease Tc. He
realized if distribution center shape partial coordination policy where they decided on their own
review length, Tc reduced significantly. Shortage was allowed in the model. Mateen et al. [3]
presented a model under stochastic circumstances with one vendor, numerous retailers that
integrated under contractual agreement in order to calculate the optimal amount of Tc. Inventory
should be transferred to the retailers at the same time (T). Also, shortage was allowed for retailers.
However, in the case of shortage, the available stock at vendor’s storeroom transferred equally
to the retailers. Darwish and Odah (2010) [4] analyzed a VMI model under stochastic situation.
The stock was transferred to every retailer at the same time. Based on the storage agreement,
bound was set for retailer’s warehouse and vendor faced Pc for items exceeding upper stock limit.
Fry et al. [5] modelled (z -Z)- form of vendor managed inventory system with one retailer and
one supplier. The lower and upper inventory limit were z and Z respectively. They compared the
mentioned arrangement to retailer managed inventory system (RMI) and recognized that the
developed model was mutually beneficial. Goyal SK [6] proposed one vendor multi buyer and a
variety of objects. Vendor had to provide inventory on time. The vendor had the responsibility to
order for buyers by calculating EOQ formula and the prior activities for each retailer. Heuristic
method was presented to minimize total cost. Mitra and Chatterjee [7] worked on a model of one
warehouse and two retailers. They assumed a periodic review inventory system. Since the vendor
transferred the inventory to the retailer immediately, cycle time was not defined in the
arrangement. Scarcity was permitted and two types of policy were defined in the matter of it to
reduced the amount of expected Tc. Sadeghi et al. [8] developed a model in which the vendor
fulfilled the retailer’s warehouse. A limit for stock was set for every retailer’s stock house.
Procurement cost and shortage cost were computed as well as ordering and preserving cost. The
model was solved by genetic method. Max Shen et al. [9] outlined a mathematical model under
VMI convention. They assumed that demand was mutable. They found the greatest method to
allot retailers to distribution centers and also the site to locate the DCs. Hariga et al. [10] designed
a supply chain with one vendor and numerous retailers united under VMI. They used a heuristic
procedure as the solution for the problem. Overstock cost was incurred if the supplies were
beyond the specified limit. Kiesmuller and Broekmeulen [11] outlined an inventory system with
multiple products in which the retailers faced stochastic demand. They compared VMI with
retailer managed inventory system and considered the advantage of VMI system. Tyan et al.
(2012) [12] deliberated the effect of applying VMI at a model with single manufacturer and
several retailers. Rout et L. [13] investigated a model to reduce the total cost and also considered
a method to decrease carbon emission. Dong and Kefeng [14] utilized a model to study on the
effect of purchase price and quantity under vendor managed inventory. Stenius et al. [15]
Page 349
SEMIT 2023
presented a model with one storeroom and multi retailers. Demand data followed a Poisson
function. The supply chain members were obliged to give their information to the storeroom.
Weraikat er al. [16] designed a model to search for the ideal number of medications in order to
reduce deficiency of medication.
3. Model formulation
In this segment, a VMI model with one vendor, numerous retailers and numerous distribution
centers is created. The vendor convey supplies to DCs and finally send to the retailers. The site
for the vendor and retailers is definite but that of the DCs are undermined. We endeavor to
minimize the total cost per unit of time (PUOT) for each party in the supply chain within
discovering the economic place to locate the DCs and determining on the best way for assigning
the DCs to the retailers. In this model we deem that vendor and retailers pay ordering cost and
holding (PUOT). Also, we assume that if the vendor make provision beyond the upper stock limit
for the retailers (the shaded triangle for retailer 2 in figure 1.), the penalty cost is levied per unit
of time. Moreover, we calculate cost for locating DCs and shipment cost between retailers and
DCs in order to detect the superlative place that. There is no limitation in vendor and the DC’s
ordering. Demand information for each of the retailers is fixed and specified.
Determining on the time of renewal and the amount of inventory send to each retailer is vendor’s
duty. So, the vendor has been given access to analyze information to get accurate figure. On the
other hand, the vendor reassurances that the data is secured.
Not only the retailers are not allowed to face shortage, they are permitted by surplus. In other
words, the vendor should provide the exact amount of inventory for each of the retailers.
The vendor supply qj for retailer j in every retailer’s unit of time (Tj ).We presume that the number
of shipments may not be equal for each of the retailers. In figure 1. retailer 1 and 2 have just taken
delivery of three (T = 3 × T1) and two (T= 2 ×T2 ) respectively.
Due to the wide area of the vendor’s warehouse more quantity is stored in compare with retailer’s
storeroom so the total holding cost for the vendor is less than that of the retailers.
The vendor tends to reduce the ordering and shipment cost by supplying large amount of
quantities to the retailers but since VMI contracts are in quest for optimal solution for every party
in the supply chain, an upper stock limit which is equal to economic order quantity (EOQ) is set
for each of the retailers. Penalty cost will levy if the vendor supplies beyond the limit.
Other assumptions include:
• Each retailer must be assigned to one distribution center.
• The replenishment cycle for each of the retailers is independent.
Page 350
SEMIT 2023
Figure 1. A supply chain with one vendor and three retailers under vendor managed inventory
contract (A) Vendor (B) the first retailer (C) the second retailer (D) the third retailer
Following notations are defined to model this problem.
4. Notations
For Retailers
Dj : Demand rate for retailer j
Aj : Ordering cost (OC) for retailer j PUOT
hj: holding inventory cost (HIC) for retailer j PUOT
Page 351
SEMIT 2023
Pj : Overstock cost for retailer j PUOT

Zj : Maximum inventory level for retailer j
For vendor:
hv: The cost for holding inventory in vendor’s warehouse PUOT
Av: The cost accrued to vendor for ordering required inventory PUOT
Dv : Vendor’s demand rate (Dv=∑Dj)
Decision variables:
Tj : replenishment cycle for retailer j
T: vendor’s replenishment cycle
nj : the proportion of consignments received by retailer j in every vendor’s replenishment cycle
qj : ordering quantity dispatched to the 𝑗 𝑡ℎ retailer (qj = Tj × Dj )
qv: ordering quantity for the vendor (qv = Dv × T)
Binary variables:
Xt : if the DC locates in spot t ( X t = 1) otherwise ( X t = 0)
Yjt : if retailer j allots to distribution center t ( Yjt = 1) otherwise ( Yjt = 0)
For distribution center:
Lt : Locating cost if the distribution center establishes at place t
Bjt : transition cost from distribution center establishes at place t to retailer j
J: the proportion of retailers {1 , … , J }
5. Mathematical model
5.1. vendor costs
In this section, the vendor’s Tc is calculated. So, we examine vendor’s OC, IHC and penalty cost.
5.1.1. Vendor’s ordering cost
Based on Eq. (1) by dividing Av which is the ordering cost for the vendor into Tv, the vendor order cost
PUOT is calculated.
Av
Vendor order cost = (1)
T
5.1.2. Vendor’s inventory holding cost (IHC)

qv
According to EOQ the average inventory stored in vendor’s warehouse is equal to 2
and because qv is
Dv ×T
equal to Dv × T so the average inventory is tantamount to 2 . Also,
the cost for maintaining each unit
of inventory is equal to hv. For this reason, the average holding can be calculated using Eq. (2) as
follows.
hv × Dv × T
Average holding cost for retailer j = 2
(2)
5.1.3. Penalty cost
In VMI contracts in order to guard retailers from overstock issues, the vendor obliges to pay penalty
cost for furnishing more than the limited level. Moreover, we assume that, the retailers who receive
excessive stock, are added to S. By means of Eq. (3), L is calculated. In fact, in Eq. (3), we compute the
Page 352
SEMIT 2023
volume of quantity which is sent to each of the retailers and compare it to the limited level. If 𝑊𝑗 is
positive integer, then that retailer belongs to L.
Wj =Tj ×Dj -Zj j∈L
𝑊𝑗 = 0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 (3)
Figure 2. indicate the retailer which receive excess inventory.

Eq. (4), indicates the penalty cost per unit of time. The proof is in Appendix A.
(qj - Zj )2 × nj
The total penalty cost = ∑jϵJ (Pj × 2×Dj ×T
) (4)
By simplifying Eq. (4) the penalty cost PUOT is equal to eq. (5) as follows.
T2 T
( × Dj 2 + (Zj )2 - 2 × × Dj ) × Zj
n2j nj
∑( ) × nj =
2 × Dj × T
jϵJ
Dj (Zj )2
∑ ( Pj × T × )+ ∑ ( nj × Pj × ))- ∑ ( Pj × Z ) (5)
2 × nj 2×T × Dj j
jϵL jϵL jϵL
So, with regard to Eq. (3) and (4) and (5) the vendor cost is computed in Eq. (6) as follows.
Av hv × Dv× Tv Dj (Zj )2
+ + ∑ ( Pj × T× )+ ∑ ( nj × Pj × ))- ∑ ( Pj × Z ) (6)
T 2 2 × nj 2 × T ×Dj j
jϵL jϵL jϵL
5.2. Retailer costs

In this segment, the total costs for the retailers are calculated. Thus, we scrutinize the retailer’s OC,
IHC.
5.2.1. Ordering cost for the retailers (OC for retailers)
By dividing ∑j∈J Aj which is the sum of OC for the retailers into Tj , the retailers ordering cost can be
shortened in Eq. (7) as follows.
Aj
order costs for the retailers =∑j∈J (7)
Tj
Due to, T= Tj ×nj, so the total ordering costs for the retailers PUOT is equal to Eq. (8)
Aj × nj
Total order costs for the retailers =∑j∈J T
(8)
5.2.2. Inventory holding cost for the retailers (IHC for Retailer)
M.Hargia et al. [1] showed that the Eq. (9) illustrated the IHC for the vendor and the retailers
hv× ∑j∈J ( Dj ) × T) hv × Dj × Tj ∑j∈J ( hj × Dj × Tj )
-∑( + ) (9)
2 2 2
j∈J
The first term in IHC mentioned above is the average IHC for the vendor thus the second and third
terms indicates the average IHC for the retailer which is equal to Eq. (10) as follows
Page 353
SEMIT 2023
1
Average IHC for retailers =2 × ∑jϵJ (hj - hv) × Dj × Tj (10)
Thus, the average IHC for the retailers is simplified in Eq. (11)
1 T
Average IHC for retailers =2 × ∑jϵJ(hj - hv) × Dj × nj
(11)
Hence, total cost for the retailers in regard to Eq. (8) and (11) can be determined as follows.
Aj × nj 1 T
∑ + × ∑(hj - hv) × Dj × (12)
T 2 nj
j∈J jϵJ
5.3. Distribution centers costs (DCs costs)

The problem is to determine the superlative site to locate the distribution centers in order to reduce the
transition costs from distribution centers to the retailers and also to decrease the FLCs.
In order to determine the way to allot distribution centers to retailers we assume Yjt as binary variable
that indicate if distribution t allot to retailer j or not. The transition cost from distribution t to retailer j
is simplified in Eq. (13).
Transition cost from DCs to the retailers =∑j∈J ( ∑t∈T Bjt × Yjt × Dj ) (13)
X t is a binary variable that indicates if the distribution center establishes at place t. Fixed locating cost
is calculated in Eq. (14).
FLC =∑( Lt × X t ) (14)
So, the total distribution center cost is provided as follows.
∑( Lt × X t ) + ∑( ∑ Bjt × Yjt × Dj ) (15)

j∈J t∈T
6. Constraint for the vendor and retailers

In this model, there is a number of constraints which are calculated from Eq. (16) to Eq. ( 20)
Eq. (16) shows that each retailer receives nj shipment during the vendor cycle time.
T=nj ×Tj (16)
𝑊𝑗 is equal to the difference between the amount of quantity that was sent by the vendor and upper
stock limit.
The number of retailers which receive overstock inventory is calculated in Eq. (17). Obviously if
supplied quantity is less than upper stock limit , 𝑊𝑗 will be zero.
Wj =Tj × Dj - Zj j∈L
Wj =0 otherwise (17)
Eq. (18) is abridged formula that indicates retailers cannot receive inventory from two or more DCs.
∑m
t=1 Yjt =1 (18)
Discernibly, a DC should be situated at place t in order to be able to handover inventory to the retailers.
Eq. (19) stipulates the mentioned constraint.
Yjt ≤ Xt (19)
Page 354
SEMIT 2023
Eq. (20) demonstrate standard integrality limitations.

Xt = {0,1}
Yjt = {0,1}
Tj ,T ≥ 0 (20)
Thus, the cost for the entire supply chain within one vendor, numerous retailers and several DCs is
considered as Eq. (21) by totalizing Eq. (6), (12), (15).
Av hv × Dv × Tv Dj
+ + ∑ ( Pj × T × )+ ∑ ( nj × Pj
T 2 2×nj
jϵL jϵL
(Zj )2
× ))- ∑ ( Pj
2×T ×Dj
jϵL
Aj × nj 1 T
× Zj ) + ∑ + ∑ ( L ×Xt )+ ∑ ( ∑ Bjt × Yjt × Dj ) + × ∑(hj -hv) × Dj × =
T t 2 nj
j∈J j∈J t∈T jϵJ
2 nj
Av + ∑j∈J Aj × nj +0.5 × ∑j∈L Pj × Zj ×
Dj
+
T
(hj-hv)×Dj Dj
hv × Dv+ ∑jϵJ + ∑jϵL Pj × m
nj nj
×T + ∑ Pj × Zj + ∑ ( L × Xt ) +
2
j∈L t=1 t
m
∑ ( ∑ Bjt ×Yjt ×Dj ) (21)
t=1
j∈J
Therefore Eq. (22) provides the objective and constraint to reduce the total cost (Tc).
nj
Av + ∑j∈J Aj × nj +0.5 × ∑j∈L Pj × Z2j ×
Dj
min 𝑇𝐶 +
T
(hj - hv)×Dj D
hv × Dv+ ∑jϵJ + ∑jϵL Pj × j m
nj nj m
×T + ∑ Pj × Zj + ∑ ( L × Xt ) + ∑ ( ∑ Bjt ×Yjt × Dj )
2 t=1
j∈L t=1 t j∈J
T=nj × Tj
Wj = Tj × Dj − Zj j∈L
Wj = 0 otherwise
Yjt ≤ Xt
m
∑ Yjt = 1
t=1
Xt = {0,1}
Yjt = {0,1}
Tj ,T ≥ 0
∀𝑗 ∈ 𝐽 , ∀𝑡 ∈ 𝑇, 𝑛𝑗 𝑖𝑠 𝑎𝑛 𝑖𝑛𝑡𝑒𝑔𝑒𝑟
T = {1,…, m} (22)
Page 355
SEMIT 2023
7. Heuristic method
M.Hargia [1] showed that, heuristic method generates a near optimal resolution. The initial phase to
calculate near optimal cost for the entire supply chain, is to assume that n1 = n2 = nj and calculate
TC. Eq. (23) is the TC function for equal nj .
Av hv × Dv × T (qj -Zj )2 × n Aj × n 1
+ + ∑ (Pj × ) +∑ + × ∑(hj -hv) × Dj ×
T 2 2×Dj ×T T 2
jϵJ j∈J jϵJ
T
+ ∑ ( L ×Xt )+ ∑ ( ∑ Bjt ×Yjt ×Dj ) (23)
n t
j∈J t∈T
Under these circumstances the optimum amount of n and the total cost are considered. Based on the
procedure, for the next reiterations by means of Eq. (24) and (25) unequal nj is calculated. Eq. (24) is
examined for the retailers who face overstock and Eq. (25) is for the retailers whose transferred supplies
are less than the defined limit. As mentioned before, the place for positioning the DCs are indistinctive
and the method is experimental. So, we randomly numbered
Xt in order to discover the economic spot for placing DCs.
Also, the method that DCs allocate to the retailers are unknown so we have to number Yjt to examine
the Tc.
hj -hv+Pj
nj ×(nj -1)≤0.5× ×Dj ×T2 ≤ nj ×(nj +1) (24)
(Zj )2
Aj +0.5×Pj × D
j
hj -hv
nj × (nj -1) ≤0.5× × Dj × T2 ≤ nj × (nj +1) (25)
Ar
By calculating the amount of nj from Eq. (24) and (25), the cost for the entire supply chain is calculated
in Eq. (26) as follows.
𝑇𝑐
nj
Av + ∑j∈J Aj × nj +0.5 × ∑j∈L Pj × Z2j × D
j
= +
T
(hj-hv)×Dj Dj
hv × Dv+ ∑jϵJ + ∑jϵL Pj × m
nj nj
×T + ∑ Pj × Zj + ∑ ( L × Xt ) +
2
j∈L t=1 t
m
∑ ( ∑ Bjt ×Yjt ×Dj ) (26)
t=1
j∈J
If the estimated nj from above equations, has not improved, the algorithm is stopped and the optimal
Tc is equal to second to last iteration (number v-1), otherwise we continue the process.
The aforementioned results indicate that the quantity of nj , is estimated from Eq. (24) and (25) and the
optimal amount of T is calculated in Eq. (27) as follows. Moreover, by replacing the amount of T in Eq.
(27) Tc is computed.
T
1
Av + ∑j∈J Aj × nj + 2 × ∑j∈L Pj ×Zj2 × nj ÷ Dj )
= √2× (27)
Dj Dj
(hv×Dv + ∑j∈J(hj − hv) × n + ∑j∈L Pj × n ))
j j
Page 356
SEMIT 2023
In other terms the process begins with equal nj . Then Eq. (24) and (25) expand the process for an
unequal nj to calculate Tc. The process continues calculating till the amount of nj and Tc has not
meliorated. At this point, the process is stopped and the optimal amount of TC is for repetition number
v -1. It is noticed that the results are approximately ideal.
In the following section we explained a brief overview of the phases in the process for disparate nj .
After calculating the amount of n, T and Tc for the first reputation we set them as initial answer.
𝑛𝑗1 = 𝑛 , 𝑗 ∈ 𝐽
𝑇𝑐1 = 𝑇𝐶 , 𝑇𝐶 ∗ = 𝑇𝐶
v=2, 𝑇 𝑣 = 𝑇
1. We calculate the amount of nvj by using Eq. (24), (25)
2. We find the amount of Xt and Yjt for r repetition (X tv and Yjtv) which 𝑋𝑡 indicates locating Dc at
place t and 𝑌𝑗𝑡 shows that which retailer is assign to 𝑣 𝑡ℎ 𝐷𝑐.
3. Computing the reorder intervals and determine the members of L
Tv
Tjv =
nvj
if Tjv × Dj > Zj , the retailer adds to L
𝑖𝑓 𝑇𝑗𝑉 × 𝐷𝑗 < 𝑍𝑗 , 𝑜𝑡ℎ𝑒𝑟 𝑤𝑖𝑠𝑒
4. We calculate Tc
If the Tc in iteration number v is less than that of in number v-1, next phase should be calculated,
otherwise if the cost for the entire supply chain sees an increase from repeat number v-1 to repeat
number v, the process stops and the optimal Tc is for repetition number v-1.
If Tc v > Tc v−1 , stop
Other wise Tc ∗ = Tc, T v = T ∗ , n∗j = nvj
5. if nvj = nv−1
j , stop
There should be an improvement in the amount of nj in every repetition. If the mentioned condition
is not fulfilled the procedure stops.
6. if the 𝑋𝑡v = 𝑋𝑡v−1 𝑎𝑛𝑑 𝑌𝑗𝑡v = 𝑌𝑗𝑡v−1 stop
If the amount of Xt and Yjt for repeat number v is equal to that of number v-1, the Tc will not develop
so the process stops.
7. set v v+1
8. We compute 𝑇 𝑣
1 2
Av+ ∑j∈J Aj ×nv-1 v-1
j + 2 × ∑j∈L Pj ×Zj ×nj ÷Dj )
v
T =√2×
Dj Dj
(hv×Dv+ ∑j∈J(hj -hv)× v-1 + ∑j∈L Pj× v-1 ))
nj nj
9. In this phase we calculate reorder intervals

Tv
Tjv =
nv−1
j
10. We decide on the retailers who belong to L. If Tjv × Dj > Zj ,the retailer j belongs to L.
Page 357
SEMIT 2023
11. Start the process again
8. An illustrative example
We provided an example with one vendor, 6 retailers and 8 distribution centers to designate the efficacy
of the model. This model is solved in heuristic procedure by PYOMO, group of Python software
packages for formulating optimization models.
Table 1. Retailer’s data for the numerical example

Retailer 𝐃𝐣 𝑨𝐣 𝒉𝐣 𝑷𝐣 𝒁𝐣
1 2000 45 8 2 150
2 1000 50 4 4 158.1139
3 3000 60 10 3 189.7367
4 2500 80 8 4 223.6068
5 3500 90 6 5 324.037
6 5000 95 10 6 308.2207
The amount of Zj is calculated by using the economic order quantity equation (EOQ).
Table 2. Vendor’s data for the numerical example
Dv Av hv
17000 119 3
Dv is equal to ∑ Dj .
Table 3. Dc’s data for the numerical example

Dc 𝐋𝐭
1 500
2 600
3 550
4 700
5 770
6 560
7 650
8 750
𝐁𝟏𝟏 = 3 𝐁𝟐𝟏 = 3 𝐁𝟑𝟏 = 2 𝐁𝟒𝟏 = 5 𝐁𝟓𝟏 = 2 𝐁𝟔𝟏 = 4

𝐁𝟏𝟐 = 4 B22 = 2 B32 = 3 B42 = 4 B52 = 3 B62 = 2
𝐁𝟏𝟑 = 2 B23 = 2 B33 = 4 B43 = 5 B53 = 3 B63 = 2
𝐁𝟏𝟒 = 3 B24 = 4 B34 = 4 B44 = 6 B54 = 3 B64 = 5
𝐁𝟏𝟓 = 5 B25 = 3 B35 = 4 B45 = 7 B55 = 4 B65 = 3
𝐁𝟏𝟔 = 4 B26 = 4 B36 = 5 B46 = 6 B56 = 4 B66 = 5
𝐁𝟏𝟕 = 3 B27 = 5 B37 = 6 B47 = 2 B57 = 4 B67 = 6
𝐁𝟏𝟖 = 2 B28 = 5 B38 = 7 B48 = 3 B58 = 4 B68 = 3
The result of repeat number 1 indicates that if Dc number 1 to number 7 locates and DC number1
allocate to retailer 6, Dc number 2 assign to retailer 4 and so on, the total cost is 88332.89. According
to the method in order to estimate the amount of Tc, at least two Tc should be calculated.
Page 358
SEMIT 2023
Table 4. The result of the first iteration solved by PYOMO
Xt Yjt nj Wj
1 𝑌16 = 1 1 0
1 𝑌24 = 1 1 0
1 𝑌33 = 1 1 20.26334
1 𝑌45 = 1 1 0
1 𝑌52 = 1 1 0
1 𝑌61 = 1 1 41.7793
0 − v=1
0 −
T = 0.07
Total cost = 88332.89
The outcome of the repetition number 2 indicates that, if Dc number 3 to 8 establishes and the Dc
number 1 and 2
allot to retailer number 8 and 7 and so on, the Tc is 83987.173. The results obtain from the first and
second repetition shows that the amount of Tc drops significantly. Therefore, we continue the process
and compute Tc for the third iteration.
Table 5. The result of the second iteration solved by PYOMO
Xt Yjt nj Wj
0 𝑌18 = 1 2 0
0 𝑌27 = 1 3 0
1 𝑌33 = 1 1 58.05134
1 𝑌46 = 1 3 0
1 𝑌54 = 1 4 0
1 𝑌65 = 1 1 104.7593
1 − v=2
−
1Total cost = 83987.173
T = 0.082596
By considering the consequences of v = 3, Dc number 3 to 8 placed and Dc number 1 and 2 allot to

retailer 3 and 5 and the amount of Tc is equal to 73731.09. Since the proportion of Tc diminish
dramatically, we continue to next iteration.
Table 6. The result of the third iteration solved by PYOMO
Xt Yjt nj Wj
0 𝑌13 = 1 4 0
0 𝑌25 = 1 4 0
1 𝑌34 = 1 1 178.1623
1 𝑌47 = 1 2 0
1 𝑌56 = 1 3 0
1 𝑌68 = 1 1 304.9443
Total cost = 73731.09
1 − v=3
1 −
T = 0.122633
\
Page 359
SEMIT 2023
The results of the fourth repetition indicates that if Dc number 3 to 8 placed and Dc number 1 allocate
to retailer 3, Dc number 2 assign to retailer 4, Dc number 3 allocate to retailer 6 and so on, the Tc is
approximately 77729.23. The amount of Tc calculated in the fourth repetition sees a momentous
increase. Based on the process the process stops when 𝑇𝑐 𝑣 > 𝑇𝑐 𝑣−1 . As the result implies in (v = 3)
the total cost is 73731.09 and in (r = 4) the total cost is 77729.23 so the optimal amount of Tc is equal
to 73731.09.
Table 7. The result of the fourth iteration solved by PYOMO
Xt Yjt nj Wj
0 𝑌13 = 1 4 0
0 𝑌24 = 1 4 0
1 𝑌36 = 1
1 173.0593
1 𝑌47 = 1
2 0
1 𝑌55 = 1
3 0
1 𝑌68 = 1
1 296.4393
1 −
v=4 Total cost =77729.23
1 −
T = 0.120932
9. Conclusion
In this paper we investigate a model with one vendor, numerous retailers and numerous distribution
centers under vendor managed inventory convention. Capacity limitation at retailer’s warehouse
constraint vendor from supply large amount of stock. In the case the vendor supplies excessive stock to
the retailers, overstock cost is levied. The optimal number of distribution centers is computed.
Moreover, we detect the best method to allot distribution centers to retailers to decrease the amount of
Tc. To solve this problem, heuristic procedure is suggested. By usage of iterative approach, the total
cost will be near-optimal. Also, we use PYOMO, to solve this problem and calculate the minimal total
cost.
Appendix A. proof of total penalty cost
Figure 2. the total penalty cost which is paid by the vendor is equal to the area of the triangles above
Zj
Wj is the excessive stock send to retailer j and is computed as follows
Wj = qj − Zj
The area of the triangles= ∑((Wj × X) / 2) × nj
By usage of Tale’s theorem, the amount of X is calculated:
Page 360
SEMIT 2023
X Wj
Tj
= qj
, X = (Tj × Wj )/qj
So the area of one mentioned triangle is (Wj × (Tj × Wj ))/(2× qj ) ,

qj
Sine Tj
= Dj , qj = Dj × Tj , by replacing qj the area of one specified triangle is as follows
Wj2 × Tj Wj2
∑( ) = ∑( )
2× Dj × Tj 2× Dj
If the area of one triangle multiply by nj , the sum of areas of triangle is calculated as follows.
Wj2
∑( )×nj
2 × Dj
For calculating the phrase for one period of vendor we divide it into Tv
W2j × nj
So the areas of the triangles is ∑( )
2× Dj × Tv
Since Tv = nj × Tj , we replace Tv and recalculate the sum of areas of triangles.

W2j × nj W2
j
∑( ) = ∑( 2× D × T )
2× Dj × nj × Tj j j
Pj × W2j
If we multiple it by Pj , (∑( 2 × Dj × Tj
)), the total penalty cost is calculated.
So, the amount of penalty cost that is paid by the vendor is equal to
Pj × Wj2
∑( )
2 × Dj × Tj
References
1. M. Hariga , M. Gumus , A. Daghfous , S.K. Goyal: A vendor managed inventory model under
contractual storage agreement, Operations Research 40 (2013) 2138–2144
2. Oded Berman , Dmitry Krass , M. Mahdi Tajbakhsh : A coordinated location-inventory model,
Operational Research 217 (2012) 500–508
3. Mateen, A., Chatterjee, A.K., Mitra, S.: VMI for Single-Vendor Multi-Retailer Supply Chains
under Stochastic Demand, (2014), doi: http://dx.doi.org/10.1016/j.cie. 2014.10.028
4. M.A. Darwish , O.M. Odah : Vendor managed inventory model for single-vendor multi-retailer
supply chains, Operational Research 204 (2010) 473–484
5. Fry MJ, Kapuscinski R, Olsen TL. Coordination Production and delivery under a (z, Z)-type
vendor-managed inventory contract. OperationsManagement 2001;3(2):151–73.
6. Goyal SK.: A one-vendor multi-buyer Integrated Inventory Model Operational Research
82(1995) 209-210
7. Subrata Mitra, A.K. Chatterjee: Leveraging information in multi-echelon inventory systems
Operational Research 152 (2004) 263–280
8. Javad Sadeghi a,Ahmad Sadeghi b, Mohammad Saidi-Mehrabad: A parameter-tuned Genetic
Algorithm for Vendor Managed Inventory Model for a Case Single-vendor Single-retailer with
Multi-product and Multi-constraint
9. Zuo-Jun Max Shen, Collette Coullard, Mark S. Daskin: A Joint Location-Inventory Model,
February 2003Transportation Science 37(1):40-55
Page 361
SEMIT 2023
10. Hariga, M., Gumus, M., & Daghfous, A. (2013a).: Storage constrained vendor managed inventory
models with unequal shipment frequencies. Omega (Accepted)
11. G.P. Kiesmuller , R.A.C.M. Broekmeulen : The benefit of VMI strategies in a stochastic multi-
product serial two echelon system, Operations Research 37 (2010) 406 – 416
12. Jonah Tyan and Hui-Ming Wee," Vendor Managed Inventory: a survey of the Taiwanese grocery
industry ", Journal of Purchasing and Supply Management, Pages 11-18, January 2003
13. Chayanika Rout, Arjun Paul, Ravi Shankar Kumar, Debjani Chakraborty, Adrijit Goswami:
Integrated optimization of inventory, replenishment and vehicle routing for a sustainable supply
chain under carbon emission regulations Operations Research 316 (2021) 128256
14. Yan Dong and Kefeng Xu, "A supply chain model of Vendor Managed Inventory",
Transportation Research Part E: Logistics and Transportation Review, Pages 75-95, April 2002
15. Olof Stenius, Johan Marklund, Sven Axsäter: Sustainable multi-echelon inventory control with
shipment consolidation and volume dependent freight costs Operations Research 267 (2018) 904–
916
16. Dua Weraikat, Masoumeh Kazemi Zanjani, Nadia Lehoux: Improving sustainability in a two-
level pharmaceutical supply chain through Vendor-Managed Inventory system Operations
Research 21 (2019) 44–55
Page 362
SEMIT 2023
Prioritization of Freight Carrier Selection Factors with the Best Worst Method in Multi-
Criteria Decision Analysis
Selin Yalçın
Ertuğrul Ayyıldız
Abstract
Efficient freight carrier selection is pivotal to modern logistics and supply chain management,
demanding a systematic approach to navigate the complexity of factors such as cost, reliability,
sustainability, and collaboration. This study presents a comprehensive investigation into the
prioritization of these factors using the Best Worst Method (BWM) within the framework of Multi-
Criteria Decision Analysis (MCDA). Drawing insights from literature, industry expertise, and
stakeholder perspectives, a criteria set is identified. These encompass economic considerations, service
quality metrics, environmental impact, collaborative capabilities, and delivery timeliness. The holistic
perspective captures the multifaceted nature of contemporary carrier selection decisions. This study
employs the BWM, one of most effective MCDA approach, to establish a structured evaluative
framework. Expert judgments and pairwise comparisons aid in the identification of both "best" and
"worst" criteria for each factor, enabling the assignment of priority weights. This research has profound
implications for logistics practitioners, supply chain managers, and decision architects engaged in
carrier selection processes. Prioritization shows a clear path for well-informed decision-making aligned
with strategic goals and sustainability objectives. Moreover, the study enriches the theoretical
understanding of carrier selection within a multifaceted criteria context. This study offers a novel
framework that elevates decision-making precision in carrier selection. It contributes to operational
efficiency, sustainability, and collaborative dynamics within contemporary logistics paradigms.
Keywords: freight carrier; Best-Worst method; sustainability; logistics
1. Introduction
With the increase of globalization, it is predicted that there will be a rise in demand for freight
transportation. In contemporary logistics, where efficiency is paramount, the movement of goods and
raw materials plays a pivotal role, accounting for over half of the logistics costs (Guo and Peeta, 2015).
However, the inherent nature of freight transportation introduces various risks into the supply chain,
making the effective management of these risks a matter of paramount importance. Ensuring the
uninterrupted transportation of products or raw materials within the supply chain reduces the fragility
of the chain. To minimize the risks, it is essential to select the right freight carrier. Businesses' freight
transportation costs include a range of cost items such as transportation fees, fuel costs, customs duties,
and transportation insurance. Cost items directly impact overall business costs, influencing pricing
strategies and competitive power. This, in turn, leads to reaching fewer customers. Another factor that
directly affects customers is service quality. Service quality is of great importance to customer
satisfaction as it ensures the timely and damage-free delivery of the product the customer desires.
Customer satisfaction also leads to increased customer loyalty, which is essential for businesses. The
formulas developed to increase the efficiency of freight carriers and reduce costs have become highly
significant with the competitiveness of freight transportation (Lin and Ng, 2012). The ability of
businesses to compete is directly related to their choice of the right freight carrier. Choosing the right
freight carrier allows for cost reduction and on-time deliveries, thereby increasing the company's market
share. Freight carriers also contribute to the creation of adverse environmental conditions. The cargoes
carried by heavy freight carriers are increased fuel consumption and carbon emissions. Fuel
consumption and the emission of greenhouse gases released into the atmosphere as a result of
combustion can lead to issues related to air pollution and climate change. The extensive use of energy
and natural resources in transportation may cause the danger of destruction of natural habitats and
rapidly exhausting natural resources. Wastes such as oil leaks spilled into the environment during
Page 363
SEMIT 2023
transportation negatively affect the environment. The proper selection of freight carriers is of vital
importance for both environmental protection and the preservation of our health.
In the modern transportation environment, decision-makers face a number of challenges, pricing due to
high competition, service quality, being subject to a series of regulations arising from environmental
protection laws, keeping up with advancing technology, complexities in transport integrations, the
necessity of timely and speedy deliveries, and security concerns (de la Torre et al., 2021). To overcome
these challenges, it is essential to determine the freight carrier selection factors.
Multi-Criteria Decision Analysis (MCDA) is an effective tool used in modeling real-life problems with
uncertainty, solving complex problems, designing and developing products or systems, optimizing the
use of resources for sustainable environments, and determining the strategic plans of businesses. The
prioritization of freight carrier selection factors can be found using MCDA. In this study, the Best-
Worst Method (BWM), one of the MCDA proposed by Rezaei to address the shortcomings of AHP, is
used to determine the importance weights of freight carrier selection factors. The BWM enables the
structured handling of complex decision-making processes, providing decision-makers with a better
understanding and evaluation of options. By direct comparisons among alternatives, it helps to see
differences between options clearer. It allows more reliable results to be obtained by systematically
handling subjective evaluations. In various fields, BWM can be used to improve the gains of businesses
in making managerial decisions. Thus, by ranking the factors to be considered in the selection of freight
carrier in the logistics sector according to their weights, it will contribute to identifying collaborations
for businesses. Businesses need a systematically structured methodology to make better decisions when
selecting partnerships. Using a systematic approach in freight carrier selection allows for the reduction
of logistics costs, regular monitoring and improvement of carrier performance, predictability of
potential risks, making better decisions with increased accessibility to information, and gaining a
competitive advantage in freight transportation.
The rest of this paper is structured as follows: Section 2 includes a brief summary about various studies
related to freight carrier selections in the literature. Section 3 explains the BWM used in the study.
Freight carrier selection factors obtained from the literature review are identified, and their importance
weights are determined by using the BWM into Section 4. The obtained results are discussed into
Section 5. The general inferences, future research directions and suggestions for further research are
discussed into Section 6.
A literature review on freight transportation and BWM based on Scopus database give a list of 13
publications. In Fig. 1, it is seen that the studies using BWM on the freight transportation are shown
their frequencies and they are applied to many subject areas such as business, management and
accounting, engineering and environmental science.
Page 364
SEMIT 2023
Figure 1. Distribution of the publications (a) based on years (b) by type of subject areas
This study's motivations can be described as (i) Critical criteria that are of vital importance in selecting
freight transportation have been identified, contributing to the transportation sector. (ii) The inclusion
of the opinions of nine experts specialized in the field of transportation in this study is crucial in terms
of the reliability of the results. (iii) The use of the BWM has allowed us to overcome various issues that
may arise from the pairwise comparisons in the AHP method, resulting in more accurate results. In
summary, this study's contribution and originality can be described as follows: (i) A novel approach is
introduced to determine prioritization of freight carrier selection factors. ii) The developed approach is
used to handle complex realistic freight carrier selection under uncertainty. (iii) The results contribute
to businesses operating in transportation activities.
The increasing global population is leads to a growing demand for both passenger and freight
transportation. Freight transportation includes significant factors such as cost, on-time delivery, and
reliability, making it crucial to have proper planning and selecting the right freight carriers. There are
different studies in the literature on the selection of freight carriers such as: Setamanit and Pipatwattana
(2015) emphasized the significance of transportation cost, efficiency, the ability to deliver products to
customers on time, and customer satisfaction in the selection of freight carriers. Additionally, Japanese
automotive companies have been found to prioritize criteria such as the reliability of service, service
quality, cost, after-sales service, and perceived capability when choosing ocean freight carriers. Čarný
et al. (2020) examined four different freight carriers to choose the right business partner in the railway
transportation market and evaluated their ability to meet the needs of customers. Guo and Peeta (2015)
mentioned the significance of rail-truck multimodal transportation due to factors such as highway
congestion, truck driver shortage, and the need for increased financing for highway infrastructure
construction and renewal. Therefore, they proposed a two-step modeling approach for rail freight carrier
decision-makers to enable collaborations. Kuzior and Staszek (2021) examined the Polish railway
undertaking to determine what measures can be taken to improve energy efficiency in railways. Soro
and Wang (2012) examined the route choice behaviors of freight carriers during peak hours and off-
hours under the constraint of delivering the cargo within a given time window. Shah and Brueckner
(2012) developed an analytical model of price and frequency competition among freight carriers. As a
result of the literature review, the importance of freight carrier selection factors has been observed
through the examination of different application areas. The evaluation of freight carrier selection factors
using the BWM has been brought a new perspective to the literature.
Page 365
SEMIT 2023
3. Best Worst Method

The Best-Worst Method (BWM) is a subjective method developed by Rezaei (Rezaei, 2015, 2016) with
the aim of addressing the shortcomings of the AHP method. Rezaei criticized AHP due to the numerous
pairwise comparison matrices for causing to inconsistencies. BWM proposes a new structure for
pairwise comparisons. It is based on a fundamental principle that involves the use of reference objects
for benchmarking purposes (Mi et al., 2019). These reference objects are identified as the most
important (𝐶𝐵 , the best) and the least important (𝐶𝑊 , the worst). The comparisons are only made with
these reference objects in the BWM. BWM simplifies the decision-making process by significantly
reducing the number of pairwise comparisons required, typically to (2𝑛 − 3) times. For 𝑛 criteria,
experts make (𝑛 − 1) comparisons with 𝐶𝐵 . On the other hand, (𝑛 − 2) comparisons are required with
𝐶𝑊 , which is already compared with the best before. All these pairwise comparisons are covered by the
ones made in the AHP. Perspectives obtained from comparisons outside of these comparisons in AHP,
in the other hand, have the potential to be derived from comparisons with references in BWM.
The basic procedure of BWM is follows as:
Step 1. Experts determine a set of 𝑛 evaluation criteria {{𝐶1 , 𝐶2 , … , 𝐶𝑁 } appropriate for the decision
problem.
Step 2. Each decision-maker specifies the criteria within the set of decision criteria based on their own
experiences, which the criteria they consider as the most important (𝐶𝐵 , the best) and the least important
(𝐶𝑊 , the worst).
Step 3. Every decision-maker evaluates the relative preference levels between the best criterion and
each individual criterion in the set, typically using a Likert scale. This assessment obtains 𝐴𝐵 = (𝑎𝐵1 ,
. . ., 𝑎𝐵𝑗 , … , 𝑎𝐵𝑛 ) vector expressed as best to other criteria. Here, 𝑎𝐵𝐽 represents the preference level
of the best criterion over criterion 𝐶𝑗 , where 𝑗 ≠ 𝐵. It's necessary to note that the 𝐴𝐵 does not include
the pairwise comparison where 𝑎𝐵𝐵 = 1.
Step 4. Each decision-makers specifies their preference levels for the criteria other than the best, this
time in comparison to the worst criterion. This yields the others-to-worst vector, 𝐴𝐵 = (𝑎𝐵1 , . . .,
𝑎𝐵𝑗 , … , 𝑎𝐵𝑛 ). In this vector, 𝑎𝑗𝑊 represents the preference level of criterion 𝐶𝑗 over the worst criterion,
where 𝑗 ≠ 𝐵 𝑜𝑟 𝑗 ≠ 𝑊.
Step 5. The obtained weights of the criteria should be based on the preference scores of the best criterion
over the others (𝑤𝐵 ⁄𝑤𝑗1 = 𝑎𝐵𝑗 ), and the preference scores of the other criteria over the worst
(𝑤𝑗1 ⁄𝑤𝑊 = 𝑎𝑗𝑊 ). This means determining a set of weights (𝑤11 , 𝑤21 , … , 𝑤𝑛1 ) that minimizes the
𝑤 𝑤1
differences |𝑤𝐵1 − 𝑎𝐵𝑗 |and |𝑤 𝑗 − 𝑎𝑗𝑊 | for all decision criteria. This can be formulated as a
𝑗 𝑊
minimization model presented in Formula 1.
min 𝜉 (1)
s.t.
𝑤𝐵
| 1 − 𝑎𝐵𝑗 | ≤ 𝜉 , ∀𝑗
𝑤𝑗
𝑤𝑗1
| − 𝑎𝑗𝑊 | ≤ 𝜉 , ∀𝑗
𝑤𝑊
∑ 𝑤𝑗1 = 1
𝑗
𝑤𝑗 ≥ 0, ∀𝑗
4. Application
The selection of criteria was informed by an exhaustive review of relevant literature within the realm
of logistics and supply chain management. This comprehensive exploration revealed the multifaceted
nature of freight carrier selection, where factors contributing to economic efficiency, operational
Page 366
SEMIT 2023
effectiveness, environmental responsibility, collaborative potential, delivery precision, security

measures, and geographical reach emerged as key considerations. Drawing from established research,
industry insights, and stakeholder perspectives, this set of criteria was carefully distilled to encompass
the diverse spectrum of variables influencing modern carrier selection decisions. Tab. 1 presents the
evaluation criteria and their brief explanations.
Table 1. Criteria to evaluate the freight carriers.

Cost Evaluate carriers based on their pricing structures, (Bean and Joubert, 2018;
including transportation fees, additional charges, and Coulter et al., 1989; Guo
overall cost-effectiveness. and Peeta, 2015)
Reliability Assess carriers' track records for on-time deliveries, (Castro and Kuse, 2005; Lin
minimal disruptions, and consistent service quality. and Yeh, 2010; Lin and
Yeh, 2012)
Sustainability Consider carriers' commitment to environmentally (Bask et al., 2016; Davis-
friendly practices, such as fuel efficiency, emission Sramek et al., 2020)
reduction, and eco-friendly technologies.
Collaboration Examine carriers' ability to engage in collaborative (Gentry, 1996)
logistics efforts, including information sharing, load
consolidation, and partnership opportunities.
Timeliness Measure carriers' performance in meeting delivery (Banomyong and Supatn,
deadlines and minimizing transit times. 2011;Ouali, 2022)
Safety Evaluate carriers' safety protocols, risk management (Fanam and Ackerly, 2019;
strategies, and compliance with industry standards. Norbis et al., 2013)
Coverage Consider carriers' reach and coverage in terms of (Kannan et al., 201; Wong et
geographic areas served, both domestically and al., 2008)
internationally
The selection of a freight carrier is of paramount importance in the realm of logistics and supply chain
management. There are several compelling reasons for this significance. Firstly, it directly impacts the
cost efficiency of a business. Different carriers come with diverse pricing structures, and making the
right selection can lead to substantial cost savings. This not only affects a company's bottom line but
also its competitiveness within the market (Izadi et al., 2020). In addition to cost considerations, the
quality of service provided by a freight carrier is a critical factor. Timely deliveries and the safe
transportation of goods are essential for maintaining customer satisfaction. Customer loyalty often
hinges on the ability to consistently meet delivery expectations. A well-chosen carrier can play a vital
role in achieving this (Ho et al., 2017).
On-time delivery is another pivotal aspect of the selection process. In today's fast-paced business
environment, any delays can have a cascading impact on the entire supply chain. Late deliveries can
lead to inventory problems, disrupted production schedules, and harm a company's reputation. The
reliability of a carrier is closely tied to service quality and on-time delivery. Reliable carriers are
consistent in their performance, reducing the risk of disruptions and delays. For businesses that rely on
just-in-time inventory systems or transport perishable goods, this is a critical consideration (de Jong et
al., 2014).
Safety is another essential facet of carrier selection. A carrier's safety record can significantly impact a
company's financial well-being. Accidents or damage to goods during transportation can result in
financial losses, liability issues, and damage to a company's reputation. Opting for a carrier with a strong
safety record is a proactive way to reduce these risks (Lingaitienė, 2008). The environmental impact of
freight transportation is also gaining increasing attention. Companies are under growing pressure to
Page 367
SEMIT 2023
adopt eco-friendly practices. Selecting a carrier with a commitment to sustainability aligns with these
environmental goals and enhances a company's image (Quak, 2011). The coverage and flexibility of a
carrier are determined by its geographic reach and specialized capabilities. Depending on the nature of
a company's operations, the right carrier selection ensures that specific logistical needs are met
effectively (Coulter et al., 1989).
Determining the weights of selection factors is of paramount importance in the decision-making
process. These weights allow decision-makers to customize the evaluation of carriers according to their
unique business objectives and the specific context of their industry. Weights are crucial because they
enable a balanced consideration of trade-offs. For instance, if a carrier offers lower costs but less
reliability, assigning appropriate weights helps in striking the right balance. Moreover, the weighting
process ensures an objective and systematic evaluation of carriers, reducing the potential for subjective
or biased decision-making. This approach fosters transparency and consistency in the selection process.
Weights also aid in resource allocation, enabling companies to focus their efforts on the criteria that
matter most, optimizing the use of time and resources during the selection process. An additional
advantage of determining weights is the ability to continuously evaluate carrier performance. By
regularly assessing carriers against the established criteria, companies can make adjustments as needed
to meet evolving business needs. Weights also play a pivotal role in risk mitigation. By giving higher
importance to factors such as safety and reliability, organizations can effectively reduce the risk of
accidents or delays during transportation. In conclusion, the selection of a freight carrier is a
multifaceted decision that impacts numerous aspects of a company's operations. The determination of
weights for selection factors is equally critical, providing a structured and tailored approach to making
the right choice in accordance with the specific needs and priorities of the organization.
In this study, the presence of a skilled team of experts holds exceptional significance in i) delineating
the nine experts forms the core of the expert group. The expert profiles are given in Tab 2.
Page 368
SEMIT 2023
Table 2. Expert profiles

Experi
Job Title Educational Degree Professional Role Department Research Focus
ence
E Logistics 12 MSc in Supply Chain Oversees carrier selection, Supply Chain Sustainable logistics,
-1 Manager years Management logistics optimization Management transportation efficiency
E Transportation 8 BEng in Industrial Analyzes carrier performance, Transportation Route planning, cost
-2 Analyst years Engineering route optimization and Logistics optimization
E Environmental 10 PhD in Environmental Assesses environmental impact of Environmental Carbon footprint, eco-friendly
-3 Scientist years Science logistics operations Sciences practices
E Supply Chain 15 MBA in Operations Manages end-to-end supply chain, Supply Chain Collaborative logistics,
-4 Director years Management strategic planning Operations partnership strategies
E 6 PhD in Business Conducts research on logistics Business Decision support systems,
Researcher
-5 years Analytics decision-making Analytics multi-criteria analysis
E 9 BSc in Occupational Ensures safety compliance in Risk management, safety
Safety Officer Health and Safety
-6 years Health and Safety transportation operations protocols
E 7 MEng in Transportation Plans efficient transportation Transportation Route optimization,

Transportation
-7 years Engineering routes, mode selection Planning congestion management
Planner
E Sustainability 11 BSc in Environmental Implements sustainable practices Sustainability and Green logistics, sustainable
-8 Manager years Management in logistics CSR supply chains
E Operations 14 PhD in Operations Applies optimization techniques Operations Mathematical modeling,
-9 Researcher years Research to logistics Research optimization algorithms
Page 369
SEMIT 2023
The application of the BWM is leveraged to integrate expert insights and assess the significance of the
criteria. The ensuing phases of the BWM are executed to compute the respective weights for each
individual criterion. In the initial step, experts engage in the evaluation process to establish the weights
assigned to the criteria. Tab. 3 shows the best and the worst criteria for different experts.
Table 3. The best and worst criteria for each expert
Expert Best Worst
E-1 Cost Collaboration
E-2 Timeliness Collaboration
E-3 Safety Collaboration
E-4 Cost Sustainability
E-5 Reliability Sustainability
E-6 Cost Sustainability
E-7 Cost Timeliness
E-8 Reliability Collaboration
E-9 Cost Collaboration
The opinions of the experts manifest notable divergences in their perceptions of criterion importance.
Notably, a consensus emerges among five experts who collectively accord highest significance to the
main criterion "Cost.". Similarly, the criterion "Collaboration" is evaluated as the worst criterion for
five experts. These views of the experts show the diversity of perspectives and the usefulness of a
multidimensional criteria assessment. After collecting expert opinions, the Best-to-Other vectors and
Other-to-Worst vectors were constructed as shown in Tab. 4.
Table 4. Best-to-Others and Others-to-Worst vectors for main criteria
Expert Best-to-Others Others-to-Worst
E-1 1,7,8,9,6,6,6 9,3,2,1,3,4,4
E-2 4,2,2,5,1,1,2 2,4,4,1,6,5,4
E-3 3,2,4,6,4,1,3 3,4,2,1,2,5,3
E-4 1,2,4,3,2,2,3 4,3,1,2,3,3,2
E-5 2,1,8,3,5,2,7 7,8,1,5,3,7,2
E-6 1,2,6,5,3,4,4 6,5,1,2,4,3,2
E-7 1,4,4,2,5,3,2 5,2,2,4,1,3,4
E-8 3,1,4,7,3,2,4 4,6,3,1,5,5,3
E-9 1,4,2,8,1,2,5 7,5,6,1,6,5,5
Then, BWM procedure is applied for each expert to determine the criteria weights. Tab. 5 presents the
criteria weights for each expert.
Page 370
SEMIT 2023
Table 5. Criteria weights for each expert

Cost Reliability Sustainability Collaboration Timeliness Safety Coverage
E-1 0.502 0.084 0.074 0.046 0.098 0.098 0.098
E-2 0.068 0.136 0.136 0.042 0.241 0.241 0.136
E-3 0.124 0.185 0.093 0.056 0.093 0.326 0.124
E-4 0.276 0.153 0.061 0.102 0.153 0.153 0.102
E-5 0.190 0.328 0.035 0.127 0.076 0.190 0.054
E-6 0.342 0.199 0.047 0.080 0.133 0.100 0.100
E-7 0.307 0.087 0.087 0.175 0.053 0.116 0.175
E-8 0.129 0.315 0.097 0.040 0.129 0.194 0.097
E-9 0.260 0.083 0.165 0.027 0.233 0.165 0.066
Tab. 5 provides valuable insight into the criteria weights assigned by different experts, each with their
unique perspective on the factors critical for freight transportation. These criteria are instrumental in
the selection of the most suitable freight carrier, and the distribution of weights among experts signifies
varying priorities within the industry.
First and foremost, "Cost" emerges as a fundamental concern across the board. E-1, E-4, E-6, and E-9
have placed a substantial emphasis on cost, signifying their commitment to minimizing transportation
expenses. Notably, E-1 allocates the highest weight to cost, indicating a strong focus on cost-efficiency.
"Reliability" also stands out as a significant factor. E-4, E-5, E-6, and E-8 attribute relatively higher
weights to reliability. This suggests that these experts highly value carriers that consistently deliver
dependable service quality and meet delivery performance expectations. "Sustainability" is gaining
prominence in the modern freight industry. E-3 and E-9 assign the highest weights to sustainability,
reflecting a growing emphasis on environmentally friendly and sustainable transportation practices.
This trend aligns with increasing concerns about the environmental impact of logistics operations.
"Collaboration" is another factor of interest to the experts. E-4 and E-7 assign higher weights to
collaboration, indicating an appreciation for carriers willing to collaborate and form partnerships or
alliances. Such collaborations often lead to cost savings and more extensive service coverage, appealing
to these experts. "Timeliness" is of paramount importance to E-2 and E-5, who have given it the highest
weights. This underscores their preference for carriers that prioritize on-time deliveries, a critical aspect
of supply chain management to meet customer demands and maintain operational efficiency. "Safety"
also holds a significant position in the criteria. E-3, E-6, and E-8 place greater importance on safety,
suggesting a strong focus on carriers that prioritize safety measures and maintain lower accident rates.
This reduces the risk of damage or loss during transportation, aligning with their priorities. "Coverage"
is seen as pivotal by E-2, E-4, and E-7, who have attributed relatively higher weights to this criterion.
Their preference for carriers with broader geographic or service coverage reflects an emphasis on
flexibility and adaptability, essential for businesses with diverse transportation needs.
These criteria weights offer a glimpse into the diverse preferences and priorities of experts in the freight
transportation field. The variations can be influenced by a range of factors, including the nature of the
cargo, business objectives, and industry trends. When selecting a freight carrier, it's essential to consider
these differing priorities and tailor your selection to align with your specific business needs and goals.
Collaboration among experts and stakeholders is invaluable in finding a balanced approach that meets
both immediate requirements and long-term supply chain objectives.
In the pursuit of deriving final criterion weights for the factors in freight transport, a process of
amalgamation was undertaken, pooling together the individual criterion weights established by each
expert. This endeavor was predicated on the assumption of uniform significance attributed to the
perspectives of all experts involved. With equal weighting assigned to each expert, the cumulative
criterion weights were computed to capture a comprehensive consensus. This approach acknowledges
the collective wisdom of the diverse expert opinions, yielding a balanced and holistic representation of
the criteria's relative importance within the context of freight transport. Fig. 2 shows the final criteria
weights.
Page 371
SEMIT 2023
Coverage 0.106
Safety 0.176
Timeliness 0.134
Collaboration 0.077
Sustainability 0.088
Reliability 0.175
Cost 0.244
0.000 0.050 0.100 0.150 0.200 0.250 0.300
Figure 2. Final criteria weights
5. The obtained results and discussion

As a result of the literature review, the freight carrier selection factors, which are crucial issue in the
supply chain management, are determined to be cost, reliability, sustainability, collaboration,
timeliness, safety, and coverage. These criteria are evaluated using the BWM by taking the opinions of
nine experts who specialized in the field. The results indicate that five experts considered cost to be the
best criterion among freight carrier selection factors, while five experts rated collaboration as the worst
criterion. The Fig. 2 demonstrates that the cost criterion is the most important factor among freight
carrier selection criteria, with safety and reliability also playing critical roles in freight carrier selection.
Collaboration is identified as the least important criterion among freight carrier selection criteria. Based
on these findings, it is evident that decision-makers in modern transportation should prioritize factors
obtained through research efforts to minimize the challenges they encounter.
Notably, "Cost" emerges as the most pivotal factor with a weight of 0.244, underlining the paramount
significance accorded to economic considerations, pricing structures, and cost-effectiveness by the
collective expertise. This prominence signifies a sharp focus on financial aspects within the intricate
landscape of logistics. "Cost" encompasses a range of considerations that profoundly influence carrier
selection. These factors include carrier pricing structures, which consist of shipping fees and additional
charges, and the broader evaluation of cost-effectiveness associated with the chosen carrier. Moreover,
"Cost" necessitates strategic alignment with budget constraints and financial goals, potentially affording
businesses a competitive edge in setting their product or service prices. It also intersects with economic
sustainability considerations, where environmentally responsible practices must be balanced against
cost efficiency. Ultimately, this emphasis on cost serves as a risk mitigation strategy, strengthening
financial resilience and adaptability in the face of dynamic market conditions. In essence, the
significance of "Cost" underscores the thorough examination that logistics professionals and decision-
makers must apply to financial implications, ensuring the efficiency of transportation operations while
strategically positioning their organizations in competitive markets and aligning with budgetary
constraints and sustainability objectives.
Following closely is "Reliability" with a weight of 0.175, reinforcing its substantial role in the decision-
making process, encompassing aspects such as on-time deliveries and consistent service quality.
Meanwhile, "Sustainability" is assigned a weight of 0.088, suggesting that while environmental
consciousness and eco-friendly practices hold importance, they are relatively less critical than cost and
reliability. "Collaboration" receives a weight of 0.077, indicating that collaborative logistics efforts,
such as information sharing and partnership opportunities, are valued but to a lesser extent. "Timeliness"
is moderately weighted at 0.134, signifying its role in ensuring timely deliveries, though it does not
supersede cost or reliability in significance. "Safety" garners a substantial weight of 0.176, underlining
the experts' collective emphasis on safety protocols, risk management, and compliance with industry
Page 372
SEMIT 2023
standards. Lastly, "Coverage" assumes a notable weight of 0.106, highlighting the importance of a
carrier's geographical reach and service coverage.
6. Conclusion
The transportation of goods or raw materials from one place to another is a dynamic process.
Minimizing the fragility of this dynamic structure is of critical importance in supply chain management.
Freight transportation contributes to the development of national economies, increased trade, customer
satisfaction, enhanced competitiveness, and the growth of employment in various sectors. In this
context, freight carrier selection factors have been analyzed. It has been revealed that particular attention
should be paid to the cost criterion, and collaborators can gain an advantage by focusing on the safety
and reliability criteria. In future studies, freight carrier selection factors can be expanded, or the results
obtained through different methods can be examined.
In the world of logistics and supply chain, choosing the right freight carriers is super important. This
study dived deep into understanding which factors matter most when making these choices. In terms of
managerial implications, the findings of this research illuminate the critical significance of these factors
and provide valuable insights for decision-makers in the field. Cost, as the foremost criterion, emerged
as the linchpin of carrier selection decisions. Its prominence reflects the financial considerations that
underlie transportation operations, emphasizing the need for cost-efficient choices while aligning with
strategic goals and budget constraints. Reliability, safety, and coverage also assumed substantial
importance, underscoring the significance of dependable services, safety protocols, and expansive
geographical reach in carrier selection. The study reveals that while sustainability and collaboration are
valued criteria, they assume relatively lower weight compared to cost and reliability. This reflects the
balance required between environmental consciousness and economic viability.
As the logistics landscape continues to evolve in response to technological advancements,
environmental imperatives, and shifting market dynamics, the insights generated by this study provide
a robust foundation for optimizing freight carrier selection in a dynamic and increasingly interconnected
world. Moving forward, this research paves the way for further investigations into real-time data
integration, dynamic decision-making, and the evolving role of technology in shaping the criteria that
govern the logistics and supply chain domain. By embracing these dimensions, logistics professionals
and decision architects can navigate the complexities of carrier selection with greater precision, agility,
and sustainability in an ever-changing environment.
Although this study has made valuable contributions, it's essential to acknowledge certain limitations.
Firstly, the prioritization of freight carrier selection factors has been applied with the BWM by
considering seven criteria in the study. By expanding the important criteria for carrier selection, the
results can be compared. The opinions of different experts in the field can be obtained and evaluated.
The method can be applied to real-life problems by considering a fuzzy environment.
References
Banomyong, R., Supatn, N., 2011. Selecting logistics providers in Thailand: A shippers’ perspective.
Eur. J. Mark. 45. https://doi.org/10.1108/03090561111107258
Bask, A., Rajahonka, M., Laari, S., Solakivi, T., Töyli, J., Ojala, L., 2016. Environmental sustainability
in shipper-LSP relationships. J. Clean. Prod. 172. https://doi.org/10.1016/j.jclepro.2017.11.112
Bean, W.L., Joubert, J.W., 2018. A systematic evaluation of freight carrier response to receiver
reordering behaviour. Comput. Ind. Eng. 124. https://doi.org/10.1016/j.cie.2018.07.030
Čarný, Š., Šperka, A., Zitrický, V., 2020. Multi-criteria evaluation of railway transport using evaluation
method. LOGI - Sci. J. Transp. Logist. 11. https://doi.org/10.2478/logi-2020-0018
Castro, J., Kuse, H., 2005. Impacts of large truck restrictions in freight carrier operations in metro
manila. J. East. Asia Soc. Transp. Stud. 6.
Coulter, R.L., Darden, W.R., Coulter, M.K., Brown, G., 1989. Freight transportation carrier selection
Page 373
SEMIT 2023
criteria. Identification of service dimensions for competitive positioning. J. Bus. Res. 19.
https://doi.org/10.1016/0148-2963(89)90040-4
de Jong, G., Kouwenhoven, M., Bates, J., Koster, P., Verhoef, E., Tavasszy, L., Warffemius, P., 2014.
New SP-values of time and reliability for freight transport in the Netherlands. Transp. Res. Part E
Logist. Transp. Rev. 64, 71–87. https://doi.org/10.1016/j.tre.2014.01.008
de la Torre, R., Corlu, C.G., Faulin, J., Onggo, B.S., Juan, A.A., 2021. Simulation, optimization, and
machine learning in sustainable transportation systems: Models and applications. Sustain. 13.
https://doi.org/10.3390/su13031551
Fanam, P.D., Ackerly, L., 2019. Evaluating ocean carrier selection criteria: perspectives of Tasmanian
shippers. J. Shipp. Trade 4. https://doi.org/10.1186/s41072-019-0042-9
Gentry, J.J., 1996. Carrier involvement in buyer-supplier strategic partnerships. Int. J. Phys. Distrib.
Logist. Manag. 26. https://doi.org/10.1108/09600039610115036
Guo, Y., Peeta, S., 2015. Rail-truck multimodal freight collaboration: Truck freight carrier perspectives
in the United States. J. Transp. Eng. 141. https://doi.org/10.1061/(ASCE)TE.1943-5436.0000792
Ho, T.C., Chiu, R.H., Chung, C.C., Lee, H.S., 2017. Key influence factors for ocean freight forwarders
selecting container shipping lines using the revised dematel approach. J. Mar. Sci. Technol. 25,
299–310. https://doi.org/10.6119/JMST-016-1227-1
Izadi, A., Nabipour, M., Titidezh, O., 2020. Cost Models and Cost Factors of Road Freight
Transportation: A Literature Review and Model Structure. Fuzzy Inf. Eng. 1–21.
https://doi.org/10.1080/16168658.2019.1688956
Kannan, V., Bose, S.K., Kannan, N.G., 2011. An evaluation of ocean container carrier selection criteria:
An Indian shipper’s perspective. Manag. Res. Rev. https://doi.org/10.1108/01409171111146661
Kuzior, A., Staszek, M., 2021. Energy management in the railway industry: A case study of rail freight
carrier in Poland. Energies 14. https://doi.org/10.3390/en14216875
Lin, D.Y., Ng, K.H., 2012. The impact of collaborative backhaul routing on carbon reduction in the
freight industry. Transp. Res. Part D Transp. Environ. 17.
https://doi.org/10.1016/j.trd.2012.08.002
Lingaitienė, O., 2008. A mathematical model of selecting transport facilities for multimodal freight
transportation. Transport 23, 10–15. https://doi.org/10.3846/1648-4142.2008.23.10-15
Mi, X., Tang, M., Liao, H., Shen, W., Lev, B., 2019. The state-of-the-art survey on integrations and
applications of the best worst method in decision making: Why, what, what for and what’s next?
Omega (United Kingdom) 87. https://doi.org/10.1016/j.omega.2019.01.009
Quak, H.J. (Hans), 2011. Urban Freight Transport: The Challenge of Sustainability, in: City
Distribution and Urban Freight Transport. Edward Elgar Publishing, pp. 37–55.
https://doi.org/10.4337/9780857932754.00008
Rezaei, J., 2015. Best-worst multi-criteria decision-making method. Omega 53, 49–57.
Setamanit, S. O., Pipatwattana, A., 2015. Criteria for Ocean Freight Carrier Selection: a Perspective of
Japanese Automotive Company in Thailand. ASBBS eJournal 11.
Shah, N., Brueckner, J.K., 2012. Price and frequency competition in freight transportation. Transp. Res.
Part A Policy Pract. 46. https://doi.org/10.1016/j.tra.2012.02.014
Soro, W. L. ; Wang, X., 2012. Effects of road pricing on freight carrier behaviors analyzed with
experimental economics: comparison of peak hour delivery and off-hour delivery. ransportation
Res. Rec. 2288, 66–73.
Page 374
SEMIT 2023
SP08: Optimization, Decision Making and Support Systems
Optimization models for production planning in the agro-industrial sector: A systematic review
of the literature
Angie del Rocio Sanchez Marquina1
1
Faculty of Mathematical Sciences, National University of San Marcos, Lima, Peru.
angie.sanchez4@unmsm.edu.pe
Abstract
The constant growth of the agro-industrial sector has generated various challenges for companies, such
as high costs, low productivity, reduced profits, loss of customers, and insufficient production levels.
In response to these issues, companies have turned to the use of optimization models in their production
planning. In this context, the objective of this research is to identify the approaches used for applying
optimization models in the production planning of the agro-industrial sector.
To achieve this goal, a systematic review of scientific publications available in online databases was
conducted, considering a period of the last decade. The results of this review reveal that the Mixed
Integer Linear Programming and Linear Programming models are the most widely used in the agro-
industrial sector for production planning. The findings also indicate the versatility of these models as
they can be adapted to different product quantities, planning horizons, and uncertainty in production.
It is worth noting that cost optimization is a common objective in production planning, but few studies
were found to incorporate uncertainty into the model. This suggests a need for further research that
delves deeper into the uncertainty commonly present in the agro-industrial sector.
Keywords: Production Planning, Optimization Model, Mathematical Optimization, Agro-industrial,

Systematic Review.
1. Introduction
According to [1], ignoring production planning can have significant consequences such as misallocation
of resources, low production efficiency, uncertainty in the work environment, and competitive
disadvantage. On the other hand, in [2], the authors point out that the profitability of companies is
affected by various factors, with production planning being one of the most important, as poor planning
can result in high production costs, non-compliance with demand, low profits and obsolete products.
However, this represents a challenge, especially in the agro-industrial sector due to the particularities
of the optimization models used, such as the type of model, solution method, planning horizon, and
uncertainty. For this reason, knowing the approaches to its application can help companies in the sector
to maintain their competitiveness and improve their decision-making.
On the other hand, despite the increase in research during the last decade, in the field of production
planning in the agro-industrial sector, it is still necessary to carry out additional studies in this field [3].
Although it has been found that in sectors such as textiles, automobiles, electronics, furniture and
fashion, the most widely used optimization models are Fuzzy Programming, Multi-Objective
Programming, Linear Programming, Integer Programming and Mixed Integer Linear Programming [4],
it is important to focus the review on the agro-industrial sector.
Therefore, the objective of this research is to analyze the optimization models used for production
planning in the agro-industrial sector. To achieve this, a systematic review of scientific articles and
theses published in the last 10 years was carried out to obtain the most current application approaches
in the field of study, in Spanish, English and Portuguese stored in databases available online.
Page 375
SEMIT 2023
This research is divided into 4 sections: Section 2 presents a general review of the conceptual
framework related to the topic of interest. Section 3 describes the research methodology used in this
review. Section 4 analyzes the results of the research review, and in section 5, the results are discussed,
and the conclusions of the research are presented.
2. Theoretical Referential
In recent years, there has been a remarkable growth in the use of optimization models in production
planning, driven by both the interest of the business sector and the scientific community [4]. Production
planning plays a fundamental role in various industries, especially in the agro-industrial sector, where
balancing production levels and capacity constraints is essential to meet demand, improve productive
performance, reduce costs and enhance competitiveness.
In [5], it is stated that production planning provides a guide to achieve production objectives, selecting
the necessary resources to carry them out, and identifying the best alternative to maximize benefits.
However, in [6] it is pointed out that production planning can be complex due to various factors such
as labor, quantity of products, raw material and demand. In this context, mathematical optimization
models, such as mathematical programming, prove to be very useful in addressing the challenges of
production planning [7]. In addition, [8] mentions that the increasing use of intensive production
systems has driven the development of more structured mathematical models for production planning.
According to [3], optimization models have been widely used to provide valuable information to
decision-makers on issues related to production planning in agri-food supply chains. In this context,
during the last decade, a significant amount of research has been dedicated to the development of
optimization models to support decision-makers in the agri-food supply chain. Therefore, optimization
models such as production planning tools are essential in the agro-industrial sector, since they allow
companies to maintain competitiveness and improve decision-making in production operations.
3. Methodology
Based on the explanation provided in the previous section, the present study will address the question:
What are the approaches to applying optimization models for production planning in the agro-industrial
sector? For this purpose, a set of keywords was defined, which were used in the search string to collect
relevant scientific publications.
To ensure an exhaustive scope, various databases available online were selected, and the search string
was applied to extract research related to the subject matter. In order to ensure the relevance of the
selected studies, inclusion and exclusion criteria were established. These criteria considered the
thematic relevance, the use of linear programming mathematical models, the language of publication
(Spanish, English and Portuguese), as well as the date of publication, covering the period between 2012
and 2022.
Subsequently, specific filters were applied to refine the selection of research aligned with the
application of optimization models in production planning in the agro-industrial sector. In this way, a
set of up-to-date and relevant publications was obtained, and will be subjected to a detailed analysis to
identify the most recurrent approaches in the application of optimization models in said context.
3.1. Definition of keywords
To answer the research question, keywords were defined and used in the search string to collect research
aligned with the topic of interest, which are detailed in Table 1.
The following search terms were used: "Production Planning", "Optimization Model", "Product Mix",
"Agro-industrial" and "Agribusiness." To improve the accuracy of the search, Boolean operators such
as "AND" and "OR" were used in the search string. In the databases available online, the search string
Page 376
SEMIT 2023
was: ((“Production Planning”) OR (“Optimization Model”) OR (“Product Mix”)) AND ((“Agro-

industrial”) OR (“Agribusiness”)).
The search string was adapted to Spanish language for the database "Acceso Libre a Información
Científica para la Innovación (ALICIA)" and to English for the databases SCOPUS, Web Of Science
(WOS), SPRINGER and SCIENCE DIRECT databases. Additionally, the search string was also
adjusted to the different formats required in each of the databases to obtain the most relevant results.
Table 1. Table of keywords and search string.

Agro-industrial
Themes Production planning
Sector
¿What are the approaches to apply optimization
Research question models for production planning in the agro-
industrial sector?
Production Planning Agro-industrial
Keywords Optimization Model agribusiness
Product Mix
(“Production Planning” OR “Optimization
Search String Model” OR “Product Mix”) AND (“Agro-
industrial” OR “Agribusiness”)
3.2. Selection of databases

The search was conducted among scientific articles available in online databases such as SCOPUS,
WOS, SPRINGER, and SCIENCE DIRECT where the search string was applied in English. Likewise,
theses available in the ALICIA database were investigated and the search string was applied in Spanish.
These databases were selected due to their importance in the academic research environment and
their wide scope to topics related to the subject of interest.
3.3. Definition of inclusion and exclusion criteria
To select the publications for review and ensure the rigor of the process, inclusion criteria were
established, which considered the thematic relevance and the focus on production planning in the agro-
industrial sector. Likewise, exclusion criteria were defined to rule out secondary research, duplicate
studies, research unrelated to the topic of interest, publications over 10 years old, among others.
These criteria were systematically applied in each of the databases consulted, thus ensuring the
selection of relevant and up to date information for analysis. Table 2 provides a detailed presentation of
the criteria used during the bibliographic search of this article in detail.
Table 2. Table of inclusion and exclusion criteria.

Inclusion criteria Exclusion criteria
- Primary research - Secondary investigations
- Scientific articles and theses available online. - Articles and theses are not available online.
- Use of mathematical models of linear - Research whose language is not Spanish,
programming. English or Portuguese.
- Studies in Spanish, English and Portuguese. - Duplicate studies.
- Studies published from 2012 to 2022. - Short articles (<5 pages).
-Studies not relevant to the investigation
- Research studies published before 2012
Page 377
SEMIT 2023
3.4. Conduct of research studies selection

The research selection process is illustrated in Figure 1, where a total of 42 publications were
identified in SCOPUS, 38 in WOS, 37 in ALICIA, 123 in SPRINGER, and 304 in SCIENCE
DIRECT. Subsequently, the selection was refined through the application of specific filters. First, the
first filter was applied to rule out studies not written in Spanish, English or Portuguese, those older
than 10 years, as well as duplicate studies identified through the use of the Zotero bibliographic
manager and secondary studies.
Secondly, the second filter was applied, where a thorough reading and analysis of the title, abstract
and content of the investigations were carried out, to exclude those short articles or investigations that
were not relevant to the study. After applying the filters in each database, a total of 1 article was
obtained in SCOPUS, 2 articles in WOS, 3 theses in ALICIA, 4 articles in SPRINGER and 5 articles
in SCIENCE DIRECT, resulting in a total of 15 publications for review.
Figure 1. Process of filters in the selection of investigations

3.5. Definition of analysis criteria
To conduct the analysis of the selected works, a series of criteria were defined, which were used as a
guide to characterize and compare the research among themselves. Firstly, criteria related to the
optimization and the solution method of the model were studied. Secondly, criteria related to
production planning were analyzed. This provides a general understanding of the approaches used in
the application of optimization models for production planning in the agro-industrial sector.
The analysis criteria related to the optimization model and its solution method are detailed below:
i. Optimization model: The optimization model used in the research was specified to determine
which models are the most employed in production planning and which one is predominant
in this field.
ii. Objective Function: The objective function of the models proposed in the research was
identified to understand the predominant focus in the literature regarding the objectives
pursued in this type of optimization models. These objectives can vary from minimizing
costs to maximizing profit or production levels, among others.
iii. Solution method: The solution method used in the research was indicated to understand
which one is most commonly used according to the complexity of the model. These methods
can be exact, heuristic-based, or scenario reduction algorithms.
Page 378
SEMIT 2023
iv. Results: The results obtained in each of the research were presented to compare the findings
achieved by different authors when using the optimization models in production planning.
Next, the criteria related to production planning applied to each research are shown below.
v. Problem: The problem faced by the research was identified by understanding the specific
problems solved and by the most frequent problems faced in the studies. These problems
may be related to inadequate production plans or the presence of uncertainty in production.
vi. Quantity of products: The quantity of products to which the models are applied was analyzed
to understand whether these models are used both in situations with a small number of
products and those with a wide variety of products, either in multi-product or single-product
scenarios.
vii. Uncertainty: Research that incorporates uncertainty in model parameters was identified to
understand how it more accurately reflect reality. This uncertainty may be related to demand,
climatic factors, environmental factors, among others.
viii. Planning horizon: The planning horizon used in the models was indicated to understand the
common approach regarding the duration and number of periods considered. This includes
both single-period models and multi-period models.
4. Analysis of results
In this section, the results of the literature review related to both the optimization model used and
findings regarding production planning will be presented. This comparison will allow the
identification of common patterns and trends among the research and will provide a comprehensive
vision of the most outstanding approaches in the field of production planning in the agro-industrial
sector.
4.1. Results related to the optimization model.
Next, the results obtained regarding the optimization model used in the selected investigations will be
analyzed. Table 3 presents the individual findings that cover aspects such as the optimization model,
the objective function of the model, the solution method employed, and the results achieved in each
study.
Table 3. Table of findings of criteria related to the optimization model.

Objective Solution
Reference Model Results
Function Method
First level: Maximize
profits from selling all
Increase of 0.5% in
the products produced in
company profits.
Mixed Integer all the mills. Second
Approximate benefits of
[9] Linear level: Maximize the Exact Method
$1000 with more than
Programming income from the sale of
91 uncertainty
all the products and the
parameters
energy exported to the
grid from each mill
Non-dominated Costs are minimized to
Mixed Integer Minimize costs and
Classification 267,938,562,000 IDR
[10] Linear maximize the level of
Genetic and 1,487,944.40 kg in
Programming service.
Algorithm service level
Maximize gross profit
Average gross margin in
Linear margin considering
[11] Exact Method test scenarios of AED
programming production risk
34,558.4
scenarios
Page 379
SEMIT 2023
Maximize productivity
Integer Linear Increased productivity
[6] in the manufacturing Exact Method
Programming by 11.61%
area.
Linear Minimize production Reduced overtime costs

[12] Exact Method
programming costs by 39%
Increase of 5.63% in the

Linear Maximize company
[2] Exact Method profit margin of the
programming profits
company
Binary Integer The total cost is

Minimize total Intuitive
[13] Linear minimized by 132,973
production costs Heuristics
Programming euros
Mixed Integer
Minimize total cost of Reduction of 16% in
[14] Linear Exact Method
seed production total production cost
Programming
Reduction of total
Mixed Integer
Minimize costs of minimum cost of
[15] Linear Exact Method
production inputs production inputs by
Programming
33.9%
Compare exact
Average 6%
Binary Integer Maximize income, solution and
improvement in total
[16] Linear subject to zoning Heuristics based
revenue with exact
Programming restrictions on
solution
decomposition
Relaxation- Production is
Maximize the
Mixed Integer correction and maximized to
[17] production of sucrose
Linear correction- 121279.92 tons of
and fiber and minimize
Programming optimization sugarcane with the exact
the stock of sucrose.
heuristics solution
Minimize total costs,
Average total cost in
Linear such as raw materials,
[18] Exact Method test scenarios of
Programming inventory, shortages,
$317,270,467.8
juice varieties.
Minimize total The total cost is reduced
Linear
[19] production and logistics Exact Method by 12% compared to the
Programming
costs initial situation
Increase in profit by
Minimize production 6.66% and reduction in
and logistics costs while transportation costs by
Linear
[20] providing financial Exact Method 12.98%, costs of semi-
Programming
benefit to continue finished products by
producing 100% and inventory
costs by 79.18%
Scenario
reduction
Maximize expected net
Mixed Integer algorithm based Increase of 21.1% in net
benefits calculated as
[21] Linear on unsupervised benefits compared to the
revenue fewer operating
Programming learning deterministic model
costs.
techniques and
exact method
Page 380
SEMIT 2023
i. Optimization model.
Regarding the optimization models used in the research, Figure 2 shows the percentage of
primary studies according to the optimization model they employ. As can be seen, it was possible
to identify that the most used optimization models in the selected publications are Mixed Integer
Linear Programming (40%) and Linear Programming (40%), which are used mostly to optimize
costs, benefits, productivity level, production level, among others.
On the other hand, it is noteworthy that most of the research that apply Mixed Integer Linear
Programming incorporate uncertainty in the model parameters, leading them to use different
optimization approaches, such us robust or stochastic optimization.
Figure 2. Research studies according to the optimization model

ii. Objective function.
Next, in the reviewed studies, different approaches were found regarding the objective function of
the models applied in the research. Figure 3 shows the percentage of investigations according to
the objective of the optimization model that they used. According to this, most of the
investigations employ an economic objective function, either to maximize profits, revenues, net
benefits, or minimize costs.
Likewise, it was identified that the predominant focus of the pursued objectives is cost
minimization, since 53% of the studies consider it, considering raw material costs, transportation
costs, inventory costs, scarcity, among others. In addition, other objectives were also considered,
such as maximizing productivity, minimizing risks, maximizing service level and maximizing
production.
Figure 3. Research studies according to the objective function
Page 381
SEMIT 2023
iii. Solution Method.

Regarding the solution methods applied in the models, a wide variety of approaches used were
observed. Among them, exact methods, algorithms based on unsupervised learning techniques for
scenario reduction, genetic algorithms, decomposition heuristics, and relaxation-correction and
correction-optimization heuristics were employed.
Figure 4 shows the distribution of the use of these methods in research. It stands out that 67% of
the studies opted for exact solution methods, followed by heuristics with a utilization rate of 20%.
The models solved by exact methods mostly considered deterministic parameters and focused on
cost minimization for multiple products. On the other hand, heuristic methods and algorithms
were especially useful for addressing computationally expensive problems in model resolution.
Figure 4. Research studies according to solution method

iv. Results.
From the analysis of the main results of each investigation, detailed in table 3, it was found that
the investigations focused on minimizing costs that applied Linear Programming models achieved
a reduction of over 12%. Meanwhile those that applied Mixed Integer Linear Programming
obtained a cost reduction of over 16%. Likewise, the studies that maximized benefits with Mixed
Integer Linear Programming obtained an increase of 0.5%.
Furthermore, it was observed that the models used were effective in optimizing different
objectives, both for those that incorporated uncertainty in the model parameters and those that did
not. This demonstrates their versatility in achieving various objectives and production planning
contexts in the agro-industrial sector.
4.2. Results related to production planning.
In this subsection, an analysis of the results obtained in the review will be conducted, focusing on the
analysis criteria related to production planning, such as the addressed problem, the number of
products considered in the production plan, the incorporation of uncertainty in the model and the
planning horizon used in each investigation. These results will be presented in table 4 below.
Page 382
SEMIT 2023
Table 4. Table of findings of criteria related to production planning.
Quantity of
Reference Problem Uncertainty Planning Horizon
products
Inadequate Uncertainty in
production planning product prices and
Multiproduct, sugar, Multiperiod, 6 months
[9] reflected in the industrial yields due
alcohol and molasses subdivided into weeks
dissatisfaction of to weather, equipment
demand failure, human error
Incorporates
Inadequate Multiproduct,
uncertainty in demand
[10] production and varieties of processed Daily multiperiod.
and shrimp yield due
distribution plan shrimp
to biological factors
Plan greenhouse Multiproduct,
Does not incorporate
[11] production without tomato, cucumber Single period
uncertainty
considering risks and pepper
Low productivity in A single product, Does not incorporate
[6] Single annual period
the production area sugar uncertainty
Multiproduct,
Inadequate
varieties of oat Does not incorporate
[12] production planning Single annual period
flakes, wheat and uncertainty
reflected in high costs
kiwicha flour
Inadequate
production planning Multiproduct, rice Does not incorporate Single period, for a
[2]
reflected in high costs varieties uncertainty semester
and low income
Multiproduct, a
Produce tomatoes, at binary variable is
Incorporates
[13] the minimum cost to associated with the Single period
uncertainty in demand
meet future demand selection of these
farms
Incorporates
Inadequate seed
uncertainty in
production plan, due
Multiproduct, seed production, due to the
[14] to not considering Single annual period
varieties variations in
uncertainty in
performance and
production.
demand.
Inadequate
production plan Multiproduct, variety Does not incorporate
[15] Single period
reflected in high of seeds uncertainty
production costs
Inadequate crop Multiproduct, variety Does not incorporate
[16] Single period,
planning of vegetables uncertainty
[17] Produce the largest Multiproduct,

Does not incorporate
amount of sugar cane varieties of sugar Monthly multiperiod
uncertainty
and energy cane cane and energy cane
It incorporates
Multiproduct, uncertainty in juice
Inadequate orange
[18] varieties of orange concentration, juice Monthly multiperiod
juice production plan
juice yield, demand and
availability of fruit.
Page 383
SEMIT 2023
Inadequate Multiproduct, Does not incorporate

[19] production and varieties of corn uncertainty in the Monthly multiperiod
logistics plan seeds parameters
Multiproduct,
Deficient production varieties of Does not incorporate
[20] Monthly, multiperiod.
and logistics plan concentrated tomato uncertainty
pulp
Inadequate It incorporates
production plan by uncertainty in the
Multiproduct, variety Daily multiperiod
[21] not incorporating demand and
of dairy products planning for a month
uncertainty in the availability of raw
parameters material
v. Problem addressed.
Regarding the different problems faced, Table 4 shows a variety of challenges addressed in the
investigations, including inadequate production planning, uncertainty in the productive
environment, low productivity, lack of consideration of risks, among others. It is notable that most
of the investigations proposed optimization models to face the problem of inadequate production
planning, resulting in high production costs, unsatisfied demand or reduced profit margins.
Then, the studies that focused on solving the problem of inadequate production planning mainly
opted for Linear Programming and Mixed Integer Linear Programming models. These models made
it possible to optimize the objectives in each specific context, whether it was reducing costs,
increasing benefits, maximizing income or improving utility.
vi. Quantity of products
Next, regarding the quantity of products, it was observed in the reviewed investigations that the
models were applied to develop production plans for a wide variety of products. Then, the vast
majority of research studies considered multiple products or multiple product varieties, such as
varieties of dairy product families, varieties of rice, processed vegetable products, varieties of sugar
cane, different concentrations of orange juice, among others.
On the other hand, few investigations where only one product was considered for the model, so it
can be said that these models can be applied both to situations where there is a variety of products
and to those in which there is only one product. It was also found that in the studies addressing the
problem of inadequate production planning, the most used optimization models were Mixed Integer
Linear Programming and Linear Programming considering multiple products.
vii. Uncertainty
Regarding the uncertainty in the model, the results revealed that most of the investigations did not
incorporate this factor into the optimization model used. However, in those studies that did consider
it, it was observed that the variability in product demand and production performance due to
climatic, biological factors, equipment failures or human errors were the main sources of
uncertainty addressed.
The fluctuations in product prices and the availability of raw material were also considered as
relevant elements for the modelling. In this sense, the Integer Linear Programming, Linear
Programming and Binary Integer Linear Programming models were the ones used to incorporate
uncertainty in production planning, especially in situations involving multiple products. By using
these optimization models, the research studies were able to improve decision-making in the agro-
industrial sector, increasing benefits and minimizing or reducing costs.
viii. Planning Horizon
It was found that there is an equitable distribution in terms of the planning horizon of the models,
as 47% of the studies used a multi-period horizon, whether it daily, monthly or semi-annual, while
Page 384
SEMIT 2023
the rest opted for a single-period horizon. Most of the studies that considered multiple periods
incorporated uncertainty in the model parameters, whereas those that considered a single period did
not, focusing on more deterministic parameters.
In addition, the models with a daily or monthly horizon were applied with the objective of making
more immediate operational and tactical decisions. While the models that considered a semi-annual
or annual planning horizon did so with the objective of making strategic decisions. This highlights
the capacity of the optimization models applied in the agro-industrial sector to adapt to different
planning horizons and production variability.
5. Discussion and conclusions

The results obtained in this review are consistent with those found by [22] where they found that Integer
Linear Programming and Linear Programming are the dominant models and optimization approaches
for planning in agribusiness. However, it differs from what was found in [4] where they found that
fuzzy programming and multi-objective programming are the most used approaches, followed by
Linear Programming, Integer Programming and Mixed Integer Programming, potentially because they
conducted a broader review including various industries. On the other hand, the results of this review
are similar to what was found by [3], where Mixed Integer Linear Programming is the most used model
in Production Planning in agri-food supply chains.
Based on the analysis of the studies included in this review, it is evident that 40% of the investigations
used the Mixed Integer Linear Programming optimization model, while another 40% opted for Linear
Programming. This was mostly to solve problems of inadequate production plan with multiple
products, either with a planning horizon of a single- period or multi-period. However, it was found that
very few studies incorporated uncertainty in the model parameters, which indicates that more research
is required to address the uncertainty commonly present in the agro-industrial sector, especially due to
variable climatic factors and raw material yield subject to variable biological factors.
This review identified various approaches to the application of optimization models for production
planning in the agro-industrial sector were identified through a review of the literature. As future work,
it is suggested to explore additional analysis criteria, such as the decision variables of the model, the
number or type of restrictions involved, and the country of application of the studies. Furthermore, it
would be interesting to include other keywords in the literature search to broaden the coverage of the
review and to examine how the topic has evolved over the years. These additional approaches could
provide a more detailed insight and further enrich the results from the field of study.
References
[1] Marrero, B. Six terrible consequences of not planning in companies. BMA Group. Recovered
fromhttps://brendamarreropr.com/6terribles-consecuencias-de-no-planificar-en-las-empresas/, 2017.
[2] J. Collantes and E. Liza, "Application of linear programming to maximize the utility of a milling
company from Lambayeque, 2020", Degree Thesis, Señor de Sipán University, Pimentel-Perú, 2022.
[3] T. Taşkıner and B. Bilgen, "Optimization Models for Harvest and Production Planning in Agri-Food
Supply Chain: A Systematic Review", Logistics, vol 5, no. 3, pp. 1-27, 2021.
[4] M. Demartini, F. Tonelli, M Pacella and Papadia, G. (2021). “A Review of Production Planning
Models: Emerging features and limitations compared to practical implementation”, Procedia CIRP, vol
104, pp. 588-593, 2021.
[5] O. Solaja, J. Abiodun, M. Abioro, J. Ekpudu, and O. Olasubulumi, “Application of linear
programming techniques in production planning”, International Journal of Applied Operational
Research 2019, vol. 9, no. 3, pp. 11-19, 2019.
Page 385
SEMIT 2023
[6] M. Monja and M. Sedán, "Application of Linear Programming in the Planning and Programming of
Sugar Production to improve the productivity of the agro-industrial company Pomalca SAA", Degree
Thesis, Señor de Sipán University, Pimentel, Peru, 2016.
[7] E. dos S. Teixeira, S. Rangel, H. de O. Florentino, and SA de Araujo, “A review of mathematical
optimization models applied to the sugarcane supply chain”, Intl. Trans. in Op. Res., vol. 30, no. 4, pp.
1755-1788, July 2023.
[8] Rajni Jain, L Malangmeih, SS Raju, SK Srivastava, Kingsly Immaneulraj and Amrit Pal Kaur,
“Optimization techniques for crop planning: A review”, Indian J. Agricultural Sci, vol. 88 no. 12, pp.
1826–35, December 2018.
[9] R.P. Paiva, C.D. Rocco, and R. Morabito, “A hierarchical supply chain model for the sugar–alcohol
energy sector with robust optimization analysis,” Int Trans Operational Res, vol. 30, no. 4, pp. 1789-
1818, July 2023.
[10] L. Herlina, M. Machfud, E. Anggraeni, and S. Sukardi.: An Integrated Production and Distribution
Planning Model in Shrimp Agroindustry Supply Chain. Ind. Eng. & Manage. Syst., vol. 21, no. 1, pp.
1-19, March 2022.
[11] E. Fathelrahman, M. Gheblawi, S. Muhammad, E. Dunn, JC Ascough, and T. Green.: Optimum
Returns from Greenhouse Vegetables under Water Quality and Risk Constraints in the United Arab
Emirates. Sustainability, vol. 9, no. 5, pp. 1-11, April 2017.
[12] V. Cieza, "Production Planning applying Linear Programming for the optimization of costs in the
company Producciones Nacionales TC EIRL", Degree Thesis, Señor de Sipán University, Pimentel-
Perú, 2021.
[13] J. Mateo, LM Pla, F. Solsona, and A. Pagès, “A production planning model considering uncertain
demand using two-stage stochastic programming in a fresh vegetable supply chain context”,
SpringerPlus, vol. 5, no. 1, pp. 1-16, December 2016.
[14] Y. Zhu, N. Shah, G. Carré, S. Lemaire, E. Gatignol, and P.M. Piccione, “Continent-wide planning
of seed production: mathematical model and industrial application”, Optim Eng, vol. 20, no. 3, pp. 881-
906, September 2019.
[15] S. Lozano, “Company-wide production planning using a multiple technology DEA approach”,
Journal of the Operational Research Society, vol. 65, no. 5, pp. 723-734, 2014.
[16] V.M. Albornoz and G.E. Zamora, “Decomposition-based heuristic for the zoning and crop
planning problem with adjacency constraints”, TOP, vol. 29, no. 1, pp. 248-265, April 2021.
[17] S.C. Poltroniere, A. Aliano Filho, A.S. Caversan, AR Balbo, and H. de O. Florentino, “Integrated
planning for planting and harvesting sugarcane and energy-cane for the production of sucrose and
energy”, Computers and Electronics in Agriculture, vol. 184, pp. 1-13, May 2021.
[18] J.R. Munhoz and R. Morabito, “Optimization approaches to support decision making in the
production planning of a citrus company: A Brazilian case study”, Computers and Electronics in
Agriculture, vol. 107, pp. 45-57, September 2014.
[19] R.A. Junqueira and R. Morabito, “Production and logistics planning considering circulation taxes
in a multi-plant seed corn company”, Computers and Electronics in Agriculture, vol. 84, pp. 100-110,
June 2012.
[20] C.D. Rocco and R. Morabito, “Production and logistics planning in the tomato processing industry:
A conceptual scheme and mathematical model”, Computers and Electronics in Agriculture, vol. 127,
pp. 763-774, September 2016.
[21] A. Guarnaschelli, HE Salomone, and CA Mendez, “A stochastic approach for integrated
production and distribution planning in dairy supply chains,” Computers & Chemical Engineering, vol.
140, pp. 1-22, September 2020.
Page 386
SEMIT 2023
[22] P. Paam, R. Berretta, M. Heydar, RH Middleton, R. García-Flores, and P. Juliano, “Planning

Models to Optimize the Agri-Fresh Food Supply Chain for Loss Minimization: A Review”, Reference
Module in Food Science, pp. 19-54, 2016.
Page 387
SEMIT 2023
Comparing OECD Countries’ Logistics Performances by Comprehensive Distance-Based

Ranking (COBRA) Method
Sedef Çalı6*[0000-0003-0091-6177] and Adil Baykasoğlu2[0000-0002-4952-7239]
1
Dokuz Eylül University, Graduate School of Natural and Applied Sciences, Department of Industrial
Engineering, Izmir, Türkiye
sedefcali@gmail.com
2
Dokuz Eylül University, Faculty of Engineering, Department of Industrial Engineering, Izmir,
Türkiye
adil.baykasoglu@deu.edu.tr
Abstract
Logistics are of crucial importance for countries to engage in international trade and build their
economies. In 2007, the World Bank (WB) developed the logistics performance index (LPI) to
assess how well countries perform in terms of logistics. The LPI score is obtained by averaging
the scores on the six components that have been weighted equally. The core components of LPI
are customs, infrastructure, services, timeliness, tracking and tracing, and international
shipments. However, it is not realistic to consider that these components are of equal importance.
Rather than taking the average of the component scores, more sophisticated techniques should
be performed to acquire the overall logistic performance scores. Considering these shortcomings,
this study aims to propose a new multi-criteria decision-making (MCDM) analysis to evaluate
the logistics performances of OECD countries by using LPI data published in 2023. The proposed
methodology determines the weights of LPI indicators thanks to the criteria importance through
intercriteria correlation (CRITIC) method, which is an objective weighting approach. The
rankings of the OECD countries are obtained by using the comprehensive distance-based ranking
(COBRA) method, which is a relatively new distance-based MCDM method. The results show
that the international shipment score is the most important indicator of LPI. The ranking indicates
that Finland is the best OECD country in logistics performance. Sensitivity analysis is conducted
to observe whether the rankings of countries differ with changes in the criteria weights. The
different rankings obtained reveal that the criteria weights should be considered in logistics
performance assessments.
Keywords: MCDM, Logistics Performance Index, COBRA, CRITIC.
1. Introduction
Logistics management is described by the Council of Supply Chain Management Professionals (2013)
as "that part of supply chain management that plans, implements, and controls the efficient, effective
forward and reverses flow and storage of goods, services, and related information between the point of
origin and the point of consumption in order to meet customers' requirements". Logistics include the
activities of supply of raw materials, material management in a factory, and distribution to customers
(Islam et al., 2013). In this manner, the five main components of logistics are inventory, transport,
information and processing, warehousing, and packaging. (Islam et al., 2013).
The logistics managers must be familiar with the degree of logistics service offered in the country from
which the goods are being sourced. The Logistics Performance Index (LPI) proposed by WB ranks the
countries on six indicators: logistics quality and competence, customs, tracking and tracing,
infrastructure, international shipment, and timeliness. Here is a definition of each indicator (Martí et al.,
2017; Arvis et al., 2023):
*
Corresponding Author
Page 388
SEMIT 2023
• Logistics quality and competence expresses the competence and quality of logistics services.
• Customs measures the timeliness, ease, and predictability of formal problems conducted by
customs control agencies in order to determine the efficiency of customs.
• Infrastructure points out the quality of trade and transport infrastructure.
• Tracking and tracing measures the capacity for tracking and tracing shipments.
• Timeliness expresses the frequency of on-time deliveries.
• International shipment evaluates the simplicity of setting up affordable shipments.
The survey conducted by WB asks the trading partners of a specific nation to rate how simple or
challenging it is to trade manufactured goods delivered in unit forms like shipping containers (Arvis et
al., 2023). Experts evaluate the eight distinct countries according to six variables on a 5-point scale, and
the results are combined to create the LPI. WB published the first LPI report in 2007. Afterward, new
reports have been released nearly every two years. However, the 2023 LPI report was published after
the last survey in 2018. The 2023 LPI survey was performed from September 6 to November 5, 2022.
It includes 4,090 country assessments by 652 experts in 115 countries in all WB regions (Arvis et al.,
2023). It is reported that the average overall score in the 2023 LPI was mostly consistent with that of
the 2018 survey although global supply chain problems have broken out due to the COVID-19 pandemic
(Arvis et al., 2023).
WB considers equal weights for these six indicators and takes the average of these indicators in
calculating LPI. The more advanced procedures should be used to obtain the overall logistic
performance scores rather than averaging the component results. On the other hand, countries aim to
increase their LPI values to obtain a competitive advantage in the global area, and it is important for
countries to know the indicator with the highest contribution to LPI. They prioritize the indicators to be
improved to reach their objectives regarding gaining sustainable logistics performance. Decision-
makers in the countries may make policies to enhance the prioritized indicators. Equal indicator weights
in the LPI give rise to obstacles for countries in developing successful policies (Yıldırım and
Mercangöz, 2020). This problem is addressed here as an example of a multi-criteria decision-making
(MCDM) problem (Yıldırım and Mercangöz, 2020).
MCDM is a research field aiming to discover the most appropriate alternative among a set of
alternatives with several conflicting criteria (Tzeng and Huang, 2011). There are several MCDM
methods in the literature. They can be categorized into four groups, namely, pairwise comparison
methods, scoring (additive) methods, distance-based methods, and outranking methods (Penadés-Plà et
al., 2016). The COBRA method is a newly developed one belonging to the group of distance-based
methods (Krstić et al., 2022). The distance-based methods rank the alternatives considering their criteria
performances’ distances from the reference points measured via various types of distance metrics. For
instance, a popular method, TOPSIS, considers only Euclidean distances from the positive ideal solution
and the negative ideal solution for ranking the alternatives (Hwang and Yoon, 1981). It is difficult to
find which metric or reference point is more convenient. However, the COBRA method assesses the
alternatives based on the integration of various distances from different reference points (Krstić et al.,
2022). These are regarded as the advantageous features of this method. Therefore, we have introduced
a novel MCDM model for the LPI via the COBRA method, and we have evaluated the logistics
performance of 38 OECD countries. Whereas WB considers equal weights for these six indicators, we
use the CRITIC method to find the weights for them.
This paper gives important contributions to the literature. This study is the first one using the COBRA
method in logistics performance assessments. In addition, it is one of the pioneer works handling 2023
LPI data.
The remainder of this study is proposed as follows: The related works are mentioned in Section
2. The methodology is introduced in Section 3. Section 4 presents the application and its results. Section
5 provides sensitivity analysis. Finally, Section 6 gives discussions and conclusions.
Page 389
SEMIT 2023
There is a growing interest among researchers in ranking countries based on LPI data by utilizing
MCDM methods. Some of them are summarized as follows:
Çakır (2017) developed a hybrid approach to evaluate the performances of the OECD countries based
on LPI published in 2014. The CRITIC is used for obtaining criteria weights, and simple additive
weighting (SAW) is used for obtaining the overall evaluation values of the countries. The country with
the highest value is accepted as the best-performing one. In addition, they proposed a fuzzy linear
regression model in which SAW’s result is dependent variable and ranked the countries according to
regression residuals.
Ulutaş and Karaköy (2019) measured the logistic performances of European Union (EU) countries
by using an MCDM procedure that aggregates stepwise weight assessment ratio analysis (SWARA)
with CRITIC to determine the weights and uses proximity indexed value (PIV) to determine the ranks
of countries. They assigned the highest weight to international shipments considering CRITIC and
assigned the highest weight to infrastructure considering SWARA.
Mešić et al. (2022) used CRITIC and measurement alternatives and ranking according to compromise
solution (MARCOS) methods to rank Western Balkan countries based on 2018 LPI data. They
performed a sensitivity analysis to observe the changes in the rankings with the different weights of
criteria. Stević et al. (2022) performed a similar analysis for all Balkan countries and attained different
rankings with different criteria weights.
Ulutaş and Karaköy (2021) analyzed the logistics performances of countries of transition economies
by using a grey MCDM model composed of grey SWARA and grey multi-objective optimization by
ratio analysis (MOORA).
Mercangoz et al. (2020) performed the MCDM analysis with the complex proportional assessment
of alternatives with grey relations (COPRASG) method through five-period LPI data of 28 EU countries
and five candidate EU countries. The weights were calculated by the fuzzy analytical hierarchy process
(fuzzy-AHP).
Yildirim and Mercangoz (2020) evaluated LPI data between 2010 to 2018 belonging to OECD
countries with fuzzy-AHP-based grey additive ratio assessment (ARASG). Based on Spearman ρ and
Kendall’s Tau correlation methods, they revealed that the obtained ranking, considering 5 years, is
highly correlated with the yearly rankings of WB.
Çalık et al. (2023) applied the AHP-TOPSIS, AHP-VIKOR, and AHP-CODAS methods, both in
their classic and fuzzy versions, to the 2018 LPI data. However, the weight determination step using
AHP is based on a survey of experts in the field. In the last step, the Borda count method was used to
generate the optimum individual ranking.
Rezaei et al. (2018) used Best Worst Method to acquire the criteria weights and multiplied these weights
by the scores in the WB’s 2016 LPI report. They ranked the countries according to weighted LPI data.
Infrastructure is found as the most important criterion.
Martí et al. (2017) utilized data envelopment analysis (DEA) to create a synthetic LPI and to benchmark
the countries on the basis of the obtained index values. DEA is utilized as an MCDM tool in the study.
The impacts of revenues and geographical location of countries on the logistics performances are
mentioned by the authors.
The other methods used are AHP-GRA (Candan, 2019), SWARA-EDAS (Kısa and Ayçin, 2019),
Entropy-ETAWOS (Görçün, 2019), and Entropy-EDAS (Orhan, 2019), as mentioned in the studies
written in Turkish.
Candan (2019) evaluated the logistic performances of 10 OECD member countries. The criteria weights
are computed by using fuzzy AHP, and the ranking is obtained via the grey relational analysis (GRA)
method. The criteria assigned the higher weights are lead time of import and export and the quality of
infrastructure in the study.
Page 390
SEMIT 2023
Kısa and Ayçin (2019) assessed the logistic performances of OECD countries by the evaluation based
on distance from average solution (EDAS) based on SWARA. They present the most important criteria
as logistics service quality and infrastructure and international shipment.
Görçün (2019) assessed the logistic performances of Central Asia Turkic Republics. They analyzed
integrated entropy and efficiency analysis technique with output satisficing (EATWOS) methods. The
entropy method is used to acquire the weights of input and output factors. The EATWOS is utilized to
compute the efficiency scores of the countries. For EATWOS, the input factors are labor force, highway
networks, fleet, railway networks, railway fleet, the number of carriages, the number of ports, logistics
investment, and import. On the other hand, the output factors are gross national product, national
income, gross national product (logistics), export, transportation load, annual cargo handled (ton),
annual cargo handled (TEU), and annual growth rate. Logistics investments is considered the most
important input factor, whereas annual cargo handled (TEU) is considered the most important output
factor. Thus, the countries are ranked according to efficiency scores.
Orhan (2019) compared Turkiye with the EU countries according to the logistic performances. To do
so, the criteria weights are determined by the entropy method at first. Afterwards, the countries are
ranked through the EDAS method. The LPI of the year 2018 is the utilized data. The criterion of
Customs has the highest weight, while the criterion of international shipments has the lowest weight in
the study.
3. Methodology
3.1. CRITIC Method
CRITIC is an objective weighting method introduced by Diakoulaki et al. (1995). It consists of three
steps explained as follows:
Step 1: The performance scores of m alternatives regarding n criteria are put into the decision matrix 𝑅.
𝑟𝑖𝑗 is the performance score of 𝑖th alternative regarding 𝑗th criterion.
𝐴1 𝑟11 ⋯ 𝑟1𝑗 ⋯ 𝑟1𝑛 (1)
⋮ ⋮ ⋯ ⋮ ⋯ ⋮
𝑅 = 𝐴𝑖 𝑖1𝑟 ⋯ 𝑟𝑖𝑗 ⋯ 𝑟𝑖𝑛
⋮ ⋮ ⋯ ⋮ ⋯ ⋮
𝐴𝑚 [𝑟𝑚1 ⋯ 𝑟𝑚𝑗 ⋯ 𝑟𝑚𝑛 ]
Step 2: The decision matrix is converted into the normalized matrix 𝑁 = [𝛼𝑖𝑗 ]𝑚𝑥𝑛 by eq. (2), where 𝐵
includes the benefit criteria and 𝐶 includes cost criteria.
𝑟𝑖𝑗 − min 𝑟𝑖𝑗 (2)
𝑖
; 𝑗∈𝐵
max 𝑟𝑖𝑗 − min 𝑟𝑖𝑗
𝑖 𝑖
𝛼𝑖𝑗 = ,
max 𝑟𝑖𝑗 − 𝑟𝑖𝑗
𝑖
; 𝑗∈𝐶
{max 𝑟𝑖𝑗 − min 𝑟𝑖𝑗
𝑖 𝑖
Step 3: The amount of information transmitted to the decision-making procedure via each criterion is
acquired by eq. (3) in which 𝜎𝑗 denotes the standard deviation of each criterion and 𝑐𝑗𝑘 denotes the
correlations of the criteria with each other. Afterward, the criteria weights 𝑤𝑗 (𝑗 = 1,2, … 𝑛) are
computed with eq. (4).
𝑛
(3)
ℎ𝑗 = 𝜎𝑗 ∑(1 − 𝑐𝑗𝑘 )
𝑘=1
Page 391
SEMIT 2023
ℎ𝑗 (4)
𝑤𝑗 = 𝑛 .
∑𝑗=1 ℎ𝑗
3.2. COBRA Method
COBRA method is a recently developed MCDM method including five steps:
Step 1: Consider the decision matrix 𝑅 = [𝑟𝑖𝑗 ]𝑚𝑥𝑛 .
Step 2: Construct the normalized decision matrix 𝑁 = [𝑛𝑖𝑗 ]𝑚𝑥𝑛 where
𝑟𝑖𝑗 (5)
𝑛𝑖𝑗 = ,
max 𝑟𝑖𝑗
𝑖
Step 3: Construct the weighted normalized decision matrix V:

𝑉 = [𝑤𝑗 ∗ 𝑛𝑖𝑗 ]𝑚𝑥𝑛 (6)
where 𝑤𝑗 is the weight of criterion 𝐶𝑗 .

Step 4: The positive ideal solution (𝑃𝐼𝑆𝑗 ), the negative ideal solution (𝑁𝐼𝑆𝑗 ), and average solution
(𝐴𝑆𝑗 ), for each criterion are specified.
𝑃𝐼𝑆𝑗 = max (𝑤𝑗 ∗ 𝑛𝑖𝑗 ) , ∀𝑗 = 1, … , 𝑛, 𝑗 ∈ 𝐵, (7)
𝑖
𝑃𝐼𝑆𝑗 = max (𝑤𝑗 ∗ 𝑛𝑖𝑗 ), ∀𝑗 = 1, … , 𝑛, 𝑗 ∈ 𝐶, (8)

𝑖
𝑁𝐼𝑆𝑗 = min (𝑤𝑗 ∗ 𝑛𝑖𝑗 ) , ∀𝑗 = 1, … , 𝑛, 𝑗 ∈ 𝐵, (9)

𝑖
𝑁𝐼𝑆𝑗 = min (𝑤𝑗 ∗ 𝑛𝑖𝑗 ) , ∀𝑗 = 1, … , 𝑛, 𝑗 ∈ 𝐶, (10)

𝑖
∑𝑚𝑖=1 (𝑤𝑗 ∗ 𝑛𝑖𝑗 ) (11)
𝐴𝑆𝑗 = , ∀𝑗 = 1, … , 𝑛, 𝑗 ∈ 𝐵, 𝑗 ∈ 𝐴,
𝑚
where 𝐵 and 𝐶 are the set of benefit and cost criteria.
Step 5: The distances from the positive ideal solution and the negative ideal solution and positive and
negative distances from the average solution are computed for each alternative via eq.(12).
𝑑(𝑆𝑗 ) = 𝑑𝐸(𝑆𝑗 ) + 𝜎 ∗ 𝑑𝐸(𝑆𝑗 ) ∗ 𝑑𝑇(𝑆𝑗 ), ∀𝑗 = 1, … , 𝑛, (12)
where 𝑆𝑗 indicates any solution (𝑃𝐼𝑆𝑗 , 𝑁𝐼𝑆𝑗 , 𝐴𝑆𝑗 ) and 𝜎 is the correction coefficient which is obtained
by
𝜎 = max 𝑑𝐸(𝑆𝑗 )𝑖 − min 𝑑𝐸(𝑆𝑗 )𝑖 , (13)
𝑖 𝑖
𝑑𝐸(𝑆𝑗 ) and 𝑑𝑇(𝑆𝑗 ) are Euclidian and Taxicab distances computed as follows:
𝑛 (14)
𝑑𝐸(𝑃𝐼𝑆𝑗 ) = √∑ (𝑃𝐼𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 )2 , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑖 𝑗=1
= 1, … , 𝑛,
𝑛
(15)
𝑑𝑇(𝑃𝐼𝑆𝑗 )𝑖 = ∑ |𝑃𝐼𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 | , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
Page 392
SEMIT 2023
𝑛 (16)
𝑑𝐸(𝑁𝐼𝑆𝑗 )𝑖 = √∑ (𝑁𝐼𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 )2 , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
𝑛
(17)
𝑑𝑇(𝑁𝐼𝑆𝑗 )𝑖 = ∑ |𝑁𝐼𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 | , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
𝑛 (18)
𝑑𝐸(𝐴𝑆𝑗 )+
𝑖 = √∑ 𝜏 + (𝐴𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 )2 , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
𝑛
(19)
𝑑𝑇(𝐴𝑆𝑗 )+
𝑖 =∑ 𝜏 + |𝐴𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 | , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
1 𝑖𝑓 𝐴𝑆𝑗 < 𝑤𝑗 ∗ 𝑛𝑖𝑗 , (20)
𝜏+ = {
0 𝑖𝑓 𝐴𝑆𝑗 > 𝑤𝑗 ∗ 𝑛𝑖𝑗
𝑛 (21)
𝑑𝐸(𝐴𝑆𝑗 )−
𝑖 = √∑ 𝜏 − (𝐴𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 )2 , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
𝑛
(22)
𝑑𝑇(𝐴𝑆𝑗 )−
𝑖 =∑ 𝜏 − |𝐴𝑆𝑗 − 𝑤𝑗 ∗ 𝑛𝑖𝑗 | , ∀𝑖 = 1, … , 𝑚, , ∀𝑗
𝑗=1
= 1, … , 𝑛,
0 𝑖𝑓 𝐴𝑆𝑗 < 𝑤𝑗 ∗ 𝑛𝑖𝑗 , (23)
𝜏− = {
1 𝑖𝑓 𝐴𝑆𝑗 > 𝑤𝑗 ∗ 𝑛𝑖𝑗
Step 5: The comprehensive distances (𝑑𝐶𝑖 ) are computed for the alternatives and they are ranked
from the lowest to highest.
𝑑(𝑃𝐼𝑆𝑗 )𝑖 − 𝑑(𝑁𝐼𝑆𝑗 )𝑖 − 𝑑(𝐴𝑆𝑗 )+ −
𝑖 + 𝑑(𝐴𝑆𝑗 )𝑖 (24)
𝑑𝐶𝑖 = , ∀𝑗 = 1, … , 𝑛
4
4. Application
In this section, we aim to rank 38 OECD countries according to their logistic performance scores. WB
employs six indicators to obtain logistic performance indices for countries. We use these indicators in
the MCDM process, which are C1: Customs score, C2: Infrastructure score, C3: International shipments
score, C4: Logistics competence score, C5: Tracking & tracing score, and C6: Timeliness score. All of
these indicators are benefit criteria to be maximized. OECD countries are the alternatives.
The performance scores of countries according to criteria are given in Table 1. Mexico has the lowest
score, whereas the highest score 4.1 is assigned to both Switzerland and Denmark in terms of Customs.
According to infrastructure, Costa Rica is the worst country while Switzerland is the best country.
According to international shipments, Finland has the best score and Chile has the worst score. While
the lowest logistics competence score is given to Costa Rica, the highest logistics competence score is
given to Denmark. Austria and Finland have the highest tracking & tracing score, but Chile, Colombia,
and Costa Rica have the lowest tracking & tracing score. The last column gives the LPI scores for the
countries. Finland has the highest LPI score, but Colombia and Mexico have the lowest LPI score. LPI
scores are rounded to one decimal point, and the rankings are based on those rounded values in the 2023
report (Arvis et al., 2023).
Page 393
SEMIT 2023
Table 1. LPI data for OECD countries

Country C1 C2 C3 C4 C5 C6 LPI
A1 Australia 3.7 4.1 3.1 3.9 4.1 3.6 3.7
A2 Austria 3.7 3.9 3.8 4 4.2 4.3 4
A3 Belgium 3.9 4.1 3.8 4.2 4 4.2 4
A4 Canada 4 4.3 3.6 4.2 4.1 4.1 4
A5 Chile 3 2.8 2.7 3.1 3 3.2 3
A6 Colombia 2.5 2.9 3 3.1 3.1 3.2 2.9
A7 Costa Rica 2.8 2.7 2.8 2.9 2.9 3.2 2.9
A8 Czech Republic 3 3 3.4 3.6 3.2 3.7 3.3
A9 Denmark 4.1 4.1 3.6 4.1 4.3 4.1 4.1
A10 Estonia 3.2 3.5 3.4 3.7 3.8 4.1 3.6
A11 Finland 4 4.2 4.1 4.2 4.2 4.3 4.2
A12 France 3.7 3.8 3.7 3.8 4 4.1 3.9
A13 Germany 3.9 4.3 3.7 4.2 4.2 4.1 4.1
A14 Greece 3.2 3.7 3.8 3.8 3.9 3.9 3.7
A15 Hungary 2.7 3.1 3.4 3.1 3.4 3.6 3.2
A16 Iceland 3.7 3.6 3.3 3.5 3.7 3.6 3.6
A17 Ireland 3.4 3.5 3.6 3.6 3.7 3.7 3.6
A18 Israel 3.4 3.7 3.5 3.8 3.7 3.8 3.6
A19 Italy 3.4 3.8 3.4 3.8 3.9 3.9 3.7
A20 Japan 3.9 4.2 3.3 4.1 4 4 3.9
A21 Korea, Rep. 3.9 4.1 3.4 3.8 3.8 3.8 3.8
A22 Latvia 3.3 3.3 3.2 3.7 3.6 4 3.5
A23 Lithuania 3.2 3.5 3.4 3.6 3.1 3.6 3.4
A24 Luxembourg 3.6 3.6 3.6 3.9 3.5 3.5 3.6
A25 Mexico 2.5 2.8 2.8 3 3.1 3.5 2.9
A26 Netherlands 3.9 4.2 3.7 4.2 4.2 4 4.1
A27 New Zealand 3.4 3.8 3.2 3.7 3.8 3.8 3.6
A28 Norway 3.8 3.9 3 3.8 3.7 4 3.7
A29 Poland 3.4 3.5 3.3 3.6 3.8 3.9 3.6
A30 Portugal 3.2 3.6 3.1 3.6 3.2 3.6 3.4
A31 Slovak Republic 3.2 3.3 3 3.4 3.3 3.5 3.3
A32 Slovenia 3.4 3.6 3.4 3.3 3 3.3 3.3
A33 Spain 3.6 3.8 3.7 3.9 4.1 4.2 3.9
A34 Sweden 4 4.2 3.4 4.2 4.1 4.2 4
A35 Switzerland 4.1 4.4 3.6 4.3 4.2 4.2 4.1
A36 Turkiye 3 3.4 3.4 3.5 3.5 3.6 3.4
A37 United Kingdom 3.5 3.7 3.5 3.7 4 3.7 3.7
A38 United States 3.7 3.9 3.4 3.9 4.2 3.8 3.8
The methodology starts with calculating criteria weights. To do so, we use the CRITIC method. First,
the decision matrix is normalized. Afterwards, the standard deviation of criteria and the correlation
between criteria are computed to obtain eq. (3). The results are given in Table 2 and Table 3.
Table 2. Correlation coefficients matrix
C1 C2 C3 C4 C5 C6
C1 1 0.9307 0.5918 0.8977 0.8154 0.7228
C2 0.9307 1 0.6473 0.9301 0.8587 0.7542
C3 0.5918 0.6473 1 0.7190 0.6836 0.7059
C4 0.8977 0.9301 0.7190 1 0.8718 0.8456
C5 0.8154 0.8587 0.6836 0.8718 1 0.8464
C6 0.7228 0.7542 0.7059 0.8456 0.8464 1
Page 394
SEMIT 2023
Table 3. Criteria weights acquired via CRITIC

Criteria C1 C2 C3 C4 C5 C6
standard
0.2714 0.2686 0.2206 0.2637 0.2992 0.2853
deviation
∑𝑛𝑘=1(1 − 𝑐𝑗𝑘 ) 1.0416 0.8790 1.6525 0.7358 0.9242 1.1252
weights 0.1688 0.1410 0.2177 0.1158 0.1651 0.1916
The most important criterion is the international shipments score (C3) according to the obtained
weights.
The methodology continues with the implementation of the COBRA method. The weights are
multiplied by the normalized decision matrix acquired by eq. (5). Then, the positive ideal
solution (𝑃𝐼𝑆𝑗 ), the negative ideal solution (𝑁𝐼𝑆𝑗 ), and the average solution (𝐴𝑆𝑗 ) are computed for
each criterion via eqs. (7-11).
Table 4. Positive ideal, negative ideal and average solutions for each criterion
C1 C2 C3 C4 C5 C6
PIS 0.1688 0.1410 0.2177 0.1158 0.1651 0.1916
NIS 0.1029 0.0865 0.1434 0.0781 0.1113 0.1426
AS 0.1429 0.1180 0.1804 0.1005 0.1431 0.1699
The distances from the positive ideal solution, negative ideal solution, and positive and negative
distances from the average solution are acquired considering both Euclidian and Taxicab distances for
each alternative. Equations (12-23) are used for the computations. In the last step, the comprehensive
distances (𝑑𝐶𝑖 ) are found. The comprehensive distance values are ranked from the lowest to the
highest. Table 6 gives all the results. It can be deduced that Finland is the best-ranked OECD country
according to logistic performance indicators.
Table 5. Results of CRITIC-COBRA methodology

Country dPIS dNIS dAS+ dAS- dC Rank
Australia 0.0668 0.0921 0.0223 0.0185 -0.0073 18
Austria 0.0296 0.1182 0.0382 0.0000 -0.0317 8
Belgium 0.0241 0.1212 0.0391 0.0000 -0.0341 4
Canada 0.0298 0.1209 0.0391 0.0000 -0.0325 6
Chile 0.1318 0.0219 0.0000 0.0668 0.0441 36
Colombia 0.1296 0.0196 0.0000 0.0662 0.0440 35
Costa Rica 0.1367 0.0135 0.0000 0.0717 0.0487 38
Czech Republic 0.0935 0.0545 0.0001 0.0362 0.0188 32
Denmark 0.0303 0.1231 0.0416 0.0000 -0.0336 5
Estonia 0.0667 0.0801 0.0131 0.0126 -0.0035 23
Finland 0.0090 0.1384 0.0561 0.0000 -0.0464 1
France 0.0390 0.1054 0.0253 0.0000 -0.0229 11
Germany 0.0253 0.1228 0.0404 0.0000 -0.0345 3
Greece 0.0543 0.0935 0.0229 0.0112 -0.0127 15
Hungary 0.1011 0.0488 0.0001 0.0438 0.0240 34
Iceland 0.0699 0.0778 0.0094 0.0128 -0.0011 24
Page 395
SEMIT 2023
Ireland 0.0642 0.0799 0.0108 0.0090 -0.0044 20

Israel 0.0607 0.0822 0.0058 0.0032 -0.0060 19
Italy 0.0585 0.0859 0.0088 0.0029 -0.0083 17
Japan 0.0480 0.1067 0.0295 0.0052 -0.0208 12
Korea, Rep. 0.0516 0.0986 0.0225 0.0006 -0.0172 14
Latvia 0.0776 0.0693 0.0083 0.0183 0.0046 27
Lithuania 0.0859 0.0610 0.0001 0.0291 0.0135 29
Luxembourg 0.0653 0.0831 0.0128 0.0167 -0.0035 22
Mexico 0.1326 0.0169 0.0000 0.0683 0.0460 37
Netherlands 0.0278 0.1198 0.0377 0.0000 -0.0324 7
New Zealand 0.0692 0.0771 0.0047 0.0109 -0.0004 26
Norway 0.0698 0.0881 0.0175 0.0212 -0.0037 21
Poland 0.0684 0.0762 0.0048 0.0091 -0.0009 25
Portugal 0.0915 0.0549 0.0000 0.0300 0.0167 30
Slovak Republic 0.0995 0.0459 0.0000 0.0357 0.0223 33
Slovenia 0.0922 0.0620 0.0001 0.0383 0.0171 31
Spain 0.0383 0.1078 0.0288 0.0000 -0.0246 10
Sweden 0.0394 0.1170 0.0378 0.0000 -0.0289 9
Switzerland 0.0273 0.1289 0.0470 0.0000 -0.0371 2
Turkiye 0.0847 0.0594 0.0001 0.0258 0.0127 28
United Kingdom 0.0578 0.0873 0.0119 0.0051 -0.0091 16
United States 0.0510 0.0987 0.0222 0.0006 -0.0173 13
5. Sensitivity Analysis
In this section, we present a sensitivity analysis which that shows how modifying the most crucial
criteria will affect the ranking outcomes. To do so, eq. (25) is employed (Erceg et al., 2019).
𝑤𝛽
𝑤𝑛𝛽 = (1 − 𝑤𝑛𝛼 ) (1−𝑤 (25)
𝑛)
where 𝑤𝑛𝛽 is the corrected value of C1, C2, C4, C5 and C6, 𝑤𝑛𝛼 denotes the reduced value of the C3
criterion, 𝑤𝛽 is the original value of the criterion under consideration, and 𝑤𝑛 is the original value of
the C3 criterion. In the first scenario (S1), the C3 criterion is reduced by 0.02, and the other values are
calculated by eq. (25). Then, the value of the C1 criterion was reduced by 5% in each succeeding
scenario (S2, S3, S4 and S5) while adjusting the values of the other criteria to satisfy the condition
∑𝑛𝑗=1 𝑤𝑗 = 1 (Erceg et al., 2019). The reduced values of C3 and the adjusted values of the others are
given in Table 6.
Table 6. Changes in the criteria weights
W1 W2 W3 W4 W5 W6
S1 0.1688 0.1410 0.2177 0.1158 0.1651 0.1916
S2 0.1731 0.1446 0.1977 0.1188 0.1693 0.1965
S3 0.1839 0.1536 0.1477 0.1262 0.1799 0.2087
S4 0.1947 0.1626 0.0977 0.1336 0.1904 0.2210
S5 0.2055 0.1716 0.0477 0.1410 0.2010 0.2332
The rankings of the countries are changed according to these five scenarios. Figure 1 shows the changes
in the positions of the countries according to the different scenarios. The countries are expressed with
the labels starting with A1 in this figure. Costa Rica (A7), Hungary (A15), Korea (A21), Latvia (A22),
and Netherlands (A26) are the countries whose positions in the rankings have not changed. In the first
Page 396
SEMIT 2023
scenario, the result of the CRITIC method, the most important criterion is the international shipments
score (C3), while the least important criterion is the logistics competence score (C4). According to these
criteria weights, Finland (A11) is in the first position in the ranking. However, Switzerland has been
replaced with Finland according to S4 and S5 and becomes the first country instead of Finland.
Meanwhile, Finland falls to the second place. The first five countries are Finland, Switzerland (A35),
Germany (A13), Belgium (A3) and Denmark (A9).
In the fifth scenario, the most important criterion becomes the timeliness score (C6), and the least
important criterion becomes the international shipments score (C3). The first five countries are
Switzerland (A35), Finland, Denmark, Sweden (A34), and Canada (A4) in this scenario. Meanwhile,
Turkiye (A36) is at the twenty-eighth position on the ranking list in the scenarios of S1, S2, and S3.
However, it becomes the twenty-ninth ranked alternative in the scenarios of S4 and S5. Thus, it can be
deduced that the criteria weights have an impact on the rankings.
Figure 1. Results of sensitivity analysis taking into account the changes in the weights
6. Discussions and Conclusions

The logistics sector is of great importance for countries since it has a vital contribution to countries’
economic growth. The WB assesses the logistics performances of the countries and proposes LPI scores.
However, there are some problems in computing these scores. The criteria weights are ignored and a
simple averaging method is utilized. Therefore, we have introduced an MCDM model to compare
countries in terms of logistics performances. In this study, we have ranked OECD countries according
to their logistics performances by using an MCDM model. We have used LPI data, including the
performances of the countries on six indicators, which are customs, infrastructure, international
shipment, logistics quality and competence, tracking and tracing, and timeliness.
Our methodology calculates the weights of LPI indicators through the CRITIC method. CRITIC
method is one of the objective weighting methods, which incorporates standard deviation and
correlation scores extracted from the data. Considering the results of the CRITIC method, international
shipment has been assigned the highest weight, whereas logistics competence has been assigned the
lowest weight. The rankings of the countries are acquired by the COBRA method which is one of the
distance-based MCDM techniques. It is a recently developed method, and more comprehensive than
simple averaging. This method evaluates the alternatives considering the integration of various
distances from different reference points such as positive ideal, negative ideal and average solutions. In
addition, it uses not only the Euclidean distance metric but also the Taxicab distance metric. Finally,
the rankings of the OECD countries are obtained. Finland is in the first position in the ranking. It has
the best logistics performance among OECD countries according to 2023 LPI data.
Page 397
SEMIT 2023
Furthermore, we conduct sensitivity analysis to discover the impact of criteria weights on the
rankings. The value of the most important criterion is reduced by the specific degrees step by step, and
the values of the remaining criteria are adjusted by considering the condition that the sum of the weights
must be one. Five scenarios with different criteria weights are considered. As a result, we found different
rankings for different criteria weights. It means that the criteria weights influence the rankings of the
countries.
This paper gives important contributions to literature. OECD countries can observe their status in the
ranking with reference to logistics performances. They can take action to develop strategies to improve
their performances.
References
Arvis, J.-F., Ojala, L., Shepherd, Ulybina, D., B., & Wiederer, C. (2023). Connecting to Compete 2023:
Trade logistics in an uncertain global economy-the Logistics Performance Index and its indicators.
World Bank. Retrieved from: https://lpi.worldbank.org/sites/default/files/2023-
04/LPI_2023_report_with_layout.pdf
Candan, G. (2019). Lojistik performans değerlendirmesi için bulanık AHP ve gri ilişkisel analiz
yöntemleri ile bütünleşik bir yaklaşım. Anemon Muş Alparslan Üniversitesi Sosyal Bilimler Dergisi,
7(5), 277-286. doi: http://dx.doi.org/10.18506/anemon.506769
Council of Supply Chain Management Professionals (2013). CSCMP Supply chain management terms
and glossary. Retrieved from:
https://cscmp.org/CSCMP/Academia_and_Awards/SCM_Definitions_and_Glossary_of_Terms/CS
CMP/Educate/SCM_Definitions_and_Glossary_of_Terms.aspx?hkey=60879588-f65f-4ab5-8c4b-
6878815ef921.
Çakır, S. (2017). Measuring logistics performance of OECD countries via fuzzy linear
regression. Journal of Multi‐Criteria Decision Analysis, 24(3-4), 177-186. doi:
https://doi.org/10.1002/mcda.1601
Çalık, A., Erdebilli, B., & Özdemir, Y. S. (2023). Novel integrated hybrid multi-criteria decision-
making approach for logistics performance index. Transportation Research Record, 2677(2), 1392-
1400. doi: https://doi.org/10.1177/03611981221113314
Diakoulaki, D., Mavrotas, G., & Papayannakis, L. (1995). Determining objective weights in multiple
criteria problems: The critic method. Computers and Operations Research, 22(7), 763-770. doi:
https://doi.org/10.1016/0305-0548(94)00059-H
Erceg, Ž., Starčević, V., Pamučar, D., Mitrović, G., Stević, Ž., & Žikić, S. (2019). A new model for
stock management in order to rationalize costs: ABC-FUCOM-interval rough CoCoSo
model. Symmetry, 11(12), 1527. doi: https://doi.org/10.3390/sym11121527
Görçün, F. Ö. (2019). Orta Asya Türk Cumhuriyetlerinin lojistik ve taşımacılık performansları ve
verimliliklerinin analizi için hibrid bir çok kriterli karar verme modeli. MANAS Sosyal Araştırmalar
Dergisi, 8(3), 2775-2798. doi: https://doi.org/10.33206/mjss.511522
Hwang, C., & Yoon, K. (1981). Multiple attribute decision making methods and applications. Springer.
doi: https://doi.org/10.1007/978-3-642-48318-9
Islam, D. M. Z., Meier, J. F., Aditjandra, P. T., Zunder, T. H., & Pace, G. (2013). Logistics and supply
chain management. Research in Transportation Economics, 41(1), 3-16. doi:
https://doi.org/10.1016/j.retrec.2012.10.006
Kısa, A. C. G., & Ayçin, E. (2019). OECD ülkelerinin lojistik performanslarının SWARA tabanlı EDAS
yöntemi ile değerlendirilmesi. Çankırı Karatekin Üniversitesi İİBF Dergisi, 9(1), 301-325. doi:
https://doi.org/10.18074/ckuiibfd.500320
Page 398
SEMIT 2023
Krstić, M., Agnusdei, G. P., Miglietta, P. P., Tadić, S., & Roso, V. (2022). Applicability of industry 4.0
technologies in the reverse logistics: A circular economy approach based on Comprehensive Distance
Based Ranking (COBRA) method. Sustainability, 14(9), 5632. doi:
https://doi.org/10.3390/su14095632
Martí, L., Martín, J. C., & Puertas, R. (2017). A DEA-logistics performance index. Journal of Applied
Economics, 20(1), 169-192. https://doi.org/10.1016/S1514-0326(17)30008-9. doi:
https://doi.org/10.1016/S1514-0326(17)30008-9
Mercangoz, B. A., Yildirim, B. F., & Yildirim, S. K. (2020). Time period based COPRAS-G method:
application on the Logistics Performance Index. LogForum, 16(2). doi:
http://doi.org/10.17270/J.LOG.2020.432
Mešić, A., Miškić, S., Stević, Ž., & Mastilo, Z. (2022). Hybrid MCDM solutions for evaluation of the
logistics performance index of the Western Balkan countries. Economics, 10(1), 13-34. doi:
http://doi.org/10.2478/eoik-2022-0004
Orhan, M. (2019). Türkiye ile Avrupa Birliği ülkelerinin lojistik performanslarının Entropi ağırlıklı
EDAS yöntemiyle karşılaştırılması. Avrupa Bilim ve Teknoloji Dergisi, 17, 1222-1238. doi:
https://doi.org/10.31590/ejosat.657693
Penadés-Plà, V., García-Segura, T., Martí, J. V., & Yepes, V. (2016). A review of multi-criteria
decision-making methods applied to the sustainable bridge design. Sustainability, 8(12), 1295. doi:
https://doi.org/10.3390/su8121295
Rezaei, J., van Roekel, W. S., & Tavasszy, L. (2018). Measuring the relative importance of the logistics
performance index indicators using Best Worst Method. Transport Policy, 68, 158-169. doi:
https://doi.org/10.1016/j.tranpol.2018.05.007
Stević, Ž., Erceg, Ž., & Kovačević, B. (2022). The impact of sensitivity analysis on the evaluation of
the logistics performance index. Novi Ekonomist, 16(1), 41-48. doi:
https://doi.org/10.7251/NOEEN2231041S
Tzeng, G. H., & Huang, J. J. (2011). Multiple attribute decision making: methods and applications.
CRC Press. doi: https://doi.org/10.1201/b11032
Ulutaş, A., & Karaköy, Ç. (2019). An analysis of the logistics performance index of EU countries with
an integrated MCDM model. Economics and Business Review, 5(4), 49-69. doi:
https://doi.org/10.18559/ebr.2019.4.3
Ulutaş, A., & Karaköy, Ç. (2021). Handbook of research on applied AI for international business and
marketing applications. In B. Christiansen & T. Škrinjarić (Eds.), Evaluation of LPI values of
transition economies countries with a grey MCDM model (pp. 499-511). IGI Global. doi:
https://doi.org/10.4018/978-1-7998-5077-9.ch024
Yildirim, B. F., & Adiguzel Mercangoz, B. (2020). Evaluating the logistics performance of OECD
countries by using fuzzy AHP and ARAS-G. Eurasian Economic Review, 10(1), 27-45. doi:
https://doi.org/10.1007/s40822-019-00131-3
Page 399
SEMIT 2023
SP09: Quality, Productivity, and Project Management
İnsansiz Hava Araçlarinda İnsan Hatasinin Etkilerinin Sistematik Analizi Ve Siniflandirilmasi-

İha Tasarimi Ve Ergonomi
İbrahim AKYÜREK
Mühendisliği Bölümü İş Sağlığı ve Güvenliği Doktora Programı
ORCID: 0000-0002-0286-7607
Prof. Dr. ERGÜN ERASLAN
Mühendisliği Bölümü
ORCID: 0000-0002-5667-0391
ÖZET
Bu çalışmada öncelikle insansız hava araçlarında insan faktörü ile ilgili yayınlanan literatür gözden
geçirilmiştir. Bir sonraki adımda, çalışmaları sınıflandırmak üzere bir içerik analizi yapılmıştır.
Literatürdeki çalışmalar sistematik olarak ana ve alt gruplarda toplanmıştır. Sınıflandırmanın ardından
literatür tartışılarak eksik alanlar ortaya çıkarılmış ve gelecekteki araştırmalara ışık tutulmaya
çalışılmıştır. 69 çalışmanın derinlemesine incelendiği bir makale üzerinden içerik analizi yapılmıştır.
Bu süreçte araştırmacıların üzerinde durdukları bazı ortak konular belirlenmeye çalışılmıştır. Beş ana
grupta toplanan sınıflandırma içinde İHA Tasarımı ve Ergonomi başlığı üzerine yoğunlaşılmıştır.
İnsansız hava araçlarının kullanıcı arayüzlerinin ve ekranlarının tasarımında insan faktörlerinin etkisi
bulunmaktadır. Operatörün bulunduğu yer kontrol istasyonunun (YKİ) tasarımı ve ergonomik
özelliklerinin insan hatasına yol açabilecek faktörleri incelenmiştir.Kullanıcı arayüzü ve ekran
tasarımları ile ergonomik yapıların iş yükü, durum farkındalığı ve karar verme gibi değişkenler
üzerindeki etkileri araştırılmıştır. İş yükünü azaltacak, durumsal farkındalığı artıracak, algıyı olumlu
yönde etkileyecek ve operatörlerin doğru karar vermesine yardımcı olacak optimum tasarım ve
ergonomik yapıların geliştirilmesine odaklanılması sağlanmıştır. Ayrıca, insansız hava araçlarının insan
faktörünün olumsuz etkilerini ortadan kaldıracak şekilde tasarlanmasıdır. Tasarım ve ergonomik
yapıdaki gelişmeler sonucunda insansız hava araçlarının operatörler tarafından daha verimli, etkin ve
güvenli bir şekilde kullanılmasının sağlanması amaçlanmaktadır.Yer kontrol istasyonlarının ekran açısı,
sıcaklık, ekran karşısındaki duruş, çalışma saati gibi hususlarda ergonomik açıdan yaşanabilecek
sorunlar ve çözüm önerileri üzerine çalışılmıştır.Örnek bir yer kontrol istasyonu tasarımı incelenmiş
ergonomik faktörler gözden geçirilmiştir. Ekranların açısı, oturma pozisyonu ,termal konfor gibi
faktörler üzerine önerilerde bulunulmuştur.
Anahtar Kelimeler: İnsansız hava aracı, yer kontrol istasyonu,tasarım,ergonomic.
A Systematic Analysis and Classification of the Effects of Human Fault in Unmanned Aviation
Abstract
In this study, first of all, the published literature on the human factor in unmanned aerial vehicles was
reviewed. In the next step, a content analysis was made to classify the studies. Studies in the literature
Page 400
SEMIT 2023
were systematically grouped into main and subgroups. After the classification, the literature was
discussed and missing areas were revealed and it was tried to shed light on future research. Content
analysis was conducted on an article in which 69 studies were examined in depth. In this process, some
common issues that the researchers focused on were tried to be determined. In the classification grouped
in five main groups, the focus is on UAV Design and Ergonomics. Human factors have an effect on the
design of the user interfaces and screens of unmanned aerial vehicles. The design and ergonomic
features of the operator's ground control station (GCS), which may cause human error, were
investigated. The effects of user interface and screen designs and ergonomic structures on variables
such as workload, situation awareness and decision making were investigated. Focusing on the
development of optimum design and ergonomic structures that will reduce the workload, increase
situational awareness, positively affect perception and help operators make the right decision. In
addition, unmanned aerial vehicles are designed to eliminate the negative effects of the human factor.
As a result of the developments in the design and ergonomic structure, it is aimed to ensure that the
unmanned aerial vehicles are used more efficiently, effectively and safely by the operators. Ergonomic
problems and solution proposals have been studied in issues such as the screen angle of the ground
control stations, temperature, posture in front of the screen, working hours. The design of a ground
control station is examined and ergonomic factors are reviewed. Suggestions were made on factors such
as the angle of the screens, sitting position, thermal comfort.
Keywords: Unmanned Aerial Vehicle, Ground Control Station, Design, Ergonomics.
1. Giriş
Uluslararası Sivil Havacılık Örgütü insansız hava araçlarını uzaktan kumandalı ve otomatik olmak
üzere iki ana gruba ayırmaktadır.
Günümüzde keşif, taarruz, diğer insansız hava araçlarına karşı savunma, hedef alma askeri eğitim ve
mayın temizleme gibi askeri görevlerde insansız hava araçları sıklıkla kullanılmaktadır.
Sivil kullanım alanları arasında doğal afetlerin ve arama kurtarma faaliyetlerinin takibi, tarım
alanlarının ve zirai ilaçlamaların izlenmesi, havadan fotoğraf ve video çekimi yer alıyor. Bunun yanı
sıra haritacılık, çevre gözlemi, kentsel yapılanma, arkeoloji, orman yangınlarının izlenmesi veya
kamusal alanlar gibi birçok faaliyet için insansız hava araçları tercih edilmektedir.
Havacılık sistemlerinde insan faaliyetinin evrimi uçağın varlığı sırasında, içlerindeki bir kişinin rolü,
otomasyonun büyümesiyle birlikte önemli değişikliklere uğramıştır. Başlangıçta, pilotun rolü manuel
kontrolden oluşuyordu, yavaş yavaş sevk kontrol ve takip operatör faaliyetlerine doğru değişti, ardından
"İnsansız hava aracı” ortaya çıkmıştır. Bir uçağı uçurmanın doğrudan insan faaliyeti, kokpit aletlerinin
kullanımı ve manuel kontrol, kokpit aletlerinin izlenmesine kadar uzaktan kumanda için yer istasyonu
araçlarının kullanımına kadar uçağı neredeyse otomatik olarak kontrol edin uçağın kontrolü olarak
tanımlanmaktadır.
Dünya askeri insansız hava araçları pazarının hacminin 2025 yılında 26,8 milyar dolara ulaşması
beklenmektedir. Dünyadaki askeri ve sivil insansız hava aracı pazarı 2019 yılında toplam 9,3 milyar
dolara ulaşmıştır. Bu pazarın yıllık bileşik büyüme oranı ile %15,5 artarak 2025 yılında 45,8 milyar
dolara ulaşması beklenmektedir. Önümüzdeki 5 yılda dünyanın her yerinde insansız hava araçlarının
sayısında artış beklenmektedir [1].
İnsansız hava araçlarının yoğun kullanımı, gerekli önlemler alınmadığı takdirde kazaların ve olayların
artmasına neden olabilir. Amerika Birleşik Devletleri Federal Havacılık İdaresi'ne (FAA) bildirilen
insansız hava aracı kazası sayısı Şubat 2014'te sadece 50 iken, bu sayı Aralık 2016'da 200'ü aşmıştır.
Kaza ve olaylara neden olan faktörler arasında insan faktörü önemli bir rol oynamaktadır. Bu bağlamda
insansız hava araçlarında insan faktörünün daha iyi anlaşılması, gelecekte olabilecek kaza ve olayların
önlenmesi açısından oldukça faydalı olacaktır.
Bu çalışmada öncelikle insansız hava araçlarında insan faktörü ile ilgili 1945-2020 yılları arasında
yayınlanan literatür gözden geçirilmiştir.
Page 401
SEMIT 2023
Bir sonraki adımda, literatüre sistematik bir bakış açısı kazandırmak için çalışmaları sınıflandırmak
üzere bir içerik analizi yapılmıştır.
Sonuç olarak literatürdeki çalışmalar sistematik olarak ana ve alt gruplarda toplanmıştır.
Sınıflandırmanın ardından literatür tartışılarak eksik alanlar ortaya çıkarılmıştır.
İHA sistemlerine özgü tasarım tavsiyeleri, ilgili ofis ve iş için literatürde tanımlanmış çevre literatürü
gözden geçirilmiş ve bu boşlukları doldurmak için kullanılmıştır.
Kontrol istasyonu üzerinde yapılan çalışmaların genel amacı, Federal Havacılık İdaresi'ne (FAA), İHA
sistemlerinin güvenli pilotluğunu sağlayacak insansız hava aracı sistemi (UAS) kontrol istasyonlarının
ergonomik tasarımı hakkında bilgi sağlamaktır. İnsansız uçak kazalarının büyük bir yüzdesi, %69'u
hem fiziksel iş istasyonu tasarımı hem de insan-makine etkileşimi dahil olmak üzere insan faktörü
konularına atfedilmiştir. Literatürde, FAA'ya uygun operatör konforunu ve performansını artıracak ve
kas-iskelet yaralanma risklerini azaltacak minimum fiziksel kontrol istasyonu tasarım önerileri
sağlamaktır. İHA sistemleri kontrol istasyonu tasarımına özgü literatürün ayrıntılı bir incelemesi
yapılmıştır. Kontrol istasyonlarıyla ilgili literatür olmasına rağmen, aslında iş istasyonunun fiziksel
tasarımına yönelik çok az çalışma vardır. Bu nedenle, kontrol istasyonları için bir ön tasarım önerileri
vermek üzere mevcut iş istasyonu tasarım literatürü yeniden gözden geçirilmiştir.
Bu tavsiyeler mevcut standartlara ve en son literatüre dayanmakla birlikte, kontrol istasyonunun
tasarımında veya içinde değişiklik yapılacak alanları belirlemek için bu tavsiyelerin resmi olarak test
edilmesi ve değerlendirilmesinin yanı sıra mevcut kontrol istasyonlarıyla karşılaştırılmasına ihtiyaç
vardır
Genel olarak, İHA'ların kontrolünün sabit bir yer ve tesiste gerçekleşmesi beklenmektedir. Bununla
birlikte, İHA'ların mobil kontrol istasyonları kullanılarak da kontrol edilmesi gerekebileceğinden, iki
ek senaryo için öneriler geliştirilmiştir: 1) hareketli bir ortamda (örn. römork, kamyonet, gemi, uçak)
bulunan kontrol istasyonu ve 2) sabit kontrol istasyonudur.
Bu itibarla çevresel her üç seviye için de faktörler (sabit kontrol istasyonu, mobil kontrol istasyonu ve
mobil operatörün taşıdığı kontrol istasyonu) dikkate alınmış ve öneriler sunulmuştur.
2. Metodoloji
İnsansız hava araçlarında insan faktörü ile ilgili çalışmaları bulmak için bir dizi arama yapılmıştır.
“insan faktörü” ve “insan hatası” anahtar kelimeleri, “insansız hava aracı”, “İHA”, “insansız hava
aracı”, “insansız havacılık sistemi”, “insansız hava aracı sistemi” ve “UAS” anahtar kelimeleri ile
birlikte aranmaktadır. Çalışmaların başlık”, “yazar anahtar sözcükleri” ve “öz” alanlarıdır. Zhang et. al.
(2018) anahtar kelime seçiminde rehber olarak kullanılmıştır [2].
Tarama ve eleme aşamalarından sonra 69 çalışmanın derinlemesine incelendiği içerik analizi
yapılmıştır. Bu süreçte araştırmacıların üzerinde durdukları bazı ortak konular belirlenmeye
çalışılmıştır. Shapel et. al. (2007), Chang ve Wang (2010) ve Wiener ve Nager (2014), çalışmaların
sınıflandırmasını tanımlamak için kullanılmaktadır. Ayrıca Zhang et. al. (2018) bu süreçte yol gösterici
olarak dikkate alınmıştır [2].
Yapılan içerik analizi sonucunda literatürdeki çalışmalar 4 ana grupta toplanmış ve daha sonra aynı
yöntem kullanılarak her grubu oluşturan alt gruplar oluşturulmuştur. İnsan Faktörlerinin Tanıtılması ilk
sıradadır. Havacılıkta insan faktörleri alanında da sıklıkla çalışılan iş yükü, durumsal farkındalık, karar
verme ve iş birliği gibi değişkenleri araştıran insansız hava araçları üzerine çalışmalar bulunmaktadır.
Daha sonra “iş yükü”, “durumsal farkındalık”, “karar verme ve özerklik”, “yorgunluk” ve “iş birliği,
takım uyumu ve koordinasyonu” olarak adlandırılan alt gruplar oluşturularak çalışmalar ilgili alt
gruplara dağıtılır.
Örneğin, bir çalışma aynı zamanda yorgunluk, stres ve uyanıklığı araştırmış olabilir. Bir diğeri ise aynı
araştırmadaki yorgunluğu, ekip boyutunu ve iş yükünü incelemektedir. Bu tür çalışmalar “Çok faktörlü
araştırmalar” adı verilen farklı bir alt grupta toplanmaktadır.
Page 402
SEMIT 2023
İHA Tasarımı ve Ergonomi başlığı bu çalışmanın ana konusunu oluşturmaktadır. Çalışma grubu,
insansız hava araçlarının kullanıcı arayüzlerinin ve ekranlarının tasarımında insan faktörlerinin etkisi
ile ilgilidir. Kullanıcı arayüzü ve ekran tasarımları ile ergonomik yapıların iş yükü, durum farkındalığı
ve karar verme gibi değişkenler üzerindeki etkileri araştırılmıştır. Bu çalışmaların birçoğu, iş yükünü
azaltacak, durumsal farkındalığı artıracak, algıyı olumlu yönde etkileyecek veya operatörlerin doğru
karar vermesine yardımcı olacak optimum tasarım veya ergonomik yapıların geliştirilmesine
odaklanmaktadır. İkinci alt gruptaki çalışmalar, benzer bir tasarım ve ergonomi yaklaşımını insansız
hava araçlarının kontrol üniteleri ve diğer sistemleri üzerinde yürütmektedir. Bu çalışmalarda, uçuş
yönetim sistemleri, navigasyon sistemleri, aviyonik sistemler gibi insansız hava aracı bileşenlerinin
tasarımının ve ergonomik yapısının insan faktörü üzerindeki etkileri derinlemesine incelenmektedir [3].
Bu gruptaki çalışmaların ortak amacı, insansız hava araçlarının insan faktörünün olumsuz etkilerini
ortadan kaldıracak şekilde tasarlanmasıdır. Tasarım ve ergonomik yapıdaki gelişmeler sonucunda
insansız hava araçlarının operatörler tarafından daha verimli, etkin ve güvenli bir şekilde
kullanılmasının sağlanması amaçlanmaktadır.
Ekip Üyeleri ise diğer bir ana grup başlığıdır. Bu gruptaki çalışmalar mürettebat konularını insan
faktörü açısından ele almaktadır. Üç alt gruba ayrılırlar. Birinci gruptaki çalışmalar, farklı koşullar ve
senaryolar altında mürettebat davranışını ve davranışsal tepkileri inceler. Ekiplerin farklı durumlara
verdiği tepkiler değerlendirilerek, insan faktörü değişkenlerine dayalı ekip davranış profillerinin
oluşturulması amaçlanmaktadır.
İkinci alt gruptaki çalışmalar, ekip seçimi, performans değerlendirmesi ve ekip yetkinliklerini insan
faktörleri perspektifinden ele almaktadır. Bu çalışmalardan bazıları, mürettebatın fiziksel yeterlilik,
tıbbi durum ve hatta dil yeterliliği açısından performansını ve yeterliliklerini ele almaktadır. Bazı
çalışmalar, insansız hava aracı pilotlarının performans ve yeterlilik değerlendirmeleri için insan faktörü
odaklı öneriler sunmaktadır. Ayrıca ekip üyesi olarak yetiştirilecek kişilerin seçiminde dikkate alınması
gereken insan faktörü konuları da bu alt grupta incelenir [4].
Operasyonel Konular son ana grup başlığıdır. Bu gruptaki çalışmalar, insansız hava araçları ile
gerçekleştirilen fiili uçuş operasyonlarını ve ilgili görevleri insan faktörü açısından incelemektedir. Bu
gruptaki yalnızca bir çalışma, operatörlerin uçuş prosedürlerine ve kontrol listelerine uyma derecesini
otomatik ve sürekli olarak kontrol edecek bir sistem önermektedir. Bu nedenle söz konusu çalışma
“işlemler ve kontrol listeleri” adı altında farklı bir alt gruba alınmıştır [5].
Diğer çalışmalar “uçuş operasyonları” adı verilen alt grupta yer almaktadır. Bu çalışmalardan bazıları,
insansız hava araçlarının paket teslim, arama kurtarma ve kamu hizmetleri gibi belirli görevlerini insan
faktörü açısından iyileştirmenin yollarını arıyor. Ayrıca, bu alt gruptaki çalışmalar, bir operatörün
kontrol edebileceği optimum uçak sayısının hesaplanması ve görev parçalarının operatörler ve sistemler
arasında dağıtılması gibi konuları araştırır [6].
Bu çalışmada yapılan sınıflandırmanın açık bir sonucu, nispeten yeni olan insansız hava araçları
alanında, insan faktörü çalışmalarına bir yol göstermesidir. Ayrıca insan faktörleri alanının ana konuları
olan iş yükü, karar verme, durum farkındalığı, yorgunluk gibi değişkenlerle ilgili çok az çalışma olması
dikkat çekicidir. Bu değişkenlerin farklı koşullar altında nasıl değiştiği, uçuş operasyonlarının etkinlik,
verimlilik ve emniyetini nasıl etkilediğinin araştırılması insansız hava araçlarının geleceği açısından
faydalı olacaktır.
Bazı yer kontrol istasyonu (YKİ), karmaşık tasarım öğelerine sahiptir. Küçük veya rutin görevleri
gerçekleştirmek için optimize edilmemiş otomasyon arayüzleri bulunmaktadır. YKİ'deki bu tasarım
hataları kokpit konumlarına tasarım ve ergonomi ilkeleri uygulanarak önlenebilir. YKİ tasarımındaki
sorunlar, geliştirilmekte olan İHA'nin optimize edilmemesine neden olabilir.
Kontrol istasyonu tasarımında;
• İHA'yı kontrol etmek için aktif olarak kullanılmadan önce pilota kontrol bağlantısının kalitesi
hakkında bilgi sağlamak
• Pilota, hangisinin aktif olduğunu belirlemek için gerekli bilgileri sağlamak (örn. seçilen
frekans, uydu vs. karasal)
Page 403
SEMIT 2023
• Bağlantı kaybolduğunda veya önemli ölçüde bozulduğunda pilotu uyarmak

• Pilotun iletişim modunu seçmesi için bir araç sağlaması (örn. karasal/uydu, frekans)
• İşitsel veya görsel uyarıların içgüdüselliği
• İş yoğunluğu
• Tutarlılık
• Kontrol girişleri hakkında geri bildirim
• Rahatsız edici uyarılar
• Acil özellikler
Seçilen İHA konsepti aşağıdaki ana maddelere dayanmaktadır;
Hava vektörleri hava ve yüzey uzayında görsel yarıçapın ötesine geçebilen, altında İHA’nın çalışması
hem hava hem de yüzey vektörleri için eşzamanlı, ayrı çalışma koşulları; operasyon ve sezgisel ve
minimum kaynak kullanımı yoluyla İHA'nın bakımı; İHA işlemi tüm aşamaları kapsar. Uçuş
planlamasından bilgilendirmeye kadar görevin boyutu, ağırlığı ve çok yönlülüğü nedeniyle İHA
konsepti mevcut bir ürüne dayanmaktadır [7]. İHA dahili bağlantıları şunlardır: güç kaynağı (dahili /
harici kaynak), video sinyali iletimi, telemetri sinyalinin iletimi, kontrol ve komut sinyalinin iletimi,
harici kullanıcılar ve bağlantılardır [8].
İHA geliştirme ekibi, tasarım aşamasından itibaren bir dizi faktörü dikkate alır, örneğin, yapı ve
geometri, donanım ve bağlantı donanımı, yazılım bileşeni ve kullanıcı arayüz dikkate alınan
parameterelerden bazılarıdır. Parametereler aşağıdaki gibi detaylandırılmıştır. Geometri ve yapı
tasarım özellikleri boyutsal olarak ve tamamen optimize edilmiştir. İHA'nın hareketliliği (sert kap,
kulplar, taşıma makaraları) gözden geçirilmiştir.
İHA operatörlerinin başlıca ergonomik sorunları şunlardır [9]
• Duyusal sinyallerde azalma (nesnelerle göz temasının bozulması (kamera bir sınırlı görüş alanı,
azalmış işitsel ve koku alma duyumları),
• Kontrol istasyonunun psikolojik ve duygusal algısı (bunlar daha çok geleneksel bir kabinden daha
kontrol odaları veya ofis iş istasyonları),
• İnsan fizyolojisi açısından gerçekçi olmayan, görevin zamanlaması (için örneğin, 24 saatten fazla)
yorgunluk ve ardından ihtiyaç ile ilişkili riskler eşlik eder. Kontrolü başka bir operatöre devretmek, bu
da hatalara yol açar;
• Uçuşun acil bir şekilde sonlandırılması ve uçağın imha edilmesi olasılığı
• Otomasyona güven ayrı bir sorundur. Bir uçağın aksine, bir İHA'nın hiçbir şekilde manuel kontrolü
yoktur.
İletişim sistemleri, sistem öğeleri ve hariciler arasında veri paylaşımı için araç sağlar. Günümüzde çoğu
İHA sistemi, veri iletmek için radyo frekanslarını (RF) kullanır. Bu sistemler yapılandırılmış uydu
iletişimleri gibi görsel aralığın ötesinde doğrudan veya dolaylı iletişim için (SATCOM) veya havadan
radyo rölesi şeklinde gerçekleşir. İletişim sistemleri, başka sistemlere entegre edilebilen birkaç
unsurdan oluşur. Sistem, modemler, vericiler, yükselticiler ve antenlerden oluşur. Antenin
entegrasyonu, aerodinamik etkiler ile anten arasında yeterli bir denge bulunmalıdır. İletişim
sistemlerinin birbirine bağlı bileşenlerinin kapsama alanı ve yakınlık verileri dikkate alınmalıdır. Anten
ve gerekli görüş alanı, entegrasyon yaklaşımını ve tasarımı etkiler. Entegrasyonu iletişim sistemi, ilk
tasarım aşamalarından itibaren tasarımda önemli bir rol oynamalıdır, çünkü kritik sistemi uyarlamaya
yönelik girişimlerin insansız uçaklar için tasarlanması zahmetli bir çalışmadır.
Bir İHA çalıştırmanın en önemli yönlerinden biri, büyük ölçüde harici kaynaklara dayanan güvenliktir.
İHA operatörleri, iletişim sinyalleri ile İHA'yı ve faydalı yüklerini doğrudan kontrol eder. Hava aracı
uçuş operatör ü(kontrolör veya yer pilotu) ve görev yükü operatörünün yer kontrol istasyonunda
Page 404
SEMIT 2023
gerçekleştirilebileceği diğer işlevleri şunlardır: (1) görev komutan, (2) iletişim operatörü, (3) akıllı
uzman ve (4) hava gözlemcisi.
İHA kayıpları, operatör hatalarına atfedilmektedir. Bir kısmı etkisiz arayüz tasarımı ile ilgilidir. Zayıf
bir yer kontrol istasyonu tasarımının bazı olası eksiklikleri, YKİ'nin; (1) aşırı iş yükü yaratması, (2)
kritik bilgileri etkili bir şekilde sunamaması, (3) acil durum oluşumu ve (4) YKİ içinde operatör stresi
oluşturan rahatsız edici bir duruma sahiptir.
Operatörün çalışma ortamı rahat olmalıdır ve arayüz etkili olmalıdır; aksi takdirde operasyonel
yorgunluk ve olası İHA kayıpları olasıdır. İnsanı ele alan bir tasarım planlandığında, ergonomik
standartların dikkate alınması gerekmektedir. İnsan faktörleri mühendisliği alanı bilimsel bilgiyi
kullanır bir insan-makine sisteminin tasarımını ve kullanımını belirlemede insan davranışı hakkında
bilgi sağlar. Amaç, insan hatasını en aza indirerek sistem verimliliğini, performansı, konforu ve
emniyeti artırmaktır. Gelişebilecek tekrarlayan hareketelrden kaynaklı zorlanmaları önlemek için
uygun ergonomik tasarıma ihtiyaç bulunmaktadır [10].
YKİ içindeki operatör, ihtiyaç duyulduğunda ilgili tüm bilgilere erişebilmelidir. Yüksek hacimli
verilerle rahatsız olmaması gerekir. Ayrıca, YKİ' deki sıcaklık ve nem, maksimum konfor ve etkinlik
sağlamak için kontrol edilmelidir. Ek olarak, eş zamanlı faaliyetlerin sayısı normal bir insandan az veya
ona eşit olmalıdır. Bu nedenle, joystick, pedal ve gösterge sayısı sınırlandırılmalıdır. Oturma ortamı ve
koltuk/masa ilişkisi ergonomik esaslara göre tasarlanmalıdır. Yazılım/algoritma, arayüzün operatöre
izin verecek şekilde menü odaklı olacak şekilde tasarlanmalıdır (Belki birkaç tuş ile gerekli bilgileri
bulmak için). Birden fazla olduğunda her operatörün görevleri açıkça tanımlanmalıdır.
Yanlış monitör konumlandırması, boyun ve göz yorgunluğuna neden olabilir. Baskı oluşturan koltuk
sırt üzerinde rahatsızlığa yol açabilir. Otururken monitör göz hizasında olmalıdır. "Görüş Konisi"
kişinin göz hizasında tepeden başlayan ve 30°'lik bir açıyla alçalan görme alanını tarayan en yakın yerde
bulunmalıdır. Monitörler çok olduğunda uzakta, insanlar iyi görmek için öne eğilme eğilimindedir. Bu,
insanlar yaşlandıkça giderek daha sık tekrarlanmaktadır. Çünkü görme neredeyse kaçınılmaz olarak
zamanla azalır. Temel kural; kişi kolunu uzatabilir ve sadece parmak uçlarıyla ekrana dokunabilirse, o
zaman doğru konuma ulaşılmıştır. Bilekleri ve kolları optimum pozisyonda tutmak, tekrarlayan hareket
riskini azaltmak, joystick, boyunduruk ve anahtarlar otururken dirseklerle aynı seviyede olmalıdır.
Herkesin standart bir boyutu yoktur, basit bir düzeltme ayarlanabilir. Düzgün oturmak %20-30 arasında
olmalıdır. Alt sırttaki baskıdan koltuk 17 inç ile 19 inç arasında bel desteğine sahip olmalıdır. Vücut,
sırtı koltuğa gelecek şekilde konumlandırılmalıdır. Paneli görmek için kendini öne doğru eğilmiş
bulursa, hareket etmesi gerekir. Bacak konumlandırma, koltuktaki genel konuma katkıda bulunur, bu
nedenle bacakların bükülü olduğundan emin olunmalı, dizlerde yaklaşık 90o açıda olmalıdır. (Şekil 1).
Bu durum, sırttaki baskıyı hafifletmeye yardımcı olur. Hareket esastır ancak sirkülasyon için, bu
nedenle konumlandırmada ince değişikliklere izin verilmeli ve ayakta durulmalı ya da en az birkaç
adım yürünmelidir. Ayaklar yere sağlam basmalıdır. Uygun bilek hizalaması için gereken oturma
pozisyonu ayakların yere ulaşmamasına neden oluyorsa, desteğin yüksekliği dizleri dik açıda tutacak
şekilde ayakları desteklemek için ayaklık kullanılmalıdır.
Page 405
SEMIT 2023
Şekil 1: Yer Kontrol İstasyonu

YKİ içinde bir dizi monitör ve ekran bulunmaktadır. Bir monitör, bir video görüntüsü taşır. Monitörlerin
altında, üzerinde kontrol çubuklarının bulunduğu kontrol güverteleri bulunur. Uçuş sırasında İHA'nın
doğrudan operatör kontrolü buradan sağlanır. Görev verilerini, yazılım güncellemelerini veya İHA'ya
yol noktalarını girmek için tuş takımı kontrolleri, kontrol güvertesinde bulunur. Uçağı başlatmak, uçuş
öncesi test verilerini seçmek ve uçak kalkışını etkinleştirmek için kontroller sağlanır. Orada olacak
kayıt ortamıyla birlikte yük durumunu ve verileri (görüntü ve/veya diğer veri türleri) gösteren ekranlar
olacaktır. İHA'da ve faydalı yükün çalışmasını kontrol etmek ve göstermek için bir ekran da dahil
edilebilir. Görüntü ekranı, İHA sensörlerinin tam çözünürlüğünden yararlanmak için operatörün
görebileceği şekilde yeterince büyük olmalıdır. YKİ operatörleri, kontrol ettikleri İHA'da hareketli bir
harita ekranında bir simge belirir ve ekranda görüntülenen faydalı yük görüntüsünü izler. Denizden
yakın mesafe İHA fırlatıldığında, kontrol istasyonu denize yakın bir gemide konumlandırılır.
İHA operatörünün faaliyetinin özellikleri şu şekilde sıralanır; İnsansız hava araçlarında insan
faktörünün tezahürleri her operasyonda ortay çıkmaktadır. Örneğin, bir uçağın pilotu fiziksel olarak
algılar ve işlerse bilgi kokpitte yer alır. Aynı zamanda İHA operatörü zihinsel olarak algılar ve bilgileri
kontrol nesnesinden uzaktan işler şeklinde yürütür. İHA operatörlerinin faaliyeti, ağırlıklı olarak bilişsel
tepki ve düşünce süreçleri ile dikkatleri yüksek olmalıdır. Beklenmedik bir durumda İHA operasyon
görevlerinin yüksek başarı oranı ve görevlerin önemli süresi, artan strese ve fazla çalışmaya neden olur.
İnsansız hava aracı sistemlerinde insan faktörü eğitimi disiplinler arasıdır. Teknolojilerin ve ekipmanın
güvenliği, insan faaliyetinin güvenliği önem arz etmektedir. Her hangi bir teknolojinin veya sistemin
yaşam döngüsü, riskleri önleme ve olası riskleri en aza indirme yeteneği insan faktörünün olumsuz
sonuçlarını azaltabilmektedir. Toplumun karmaşık sistemlerinin (ulaşım, enerji vb.) güvenliğini
sağlama ihtiyacı üzerinden eğitim uzmanları insan faktörünün doğası üzerine yoğunlaşmaktadır.
İnsansız sistemlerin tasarımında insan operatörünün psikofizyolojik ve bilişsel özellikleri arayüz ve
bunların teknik nitelikleri ve yetenekleri dikkate alınmadan yapılan geliştirme faaliyetlerinde güvenlik
sorunu yaratacaktır. . Bu nedenle günümüzde İHA'ları da içeren karmaşık sistemlerin tasarımında ve
işletilmesinde, insanın bireysel yeteneklerini ve sınırlarını dikkate almak gerekir [12].
Bir İHA için YKİ tasarlanırken,YKİ tipini seçmeli, ölçüm cihazlarını, operatör sayısını seçmeli ve
ardından bazı hesaplamalar ve analizler yapılmalıdır. Genel olarak, YKİ tasarımı için birincil kriterler
şunlardır:(1) üretim teknolojisi, (2) sistem validasyonu, (3) görev, (4) hava durumu, (5) güvenilirlik,(6)
yaşam döngüsü maliyeti, (7) İHA konfigürasyonu, (8) insan faktörleri, (9) sürdürülebilirlik, (10)
dayanıklılık, (11) iletişim sistemi, (12) ağırlık ve (13) kontrol seviyesi [13].
Genel olarak tasarım süreci, net bir çizgi oluşturmak için bir ödün verme çalışmasıyla başlar. Maliyet,
performans (yani doğruluk) gereklilikleri arasında devam eder ve optimizasyon ile sona erer [14].
Tasarımcı iki öğe hakkında karar vermelidir. YKİ tipini ve ya istenen ekipmanı seçmelidir. Sonrasında
hesaplama sürecini yürütürken, tasarım gerekliliklerinin sağlandığından emin olmak için kontrol
edilmelidir. Tasarım sürecinin çok önemli bir parçası YKİ’yi hava aracına entegre etmektir. Eğer tüm
Page 406
SEMIT 2023
ekipman/ekranlar seçilir/tedarik edilir, entegrasyon süreci devam etmelidir. Aksilik ve hatanın

minimum düzeyde olduğundan emin olmak için operatörlerin konfor düzeyi ve etkinliği (ergonomik
göstergeler) sırasıyla belirlenmelidir.
Çoğu YKİ tasarımının, İHA uçuş bilgilerinin yanı sıra yeterli düzeyde operasyon alanı bilgisi sağladığı
bulunmuştur. Operatör reaksiyon hızı ve performansı da çok önemlidir. Tehditlerin hızlı bir şekilde
tanımlanmasını sağlarken, savaş alanı durumlarının iyileştirilmiş görünürlüğü, gelişmiş grafik
teknikleri kullanılarak gerçekleştirilmiştir [15].
Tehdidin büyüklüğünün erken tespiti ve tanımlanması, hızlı karar alınmasını sağlar. İnsan arayüzü, hızlı
tepki ve gerekli bilgilerin geniş sunumu için dokunmatik ekran ve çoklu ekran sistemine dönüşmüştür.
Bir görüntü ekranının konfigürasyonu açısından operatörün bireysel rahatlığı, özelleştirme ile
gerçekleştirilmiştir [16].
Ekranın boyutu ile birlikte ekran sayısı da artma eğilimindedir. Operatörün yorgunluğunu azaltmak için
ekranı yukarı ve aşağı yerleştirme eğilimi vardır. Ayrıca, savaş alanı durumunu ayrıntılı olarak ifade
etmek için üç boyutlu bir taktik harita kullanır. Farklı bilgiler için sembol renklere göre
sınıflandırılmıştır. Pilotların iş yükünü azaltmak için sistemi çalıştıracak ek operasyonel personel
mevcuttur [17].
İnsansız sistemler için yer kontrol istasyonu geliştirmek için, tasarım aşamalarından yerleşim, bilgi
bileşeni, temsil şeması ve insan operasyon yöntemleri gibi birçok faktörün dikkate alınması gerekir. Bu
kadar çok faktörün çok zor olabileceğini göz önünde bulundurarak, dünyanın dört bir yanındaki büyük
İHA istasyonlarından tasarım faktörlerinin derinlemesine bir araştırması yapılmaktadır [18]. Tasarım
özelliklerini ve özelliklerini analiz edilmelidir. Söz konusu parametereler, ekran görüntüleme bilgileri,
görüntüleme yöntemi, ekran sayısı, taktik durum bilgisi, taktik harita ekipmanının çalışması, operatör
sayısı, operatörün konumu, düzeni ve operasyon ortamını içerir.
3. Sonuç ve Tartişma
İnsansız hava araçlarının birçok alanda kullanılmaya başlanmasıyla, son yıllarda insan faktörünün bu
alandaki rolü hızla önem kazanmaktadır. İnsan faktörlerinin insansız hava aracı operasyonları
üzerindeki etkisini bilmek, güvenlik açısından giderek daha kritik hale gelmektedir. Bu çalışmada,
insansız hava araçlarında insan faktörü konularına ilişkin literatüre sistematik bir bakış sunulmaktadır.
Çalışmalar, ele aldıkları konulara göre gruplara ve alt gruplara ayrılmaktadır. Bu sayede ne tür
çalışmalar yapıldığı, hangi konuların incelendiği ve gelecekte ne tür çalışmalar yapılabileceğinin
anlaşılması amaçlanmaktadır.
İnsansız hava sistemleri, ergonomik problemlere sahiptir, İHA operatörünün özellikleri ve İHA'nın
teknik özellikleri karmaşık sistemlerdir. Operatörlerin tasarımında, çalıştırılmasında ve eğitiminde
insan faktörü önemli yer almaktadır. Ayrıca ergonomiyi zorunlu bir disiplin olarak uygulamaya sokmak
acil bir ihtiyaçtır. İnsan-makine etkileşiminin karmaşık süreçlerinin anlaşılmasıdır; riskleri tahmin etme
yeteneği ve bu riskleri önceden göz önünde bulundurarak sistemlerin geliştirilmesi insan hatalarının
azaltılmasında büyük pay sahibi olacaktır.
İnsansız hava araçlarında insan hataları faktörleri içinde İHA tasarımı ve Ergonomi başlığı içinde yer
kontrol istasyonları gözden geçirilmiştir. Yapılan literatür araştırması sonunda insan faktörü açısından
önemli bir yeri olan ergonomi unsuru üzerinden pilotun yer aldığı ve uzun saatler geçirdiği yer kontrol
istasyonunda sistem tasarımı ve yerleşimi konusunda önerilerde bulunulmuştur.
Pilotun duruşu, koltuğu, kullandığı klavye mause, joystick vb. tüm aparatlara erişimi uzun saatler
çalışıldığı düşünüldüğünde eklem ve kol kaslarını yormayacak şekilde tasarlanmalıdır. Ekran açısı ve
yüksekliği boyun sinirlerine bası yapmayacak şekilde ayarlanmalıdır. Ayakların posziyonu ile kapalı
alanda termol konfor şartları gözden geçirilmelidir. Arazi şartlarında kabin içi sıcaklık kontrollü olmalı
ve oda sıcaklığı şartları sağlanmalıdır.
Page 407
SEMIT 2023
Çalışma, gelecekte yapılacak tasarım ve ürün geliştirme çalışmalarına ön bir gereklilik sağlaması
açısından önemlidir. İnsansız hava araçlarının gelecekte çok daha sık kullanılacağı düşünüldüğünde yer
kontrol istasyonlarının tasarımı ve ergonomisi üzerine çalışılması gereken konulardandır.
4. Kaynaklar
[1] A. C. Watts, V. G. Ambrosia, and E. A. Hinkley, “Unmanned Aircraft Systems in Remote Sensing
and Scientific Research: Classification and Considerations of Use,” Remote. Sens., 2012, doi:
10.3390/rs4061671
[2] X. Zhang, G. Jia, and Z. Chen, “The Literature Review of Human Factors Research on Unmanned
Aerial Vehicle –What Chinese Researcher Need to Do Next?,” in Cross-Cultural Design. Methods,
Tools, and Users, Cham, 2018, pp. 375–384, doi: 10.1007/978-3-319-92141-9_29.
[3] “List of unmanned aerial vehicle applications,” Wikipedia. Jun. 29, 2020, Accessed: Aug. 05, 2020.
[Online]. Available: https://en.wikipedia.org/w/index.php?title=L
ist_of_unmanned_aerial_vehicle_application s&oldid=965129299.
[4] “UAS by the Numbers.” https://www.faa.gov/uas/resources/by_the_n umbers/ (accessed Aug. 05,
2020). [6] “Military Drones Market Size, Growth, Trend and Forecast to 2025 | MarketsandMarkets.”
https://www.marketsandmarkets.com/Marke t-Reports/military-drone-market221577711.html
(accessed Aug. 05, 2020).
[5] “Military Drones Market Size, Growth, Trend and Forecast to 2025 | MarketsandMarkets.”
https://www.marketsandmarkets.com/Marke t-Reports/military-drone-market221577711.html
(accessed Aug. 05, 2020).
[6] “Unmanned Aerial Vehicle Market, UAV Size, Share, system and Industry Analysis and Market
Forecast to 2024 | MarketsandMarketsTM.” https://www.marketsandmarkets.com/Marke t-
Reports/unmanned-aerial-vehicles-uavmarket-662.html (accessed Aug. 05, 2020).
[7} Ground control station. Benchmark Study, 2015, available at
http://dronecode.github.io/UXDesign/Research/Benchmark/GCSBenchmark.pd
Hong, J., Kim, S. and Kwon, Y. (James) (2015) Analysis of Modern Design Approach for Anti-Air
Radar Screen. Technical Report of Ajou University, Suwon, South Korea.
[8] Jongsoon, I., et al. (2006) Study of Aircraft Landing Loads and Ground Handing Loads.
Proceedings of The Korean Society for Aeronautical and Space Sciences, Fall Conference, 189-197.
[9] Jang, J., et al. (2011) Development of GCS for Small UAV with Automatic Takeoff and Landing
Mode. Proceedings of The Korean Society for Aeronautical and Space Sciences, Fall Conference, 1637-
1641.
[10] Backes, P.G., et al. (1995) Ground Control Station Development for the Dexterous Orbital
Servicing System. Proceedings of the IEEE International Conference on Robotics and Au-tomation, 21-
27 May 1995, 1259-1264. http://dx.doi.org/10.1109/robot.1995.525454
[11] Aydemir, M.E., et al. (2013) Ground Station Design Procedures for CANSAT. The 6th In-
ternational Conference on Recent Advances in Space Technologies, 909-912.
[12] Olmos, O., et al. (2014) Tactical Displays for Combat Awareness: An Examination of Di-
mensionality and Frame of Reference Concepts and the Application of Cognitive Engi-neering. The
International Journal of Aviation Psychology, 10, 247-271.
http://dx.doi.org/10.1207/S15327108IJAP1003_03
[13] Hobbs, A., & Lyall, B. (2016). Human factors guidelines for unmanned aircraft systems.
Ergonomics in Design, 24, 23-28. DOI:10.1177/1064804616640632
Page 408
SEMIT 2023
[14] Pedersen, H. K., Cooke, N. J., Pringle, H. L., & Connor, O. (2006). UAV human factors: Operator
perspectives. In N. J Cooke, H. L. Pringle, H. K. Pedersen, & O. Connor (Eds.), Human factors of
remotely operated vehicles. San Diego, CA: Elsevier, 21–33.
[15].Shneiderman, B., & Plaisant, C. (2005). Designing the user interface: Strategies for effective
human–computer interaction. Boston, MA: Pearson.
[16] Sanders, M. S., & McCormick, E. J. (1993). Human factors in engineering and design (7th ed.).
New York, NY: McGraw-Hill
[17] Mygal, G., Protasenko, O. (2020). Human resources are a factor in applying of man-machine
systems safety. Municipal Economy of Cities 6(159), 139-146. DOI:10.33042/2522-1809-2020-6-
159-139-146
[18] Parasuraman, R., Mehta, R. (2013). Neuroergonomics: a review of applications to physical and
cognitive work. Frontiers in Human Neuroscience, 7(889), 1-10. https://doi.org/10.3389/
fnhum.2013.00889.
Page 409
SEMIT 2023
The Quality Assessment of a Company in Defence Industry: A Case Study Approach
Vildan Tezcan
Harika Ergün
Enes Öztürk
Güler Zengin
Deniz Efendioğlu
Summary
In order to address the issue of sub-component loss faced by a company in the defense industry, this
article explores SPC procedures. The missing components were found to have passed quality checks,
and records were retained. To address the issue, methods including root cause analysis, pareto analysis,
and the PDCA cycle are recommended. The study also contains information about the organizational
structure of the business and the workflow. The loss of time due to the problem is emphasized. The
paper suggests employing quality assurance methods to identify the best option.
Keywords: PDCA, SPC, Lean Production, 5 Why, Pareto Analysis
Introduction
The company studied has achieved significant success in the defense industry. The company's activities
include manufacturing, R&D, machine design and production. In order to better respond to the demands
of its customers and meet their needs, it strengthens both the design and production team and acts with
a broad perspective. The company manufactures and designs in accordance with the quality
management system, thus offering high quality products to customers. By implementing Quality
Assurance techniques, the company can ensure that all sub-components are thoroughly inspected and
accounted for before they are incorporated into the production process. This will help identify any
missing parts early on and allow for timely replacements or reordering, minimizing disruptions and
delays. Additionally, the company can establish a strong relationship with reliable suppliers to ensure
consistent and timely delivery of sub-components, further reducing the risk of losses. The acquisition
of supporting components kicks off the procedure. The business enters into a contract with the supplier
businesses to provide the sub-components. After that, incoming parts are examined and noted in the
area designated for quality controls. The parts are loaded onto the shelf system, stocked, and lowered
to the lower floor following the completion of the quality checks. The assembly process, which
comprises all the procedures required to create the products, is carried out on the lower floor. When a
component is missing throughout this procedure, the suppliers are contacted, and the process is then
continued by making the necessary adjustments. It adheres strictly to quality control and inspection
mechanisms at every level of the process in order to provide its clients with goods that are high-quality
and dependable.
In this area, different studies have been made about the methodologies studied for this article. Bamber
& Dale (2010) are reported the findings of a research study on the application of lean production
methodologies to a typical aerospace manufacturing organization. Forza (1996) proposes a framework
(derived from a review of the current literature) for investigating the connections between work
organization and lean production processes. Sydnes et. al. (2022) indicate that lean practices such as
lean meetings, the use of lean tools, and a lean production-inspired organizational culture enhance the
development of organizational capacities that foster anticipation, coping, and adaptation. Kidane &
Sharma (2017) studied the cultural difference of Ethiopian and Japanese with the implementation of
Lean Manufacturing. Siju et. al. (2022) highlights options for process improvement that will lead to a
Page 410
SEMIT 2023
reduction in non-conformances per engine. Widodo, A., & Soediantono (2022) studied DMAIC in the
defence industry, while Brata & Soediantono (2022) studied TQM with literature review.
Methodology & Results

Process Components
Supply Process
• Contracted Suppliers: Supplier companies from which sub-components are supplied.
• Orders: Orders for sub-components to start the procurement process
• Goods Acceptance: The stage where incoming parts are checked and recorded.
• Stocking and Storage: The parts that pass quality control are placed on the shelf system and
stocked or lowered to the lower floor.
Assembly Process
• Sub-Components: Parts used in the manufacturing phase are assembled.
• Assembly Stations: Specialized workstations are used to complete products.
• Quality Control: Inspecting the conformity of the products to the quality standards during the
assembly phase.
• Communication: Providing necessary corrections through communication with suppliers
Quality Control
• Control Equipment: Tools and devices used to measure and control the quality characteristics
of products.
• Quality Control Processes: Steps taken to check the conformity of parts, assemblies, and
products to quality standards.
• Recording and documentation: The process by which quality control results are recorded and
documented.
Figure-1: The flow of the process

With the information received from the organization, statistical methods (SPC) are employed to monitor
and regulate the performance of the process. To establish if the process is under statistical control, this
model analyzes statistical data to identify deviations in the process.
Page 411
SEMIT 2023
Pareto Analysis
Upon analyzing the Pareto chart, we can observe the following:
• The top three categories, namely December, June, and October, contribute to approximately
one-third (31.85%) of the total lost parts count. These categories appear to have the highest
impact on the overall issue.
• The top five categories, including September and May, account for nearly half (50.77%) of the
cumulative lost parts. These categories also hold significant importance in addressing the
problem.
• The cumulative percentage steadily increases as we move down the chart, indicating a
diminishing contribution from each subsequent category.
• Categories such as August, March, and November make up a considerable portion of the
cumulative total but have relatively lower significance compared to the top-ranked categories.
• Categories like April, January, and July have a relatively smaller contribution to the cumulative
total, suggesting they have less impact on the overall issue.
Based on this analysis, it is clear that focusing on resolving issues related to December, June and
October would yield the most significant improvements. Allocating resources and efforts toward these
categories would have the greatest impact on reducing the overall count of lost parts.
Table-1: The Average Number of Missing Parts
Page 412
SEMIT 2023
Figure-2: Pareto Chart of the Process
Regression Analysis
Based on the provided OLS regression results, here's an evaluation:
• R-squared (R2) value: 0.043. The R-squared value represents the proportion of the variance in
the dependent variable (y) that can be explained by the independent variable (x1) in the model.
In this case, the R-squared value is very low (0.043), indicating that only a small percentage of
the variation in the average lost part count can be explained by the month.
• Adjusted R-squared value: -0.083. The adjusted R-squared value accounts for the number of
predictors in the model and adjusts the R-squared
value accordingly. A negative adjusted R-squared value suggests that the model is not well-
suited for the data.
• F-statistic value: 0.3412. The F-statistic tests the overall significance of the regression model.
In this case, the F-statistic is low, and the associated p-value (0.575) is larger than the typical
significance level of 0.05. This indicates that the model is not statistically significant overall.
• Coefficient estimates: Constant (const): The estimated coefficient is 2.4083, indicating the
average lost part count when all other variables are held constant. It is statistically significant
(p-value < 0.001).
• Independent variable (x1): The estimated coefficient is -0.0233, suggesting a slight negative
relationship between the month and the average lost part count. However, it is not statistically
significant (p-value = 0.575).
In conclusion, based on these results, the regression model does not provide strong evidence of a
significant relationship between the month and the average lost part count. The low R-squared and
adjusted R-squared values indicate that the model explains only a small amount of the variability in the
data. Additionally, the non-significant p-value for the independent variable suggests that the month may
not be a meaningful predictor of the average lost part count in this context. Further analysis or the
inclusion of additional variables may be necessary to better understand the factors influencing the lost
part count.
Page 413
SEMIT 2023
Figure-3: Pareto Chart of the Process

5 Why?
The 5W rule is a problem identification method that involves focusing on five different "whys" to
understand problems and their causes. Let's apply the 5W rule on this problem:
• Why (Question: Why do parts disappear?)
Problem: Parts are lost after quality control.
Cause: An error or omission may have occurred between the quality control process and the
scrapping of parts.
Potential Causes: Parts were not placed on the shelf correctly or were dropped, parts were
placed in the wrong places because they are not recorded or labeled. Employees accidentally
moved or lost parts.
• Why? (Question: Why are parts not racking correctly?)
Problem: Parts are not racking properly after quality control.
Cause: There may have been a deficiency or error in the rack placement process
Potential Causes: Workers not receiving adequate training or instruction Insufficient shelf
arrangement or inadequate labeling system. Lack of attention due to time pressure or workload.
• Why (Question: Why did the workers not receive adequate training or instruction?)
Problem: Workers do not receive adequate training or instruction to properly rack parts.
Cause: Lack of training or instruction prevents employees from doing business properly.
Potential Causes: Lack or inadequacy of the training program. Insufficient direction or
mentoring of new employees. Educational materials are not descriptive or understandable.
• Why (Question: Why are the training program or instructions inadequate?)
Problem: The training program or instructions are not effective enough to get employees to do
the job properly.
Cause: There is a problem where the training program or instructions are missing or insufficient.
Potential Causes: The training program has not been updated or revised. The instructions are
complex or difficult to understand. Unidentified or overlooked training needs.
Page 414
SEMIT 2023
• Why (Question: Why aren't training needs identified?)

Problem: Inadequate training programs or instructions are provided without identifying training
needs.
Cause: There is a gap in the process of identifying training needs.
Potential Causes: The lack of a feedback mechanism to identify training needs The lack of
consideration for employee’s opinions or suggestions.
Fishbone Diagram
A fishbone diagram, also known as a cause-and-effect diagram, gets its name from its shape, which
resembles the skeleton of a fish. The problem or effect being analyzed is typically represented as the
"head" of the fish, while the potential causes are represented as the "bones" or branches extending from
the main line. The diagram provides a structured way to brainstorm and categorize potential causes,
leading to a better understanding of the problem and its underlying factors. Overall, the Fishbone
diagram serves as a powerful visual tool to explore the multiple causes of a problem and supports the
development of effective solutions by addressing the underlying issues.
Figure-4: Fishbone Diagram of Process
Lean Methodology
The core principle of lean methodology is to maximize value for the customer while minimizing waste.
It focuses on eliminating activities that do not add value to the final product or service, thereby
increasing efficiency, reducing costs, and improving overall quality. Lean methodology promotes a
culture of continuous improvement and empowers employees at all levels to identify and eliminate
waste in their work processes.
Cause: Employee Carelessness
Creating standardized business procedures: Unusual corporate procedures frequently lead to
carelessness. Employees can clearly define their tasks thanks to the standardization of business
procedures. Expected conduct and business practices can be made clear by creating standard operating
procedures and communicating them to staff members.
Visual Guidance: Visual orientation is the use of visual materials that clearly demonstrate instructions,
workflow, and correct behavior in the workplace. This can help employees avoid carelessness. For
example, visual posters or charts showing the workflow can be used for instructions.
Error Sharing and Learning: It is important to encourage a culture of error sharing and learning among
users so that careless mistakes are not repeated. Error analysis should be performed to understand the
Page 415
SEMIT 2023
causes of errors, identify shaping opportunities, and prevent the recurrence of similar errors. In this
process, working groups should be encouraged to share and offer solutions.
Cause: Insufficient Training of Employees
Value Stream Map: A value stream map helps you create a map that shows the steps from the beginning
to the end of business processes and valuable and worthless activities in the process. Using this map, it
is possible to identify what skills employees need and where they are lacking. Targeted training
programs can then be planned to address these shortcomings.
Standard Training Process: A standard training process is a guideline for correctly performing a
particular job task or process. This guide shows employees step-by-step how to act on their
shortcomings. By establishing a standardized training process, standardization of work processes and
job skills can be achieved, and the training process of new employees can be facilitated.
Visual Orientation in the Work Environment: Visual orientation is the use of visual materials that show
instructions and processes in the workplace. These materials help employees clarify how they should
act on what they lack. Visual posters, workflow charts, instructions, warning signs, etc. Employees can
be provided with information quickly.
Training Materials and Videos: By creating training materials and videos based on lean principles,
employees can be provided with training on the subjects they lack. These materials can be used to teach
new skills to employees, prevent mistakes, and increase productivity. For example, training videos,
animations, or interactive training modules can be prepared.
Error Analysis and Continuous Improvement: It is important to perform error analysis in order to
identify the missing points among the employees and make improvements. Continuous improvement
methods based on error analysis can be used to understand the causes of errors, identify training needs,
and address problems in business processes.
Cause: Process Management
Identifying Non-Valuable Activities: Identify non-valuable activities (muda) in your process. This can
include sources of waste such as waiting times, excess inventory, unnecessary movements, and faulty
production. Identifying these activities allows you to find opportunities for improvement in your
process.
Elimination of Waste: Steps can be taken to eliminate or reduce unvalued activities. Lean tools such as
5S, accelerated teamwork, and error prevention methods can be used to reduce waste. These steps help
reduce errors and waste while increasing the efficiency of the process.
Ensuring continuous flow: Focusing on optimizing the process flow and providing continuous flow.
Continuous flow can be supported by reducing types of waste such as overproduction, overprocessing,
and overhandling. Your process can be arranged to provide the required product or service at the right
time, in the right quantity, and with the right quality.
Cause: Goods Acceptance Stage
Identification of Valuable Works: Valuable works in the goods acceptance process are determined.
Valuable jobs are activities that add value to the customer and accelerate the flow of the process. For
example, activities such as accurate and complete material control, fast record keeping, and material
placement are valuable work. Identifying these jobs provides a focal point for optimizing the process.
Proximity of Resources: During the goods receipt process, it is important to store materials as close to
the workplace as possible. This helps reduce movement times and errors. Storing materials in a properly
organized and labeled warehouse reduces search times and speeds up the goods receipt process.
Visual Guidance: Errors and carelessness can be reduced by using visual directions in the goods receipt
process. Visual orientation includes posters, work flow charts, warning signs, etc. that clearly show
instructions, workflow, and correct behavior in the workplace includes using this way, employees make
fewer mistakes when following the right steps.
Page 416
SEMIT 2023
Cause: Considering Incoming Products as Garbage Incorrectly

Collaboration and Communication: In the goods acceptance process, cooperation and effective
communication can be established with suppliers and other stakeholders. Agreements can be made with
suppliers on standards, labeling requirements, and accurate documentation. In addition, errors can be
detected and corrected in a timely manner through cooperation and communication channels.
Feedback and continuous improvement: Feedback from employees can be collected and included in the
continuous improvement process. Root cause analysis can be done to understand the causes of
accidentally discarded products. Then, remediation measures and controls can be developed to prevent
errors in the process. The process can be continuously monitored, feedback can be evaluated, and work
can be done for continuous improvement.
5S Application: By using the 5S (Seiri, Seiton, Seiso, Seiketsu, and Shitsuke) principles, order and
cleanliness can be achieved.
Cause: Inadequate Labeling System
Value Stream Analysis: First, a value stream analysis can be performed to identify disruptions in the
labeling and barcoding processes and causes of disposal. Value stream maps can be created to
understand how the process works and which steps have problems.
Standard Business Operations: Labeling and barcoding processes can be standardized. Standard
business procedures can be established, and these standards can be disclosed to employees. The labeling
and barcoding procedures that should be used in each step can be determined. This will reduce
inconsistencies in the process and prevent errors.
Technological Innovations: Existing labeling and barcoding systems can be reviewed, and technological
innovations can be searched for. By using solutions such as advanced barcode readers, automatic
labeling systems, or data management software, the process can be made faster and error-free.
Technological innovations can increase efficiency and reduce errors.
Plan-Do-Check-Act
The PDCA method offers a systematic approach to solving the missing parts problem of the company.
By following the steps of planning, implementing, checking, and correcting, you can identify the causes
of the problem, take action, and continuously improve the process.
Plan: Do a detailed analysis to fully understand the problem. Identify the causes of missing parts and
set goals for solving this problem. Review the inventory management of sub-parts and update as needed.
Ensure that critical parts are in stock in sufficient quantities by reviewing the policy for spare parts.
Identify problematic processes and mechanisms that lead to missing parts. For example, there may be
deficiencies in areas such as keeping records in the quality control process and tracking parts. Plan the
necessary measures and improvement steps. For example, measures such as establishing an asset
tracking system, strengthening inventory management processes, and arranging storage areas can be
taken.
Do: Implement the planned measures. Complete the parts tracking system setup and update inventory
management processes. Implement necessary controls in processes to reduce the risk of loss of parts.
Strengthen the quality control process and track part movement more systematically. Raise awareness
among employees about missing parts by organizing training and awareness programs. Clearly explain
to employees their responsibilities and procedures, especially in areas where part-time movement
occurs.
Check: Monitor and evaluate the effectiveness of the measures implemented. Regularly check the parts
tracking system and update inventory management processes.
Monitor the number of missing pieces and assess the decreasing trend. If losses persist, analyze problem
areas and processes in more detail. Collect customer feedback and measure customer satisfaction.
Follow up on customer complaints due to missing parts and contact customer relations as needed.
Page 417
SEMIT 2023
Act: Problems detected during the checking phase. If the measures are not sufficient or the expected
results are not being achieved, return to the planning stage and develop new measures.
Evaluate feedback and experiences for continuous improvement. Optimize processes by considering
suggestions and ideas from employees. Maintain measures and continually review the process so that
the problem of missing parts does not recur. Continuing improvement efforts for a better inventory
management and parts tracking system.
Conclusion
Part loss problem is a major challenge faced by the company. Various approaches and techniques have
been used to solve this problem and improve process performance. First of all, the process components
method was applied to analyze the components in the production process. This analysis is important to
identify the root causes of the problem and identify areas to focus on.
In addition, the continuity and variability of part loss were investigated using the control chart and
regression model. Control chart analysis helped identify the causes of the problem by identifying
statistical variability in the process.
A regression model was used to determine the factors affecting the part loss. The 5 Whys method and
root cause analysis were used to identify the root causes of the problem. Thanks to this approach, the
real causes of the problem were reached by asking "why" questions consecutively. Identifying root
causes is an important step towards producing more effective solutions.
Lean, PDCA, and Six Sigma methods are also approaches used to solve the problem. By applying Lean
principles, waste is reduced, process efficiency is increased, and errors are minimized. Continuous
improvement was achieved using the PDCA cycle, and errors were detected and reduced with the Six
Sigma methodology. These methods offered a systematic and efficient approach to solving the problem.
As a result, the part loss problem has been tried to be solved by a combination of the above approaches.
This analysis and solution process includes important steps for improving process performance and
raising quality standards.
References
Bamber, L., & Dale, B. G. (2000). Lean production: a study of application in a traditional manufacturing
environment. Production planning & control, 11(3), 291-298.
Brata, J., & Soediantono, D. (2022). Total quality manufacturing (TQM) and recommendations for its
application in the defense industry: A literature review. International Journal of Social and Management
Studies, 3(3), 50-62.
Forza, C. (1996). Work organization in lean production and traditional plants: what are the
differences?. International Journal of Operations & Production Management, 16(2), 42-62.
Kidane, T. T., & Sharma, R. R. K. Cultural considerations when implementing lean Manufacturing:
Ethiopia Perspective.
Siju, V., Ram, C., & Krishnan, R. R. Pareto Analysis of Non-Conformances Observed During the
Fabrication of Rocket Engines and Suggestions to Prevent their Occurrence.
Sydnes, M., Gausdal, A. H., & Åssveen, I. D. (2022). Lean Production: One path to Organizational
Resilience?. Beta, 36(1), 1-22.
Widodo, A., & Soediantono, D. (2022). Benefits of the six sigma method (dmaic) and implementation
suggestion in the defense industry: A literature review. International Journal of Social and Management
Studies, 3(3), 1-12.
Page 418
SEMIT 2023
Application of Best Project Management Practices: Case Study Application in Social Projects
Verónica Moreno Ponce de León1[0000-0002-8072-3991] and Agatha Clarice da Silva-Ovando1[0000-0002-
5779-4003]
1
veronicamoreno1@upb.edu
Abstract
This article pretends to highlight the benefits of applying the best project management projects,
presenting the experience of Fundación Educación para el Desarrollo FAUTAPO. This foundation
works with Bolivia's social projects and shows how those contribute to improving execution times,
achievement of objectives, and knowledge generation. The paper carries out a bibliographical and
documentary review and a survey coordination meeting with agents involved in the projects to
complement the information gathered from historical data. The study uses the descriptive method to
illustrate the experience of implementing best project management practices to a social project executed
over three years. Adopting the best project management practices in a social context improves objective
achievement and compliance with deadlines. Project communication management and project
stakeholder management are vital aspects of highly complex projects. The transference of knowledge
and lessons learned has proven essential to reduce the learning curve. There was no access to budget
information, so achieving objectives and deadlines were taken as analysis variables. The paper enriches
the documentation of experiences of best practices implementation in social projects to promote their
use. It contributes to the review and documentation of the experience gained through applying best
project management practices to a social project developed in a Latin American context.
Keywords: project management, best practices, project planning, social projects, project
accomplishment.
1 Introduction
We live in a materialized world thanks to projects, which is evident by looking at everything we have
at hand and considering their impact on all aspects of life, such as health, education, communication,
work, and entertainment. A project is a temporary endeavor to produce a unique product or result [17]
and requires adequate management to achieve its goals.
Literature shows that evaluating a project's success has a multivariate nature [2] [13]. On the one hand,
it measures compliance with the scope, time, and cost (iron triangle) [15]. Conversely, it evaluates the
project's contribution to fulfilling the organization's strategic objectives [7]. The first case is related to
the efficiency in the use of resources, and the second, is related to effectiveness [18]; proper project
management contributes mainly to the first case.
Good practices or best practices refers to those that contribute to the project's success [4]. Applying
practices, methods, techniques, and procedures allows the achievement of the best project management
[16], which are denoted as best practices in reference frameworks (Knowledge Guides) that are prepared
and updated by volunteers from well-known international organizations [1]. PMBOK® Guide is one of
these cases, which presents the best project management practices used in organizations of different
industries, sizes, and contexts [17], which must be selected and adapted according to the context of each
project due to its specificity [19].
Applying best project management practices reduces delays [5], increasing the probability of success.
Project Management Institute [11] found that organizations that apply them waste 28 times less money
than those that do not. More and more organizations worldwide added to these practices, positively
impacting statistics. From 2019 to 2021, the percentage of projects that managed to maintain and meet
their original objectives and business intentions increased from 68% to 73%; the percentage of those
that were completed within the budget went from 57% to 62%, and those that met the deadline rose
Page 419
SEMIT 2023
from 51% to 55%. On the other hand, there was a reduction in the percentage of projects that failed
(from 37% to 35%), those that presented corruption or scope slippage (from 37% to 34%), or finally,
the ones that had considerable failures (from 15% to 12%) [12].
Although since 2019, the performance of the projects has been improving, there is still concern that
almost 30% of the projects do not meet their original objectives, about 45% end after the deadline, and
roughly 35% failed and lost their financing; all this causes a loss of economic resources, affects the
population negatively and produces a loss of opportunities. The situation in Latin America is similar to
the one presented globally. The publication Pulse of Profession® Latin America [12] indicates that 29%
of the projects completed in the region did not meet the original objectives they were designed for, 37%
exceeded the original budget assigned, and 47% did not meet the deadline.
It is known that projects have a particular specificity that makes them unique, that best practices
generate a positive impact, and that, although these are embodied in the Knowledge Guides, they do
not establish which ones should be used in a specific context. This broad scope of possibilities means
that the research carried out to date aims to show the results obtained in a specific field or apply a group
of specific processes or areas of knowledge [1].
There is literature that shows the practices used and the results obtained in projects, mainly in the areas
of software development (IT), construction, and engineering [1] [14] [10]. However, little
documentation was found about experiences in social projects, education, or the manufacturing sector
[8]. Thus, in the existing literature, we can identify the best practices regarding the establishment of
baselines, procurement, requirements, changes, communication, as well as lessons learned, and
estimation of cost and time, among others [1, 8, 9, 15].
Given the importance of studying the impact of best practices in areas scarcely studied and the limited
literature for social projects, especially in the Latin American context, there is a real need to share the
experience of applying these practices in social projects. This study shows the best project management
practices, as established in the 6th Edition of the Guide to the Fundamentals of Project Management
(PMBOK® Guide), implemented in a social project and how this impacted meeting the deadlines. This
study will answer the following questions: Which best project management practices were applied in
the social project? What is the impact of implementing best practices in social projects over time and
scope?
2 Theoretical background
There are organizations and institutions in the world that generate their knowledge framework and
guides that support project management practices (e.g., the PMBOK® Guide, the APMBoK, and
PRINCE2 Guide) [13]. How a project is managed determines its success and accomplishment of its
objectives. The PMBOK® Guide is the most renowned worldwide and provides the basis for its use in
projects from all industries.
The PMBOK® Guide [17] presents different best project management practices in 49 processes
categorized into ten areas of knowledge: Integration, Scope, Schedule, Costs, Quality, Resources,
Communications, Risks, Procurement, and Stakeholders and five process groups: Initiation, Planning,
Execution, Monitoring and Controlling, and Closure. These are shown in Table 1.
Page 420
SEMIT 2023
Table 2. Processes, Process Groups, and Knowledge Areas
Groups of Processes
Knowledge Initiation Planning Execution Monitoring and Closure
areas Control
- Develop - Develop Project - Direct and -Monitor and - Close
Project Charter Management Plan Manage Project Control Project Project or
Management
Work Work Phase

Integration
- Manage Project - Perform

Project
Knowledge Integrated Change

Control
- Plan Scope - Validate Scope
Management
- Control Scope
Project Scope
- Collect
Management
Requirements
- Define Scope
- Create WBS
- Plan Schedule - Control
Management Schedule
- Define Activities
Project Schedule
- Sequence Activities
Management
- Estimate Activity
Durations
- Develop Schedule
- Plan Cost - Control Costs
Management
Project Cost
Management
- Estimate Costs
- Determine Budget
- Plan Quality - Manage - Control Quality
Project Quality
Management Quality
Management
- Plan Resource - Acquire - Control

Management Resources Resources
Management
- Estimate Activity - Develop Team

Resource
Project
Resources
- Manage Team
Page 421
SEMIT 2023
- Plan - Manage - Monitor

Communications Communications Communications Communications
Management Management
Project
- Plan Risk - Implement Risk - Monitor Risks

Management Responses
Project Risk Management
- Identify Risks
- Perform Qualitative
Risk Analysis
- Perform
Quantitative Risk
Analysis
- Plan Risk Responses
- Plan Procurement - Conduct - Control
Management Procurements Procurements
Management
Procurement
Project
- Identify - Plan Stakeholder - Manage - Monitor

Stakeholders Engagement Stakeholder Stakeholder
Management
Engagement Engagement
Stakeholder
Project
Source: Adapted from [17]

The Project Director must select and adapt the processes to be used according to the dynamics and
nature of the project. However, the choice of the processes and areas of knowledge that will be applied
depends on the team and the Project Director's time, experience, and knowledge [15], which affects the
probability of success of the projects and how those would generate value for the client. The study also
identified that out of the knowledge areas listed in the 5th edition of the PMBOK®, the ones that
organizations apply the most are integration, schedule, and scope, but the areas that can generate the
most significant impact on successful project management are schedule, scope, and risks.
Ibbs & Kwak [6] also identified that the communications knowledge area is the least used. However,
the sixth and seventh editions of the PMBOK® Guide [17, 19] highlight the importance of the
knowledge areas related to communications and stakeholders, revealing their high impact on achieving
appropriate requirements management for the interested parties, directly affecting the projects' success.
Literature shows experiences using certain best project management practices for specific areas, which
can produce positive results for those, but do not affect others, even if those are from related areas [5].
For example, Gurmu et al. [5] carried out research that identified the practices that generated the most
significant positive effect on the productivity of projects in the construction industry; these were related
to the integration of the schedule with the work, planning of material procurement, distribution of
materials and machinery, management of subcontractors, schedule management, project launch, project
execution plan, and innovation and emerging technologies.
Page 422
SEMIT 2023
Following this line of research in the construction field, [14] established that setting the scope baseline
early and change management positively impact schedule and cost. As a negative aspect, they identified
that the construction industry is worst at applying lessons learned related to change management in
projects.
In another area, Meier [9] addresses the experience of applying best practices in projects of the
American Government, highlighting the importance of setting reasonable margins when carrying out
time and cost estimates so that it is possible to obtain more realistic baselines before submitting a
proposal to the Congress to be considered for approval. This practice increased the number of projects
executed and delivered on time and within budget.
Related to software development, Nelson [10] retrospectively reviewed 99 projects and established
project management practices that could have prevented problems and breaches. The leading causes
identified were grouped into four categories: people, processes, products, and technology, out of which
a set of best practices were established. This set shows that applying the lessons learned as best practices
will reduce the probability of making the same mistakes and allow us to apply all the positive aspects
identified.
In the field of manufacturing, Massa et al. [8] showed how implementing best project management
practices positively impacted projects, reducing the time it took to launch new products in Brazil's
highly competitive fashion market. They allocated these best practices into three groups: coordination,
leadership, and process management, presented in Table 2.
Table 3. Best practices identified by Massa et al. [8] in the field of fashion.
Coordination Leadership Process Management
- Establish the best - Develop the staff's - Adhere to the schedule and
communication. adaptability skills and budget.
- Apply effective emotional intelligence. - Use target costing in
communication channels. - Create a collaborative and product design.
- Achieve team cohesion. friendly work environment. - Apply cost aggregation
- Hold regular meetings and - Manage change processes techniques.
review project progress. effectively. - Define and describe the
- Hold follow-up meetings - Perform proper resource scope clearly: product
with the staff. management. characteristics.
- Generate work integration - Establish a climate of trust. - Select and train assertive
between departments. - Use analytical thinking. teams.
- Develop team spirit. - Have a leader as a problem - Focus on core
- Use project management solver, execution competencies.
software. accelerator, and work
- Apply idea-generation facilitator.
techniques. - Have and maintain the best
image of the team's leader.
Source: Adapted from [8]
As evidenced in the previous paragraphs, there is a diversity of best practices translated into tools,
techniques, and practices used in different industries or areas. There are some coincidences, but there
are also those specific to some of them, making it evident that adaptability is necessary and valuable. A
general description of the best practices used the most is highlighted in the 6th Edition of the PMBOK®
Guide.
The expert consultation is carried out in the planning stage of a project whenever there is any difficulty
or the opportunity to draw upon the experience and knowledge of people involved in similar projects
or leaders in their area. Likewise, performing a document review of the regulatory framework and
lessons learned from similar projects is crucial to take advantage of the best practices applied and avoid
repeating mistakes.
Page 423
SEMIT 2023
A project is developed in a particular context, so the project manager must identify and consider the
environmental factors and process assets that encompass it, along with those within the organization
and the team, throughout the project lifecycle. This consideration avoids administrative problems within
the organization, improves team integration, improves the work environment, and generates the best
relationship with external actors in the project.
The project charter is an element through which the sponsor gives the project legitimacy, helping clarify
its objectives and narrow its scope. Socializing it with the project's director and the team is helpful to
make it understandable; it helps size the project and is an essential input when planning the project.
During scope planning, the team's participation is vital to performing the requirements gathering,
developing the scope definition, and creating the WBS (Work Breakdown Structure). A point to
consider in this process is incorporating elements of project management, such as reports, meetings,
and the management of lessons learned. On the other hand, it must be structured in such a way that the
execution of the project starts to generate value at an early stage as soon as the first deliverables are
completed.
The team's participation will also be crucial when identifying the activities necessary to materialize
deliverables, their relationship, and the estimation of activity durations in identifying, measuring, and
valuing resources. This way, both the schedule and budget are more realistic.
When discussing resources, it is essential to consider their type, nature, and characterization. Managing
human resources is critical, so it is necessary to identify the skills and knowledge required for the project
and the requirements for the team's development, along with considering their availability throughout
the project lifecycle. Additionally, it is crucial to create an environment of cooperation and trust so that
they can perform their job adequately.
The involvement of stakeholders from the beginning of the project is of vital importance. Early and
constant identification will help to know their position regarding the project and adopt the best strategy
to get them to the point at which the project requires them. Stakeholder engagement management needs
to be accompanied and aligned with communications management. Establishing the communications
plan, which incorporates the primary stakeholders' communication requirements, will help establish the
best relationship and avoid misunderstandings and conflicts with the proper fluency.
When facing a volatile, complex, ambiguous, and uncertain world, constant risk identification and
evaluation is essential; besides team members, key stakeholders should be alert to changes in the
environment or within the project that may affect it positively or negatively. The risk response plan
requires constant updating and establishing those responsible for monitoring its evolution to give a
timely response at the time of its activation.
Throughout the project's lifecycle and mainly during execution, there will be requests to change
planning as a response to changes in the environment, corrections to the planning, or changes related to
project specifications. Therefore, there must be appropriate change control before making any decision;
this way, both the sponsor and the main stakeholders will know the possible impact of such changes.
The execution of the project generates knowledge that is useful for the planning and execution of this
or future projects. The team must get used to analyzing and recording, as lessons learned, the practices
that did or did not produce the expected results. This knowledge management is incredibly beneficial
when undertaking new projects and if team members change.
The literature identified and reviewed focuses on best management practices of projects applied mainly
to the construction and software development industries, showing best practices particular to certain
areas of knowledge or a specific aspect. This fact illustrates the need to continue exploring the
application of these in other sectors. To reduce this gap, this article presents a case that will contribute
to identifying best practices that were applied in a social project.
Page 424
SEMIT 2023
3 Research approach and context

This research is presented as a case study since there is little control over the events described.
Additionally, this study intends to show the application of best management practices in a specific social
project implemented by the Fundación Educación para el Desarrollo FAUTAPO in Cochabamba,
Bolivia, for the lapse of three years. This research used a qualitative descriptive methodology for the
point of view. Figure 1 outlines the methodology used in this study.
Fig. 1. Methodology applied.

The research began with a literature review to identify experiences in implementing best project
management practices in different areas. It was possible to find information, especially in construction,
engineering, and information technologies. A search was carried out for scientific articles on best
project management practices in the Project Management Journal database and the EBSCO platform
research database, to which the researchers had full access. Also, the 6th Edition of the PMBOK®
Guide was reviewed to examine its content. All this information allowed us to establish the ideal
framework in which a project should be executed and also to identify the information necessary to
structure a survey and establish guidelines for the evaluation meeting to be held with the Regional and
Local Coordinators of the FAUTAPO foundation, the team members, and one of the technicians of the
Gobierno Autónomo Municipal de Cochabamba (which acted as the counterpart of the project), all of
which participated in the execution of the project. Their names will not be mentioned for confidentiality.
Simultaneously, there was the revision of progress reports produced by one of the team members that
had been generated during the three years of execution of the project to identify the problems raised,
the best practices used, and the achievements obtained.
After collecting the information, the next phase was to prepare a survey addressed to the main
stakeholders of the project; therefore, three types of surveys were prepared: one for the Local
Coordinator, the second for the Regional Coordinator La Paz and her team members, and the last survey
addressed to officials of the Gobierno Autónomo Municipal de Cochabamba.
The questions included in the surveys aimed to identify problems presented during the initiation,
planning, execution, monitoring, control, and closure phases of the first year of execution from the
Page 425
SEMIT 2023
diverse perspectives of the respondents. There were sections added to identify the impact of these best
practices on the schedule and scope generated, the achievements during all three years and to spot the
best project management practices implemented to achieve the results during the second and third years
of execution.
The answers obtained through the surveys were reviewed, deepened, and complemented through an
evaluation meeting. Individual meetings were later held with each respondent, seeking to expand the
information found through the questionnaire.
There was a comparison between the project management practices applied during the first year of
execution of the project and the practices identified in the bibliographic review, considering the results
of the surveys, meetings, and reports. The changes in the second and third-year results referred to the
fulfillment of the execution time, the number of beneficiaries reached, and the percentage of desertion.
Finally, there was a general meeting to present and communicate these results.
4 Case study
This section presents the case to be dealt with and its background. The first part of this section describes
the work carried out by the foundation and the project. Afterward, it is contextualized.
4.1 Fundación para el Desarrollo FAUTAPO
There is a diversity of foundations and non-governmental organizations that execute programs and
projects that aim to improve the living standards of vulnerable populations in Bolivia, both in rural and
urban areas. They seek to generate the most significant impact possible, implementing actions that tend
to contribute to public policies established by the government. These institutions structure, design, and
plan projects presented to international cooperation organizations to request funding. If these projects
are aligned with the areas of interest, the potential financiers evaluate them and approve funds for their
execution. These limited funds must be efficiently managed to achieve the objectives. Therefore, project
management is crucial to guarantee efficiency and generate the value intended.
Among these foundations, there is Fundación Educación para el Desarrollo FAUTAPO, whose purpose
is to "stimulate production and employment by strengthening processes related with technical training
and thus contributing to achieve equitable, inclusive and sustainable social and economic development,
for the benefit of the Bolivian population" [3]. It runs projects and programs throughout the country in
rural and urban areas, designed and executed by a multidisciplinary team that applies an integrative
approach. In addition, it applies a cross-cutting approach that covers gender equality and environmental
protection [3]. As it measures the project's success according to how well the objectives have been
achieved (effectiveness evaluation), the donor also asks them to comply with the budget and schedule
(efficiency evaluation). This article shows how applying best project management practices helped
improve time and scope compliance, but it will not cover the evaluation of compliance costs.
The case study refers to the execution of the project "Socio-economic empowerment of women and
men within the framework of a shared vision of sustainable local productive development, with a gender
approach." This project is related to economic development, with an integrated approach in the business
field that, in addition, addresses and support the personal and family-related aspects of its participants.
It is national in scope, with a presence in eight of the nine departments in Bolivia, and its execution
adapts to particular characteristics and dynamics of each of the three regions in the country. This
investigation considered the execution of a project in the city of Cochabamba.
4.2 Project Description
The project implementation in Cochabamba received international financing, with the Gobierno
Autónomo Municipal de Cochabamba as a local counterpart that gave economic support. The vision
was that, after the first three years of execution, the project would be entirely transferred to this public
instance, becoming part of the Municipality's public policies and ensuring its sustainability over time.
It was part of a national project considering the geographical regions' differences. There is a national
Page 426
SEMIT 2023
coordination instance, as well as a local one. Numerous actors participated, which generated high
complexity and a high level of uncertainty.
The purpose of the project in Cochabamba, for its three years of execution, was the socio-economic
empowerment of 290 individuals from vulnerable populations with their production units. The project
accepted a 20% dropout during the three years of execution. The project's beneficiaries belonged to the
following categories: food production, textile manufacturing, agricultural production, wooden furniture
manufacturing, metal goods manufacturing, and handicrafts, among others.
The project team comprised a project manager and three consultants (two responsible for strengthening
business capacities and one for the social area). The hiring of consultants was done annually. Figure 2
displays the project's organizational chart.
Fig. 2. Project Organizational Chart for Cochabamba.

Among the main functions of the Regional Coordinator of La Paz is to provide the guidelines for the
project executed in El Alto and Cochabamba cities, manage the project budget of both cities, and
perform the corresponding coordination and monitoring.
The Local Coordinator is Cochabamba's Project Manager, who directs the team, leads it, directs the
planning, monitors, controls, and coordinates with the Regional Coordinator, La Paz.
Business Strengthening Consultants oversee recruiting and selecting the beneficiaries, diagnose their
production unit, generate the improvement plan, and guide and accompany the implementation of
business tools.
The Social Area Consultant is in charge of identifying if the beneficiaries have any family, health, or
legal problems to refer to the respective institution that can provide a solution and carry out the
respective accompaniment.
The project consisted of three stages: 1) Recruitment and selection, 2) Diagnosis, and 3) Improvement
(Figure 3). The first stage consisted of identifying, recruiting, and selecting the potential beneficiaries
of the project, who were supposed to have a production unit under specific criteria. It is the most critical
stage as it highly affects the dropout percentage. In the second one, there was a diagnosis to establish
the conditions of that production unit, considering issues of production, administration, accounting and
marketing, the social environment in which the individuals lived, and their entrepreneurial capacities.
With the results of that diagnosis, the third stage covered the development of an improvement plan and
its implementation. In this way, the empowerment of the beneficiaries is generated. The plan developed
considers six instances:
Training: Training group sessions on business management skills (management, marketing, and
production) and workshops related to empowerment and self-esteem, aiming at empowering them,
allowing knowledge transfer, and generating a space in which supporting networks are created, and the
beneficiaries can share the problems and experiences in their productive units
Page 427
SEMIT 2023
Administrative, technical support: Accompaniment and guidance in implementing business concepts

and tools (management, accounting, marketing, and production) obtained in the training sessions. This
accompaniment has two modalities: an individualized follow-up carried out in the business initiative
and a group accompaniment in the foundation's offices, seeking to improve their management skills.
Technical support in the area: Technical training in topics specific to the area in which the production
unit operated, pursuing to improve product quality. A technical professional gives this technical training
to a small group of beneficiaries with similar quality problems. In this way, technical support networks
are generated.
Social support: Guidance and accompaniment in legal, domestic, and health issues, empowering the
participants, supporting the beneficiaries, and creating a calm environment that allows the beneficiaries
to strengthen and develop their productive units.
Procurement: Provide specific tools or equipment necessary for their operations, aiming at improving
product quality.
Increase Sales: Host different activities, such as fairs or virtual showcases, for the beneficiaries to
present their products, attract new clients, and acquire sales skills. An approach and accompaniment to
using technological platforms are carried out to boost the sales of the productive units.
Fig. 3. Project Scope Structure.

4.3 Background
Although the institution has experience in planning and executing projects, introducing the best project
management practices will further strengthen the efficiency of achieving results by using time and
resources better. After the first year of execution, and since multiple variables were involved in the
project, the Regional Coordinator decided to include best project management practices to improve the
quality and execution time.
Following are the findings obtained after conducting the surveys and interviews.
5. Results
This section presents the results of the research with the methodology proposed, highlighting the best
project management practices applied during the three years, besides the problems and achievements
presented in the project.
5.1. Situation Year 1
The project began with the presentation of the logical framework to the financier, showing the
objectives and desired outcomes of the project at a national level. These were distributed to each
Page 428
SEMIT 2023
regional office to establish a general plan (high-level) based on the funder's requirements. The logical
framework considered the specific needs of the different actors who participated in the process.
The new team of job formed was not aware of the organizational culture, internal procedures, or the
characteristics of the population with which they would work, so there was a learning curve involved
that had to be assumed for the first year. However, the foundation's internal team already had experience
with similar populations and handling projects with the same components; therefore, internal
knowledge was applied. However, there were problems at the beginning of the execution delaying the
knowledge transfer to the team due to a medical leave of the local coordinator.
Although there were meetings to socialize the project and its components with the local team, they
could not dimension and understand the scope to its full extent. It was not possible to go beyond the
general planning into a detailed one that would involve the project team, identify activities, and produce
time estimates that were more real and precise to the environment, the regional dynamics, and the
uncertainties of that moment. The requirements related to financing in terms of scope and time were
narrow and non-negotiable during the three years of execution. However, there was the flexibility to
adjust the ways of complying with these requirements, so the planning followed a trial-and-error
strategy during the whole implementation at a national level. There was no integration between the tight
requirements of time established by the project, the technical team's availability, and the budget, making
it challenging to cover the scope in the required time.
Best project management practices implemented during the first year.
These were the best project management practices applied during the project's first year of execution.
Unfortunately, there was poor culture management, which made it difficult to understand the work
expected (making it necessary to apply changes to generate improvements for the benefit of the project),
the procedures of the organization, and the relationship between the different stakeholders; all of which
generated delays and reprocessing. There were coordination meetings during the execution, improving
internal communication and communication with the regional instance. Table 3 lists the best practices
that were applied during the first year.
Table 4. Best project management practices applied during the first year.
Initiation Planning Execution Monitoring and Closure
Control
• High-level • High-level • Meetings to • Weekly team • Evaluation
identification planning using the socialize meetings to of the
of logical objectives and monitor and execution
stakeholders. framework. scope. control scope during the
• Inclusion of • High-level time • Empowerment of and schedule. first year.
knowledge estimation. the project team. • Change control
from similar • Meetings to • Clear at a managerial
projects. conduct short- establishment of level.
• Development term planning. roles for team • Take action to
of the logical • The methodology members. meet delivery
framework as was established to • Identification of dates.
a business guarantee the high lessons learned • Openness to
case to obtain quality promised since the delivery receiving
financing. to the financier. of the first report. suggestions for
• Clear • Documentation changes that
establishment of of the knowledge generate
the project stages generated. improvement.
to be applied. • Risk
• Response identification.
planning of
Page 429
SEMIT 2023
undocumented • Establishment of
risks. communication
channels with the
project team.
• Constant
interaction with
financiers.
Project management impact during the first year.

By the end of the first year of execution, the project presented several achievements but faced specific
difficulties that affected the project's implementation, which is detailed below.
Achievements identified during the first year of execution.
The project management practices used during the first year contributed to the following.
Achievement of objectives: Completing all the components related to supporting the beneficiaries
during the first year. Identification, examination, and execution of action items to provide support to
100 beneficiaries.
Value creation for the main stakeholders: Through self-knowledge and awareness of their rights and
duties, acquiring technical business knowledge, and putting it into practice, beneficiaries received
empowerment that helped them increase their income and open new markets.
Team strengthening: A team was created, and members received training related to gender issues,
developed cohesion, and managed to implement extensive work.
Knowledge and experience generation: Meetings allowed sharing of the experiences generated from
different points of action; lessons learned were used to adjust the planning and were considered in the
planning for the following two years.
Establishment of communication channels: The team attained efficient communication channels
between their members, the regional coordinator, and the different stakeholders.
Difficulties identified during the first year of execution.
The difficulties that arose during the first year were:
Delay in the closure of the first year: It took about 40% of the execution time of the first year to carry
out the recruitment, selection, and diagnosis, causing a 2-month delay in the first year. This delay was
due to the team's learning curve, the slow knowledge transfer of experiences, and the uncertainty
generated by having general planning with tight execution times.
Lack of clarity in scope, activities planned, and schedule: The team did not establish instances to
perform short-term planning, and there was difficulty in carrying out knowledge transfer about the
beneficiaries' experiences and characteristics.
Desertion of beneficiaries who had been selected: 38% of the participants deserted during the first year,
originated by the delay in implementing improvement actions for the beneficiaries, for not providing
the corresponding information about this delay, and related to the situation the target population was
going through.
Project team burnout: This was caused by a lack of clarity in planning.
Late socialization of the format of the first report: Certain information required for the preparation of
the report was not collected or processed on time.
5.2 Situation Years 2 and 3
Best project management practices implemented during the second and third years.
Page 430
SEMIT 2023
Applying the best project management practices during the second and third years was possible. There
was also improvement in the ones applied during the first year. These findings are detailed below.
Initiation:
• Clear scope setting.
• Preparation of the project charter before starting the second and third years.
• High-level identification of stakeholders.
• High-level identification of risks.
• Inclusion of knowledge generated during the first year.
• Transfer of tacit knowledge to the local team.
Planning:
• Evaluation of the first phase of the project; review the reports generated.
• Inclusion of lessons learned into the planning process.
• Hold planning meetings with the project team before starting execution.
• Planning adjustments based on project monitoring.
• Communication with the beneficiaries to reduce the response time between the moment they are
contacted in the diagnosis and when the improvement plan is launched.
• Preparation of a WBS and a more realistic schedule.
• Risk identification and response planning.
• Advanced planning of the procurement stage to obtain better suppliers' prices and develop it more
orderly.
• Inclusion of action items to involve technicians from the Gobierno Autónomo Municipal.
Execution:
• Identification of lessons learned.
• Inclusion of lessons learned during execution.
• Execution based on prior planning.
• Clear establishment of the roles of team members.
• Documentation of the knowledge generated in a manual.
• Environmental analysis to identify risks.
• Constant communication with the project team about what needed to be done.
• Constant interaction with financiers.
• Execution of actions to involve the technicians of the Gobierno Autónomo Municipal.
• Project culture management.
• Socialization of internal procedures that affect execution.
• Improvement of the methodology and tools used to increase efficiency and generate value for the
beneficiaries since the early stages.
• Accompaniment in the execution of a pilot test by the Gobierno Autónomo Municipal.
Monitoring and Control:
• Weekly meetings to monitor and control the schedule and progress of the project.
• Change management meetings at a local level to assess the impact of requests presented.
• Weekly team meetings to monitor and control the project.
• Change control at a managerial level.
• Take action to meet delivery dates.
• Openness to receiving suggestions for changes that generate improvement.
Closure:
• Evaluation of the execution of the second and third years.
• Documentation of lessons learned.
Impact achieved during the second and third years.
Page 431
SEMIT 2023
The formal incorporation and application of additional best project management practices allowed to
maintain and enhance the achievements obtained during the first year of the project's execution. The
achievements are listed below.
Achievement of objectives: It was possible to achieve the objectives set by the sponsor. Two hundred
ninety beneficiaries were selected; 238 finished the program, and there was an average dropout of 18%,
less than the 20% required. The beneficiaries who finished the program improved their income and
strengthened their market positions.
On the other hand, the beneficiaries received support in the social area so that they know and exercise
their rights and obligations and receive the necessary legal, health, and psychological advice. The
project executed all the improvement components and was able to transfer the execution process to the
Gobierno Autónomo Municipal de Cochabamba, as expected.
Knowledge generation: Planning specifications and lessons learned were documented in a project
execution manual, which is a valuable instrument for the planning of similar projects in the future, and
to enable knowledge transfer to other teams.
Reducing execution time: The most significant impact generated by the project was the reduction of
execution time. It took 11 months to complete the project in the second year and only ten during the
third. This way, the project complied with the date and scope established by the sponsor and reduced to
10% and to 0% the percentage of time exceeded in the second and third year, respectively.
Reduction of the dropout rate: Improvement in project management, mainly related to stakeholders and
communications management, reduced beneficiaries' desertion rate from 34% to 10% during the second
year and only 8% in the third year.
Team building: The team managed to achieve team cohesiveness, promote, and implement
improvement actions for the execution of the project. It also managed to consider the institutional
culture in its activities.
Quality improvement: By reducing the time necessary for the selection process and diagnosis of the
beneficiaries, there was more time to carry out improvement actions and empowerment, directly
impacting the quality and compliance with the scope.
Effective communication: Formalizing specific communication channels for each interested party was
possible, establishing fluid communication, especially between team members.
Generation of formats and documents for future projects: The format of different project reports and
the respective information requirements that need to be collected during execution were established. It
was also possible to establish the format of administrative documents such as forms, minutes, and tools.
Formal incorporation of best project management practices: A positive precedent was left over from
this experience; the project team acquired project management skills to replicate and implement best
project management practices in future projects.
5.3 Results Summary
The case analysis shows that as best project management practices were incorporated, there were
improvements in project execution duration, the dropout rate and the percentage of exceeded time
decreased, and the number of beneficiaries was achieved. Table IV summarizes the results obtained
during each of the three years.
Table 5. Results obtained.
Evaluation criteria First Year Second Year Third Year
Planned duration in 10 10 10
months
Executed duration in 12 11 10
months
Page 432
SEMIT 2023
Percentage of exceeded 20% 10% 0%

time
Beneficiaries reached 100 100 90
Beneficiaries that 66 90 82
finished the program
Dropout rate (%) 34% 10% 8%
6. Conclusions and future work
The content of this article shows the outstanding level of complexity of social projects when there are
many parties involved, with the resultant uncertainty related to their actions.
Considering it is a case study, we cannot generalize the practices used; however, it presents an
experience that drives the use of these practices following appropriate adaptation to comply with the
delivery times of each phase and achievement of objectives.
The bibliographic review emphasizes how the evaluation of a project's success is multivariate.
Efficiency and effectiveness are two of its fundamental elements. Incorporating best project
management practices is how efficiency will be improved, helping to complete the project.
Given the impact generated by the incorporation of best practices, organizations worldwide identify and
update them according to their application in organizations of different areas, as well as with different
levels of complexity. Likewise, these promote their adjustment and implementation, considering the
improvement in indicators. Despite this, many projects still fail to meet the time, scope, and budget
constraints.
It was identified that there is plenty of experience in using best practices in engineering, construction,
and software development and that exploring experiences in the social sphere is necessary.
The research showed the results of implementing best project management practices in social projects
involving high complexity and multiple stakeholders.
Communications management and stakeholder management are presented as the most valued aspects
to achieve success with a project. Likewise, knowledge management and its timely transfer contribute
to better project planning. Culture management was also highlighted as essential for the team to adopt
adequate work styles and assimilate internal processes.
The case presented shows that the experience gained, along with the lessons learned, contribute to
incorporating critical aspects into the planning and execution of the project to reduce the probability of
risk occurrence. It also depicts that planning contributes to achieving better execution of the project,
providing a clearer horizon, and reducing the distractions of reprocessing. In addition, it contributes to
generating value in earlier stages, generating a goal-oriented environment.
The use of best practices in the project since the first year of execution contributed to fulfilling the
objectives, generating knowledge, strengthening the project's team, and improving communication.
It was evident that the situation that the project's target population is going through in personal aspects,
with their family, and related to their business skills, makes it difficult to remain within the plan for the
different actions. In many cases, it limits the continuity of the actions proposed in the project. Although
this fact was considered a risk and actions were taken to mitigate it, it is an element that must be studied
further.
This study allowed us to show practical experience related to the best project management practices
applied to a social project in Latin America. Social projects, besides generating an impact in the short
term, also do so in the long term; therefore, they provide the opportunity to carry out studies to evaluate
their impact when they are being implemented and foster research on the implementation of best project
management practices in the social field and the impact generated in the budget over time.
This study also opens the possibility to research the process of implementing best project management
practices in projects executed in a public institution such as the Gobierno Autónomo Municipal de
Page 433
SEMIT 2023
Cochabamba. It will be essential to identify the barriers that the institution encounters and analyze the
behavior framed in its culture. Likewise, replicating these best practices in other social projects, such
as educational projects, will be an opportunity.
Since the project was implemented in various Bolivian cities, it is possible to compare the results
obtained in Cochabamba with those obtained in cities with similar characteristics.
Acknowledgments
Special thanks to Fundación Educación para el Desarrollo FAUTAPO for allowing us to share the best
management practices implemented in their project.
References
1. Besner C, Hobbs B (2013) Contextualized Project Management Practice: A Cluster Analysis
of Practices and Best Practices. Project Management Journal 44:17–34. doi: 10.1002/pmj.21291
2. Dvir D, Lipovetsky S, Shenhar AJ, Tishler A (2004) What is really important for project
success? A refined, multivariate, comprehensive analysis. International Journal of Management
and Decision Making
3. Fundación FAUTAPO Fundación FAUTAPO – Educación para el Desarrollo. In: Homepage.
http://www.fundacionautapo.org/. Accessed 15 Jul 2022
4. Georgieva S, Allan G (2008) Best Practices in Project Management Through a Grounded
Theory Lens. 6
5. Gurmu AT, Aibinu AA, Chan TK (2016) A study of best management practices for enhancing
productivity in building projects: construction methods perspectives. CEB 16:1–19. doi:
10.5130/AJCEB.v16i3.4882
6. Ibbs CW, Kwak YH (2000) Assessing Project Management Maturity. Project Management
Journal 31:32–43. doi: 10.1177/875697280003100106
7. Ika LA (2009) Project success as a topic in project management journals. Project Management
Journal 40:6–19. doi: 10.1002/pmj.20137
8. Massa N, Figueiredo PS, Loiola E, Passos FU (2020) MELHORES PRÁTICAS EM GESTÃO
DE PROJETOS DE NOVOS PRODUTOS APLICADAS A UMA EMPRESA DA ÍNDÚSTRIA
DA MODA. G&P 21:512–530. doi: 10.21714/2178-8030gep.v.21.5441
9. Meier SR (2008) Best Project Management and Systems Engineering Practices in the
Preacquisition Phase for Federal Intelligence and Defense Agencies. Project Management Journal
39:59–71. doi: 10.1002/pmj.20035
10. Nelson RR (2007) Failures, Classic Mistakes, and Best Practices. IT Project
Management 6:66–78
11. Project Management Institute Inc. (2017) Aumento de las tasas de éxito. La
transformación del alto costo de un bajo desempeño. Pulse of the Profession
12. Project Management Institute Inc. (2021) El nuevo ecosistema de Trabajo. Pulse of the
Profession Ameríca Latina
13. Rode ALG, Svejvig P, Martinsuo M (2022) Developing a Multidimensional
Conception of Project Evaluation to Improve Projects. Project Management Journal 53:416–432.
doi: 10.1177/87569728221095473
14. Zou Y, Lee S-H (2009) Implementation of project change management best practice in
different project environments. Can J Civ Eng 36:439–449. doi: 10.1139/L08-138
Page 434
SEMIT 2023
15. Zwikael O (2009) The Relative Importance of the PMBOK ® Guide's Nine Knowledge
Areas during Project Planning. Project Management Journal 40:94–103. doi: 10.1002/pmj.20116
16. (2013) La Guía de los Fundamentos para la Dirección de Proyectos (Guía del
PMBOK), Quinta Edición. Project Management Institute Inc, Pennsylvania
PMBOK), Sexta Edición. Project Management Institute, Inc, Pennsylvania
18. (2021) Call for Papers: Special Issue on Project Success. International Journal of
Project Management 39:213–215. doi: 10.1016/j.ijproman.2021.01.007
PMBOK), 7ma Edición. Project Management Institute Inc, Pennsylvania
Page 435
SEMIT 2023
Analysis of Investment Feasibility Using the Life Cycle Cost (LCC) Method in Implementing
Green Retrofits in High-Rise Office Buildings
Aviva Cantika Alfatihanti1, Yusuf Latief1, Bernadette Detty Kussumardianadewi 1
1 Civil Engineering Department, Faculty of Engineering, Universitas Indonesia, Kampus UI Depok
16424 Indonesia
Abstract
As one of the world's largest carbon gas emitters, construction activities contribute significantly to
global warming. Various countries in the world have joined forces to achieve the Net Zero Emission
target by 2050, and Indonesia is also contributing by implementing green building concepts in existing
buildings, including office buildings. There are challenges in implementing green building concepts
in existing buildings due to the lack of public understanding of the investment costs supported by the
high costs required. This research aims to help provide an understanding of investment costs using the
life cycle cost (LCC) method by providing information on the value of investment feasibility through
the application of green retrofits that refer to the new certification by the government. This research
involved five experts and more than ten respondents to provide opinions on qualitative assessment.
At the same time, the regulation of the Minister of Public Works and Public Housing No. 21 of 2001
was used. Investment cost simulation was carried out for ten years to get the results of profit and
benefits.
Keywords: Life Cycle Cost, Investment Feasibility, Green Retrofit, Regulation of the Minister of
Public Works and Public Housing No. 21 of 2021, Office Building.
1. I. Introduction
As the world's population increases with the number of human activities, it has an impact on
increasing carbon/greenhouse gas (GHG) emissions, one of which is caused by energy consumption.
The highest achievement is known in 2018 due to the biggest contribution from China, the United
States, and India by 70% with Indonesia ranked 8th [1]. In 2020, the building and industrial construction
sector accounted for 10% of the 37% of global energy consumption directly related to GHGs.
Indonesia's primary energy consumption generated by the building sector in 2018 was almost 20% [2]
[3]. The increase in GHG then affects the temperature change which is quite extreme in Indonesia and
is still increasing. Recorded in January 2023, there was a positive anomaly in Indonesia's temperature
of 0.1ºC from the normal monthly average temperature 1991-2020 [4]. Also supported by a report
issued by IPCC found various adverse impacts in the form of extreme fluctuations in water flow, severe
drought, and the known impact of sea level rise on 50 million people [5].
The current natural conditions of increasing earth temperature encourage Indonesia to
contribute to the Net Zero Emission 2050 (NZE) movement through zero energy buildings carried out
with a three-aspect approach, namely energy, economy, and environment by keeping the earth's
temperature under 2.0ºC followed by an optimistic target under 1.5ºC collaboratively to achieve the
goal of climate-neutral conditions in 2050 [6] [7]. One of the main breakthroughs listed in the Green
Building Performance Scheme Roadmap is the retrofitting of buildings managed by the Ministry of
Public Works and Housing through Government Regulation No. 16 of 2021 [8]. green retrofitting is a
key factor that can be applied in Jakarta, which consists of 42% of office buildings out of 382 high-
rise buildings [9].
One of the main factors why Indonesia has few green-certified buildings is due to the lack of
knowledge of building values, especially regarding green new buildings and green retrofits. With the
Page 436
SEMIT 2023
LCCA method, building owners can know the impact of short design stage decisions on the economic
value that will benefit over the life of the building when changing the building planning to a green new
building or green existing building [10]. There is also another consideration, the cost of GREENSHIP
certification with 3 categories, each costing about 3-10% more than usual [11]. The existence of this
burden on building owners and developers is supported by the fact that there is no incentive by the
government for green investment in the building construction sector [12]. The opportunity to minimize
costs can be achieved if the design brief stage recommends using energy-efficient materials. Proven in
2019, a high-rise office building was able to save electricity costs by 18% by saving lighting energy
with payback costs falling in year 3 [13].
Considering Indonesia's GHG condition as one of the largest contributors of gas emissions
through construction activities, and supported by the lack of knowledge about green building
investment, it is necessary to do a study to provide information about green building financing,
especially green retrofitting in high-rise office buildings in Jakarta using the LCC method. It is
important to use more effective and efficient certification in order to reduce the amount of gas
emissions and accelerate the progress of achieving Net Zero Emissions in 2050.
2.1. The Concept of Green Retrofitting in High-Rise Buildings
High-rise buildings are commonly found in every country with characteristic heights ranging
from 45-300 meters [14]. As a country with the 8th highest number of high-rise buildings in the world,
Indonesia has made a major contribution to the development of construction over the past 30 years,
especially in the city of Jakarta [15]. This contribution is also accompanied by the large production of
carbon emission gases produced by high-rise buildings, one of which is used as an office building.
Throughout the entire process of preparing, constructing, and maintaining high-rise
buildings, the consumption of energy is the main culprit behind the production of carbon emission
gases. The sustainability of such buildings poses a challenge that drives advancements in science and
technology aimed at reducing excessive carbon emissions. High-rise constructions in particular can
play a significant role in addressing the pressing need for change in the building sector, which currently
accounts for a staggering 20% of primary energy consumption in Indonesia [16].
The existence of green buildings is now considered incompatible with the availability of
available space in Indonesia, especially in Jakarta. As the most populous city in Indonesia, high-rise
buildings have been spread evenly throughout the region and are not recommended for planning new
building construction. The low possibility of this creates new limitations on the concept of green
building, thus creating a new breakthrough in the form of green building modification or green
retrofitting. Green retrofitting is any form of development of an existing building that is used in part
or in full as a form of improving energy and environmental performance, minimizing excess water use,
maximizing the comfort and quality of indoor natural lighting, improving air quality, and reducing
noise in a way that is financially beneficial to the owner [17].
The Ministry of Public Works and Public Housing is one of the institutions that has developed
a green building customization rating system. A part of this system is aimed at improving the
performance of existing buildings to meet green customization and involves making efforts to adjust
the performance of utilized buildings. According to Ministry of Public Works and Public Housing
Regulation no. 21 of 2021, is an effort to adjust the performance of utilized buildings in order to fulfill
green building requirements. Financially, green buildings, especially green retrofit buildings, can save
up to 40% of the cost of existing (conventional) buildings.
Page 437
SEMIT 2023
Table 2. Ministry of Public Works and Public Housing Regulation no. 21 of 2021 Rate
Source: Ministry of Public Works and Public Housing Regulation no. 21 of 2021
No. Requirements Point BGH Pratama BGH Madya BGH Utama
performance achievements based on

Organization and
regulation of the minister of public
regulation of the minister of public

according to the parameters of the
according to the parameters of the

building performance assessment
building performance assessment

certificate of performance (SLF)
works and public housing green
works and public housing green

A governance of green 83
building
More than 80% - 100%

More than 65% - 80%
Customisation
B 26
45% - 65%
construction process
Maintenance of green
building performance
C 50
during the utilization
period
Role of occupants /
D 6
users of green buildings
Total 165
2.2. Life Cycle Cost Application In Green Retrofits

Cost is one of the main factors that are very influential in sustainability, especially in running
a project. There are several stages of the process that need to be supported by costs to be able to produce
products during the predetermined life of the plan. This is evidenced in studies with the term life cycle
cost (LCC). In the ISO 15686 Part 1 standard regarding Life Cycle Planning, LCC is defined as a
comparative cost assessment technique of all cost aspects by considering all economic factors within
a certain period of time [18]. LCC is a method of evaluating cost requirements that apply to the
operation, management, maintenance, and disposal stages of a component in an economic manner to
make decisions [19]. So, it can be interpreted that LCC is a method of analyzing the costs required
during the life cycle of a project.
In analyzing construction costs, LCC (Life Cycle Cost) considers four types of costs:
investment costs during the pre-construction and construction phases, as well as post-construction
costs such as operation and maintenance costs, replacement costs, and demolition costs [20]. This
comprehensive approach to cost analysis ensures that all relevant costs are considered. In addition, the
cost conditions used in the LCC are based on current value conditions, which take into account changes
in the value of money over time. This detailed approach to cost analysis helps ensure that construction
projects are planned and executed efficiently and effectively.
Life cycle costing (LCC) calculations are essential in determining the financial viability of a
project over its lifetime. These calculations are usually performed for a period of 30 years, as
recommended by previous literature studies. However, Iain Campbell's research on green retrofits
aimed to achieve cost recovery as one of the return on investment objectives. To achieve this, he
conducted LCC simulations for shorter time horizons ranging from 2 to 15 years [17]. Campbell's
approach provides a more detailed understanding of the financial performance of green retrofit
projects, and how they can achieve cost recovery in shorter timeframes than traditionally assumed.
2.3. Investment Performance
Investment performance represents active actions aimed to achieve investment feasibility.
The results of investment performance are goals that become the output of investment feasibility.
Referring to the Global Investment Performance Standard (GIPS), the investment performance
calculation report is presented in the form of a return on investment (ROI) presentation which can be
assessed skinatively through realized and unrealized returns [21].
Page 438
SEMIT 2023
A reliable form of return is a realized return, which is a profit that has been recovered by
investors. Meanwhile, the unrealized return is an expense that continues to be used by the owner and
is considered weak or bad. In order to prevent weak returns, there are several efforts that can be taken
[21], such as:
1. Calculate the elements of investment return comprehensively.
2. Maintain the stability of investment value by triggering a reasonable level of trading level as
reasonable as possible.
3. Implementing cut offs to prevent prolonged large losses.
To achieve a good investment performance, investment feasibility is required by conducting
a cost analysis to produce an efficient value. NPV value is required by taking into account costs and
benefits to analyze investment options. The conditions of investment decision making with this method
if the NPV value is negative means that the investment carried out is losing and not feasible to continue,
while if the NPV value is positive means that the investment carried out is profitable and feasible to
implement.
(𝐵𝑡 +𝐶𝑡𝑖 )
𝑁𝑃𝑉𝐼 ∑𝑇=𝑁
𝑡=0 (1+𝑟)𝑡
(1)
1
𝐷𝑖𝑠𝑐𝑜𝑢𝑛𝑡 𝐹𝑎𝑐𝑡𝑜𝑟 = (1+𝑟)𝑡
(2)
Furthermore, the BCR value is needed to evaluate the relationship between the costs and
benefits of a project. Through this calculation, a condition is obtained in the form of BCR < 0, then
NPV < 0 so that the investment is not feasible because it can cause losses, if BCR> 1, then NPV> 1 so
that the investment is feasible because it can be profitable.
There is a payback period which is a period to be able to recover investment expenditures by
using net cash flow to determine the number of years needed so that all investments that have been
issued can return. Thus, the payback period of an investment describes the length of time required for
the funds invested in an investment to be fully recovered.
After knowing all cost components through LCC analysis, Internal Rate of Return (IRR). To
determine whether a project is worth investing or not is to compare the IRR value obtained with the
Minimum Attractive Rate of Return (MARR) and Weighted Average Cost of Capital (WACC). MARR
is the interest rate set by a company to evaluate and select project alternatives. There are decisions to
determine whether a project is financially feasible, namely IRR ≥ MARR, IRR > WACC means the
project is considered profitable and the IRR > WACC means the project is considered profitable.IRR
can be calculated by the formula below.
𝑁𝑃𝑉1
𝑖1 + (𝑖 − 𝑖1 ) = 𝐼𝑅𝑅 (3)
(𝑁𝑃𝑉1 −𝑁𝑃𝑉2 ) 2
Or
𝐶
0 = 𝑁𝑃𝑉 ∑𝑇𝑡=1 (1+𝐼𝑅𝑅)
𝑡
𝑡 − 𝐶0 (4)
𝑓𝑢𝑡𝑢𝑟𝑒 𝑐𝑎𝑠ℎ𝑓𝑙𝑜𝑤 1/𝑛

𝑑𝑖𝑠𝑐𝑜𝑢𝑛𝑡 𝑟𝑎𝑡𝑒 = ( 𝑝𝑟𝑒𝑠𝑒𝑛𝑡 𝑣𝑎𝑙𝑢𝑒
) −1 (5)
2.4. Component of Green Retrofitting Life Cycle Cost

Initial cost or also called initial/original cost consists of planning costs such as design costs,
construction costs, implementation costs such as contract costs, certification costs, and other expenses.
Initial costs consist of production costs, development costs in the form of design and testing,
implementation costs including contract administration, inspection, and training, as well as other costs
[22].
Page 439
SEMIT 2023
Then maintenance is a series of policies needed to maintain or restore an item in an effective

operational state [23]. All necessary actions are able to maintain product quality so that there is no
damage or interference with the machine to produce products that can reach the target desired by the
company. So that operating costs are the costs used to maintain and maintain the quality of the building.
Furthermore, replacement cost is the cost associated with replacing an asset or service that
will occur future costs at the time of replacement. The concept of replacement cost is useful for the
preparation of budgets and standard costs and capital budgeting such as re-planning, addition,
replacement, or removal of fixed assets [24].
Initial Cost Operational and Replacement Salvage

LCC = + Maintenance Cost + Cost - Value
Figure 1. Life Cycle Cost Component (Source: Self-Administered)
2.5. Energy Efficiency Aspect

Efficiency is a form of effort to reduce on the basis of environmental imbalance due to
excessive activity to return to normal conditions. However, energy efficiency itself means an effort to
reduce the amount of energy use through the process of energy utilization or conversion [25].
Based on a report issued by Climate Transparency, Indonesia still needs to improve energy
efficiency in the building sector due to the fantastic increase in gas emissions [26]. The results of
energy efficiency not only impact the appliance technology and physical building but also affect air
temperature and lighting conditions which can improve work comfort and productivity [27].
There is a legal basis in Government Regulation No. 70 of 2009 Articles 10-12 which states
that energy providers and entrepreneurs in the form of individuals, business entities, and permanent
businesses are required to carry out energy conservation, and energy users are required to save and
improve energy efficiency. Supported by DKI Jakarta Governor Regulation No. 60 of 2022 Article 9,
energy efficiency in buildings includes the following [28].
1.Building envelope.
2.Air conditioning system.
3.Artificial lighting system.
4.Transportation system in the building.
5.Electrical system.
6.Renewable energy system in the building.
Furthermore, there is a demolition stage that is carried out with due regard to environmental
conditions, this stage consists of:
1. Area requirements.
2. Consideration of building components.
3. Consideration of building materials.
4. Consideration of facilities and infrastructure.
3. Research Methodology
These research goals are answered by using the LCC method. Materials were collected
through case studies of office buildings and structured surveys. Data collected using a standardised
questionnaire with five experts with more than five years of experience in green retrofit applications
Page 440
SEMIT 2023
and occupants of green retrofit office buildings. This data was later processed through critical index
analysis. Further data was collected after being supplemented with interview sessions with the building
management and visualized through BIM 5D. From the two types of data that have been validated, the
research can proceed using life cycle cost analysis.
Start The following is an overview of the structured
research flow used in the study:
Problem Phenomena
Problem Identification Literature Review
Research Hypothesis
Research Purposes
Problem Formulation
Study Literature Defining Green Retrofitting LCC

(Ministry of Public Works RQ1 Component
and Housing Regulation
No.21 of 2021)
Quetionnaire Distribution Quetionnaire Draft
Expert Judgement Index Analysis
Selecting Case Study Buildings
Collecting Building Data

No
LCC Analysis
Yes
Running Scenario Plan RQ2 Building Cost No
Study Literature
(Ministry of Public Works Green Retrofitting Work Assessment
and Housing Regulation
No.21 of 2021)
LCC Financial
Feasibility
Analysis
Yes
Running Scenario Plan RQ3 Investment Feasibility
Quetionnaire Distribution Questionnaire Draft
Descriptive Analysis
Literature Review
and Stakeholders RQ4 Define Green Retrofitting Benefit
Judgement
Define New Green Retrofitting Strategies
Conclusion & Recommendation
Finish
Page 441
SEMIT 2023
Figure 3. Conceptual Framework of the relationship between green retrofitting and life cycle cost
Source: Self-Administered
Page 442
SEMIT 2023
4. Conceptual Framework
The Relationship Model between Variables in the Relationship between Green Retrofitting,
Life Cycle Cost, Green Retrofitting Job Appraisal, and Investment Feasibility in the study is
represented in the form of a conceptual framework as follows:
Figure 3. Conceptual Framework of the relationship between green retrofitting and life cycle cost
Indicator variables are categorical analyses where factor variables must be converted to a set
of indicator variables. Once the WBS standard is attained, the next step is identifying the risks that
affect the cost accuracy of Green Retrofitting office high-rise building based on BIM-5D. The risk
variables and cost model [29]. in this research are explained below:
Table 2. Research Indicator Variables (Source: Self-Administered)

VARIABLE INDICATOR Reference
GREEN X1. 1 Utilization Stage [25]
X1
RETROFITTINGX1. 2 Disposal Stage [25]
X2. 1 Initial Cost [19]
X2 LIFE CYCLE COST X2. 2 Operation and Maintainance Cost [20]
X2. 3 Replacement Cost [21]
X3. 1 Landscape Treatment [25]
X3. 2 Energy Use Efficiency [25]
GREEN X3. 3 Water Use Efficiency [25]
X3 RETROFITTING WORK
ASSESSMENT X3. 4 Air and Space Quality [25]
X3. 5 Waste Management [25]
X3. 6 Waste Water Management [25]
Y. 1 Net Present Value (NPV) [26]
INVESTMENT
Y. 2 Internal Rate of Return (IRR) [26]
FEASIBILITY
Y Y. 3 Benefit and Cost Ratio (BCR) [26]
Y. 4 Payback Period (PP) [27]
Page 443
SEMIT 2023
5. Result
Through the data provided by the management of the related case study building, there are
financial data inputs in the form of operation and maintenance costs for the first 3 years of green
certification using greenship certification by GBCI. These data consist of carbon dioxide (CO2)
sensors, ac retrocommissioning, clean water testing, addition of vertical gardens, laboratory testing for
waste, cfc - non cfc ac revitalisation, led bulbs, and three-category bins. This cost then underwent
regulatory adjustments in accordance with the regulations used during the research, initially using
GBCI regulations and then using the regulations of the Ministry of Public Works and Public Housing
number 21 of 2021.
In the adjustment process, there are findings of differences between regulations, which are located
in the duration of recertification for existing buildings and the parameters used. Regarding duration,
greenship adheres to a 3-year period for recertification, while the regulation of the minister of public
works and public housing no. 21 of 2021 adheres to a 4-year period for recertification. Also supported
by the parameters used, the regulation of the minister of public works and public housing no. 21 of
2021 uses four assessment parameters at the utilization stage and two assessment parameters at the
demolition stage. Since the case study building does not carry out the demolition process, researchers
only need to focus on the utilization stage.
In the process of distributing the first stage of the questionnaire to the experts, the green retrofitting
life cycle cost component can be determined into five main points referring to the regulatory standards
being used. It was found out that there is a dominance in the physical inspection of the building at the
utilisation stage, followed by the consideration of building components to be dismantled at the
demolition stage. These two points have a major impact on the realisation of green building conversion
in existing high-rise office buildings. In the third order there is an inspection of building technical
documents and inspection of operational procedures at the utilisation stage, in the fourth order there is
a consideration of the building at the demolition stage, and in the fifth order there are area requirements
and consideration of facilities and infrastructure at the demolition stage. All of these points illustrate
the priority scale from the highest to the lowest to then be considered by investors to implement green
retrofits in high-rise office buildings.
Furthermore, during the interview process with the building management, the most convenient
maintenance aspect that can be implemented at this time is the upgrading of the energy efficiency
aspect. Also referring to the literature study, the researcher decided to focus on the energy efficiency
aspect as the main data input into the research with a breakdown of costs as in the following table.
Table 2. Cost for Energy Eficiency Aspect (Source: Self-Administered)

Cost for Energy Eficiency Aspect
No Items Cost (Rp)
1 SOLAR PANEL 169,740,000
2 AC RETROCOMSSIONING 20,050,000
3 CO 2 SENSOR 3,458,800
4 LED LIGHT BULP 36,000,000
5 CO 2 SENSOR 1,076,000
8 CFC - NON CFC AC REVITASALISATION 260,000,000
9 CFC - NON CFC AC REVITASALISATION 326,000,000
Total 927,624,800
Page 444
SEMIT 2023
Since the investment time period used is for 10 years, researcher tried to combine the financing
data with the mandatory requirements that need to be done by investors in order to make their existing
building green. So there are additional financing items in the form of laboratory tests to identify
bacteria levels, and the addition of water taps. Refined with income costs in the form of base transceiver
stations (BTS) and leasing areas for automatic teller macchine (ATM), in this case the building
management as the data provider does not consider tenant rental costs due to unstable incoming costs.
The calculation of life cycle costs is simulated in the form of cahsflow with the effect of inflation
of 4.33% in accordance with the year of implementation of green retrofits in the case study of high-
rise office buildings. Through the data processing, it was found that the cost of green retrofit of high-
rise office buildings is Rp1,429,238,543.00 (one billion four hundred twenty nine million two hundred
thirty eight thousand five hundred forty three rupiah). Through the annual electrical energy usage data
provided by the building management, there are energy savings resulting from the implementation of
green retrofit in their high-rise office building. It is estimated that there is a 13% saving electrical
energy use by comparing the conventional existing building (before the green retrofit) with the green
retrofit existing building.
The value of investment feasibility is known in the form of internal rate of return (IRR) as the
value of the cost efficiency of an investment, minimum attractive rate of return (MARR) as the
minimum return interest rate, benefit cost ratio (BCR) as an information provider whether the
investment is feasible to continue or cause losses through the ratio, and net present value (NPV) as the
difference between the initial cashflow value and the current cashflow value whether positive or
negative.
Table 3. Investment Feasibility Value

IRR 20,72%
MARR 5,65%
BCR 1,77
NPV Rp599.106.581,42
The IRR value is greater than MARR, which means that the investment value is effective and
economical to implement. Then for the BCR value is more than one which means the investment is
profitable and can be implemented. And the last value of NPV is positive with the amount of
Rp599,106,581.42 (five hundred ninety-nine million one hundred and six thousand five hundred and
eighty-one rupiah). In addition, the data is also calculated to obtain the payback period value, the
positive value that appears after the entire negative value is in the 6th year. With the known amount of
costs and the feasibility value of green retrofit high-rise office building investment, the predicate BGH
Utama is obtained.
Through the analysis conducted on the second questionnaire, it is known that all the indicators of
benefits of green retrofit high-rise office buildings are accepted. This means that there are 13 non-
financial benefits experienced by building occupants, building management, and green building
consultants. However, there are no non-financial benefits obtained by the building owner (owner) as
an investor by not filling out the questionnaire by the building owner. benefits in the form of building
occupant comfort which can increase occupant productivity, increase body health stability, reduce
daily stress levels, increase cognitive intelligence, create a good working atmosphere, increase a sense
of comfort, a healthier work atmosphere, reduce daily stress levels, get a certificate to participate in
tenders, a cleaner office atmosphere, a greener and friendlier building for pedestrians and cyclists,
contribute to saving resources & environmental safety awareness, and increase energy efficiency.
However, through financial benefits, the owner is expected to benefit from year 6 to year 10 with a
Page 445
SEMIT 2023
total cost of Rp1,136,827,770.00 (one billion one hundred thirty six million eight hundred twenty
seven thousand seven hundred seventy rupiah) cumulatively.
6. Conclusion
The results of this research by applying the LCC method to the retrofit concept based on GBCI
and Permen PUPR No. 21 of 2021, are expected to increase the number of green retrofit buildings
and the awareness of building owners to implement green retrofits. It was found that using
government regulations will make it easier for most office buildings in Jakarta to obtain green
certification. In this case study, it costs Rp1,429,238,543.00 with 65% of the cost of customizing the
energy efficiency aspects. The resulting payback period in year 6 and a positive NPV value with a
10-year investment simulation means that it is quite effective. Through this calculation, a profit of
Rp1,099,268,970.00 is obtained and 13 benefits in the form of building occupant comfort which can
increase occupant productivity, increase body health stability, reduce daily stress levels, increase
cognitive intelligence, create a good working atmosphere, increase a sense of comfort, a healthier
work atmosphere, reduce daily stress levels, get a certificate to participate in tenders, a cleaner office
atmosphere, a greener and friendlier building for pedestrians and cyclists, contribute to saving
resources & environmental safety awareness, and increase energy efficiency. This green retrofit
concept needs to be supported economically in the form of incentives and socialization of the ease
and benefits of certification made by the government. there is a 13% saving electrical energy use by
comparing the conventional existing building (before the green retrofit) with the green retrofit
existing building.
Bibliography
[1] UNEP. (2021). GLOBAL STATUS REPORT FOR BUILDINGS
AND
CONSTRUCTION Towards a zero-emissions, efficient and resilient buildings and
construction sector. www.globalabc.org.
[2] Badan Pusat Statistik. (2022). STATISTIK INDONESIA 2022. BMKG.
[3] IPCC. (2022). Mitigation of Climate Change Climate Change 2022. www.ipcc.ch
[4] IEA. (2021). Buildings.
[5] F Purbantoro, & M Siregar. (2019). Design of Net Zero Energy Building (NZEB) for Existing
Building in Jakarta. Earth and Environmental Science.
[6] United Nations Framework Convention on Climate Change. (n.d.). The Paris Agreement.
[7] Hu, M. (2022). An evaluation of the retrofit net zero building performances: life cycle energy,
emissions and cost. Building Research and Information.
https://doi.org/10.1080/09613218.2022.2142497
[8] Direktorat EBTKE. (2022). PELUNCURAN PETA JALAN EFISIENSI ENERGI DI SEKTOR
BANGUNAN GEDUNG DAN KATALOG DESAIN NET ZERO ENERGY BUILDING
(NZEB).
[9] Wafa, M. A. (2020). Pertumbuhan Gedung Bertingkat di Jakarta.
[10] Gunawan, B., Budihardjo, Juwana, J. S., Priatman, J., Sujatmiko, W., & Sulistiyanto, T. (2012).
Buku Pedoman Energi Efisiensi untuk Desain Bangunan Gedung di Indonesia.
[11] WBP. (2012). Pengembang : Biaya Pembangunan Properti Hijau Lebih Mahal.
[12] Finaka, A. W., Negara, S. B., & Devina, C. (2021). Potensi Investasi Bangunan Hijau Ramah
Lingkungan.
Page 446
SEMIT 2023
[13] Purbantoro, F., & Siregar, M. (2019). Implementation of green building concept in office
building Jakarta. Journal of Physics: Conference Series, 1424(1), 012023.
https://doi.org/10.1088/1742-6596/1424/1/012023
[14] Oktaviani, A. D. (2021). PENGEMBANGAN METODE PRODUKSI PRECAST CONCRETE
BERBASIS TEKNOLOGI INTERNET OF THINGS PADA PEMBANGUNAN GEDUNG
BERTINGKAT TINGGI UNTUK MENINGKATKAN PRODUKTIVITAS PROYEK.
[15] CTBUH. (2022). Indonesia Overview.
[16] Kementerian ESDM. (2018). Peraturan Menteri ESDM Nomor 49 Tahun 2018 tentang
Penggunaan Sistem Pembangkit Listrik Tenaga Surya Atap, Tahun 2018.
[17] Urban Land. (2009). green retrofit definition.
[18] Heralova, R. S. (2017). Life Cycle Costing as an Important Contribution to Feasibility Study in
Construction Projects. Procedia Engineering, 196, 565– 570.
https://doi.org/10.1016/j.proeng.2017.08.031
[19] Fuller, Sieglinde. K., & Petersen, Stephen. R. (1996). Life Cycle Costing Manual For The
Federal Energy Management Program, NIST Handbook 135.
[20] Marszal, A. J., & Heiselberg, P. (2011). Life cycle cost analysis of a multi -storey residential
Net zero energy building in Denmark. Energy, 36(9), 5600–5609.
[21] Rakhmawati, S. (2022). Standar Benchmark Pengukuran Kinerja Investasi. Regína
[22] Utami, S. R. (2013). PENGEMBANGAN DESAIN GREEN BUILDING MENGGUNAKAN
STUDI VALUE ENGINEERING DALAM MEWUJUDKAN NEAR ZERO ENERGY
BUIILDING UNTUK MENINGKATKAN NILAI EKONOMI BANGUNAN (thesis).
[23] Blanchard, B. S., Verma, D., & Peterson, E. L. (1995). Maintainability: A key to effective
serviceability and Maintenance Management. Wiley.
[24] Supriyono, R. A. (1999). Akuntansi Biaya: Perencanaan dan Pengendalian Biaya serta
Pembuatan Keputusan. BPFE-UGM.
[25] Azahra, M. (2018). Efisiensi Energi dan Konservasi Energi, apa bedanya?
[26] Climate Transparency. (2021). CLIMATE TRANSPARENCY REPORT: COMPARING G20
CLIMATE ACTION TOWARDS NET ZERO.
[27] Handayani, T. (2010). EFISIENSI ENERGI DALAM RANCANGAN BANGUNAN Energy
Efficiency in Building Design (Vol. 1, Issue 2).
[28] Björnsdóttir, A.R. (2010). Financial Feasibility Assessments. Building and Using Assessment
Models for Financial Feasibility Analysis of Investment Projects.
Farr, J. V., & Faber, I. J. (2023). Engineering Economics of Life Cycle Cost Analysis.
https://doi.org/10.1201/9781003254782.
Page 447
SEMIT 2023
SP10: Other Fields of Study Related to EM and IT
The Impact of Technostress and COVID-19 Stress on Employee Burnout: The Mediating Role
of Resilience
Nour El Hoda Tarabah [0000-0001-7864-2394] and Uğur Şener [0000-0001-5524-1707]
1
Ibn Haldun University, Istanbul, Türkiye
2
Istanbul Aydın University, Istanbul, Türkiye
nour.tarabah@stu.ihu.edu.tr
Abstract
In today’s rapidly evolving work landscape, employees are challenged with many stressors that can
harm their job performance and overall well-being. The COVID-19 pandemic has put employees under
so many stressors compounding existing stressors faced by employees in many sectors. Compliance
with COVID-19 interventions has also accelerated the use of technological advancements which has
also put employees under pressure to adapt to these technologies to thrive in their jobs especially since
organizations are heavily relying on technology to carry out their daily tasks. Thus, a modern disease
caused by technological advancements “Technostress” can also occur in employees. Such stressors
might lead to burnout if the necessary resources are not built and fostered. In this matter, resilience can
play an important role in mediating these stressors. Guided by the theoretical framework of the Job
Demands-Resources (JD-R) model, this study explores how job demands (technostress and COVID-19
stress) and job resources (resilience) influence the context of these diverse industries.
It employs a survey adopted from existing scales about each variable to gather data from white-collar
workers in the real estate, medical tourism, and education sectors in Istanbul, and mediation analysis
was carried out. COVID-19 Stress and Technostress had a direct impact on Burnout. However,
resilience could only mediate the relationship between COVID-19 Stress and Burnout.
The findings of this study provide theoretical and practical implications to foster a resilient workforce
that can create a sustainable and supportive work environment that can benefit employees and
employers.
Keywords: Technostress, COVID-19 Stress, Employee Burnout, Resilience
1. Introduction
In today's interconnected and globally integrated world, individuals and employees encounter numerous
daily challenges and try to adjust to the requirements of this digitally connected and globally integrated
environment. In such a fast-paced and technology-driven world, white-collar workers in multiple
industries face many stressors that can highly impact their well-being and job performance. The
coronavirus pandemic has increased these challenges adding concerns and uncertainties about one’s
health, work-life balance, safety, and job performance. It brought many stressors to employees in many
aspects, and it caused “COVID-19 Stress”. Compliance with COVID-19 prevention has accelerated
digitalization more than ever before. Employees are no longer dealing with only familiar technologies
but with many new technologies that came front after the pandemic. Thus, among many other stressors,
technostress which is stress caused by technology and comes with many facets might increasingly arise
in such a time of rapid technological changes and integration and can be a critical factor affecting
employees' experiences.
Page 448
SEMIT 2023
The present research aims to investigate the complex relationships between technostress, COVID-19
stress, employee burnout, and resilience among English-speaking white-collar workers in the real estate,
tourism, and education sectors. Guided by the Job Demands-Resources (JD-R) model, this study
explores how job demands (technostress and COVID-19 stress) and job resources (resilience) influence
employee burnout in the context of these diverse industries. The research questionnaire assesses
participants' experiences of COVID-19 stress, technostress, employee burnout, and resilience levels.
COVID-19 stress is examined through worries related to the virus, while technostress is explored by
assessing participants' perceptions of technology's impact on their work. Employee burnout is evaluated
based on emotional exhaustion, depersonalization, and reduced personal accomplishment. Resilience,
as a mediating factor, is measured to understand its role in moderating the relationship between stressors
and burnout.
The findings from this research hold significant implications for both theoretical and practical domains.
The results will contribute to the existing literature on employee well-being, technostress, COVID-19
stress, and Organizational Behavior in general, offering insights into the unique challenges faced by
white-collar workers in the real estate, tourism, and education sectors. Additionally, understanding the
mediating role of resilience can provide valuable insights for organizations that aim to have
interventions to support employees' coping mechanisms, enhance their resilience, and alleviate burnout.
1.1. Theoretical Background and Hypotheses Development
The Job Demands-Resources (JD-R) model, which was formulated by Bakker and Demerouti in 2007,
provides a valuable framework for understanding the relationship between job characteristics, employee
well-being, and performance. They emphasize that work-related demands and resources have distinct
impacts on employees. According to their model, job demands refer to the psychological, physical, or
social organizational aspects of work that require continuous effort and are linked to specific costs
(Bakker & Demerouti, 2007). These demands can include time pressure, workload, role ambiguity, and
emotional labor. On the other hand, job resources are the psychological, physical, social or
organizational elements of a job that contribute to achieving goals, alleviating job demands and
fostering individual development and growth (Bakker & Demerouti, 2007). In the context of COVID-
19 stress and technostress, the JD-R model provides a relevant framework for understanding the impact
of these stressors on employee burnout. COVID-19 stress, resulting from the ongoing pandemic,
introduces additional demands on employees, such as fears of infection, uncertainties, and socio-
economic consequences (Taylor et al., 2020a; UNDP, 2020). Technostress, on the other hand, arises
from the challenges and negative consequences of individuals' interactions with technology (Brod,
1984). Resilience, as an individual asset, plays a vital role in mediating the relationship between job
demands and employee burnout within the JD-R model. Resilience refers to individuals' ability to
positively adapt and recover from adverse events (American Psychological Association, 2014). It
enables individuals to effectively cope with and bounce back from stressors, reducing the negative
impact on well-being and performance. By understanding the JD-R model and the mediating role of
resilience, organizations can implement strategies to mitigate the impact of COVID-19 stress and
technostress on employee burnout. This can include providing social support, promoting autonomy,
fostering positive work environments, and offering resilience-building interventions.
The formulation of the research topic and hypotheses was informed by conducting a comprehensive
literature review. Table 1 provides a summary of relevant studies that have investigated the relationship
between the study variables from various theoretical perspectives.
Page 449
SEMIT 2023
Table 6. Summary of Selected studies for Hypotheses Development

Study Hypotheses Variables
Shrom et al. (2020) H01
COVID-19 stress Employee burnout
Bakker et al. (2020)
Rosen et al. (2013) H02
Shen et al. (2014)
Schneider (2016) Technostress Employee burnout
Lee (2017)
Zohar and Luria (2017)
Mahapatra and Prakash Pati (2018) H03
Lin et al. (2020) COVID-19 stress Resilience
Feng et al. (2020)
O'Brien and Jones (2016)
Shih et al. (2018)
Al Kathani and Alghamdi (2021)
H04 Technostress Resilience
Azami et al. (2020)
Du and Liu (2020)
Kim and Kim (2021)
Yıldırım & Solmaz (2020) COVID-19 stress Resilience
H05
Tarabah (2021) Employee burnout
Feng et al. (2020)
Lin et al. (2020)
Wu and Chi (2020)
Zhang and Cheng (2020)
Kim and Lee (2021)
H06 Technostress ➡ Employee burnout
Shen et al., (2014)
Diller (2016)
Chori et al. (2021)
Schneider and Raucchenbach (2016)
Tarabah (2021)
Thus, this model along with the comprehensive literature review of relevant research has helped in
developing the following hypotheses and research framework (Fig.1).
Hypotheses:
H01: There is a significant relationship between COVID-19 Stress and Employee Burnout
H02: There is a significant relationship between Technostress and Employee Burnout
H03: There is a significant relationship between COVID-19 Stress and Resilience
H04: There is a significant relationship between Technostress and Resilience
H05: Resilience mediates the impact of COVID-19 Stress on Employee Burnout
H06: Resilience mediates the impact of Technostress on Employee Burnout
Resilience
COVID-19 Employee
Stress Burnout
Technostress
Fig. 8. Conceptual Model
Page 450
SEMIT 2023
2.1. Technostress
Technostress, as a concept, has evolved to capture the challenges and negative consequences arising
from individuals' interactions with technology in modern society. Being an interesting phenomenon,
many research studies were done to develop a better understanding of its inhibitors and consequences
(Li &Wang 2020; Ennis, 2005). It is a modern disease triggered by the incapability to deal with recent
technologies’ complexities and demands in a healthy way which can manifest in physiological and
psychological strain (Brod, 1084). It can occur in many forms such as Techno-overload, Techno-
invasion, Techno-complexity, Techno-insecurity, and Techno-uncertainty.
Techno-overload refers to the experience of being overwhelmed by a wide volume of technology-
related tasks, information, and communication. (Brod, 1984). Techno-invasion relates to the intrusion
of technology into personal and professional lives, blurring boundaries and creating a constant
connectivity that can be disruptive and invasive. (Ragunathan et al., 2008; Ayyagari, Grover & Purvis,
2011). Techno-complexity refers to the challenges and difficulties individuals face in using and
understanding new technologies such as frequent software updates (Tarafdar, Ragu-Nathan, & Ragu-
Nathan, 2007). Techno-insecurity reflects the feeling of vulnerability and lack of control over personal
or professional information in the digital realm which manifests in fear, anxiety, and uncertainty due to
technology-related risks, such as data breaches and cyber-attacks (Li & Wang, 2020; Ragu-Nathan,
2008). Techno-uncertainty refers to the constant change and unpredictability associated with technology
which leads to uncertainty in adapting to and coping with new technologies, as well as the anxiety
arising from the rapid pace of technological advancements (Chen & Lu, 2016).
These types of technostress creators contribute to the overall technostress experience, leading to
detrimental effects on individuals' well-being and performance. Understanding these different facets of
technostress is crucial for addressing the challenges posed by emerging technologies and promoting
healthy technology usage.
2.2. COVID-19 Stress
Coronavirus is a respiratory illness declared a pandemic by the World Health Organization
(WHO, 2020) has enormously impacted human beings, the economy, government systems, and
organizations worldwide. It has caused a new type of stress among many individuals denoted as
“COVID-19 Stress”. It is psychological and emotional distress experienced by individuals because of
the ongoing COVID-19 pandemic. This stress has evolved as the pandemic unfolded, starting with
initial concerns about the virus and its spread, followed by prolonged periods of uncertainty, isolation,
and significant changes to daily routines and social interactions.
COVID-19 Stress encompasses several dimensions that reflect the psychological and emotional impact
of the pandemic on individuals. These dimensions are Danger, 2 Socio-economic consequences,
xenophobia, contamination, 5 traumatic stress, and compulsive checking (Taylor et al., 2020a). Where,
danger involves fears and concerns about the risk of infection, doubts regarding healthcare systems'
ability to protect one's health and loved ones, and worries about the effectiveness of preventive measures
(Taylor et al.2020b). It can lead to depression, increased anxiety, and thoughts of self-harm (Vatansever
et al., 2020). Additionally, the stress regarding the socio-economic consequences is characterized by
fears of food and supply shortages, store closures, and the distressed economic and social impacts of
the pandemic (UNDP, 2020; WHO, 2020; Taylor et al., 2020b). However, Xenophobia arises from fears
of foreigners spreading the virus, including concerns about foreign food restaurants and hygiene (Taylor
et al., 2020a; Lin, 2020). However, contamination stress is associated with the fear of touching objects
in public areas and the avoidance of potentially contaminated surfaces (Luceño-Moreno et al., 2020;
Taylor et al., 2020a; Taylor et al., 2020b). Moreover, traumatic stress involves concentration troubles,
physical reactions, nightmares, negative mental images and thoughts concerning the virus, and sleep
disturbances. (Schredl & Bulkeley, 2020; Zhang et al., 2020; Wang et al., 2020). However, Compulsive
checking is manifested through seeking healthcare advice, reassurance from others, and constantly
checking COVID-19 related information (Bento et al., 2020; Taylor et al., 2020a; Taylor et al.2020b).
Page 451
SEMIT 2023
2.3. Resilience
Resilience, a multidimensional concept, has garnered significant attention in various fields,
including psychology, neuroscience, medicine, mental health, sociology, management, education,
biology, anthropology, and more. Extensive research has explored resilience in diverse disciplines, with
contributions from scholars such as Southwic (2014), Salehzadeh (2019), VanBreda (2001), Russo et
al. (2012), and McCormac et al. (2018). The term "resilience" has seen a significant increase in research
since the 1990s, with a growing number of scientific studies focusing on this topic (Longstaff, 2013).
It is defined in various ways in the literature. It refers to individuals' ability to positively adapt and
recover from adverse events (Britt et al., 2013; American Psychological Association, 2014). It involves
effective adaptation to trauma, adversity, tragedy, threats, or significant stressors (American
Psychological Association, 2014). Individuals' resilience is shaped by how they engage with their
surroundings, encompassing social, cultural, psychological, and biological elements (McCormac et al.,
2018). It is linked to thoughts, behaviors and actions that enable individuals to maintain normal physical
and psychological functioning in the face of extreme stress and trauma (Russo et al., 2012). In the
context of organizational behavior, resilience plays a crucial role in individuals' and teams' ability to
navigate challenges, adapt to change, and thrive against of adversity. Organizational resilience refers to
an organization's capacity to anticipate, respond to, and recover from disruptive events or challenges
(Lengnick-Hall et al., 2011). Thus, by promoting resilience at various levels, organizations can foster a
positive work environment, enhance employee well-being and engagement, and improve overall
organizational performance.
2.4. Employee Burnout
Employee burnout is a phenomenon that has gained significant attention in organizational research and
has evolved over time. The concept of burnout originated in the 1970s and was first coined in the
literature by psychologist Herbert Freudenberger as a state of chronic emotional and physical
exhaustion resulting from excessive work demands (Freudenberger, 1974). Since then, numerous
studies and theoretical frameworks have expanded our understanding of burnout. It is marked by
sensations of emotional, cognitive, and physical exhaustion along with pessimism towards one's job
due to severe stress (Gill et al., 2006). WHO (2019) defines burnout as an occupational phenomenon
caused by chronic workplace stress that is not effectively managed. One prominent theory is the
Maslach Burnout Model proposed by Christina Maslach and Susan Jackson, which highlights three
dimensions of burnout: emotional exhaustion, depersonalization, and reduced personal accomplishment
(Maslach et al., 2001). Emotional exhaustion refers to a state of increased exhaustion, negative attitudes
towards others, and a depletion of energy and emotional resources (Yıldırım & Taşmektepligi, 2012;
Ulutaşdemir, 2012). Depersonalization involves distancing oneself from work, perceiving it as
unimportant, and avoiding job responsibilities (Maslach, 1996; WHO, 2019). Reduced personal
achievement is characterized by negative self-evaluation and difficulties in meeting work-related
demands (Koçak & Gürsoy, 2018). These dimensions collectively capture the experience of burnout,
including exhaustion, cynicism, and reduced feelings of accomplishment. These dimensions
collectively contribute to employee burnout, highlighting the complex and multifaceted nature of this
phenomenon and its detrimental impact on both individuals and organizations.
2.5. Principal Component Analysis (PCA)
Principal Component Analysis is a widely used technique in statistics and machine learning for
dimensionality reduction and data analysis. Its origins are traced back to Karl Pearson (1901) and later
to Harold Hotteling (1930) and it is originated in the field o linear algebra and multi-variate statistics).
It is considered one of the most powerful and significant techniques in wealth of areas (Bro and Smilde,
2014). It is a linear dimensionality reduction technique that can transform dataset with correlated
variables into a new coordinate system and these variables are uncorrelated and ordered by their
significance, where the most variance is captured first (Hotelling, 1993; Joliffe, 2002). These principal
components are referred to as PC1, PC2 etc. In other words, PCA aims to extract the significant
dimensions and represent them as a set of new orthogonal variables referred to as “Principal
Components” where the patterns of similarity of observations and the variables are displayed as points
in maps (Abdi and Williams, 2010). PCA has four main goals (1) Dimensionality reduction (Jollife
Page 452
SEMIT 2023
2002), Decorrelation (Abdi and Williams, 2010), Variance Maximization (Jolifee, 2002), and
Simplification and Interpretability (Dunteman, 1989). These four goals can lead to the reduction of
number of variables while retaining the significant information in the process, transforming the original
variables into uncorrelated principle components, arranging principal components to capture the most
variability in the data, and lastly, simplifying the complex data for less complex interpretation and
understanding specially that, PCA helps in visualizing high-dimensional data through projecting them
onto a lower dimensional space (Wold et al., 1987). In the context of Business research, PCA can have
high significance as it can be applied in multiple disciplines including Employee Satisfaction (Sarros et
et al.2002), Supply Chain Optimization (Liu et al., 2005), Risk Management (Hull, 2015), Marketing
Research (Hair et al., 2010) and many other others. Given its benefits of Dimensionality reduction
(Hotelling, 1993), Visualization and Interpretability (Wold et al., 1987), Multicollinearity Mitigation
(Abdi and Williams, 2010), Feature Selection (Jollifee, 2002), improved generalization (Bishop, 2006)
and the simplification of complex data (Duneteman, 1989), PCA was carried out in this research.
3. Methodology
3.1. Sample
The sample consisted of English-speaking white-collar workers from the medical tourism, real estate,
and academic sectors. These sectors were chosen due to their face-to-face interactions with
customers/students and moderate risk during the pandemic. Also, these sectors encompass diverse
stressors, ranging from high-pressure sales targets in real estate, uncertainties due to travel restrictions
in tourism, to challenges in adapting to online teaching in education. Additionally, technology
increasingly integrates into these industries, employees are likely to experience technostress from rapid
technological changes and the need to adapt to digital tools. Moreover, studying English-speaking
workers allows for potential cross-cultural comparisons, enriching the study's applicability to a broader
audience. Out of approximately one million potential participants, data was collected from 420
respondents. After eliminating inconsistent responses, 355 respondents were considered. The
confidence level achieved was 94%, meeting the objective of a 95% confidence level with 384
respondents.
3.2. Data Collection
Online survey was conducted using a 5-point Likert scale to collect data from the targeted sample.
The survey included demographic questions and measures for the study's variables.
3.3. Measurement of Variables
For Technostress, we used Nimrod's (2018) Technostress Scale, consisting of 14 items and five
dimensions: overload, invasion, complexity, privacy, and inclusion. The COVID-19 scale adopted by
Taylor et al. (2020) comprised 36 items and five dimensions: (1) Danger and contamination fears, (2)
fears about economic consequences, (3) xenophobia, (4) compulsive checking and reassurance-seeking,
and (5) traumatic stress symptoms about COVID-19. To assess resilience, we employed the Brief
Resilience Scale (BRS) developed by Smith et al. (2008), which consisted of 6 items covering
dimensions of social relations, personal characteristics, health, and coping. For Employee Burnout, we
used the Oldenburg Burnout Inventory Scale developed by Demerouti et al. (2010), which consisted of
16 items and included dimensions of disengagement and exhaustion.
3.4. Data Analysis
The study focused on English-speaking white-collar workers in Istanbul, estimated to be 1-3
million in population. Convenience sampling was used to select 450 subjects, aiming for a 94%
confidence level. Data is collected via Google Forms, and after eliminating 95 respondents due to low
standard deviation, 355 data points are used for analysis. Statistical Package for the Social Sciences
(SPSS) with Process 3 Package was used for data analysis. Pearson correlation tested validity and
relationships between variables, while Cronbach's coefficient alpha assessed reliability. The research
model employs Hayes' (2017) simple mediation model with two independent variables (Technostress
and COVID-19STRESS), a mediator variable (Resilience), and a dependent variable (Employee
Page 453
SEMIT 2023
Burnout). Bootstrapping was used to test the paths in the model. Demographic information like age,
years of experience, employment level, educational level, and marital status is analyzed using frequency
statistics, while descriptive statistics were used for mean and standard deviation analysis.
4. Results
4.1. Principal Component Analysis (PCA)
Item reduction is conducted with PCA before calculating the variables from items. Firstly, all items in
a variable are forced into a single component in order to see their contributions, and then elimination is
done. Then PCA analysis is done with allowing multiple components for each variable. According to
results, multiple components are observed in the Covid-stress variable since it is expected considering
the multiple dimensions of the variable. Two components are observed in the resilience where one of
them resulted from the reverse questions which are eliminated from the variable. Component plot of
the resilience also (see Figure 3) supports this decision. Technostress has two components and also a
single item separated from them which is eliminated, and two components are evaluated as dimensions
of the variable similar with the resilience variable.
Table 7. KMO and Bartlett's Test: Resilience

KMO Sampling Adequacy. .825
Bartlett's Test of Sphericity Chi-Square 1124.2
Df 15
Sig. .000
The Kaiser-Meyer-Olkin Measure indicates a successful sampling process for analysis (0.825)
suggesting that there is strong partial correlation among the items. Bartlett's Test of Sphericity suggests
that the variables are significantly correlated (p < 0.001), making PCA analysis appropriate.
Table 8. Total Variance Explained: Resilience

Initial Eigenvalues Extraction Sums of Squared Loadings
Comp Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.591 59.854 59.854 3.591 59.854 59.854
2 1.116 18.602 78.456 1.116 18.602 78.456
Component 1 and Component 2 explain 59.854% of the variance, and 18.602% respectively and
together they explain a total of 78.456% of the variance in the data. The other components contribute
less to the overall variance and can be considered for emissions.
Table 9. Component matrix: Resillience

Component
Item 1 2
R6_R .808 .392
R1 -.794 .364
R2_R .788 .399
R4_R .779 .461
R3 -.768 .457
R5 -.702 .500
Page 454
SEMIT 2023
Fig. 9. Scree Plot: Resilience
The scree plot analysis shows a clear 'elbow' point at the second component. This indicates that the first
two components capture the most significant variance in the data. Thus, these two components are
retained.
Fig. 10. Component Plot in Rotated Space: Resilience
In the rotated component plot, two different clusters have emerged. R1, R3, and R5 are closely
associated with component 1. However, R2, R4, and R6 show strong loadings on component 2. This
indicates two distinct underlying patterns or constructs in the data.
Table 10. KMO and Bartlett's Test: Burnout

Bartlett's Test of Chi- 3079.2
Sphericity Square
Df 120
Sig. .000
The Kaiser-Meyer-Olkin Measure (KMO) indicates that the dataset is suitable for analysis suggesting
that there is enough shared variance among the variables. Bartlett's Test of Sphericity with a chi-square
value its significance suggests that the variables are significantly correlated, supporting the
appropriateness of PCA analysis. Burnout items represent two components but considering their
contributions to the variable when all items are forced to single component, both components are
included to the variable. The situation here is evaluated different from the case in the resilience.
Page 455
SEMIT 2023
Table 11. Total Variance: Burnout
Initial Eigenvalues
% of Cumulative
Comp Total Variance %
1 6.773 42.334 42.334
2 2.598 16.236 58.570
Components 1 and 2 account for 42.334% and 16.236% of the variance, respectively, and together
explaining 58.570% of the variance in the data. The other components (3 through 16) contribute less to
the overall variance. Thus, the first two components capture the majority of the variance in the dataset,
making them the primary focus for analysis, while the remaining components may have limited
relevance and could be considered for potential reduction or exclusion from the analysis.
Table 12. Component Matrix: Burnout

Component
Item 1 2
B3 .760
B16_R -.713 .418
B9 .706 .344
B12 .703 .416
B15_R -.702 .465
B7_R -.702 .457
B8 .695 .404
B6 .684 .338
B11 .675 .376
B4 .635 .427
B2 .616 .434
B5_R -.602 .414
B1_R -.588 .474
B14_R -.541 .484
B10_R -.532 .301
B13_R -.487 .336
Fig. 11. Scree Plot: Burnout
Page 456
SEMIT 2023
The scree plot analysis reveals a distinct "elbow" point at the second component, signifying a significant
shift in the eigenvalues' rate of decline. This, indicate that the first two components capture the most
significant variance in the data. Thus, these components are retained for further analysis.
Fig. 12. Component Plot in Rotated Space: Burnout
In the rotated component plot, two different clusters have emerged. B2, B3, B4, B6, B8, B9, B11, and
B12 are closely associated with component 1. However, B1, B5, B7, B10, B13, B14, B15, and B16
show strong loadings on component 2. This indicates two distinct underlying patterns or constructs in
the data.
Table 13. KMO and Bartlett's Test: Technostress

Df 91
Sig. .000
The Kaiser-Meyer-Olkin Measure (KMO) of 0.901 indicates that the dataset is highly suitable for
analysis, signifying a substantial amount of shared variance among the variables. Bartlett's Test of
Sphericity, with an approximate chi-square value of 3144.661 and a significance level of 0.000, strongly
supports the notion that the variables are significantly correlated, affirming the appropriateness of
conducting factor analysis on this dataset.
Table 14. Total Variance: Technostress

Initial Eigenvalues
Component Total % of Variance Cumulative %
1 6.531 46.648 46.648
2 2.218 15.844 62.492
Components 1 and 2 collectively explain 58.570% of the variance in the dataset, with individual
contributions of 42.334% and 16.236%, respectively. The subsequent components (3 through 16)
gradually contribute less to the overall variance. As such, the primary emphasis in the analysis should
be on the first two components, as they capture the majority of the variance. The remaining components,
from 3 to 16, offer less contributions and can be excluded from the analysis due to their limited
relevance.
Page 457
SEMIT 2023
Table 15. Component Matrix: Technostress

Component
Item
1 2
TS8 .817
TS7 .812 -.360
TS6 .805 -.403
TS3 .780
TS5 .738 .336
TS9 .722 .475
TS10 .719 .506
TS4 .708 .390
TS13 .675 -.354
TS11 .645 .537
TS14 .643 -.435
TS1 .498 -.415
TS2 .472 .467
TS12_R -.325
Fig. 13. Scree Plot: Technostress
The scree plot analysis shows a clear 'elbow' point at the second component. This indicates that the first
two components capture the most significant variance in the data. Thus, these two components are
retained.
Fig. 14. Component Plot in Rotated Space: Technostress
In the rotated component plot, TS2, TS10, TS11, TS9, TS4, and TS5 are tightly clustered in the upper-
right quadrant, leaning towards the left side, indicating strong intercorrelations among them. However,
TS3, TS8, TS7, TS13, TS14, TS1, TS6, and TS7 form another cluster in the upper-right quadrant but
are positioned closer to the lower right side, suggesting a separate set of interrelated variables. TS12
stands alone in the left-middle area, suggesting it has distinct characteristics and may not strongly
correlate with either of the two main clusters.
Page 458
SEMIT 2023
Table 16. KMO and Bartlett's Test: COVID-19 Stress

Df 630
Sig. .000
The Kaiser-Meyer-Olkin Measure (KMO) of 0.935 indicates that the dataset is highly suitable for
analysis, with significant shared variance among the variables. Bartlett's Test of Sphericity, with an
approximate chi-square value of 11719.127 and a significance level of 0.000, supports the belief that
the variables are significantly correlated, confirming the suitability for analysis.
Table 17. Total Variance: COVID-19 Stress

Initial Eigenvalues
Component
Total % of Variance Cumulative %
1 14.045 39.014 39.014
2 5.803 16.119 55.134
3 2.549 7.082 62.215
4 2.124 5.900 68.115
5 1.243 3.453 71.568
6 1.161 3.226 74.794
Components 1, 2, and 3 collectively explain 62.215% of the variance in the dataset, with individual
contributions of 39.014%, 16.119%, and 7.082%, respectively. Therefore, the primary focus in the
analysis should be on the first three components, as they capture the majority of the variance. The
remaining components, from 4 to 36, offer less contribution and may be considered for reduction or
exclusion from the analysis due to their limited relevance.
Table 18. Component Matrix: Covid-Stress

Component
Item 1 2 3 4 5 6
CS26 .766
CS25 .759 -.323
CS27 .749 -.318
CS29 .734
CS32 .727 -.300 -.327
CS28 .724 -.315
CS35 .723 -.370
CS22 .711 -.332 -.313
CS34 .690 -.396
CS15 .689 .319 -.522
CS18 .682 -.369
CS31 .682 -.369
CS23 .679 -.430 -.307
CS30 .669 -.371
CS33 .668 -.344
CS16 .653 -.515
CS14 .649 -.375
CS17 .648 -.449
CS24 .644 -.479
CS13 .598 -.500
CS5 .582 .499 .309
Page 459
SEMIT 2023
CS10 .574 -.558 .419

CS19 .568 .550
CS9 .563 -.559 .417
CS4 .557 .533
CS8 .556 -.519 .432
CS7 .549 -.449 .440
CS36 .497 .324
CS21 .573 .611
CS20 .502 .580 -.315
CS1 .544 .579
CS12 .504 -.554 .443
CS3 .406 .541 .384
CS6 .489 .539 .317
CS11 .452 -.499 .375
CS2 .459 .464 .304
Fig. 15. Scree Plot: COVID-19 Stress
The scree plot analysis shows a clear 'elbow' point at the third component. This indicates that the first
two components capture the most significant variance in the data. Thus, these three components are
retained.
Fig. 16. Component Plot in Rotated Space: COVID-19 Stress
In the component plot, the dimensions are arranged into distinct clusters. Dimensions clustered on the
right side are likely to have positive correlations and shared similarities, indicating strong relationships
or common underlying factors among them. Similarly, dimensions on the left side have similar
characteristics. The dimensions located along the middle line may have unique qualities or do not
strongly correlate with either of the main clusters. These patterns provide valuable insights into the
relationships among the dimensions in the dataset.
Page 460
SEMIT 2023
4.2. Descriptive Analysis
Table 19. Descriptive Statistics of Scales Factors

Variable Mean Std. Deviation Variance Skewness Kurtosis
COVID-19 Stress 1.2902 0.8189 0.671 0.947 0.493
Tecnostress 3.3305 0.8824 0.779 -0.053 -0.593
Burnout 2.8531 0.76891 0.591 0.244 -0.319
Resilience 3.3653 0.85852 0.737 0.008 -0.292
For COVID-19 Stress, the mean score is 1.2902, indicating a relatively low level of COVID-19 related
stress among the participants. The standard deviation (0.8189) suggests some variability in responses,
with a variance of 0.671. The skewness value (0.947) indicates a slightly positively skewed distribution,
while the kurtosis value (0.493) suggests a moderate peak in the distribution. For technostress, the mean
score is 3.3305, indicating a moderate level of technostress experienced by the participants. The
standard deviation (0.8824) shows some variability in responses, with a variance of 0.779. The
skewness value (-0.053) is close to zero, suggesting a relatively symmetric distribution, while the
kurtosis value (-0.593) indicates a flatter distribution. For burnout, the mean score is 2.8531, suggesting
a moderate level of burnout experienced by the participants. The standard deviation (0.76891) indicates
some variability in responses, with a variance of 0.591. The skewness value (0.244) indicates a slightly
positively skewed distribution, and the kurtosis value (-0.319) suggests a relatively flat distribution. For
resilience, the mean score is 3.3653, suggesting a moderate level of resilience among the participants.
The standard deviation (0.85852) indicates some variability in responses, with a variance of 0.737. The
skewness value (0.008) indicates a nearly symmetric distribution, and the kurtosis value (-0.292)
suggests a relatively flat distribution.
4.3. Mediation Analysis
Technostress is included as explanatory variable besides control variables according to the simple
mediation model with k antecedent variables of Hayes (2018). Mediation analysis is a statistical method
that tests if a third variable can influence the relationship (Ukah & Şener, 2021; Şener & Elkedra, 2020)
is done in this study using a regression-based bootstrap approach and the PROCESS macro for SPSS.
The used Simple Mediation Model of Hayes (2017:194) which is his Model 4, with k antecedent
explanatory variables.
Table 20. Simple Mediation Model First Relationship

Variable B T Sig LLCI ULCI R R-Sq F Sig
Resilience = B0 + B1Covid-19 Stress + Control variables + B8Technostress
Constant 3.485 14.683 .000 3.018 3.952 .632 .399 25.446
Covid-19 Stress -.171 -2.622 .009 -.299 -0.43
Gender .108 1.148 .252 -.077 .293
Age -.001 -.027 .979 -.098 .095
Education -.150 -2.440 .015 -.272 -0.29
Employment .037 .736 .462 -.063 .138
Experience .081 1.163 .246 -.056 .218
Marital Status .013 .112 .911 -.212 .237
Technostress* .059 .961 .337 -.062 .179
Burnout = B0 + B1 Covid-19 Stress + B2Resilience + Control variables + B9Technostress
Constant 3.920 18.592 0.00 3.505 4.335 .632 .399 25.446
Covid-19 Stress .043 .940 .348 -0.47 .134
Resilience -.511 -13.635 .000 -.585 -.437
Gender -.020 -.302 .764 -.149 .109
Age .015 .443 .658 -0.52 .082
Education .025 .585 .559 -0.60 .111
Employment -.051 -1.443 .150 -.121 .019
Page 461
SEMIT 2023
Experience .008 .170 .865 -.087 .104

Marital Status -.069 -.871 .384 -.226 .087
Technostress* .191 4.458 .00 .107 .275
Burnout = B0 + B1 Covid-19 Stress + Covariates + B8Technostress
Constant 2.139 10.433 .000 1.736 2.542 .274 .075 3.511
Covid-19 Stress .131 2.318 .021 .020 .242
Gender -.075 -.924 .356 -.235 .085
Age .016 .374 .709 -.068 . 099
Education .102 1.920 .056 -.002 .207
Employment -.070 -1.602 .110 -.157 .016
Experience -.033 -.551 .582 -.151 .085
Marital Status -.076 -.770 -.270 .118
Technostress .161 3.035 .003 .056 .265
Indirect Effect
Effect BootSE BootLLCI BootULCI
Resilience 0.087 0.032 0.028 0.153
In Table 15, the regression results of the first model are presented, examining the relationships between
COVID-19 Stress (X), Resilience (M), and Employee Burnout (Y), while controlling for demographic
variables and Technostress. The findings indicate that COVID-19 Stress has a significant negative effect
on Resilience (B=-0.171, p=0.009), showing that higher COVID-19 Stress is associated with lower
Resilience levels. Among the demographic variables, only education has a significant negative impact
on Resilience (p=0.0160), suggesting that higher education is linked to decreased Resilience. In the
second regression equation, Employee Burnout (Y) is regressed over COVID-19 Stress (X) and
Resilience (M). The results reveal that both COVID-19 Stress (B=0.043, p=0.348) and Resilience (B=-
0.511, p=0.0000) significantly influence Employee Burnout. Higher COVID-19 Stress is associated
with increased Burnout, while higher Resilience is associated with decreased Burnout. Technostress
also significantly predicts Employee Burnout (B=0.191, p=0.00). In the third regression equation,
Employee Burnout (Y) is regressed solely over COVID-19 Stress (X). The findings indicate that
COVID-19 Stress has a significant positive effect on Employee Burnout (B=0.131, p=0.021), meaning
higher COVID-19 Stress is associated with increased Burnout. While none of the demographic variables
are significant predictors of Burnout, education shows a borderline significance (p=0.0503). The
indirect effect of Resilience on the relationship between COVID-19 Stress and Employee Burnout is
estimated to be 0.087. The bootstrapped standard error (BootSE) for this indirect effect is 0.032, and
the 95% bootstrap confidence interval (BootLLCI to BootULCI) ranges from 0.028 to 0.153. These
results suggest that Resilience partially mediates the relationship between COVID-19 Stress and
Employee Burnout. The indirect effect indicates that a higher level of Resilience is associated with a
reduction in the impact of COVID-19 Stress on Employee Burnout. The confidence interval indicates
the range within which the true indirect effect is likely to lie, and it does not include zero, further
supporting the significance of the mediation effect.
Table 21. Simple Mediation Model Second Relationship

LL
Variable B t Sig. ULCI R R-Sq F Sig
CI
Resilience = B0 + B1Technostress + Control variables + B8Covid-19 Stress
3.01 .219 .048 2.170
Constant 3.485 14.683 .000 3.952
8
-
Technostress .059 .961 .337 .179
.062
-
Gender .108 1.148 .252 .293
.077
Page 462
SEMIT 2023
-
Age -.001 -0.27 .979 .095
.098
-
Education -.150 -2.440 .015 -.029
.272
-
Employment .037 .736 .462 .138
.063
-
Experience .081 1.163 .246 .218
.056
-
Marital Status .013 .112 .911 .237
.212
-
Covid-19 Stress -.171 -2.622 .009 -.043
.299
Burnout = B0 + B1 Technostress + B2Resilience + Control variables + B9Covid-19 Stress
3.50 .632 .399 25.45
Constant 3.920 18.592 .000 4.335
5
Technostress .191 4.458 .000 .107 .275
-
Resilience -.511 -13.635 .000 -.437
.585
-
Gender -.020 -.302 .763 .109
.149
-
Age .015 .443 .658 .082
.052
-
Education .025 .585 .559 .111
.060
-
Employment -.051 -1.443 .150 .019
.121
-
Experience .008 .170 .865 .104
.087
-
Marital Status -.069 -.871 .384 .087
.226
-
Covid-19 Stress .043 .940 .348 .134
.047
Burnout = B0 + B1 Technostress + Covariates + B8 Covid-19 Stress
1.73 .274 .075 3.511
Constant 2.139 10.433 .000 2.542
6
Technostress .161 3.035 .003 .056 .265
-
Gender -.075 -.924 .356 .085
.235
-
Age .016 .374 .709 .099
.068
-
Education .102 1.920 .056 .207
.002
-
Employment -.070 -1.602 .110 .016
.157
-
Experience -.033 -.551 .582 .085
.151
-
Marital 2Status -.076 -.770 .442 .118
.270
Covid-19 Stress .131 2.318 .021 .020 .242
Indirect Effect
BootLL
Effect BootSE BootULCI
CI
-.030
Resilience 0.033 -.094 0.037
Page 463
SEMIT 2023
In the first regression equation, Technostress is regressed over Resilience. The unstandardized
coefficient for Technostress is 0.059, with a p-value of 0.337, indicating that Technostress is not
significant in predicting Resilience. Among the demographic variables, only education is significant
(p=0.015) in this equation. In the second regression equation, Employee Burnout is regressed over
Technostress and Resilience. Technostress shows a significant positive effect on Employee Burnout,
with an unstandardized coefficient of 0.191 and a p-value of 0.000. Resilience also has a significant
negative effect on Employee Burnout, with an unstandardized coefficient of -0.511 and a p-value of
0.000. None of the demographic variables are significant in this equation. In the third regression
equation, Employee Burnout is regressed solely over Technostress. The unstandardized coefficient for
Technostress is 0.161, with a p-value of 0.003, indicating that Technostress is significant in predicting
Employee Burnout. Among the demographic variables, education shows a borderline significance
(p=0.056). The bootstrap confidence intervals were estimated for better accuracy. The results of the
indirect effect of Resilience in the relationship between Technostress and Employee Burnout were (-
0.094, 0.037), indicating that Resilience is not a significant mediator in this relationship.
5. Findings and discussion

The findings from the regression analysis shed light on the relationships between COVID-19 Stress,
Technostress, Resilience, and Employee Burnout among the studied population of English-speaking
white-collar workers in Istanbul. The results show that COVID-19 Stress and Technostress both have
a significant impact on Employee Burnout, thus, H01 and H02 are accepted. These findings are also
supported by other research works such as Shirom et al. (2020), Bakker et al. (2020) Rosen et al. (2013),
Shen et al. (2014), Schneider (2016), Lee (2017), Zohar and Luria (2017), and Mahapatra & Prakash
Pati (2018). Also, COVID-19 Stress has a significant impact on Resilience which is also supported by
another research such as Mahapatra & Prakash Pati (2018), Yıldırım & Solmaz (2020), Feng et al.
(2020) and Lin et al. (2020). However, Technostress has no significant impact on Resilience which is
supported by O’Brein and Jones (2016), Shih et al. (2018), Al Kathani and Alghamdi (2021), Azami et
al. (2020), Du and Liu (2020), and Kim and Kim (2021). Thus, H03 is accepted and H04 is rejected.
Moreover, Resilience could mediate the impact of COVID-19 Stress on Employee Burnout which is
supported by othe research such as Yıldırım & Solmaz (2020), Feng et al. (2020), Lin et al. (2020), Wu
and Chi (2020), Zhang and Cheng (2020), and Kim and Lee (2021). However, resilience couldn’t
mediate the impact of Technostress on Employee Burnout which contradicts other research such as
Shen et al. (2014), Diller (2016), Du and Liu (2020) Chori et al. (2021), and Schneider and
Rauschenbach (2016). Thus, H05 is accepted but H06 is rejected.
In this matter, COVID-19 Stress has a significant effect on both Resilience and Employee Burnout.
Higher COVID-19 Stress levels are associated with lower Resilience, which, in turn, contributes to
increased Employee Burnout. This suggests that the ongoing pandemic has affected employees' ability
to cope with stress and has resulted in higher levels of burnout.
Interestingly, Technostress does not show a significant relationship with Resilience. This may imply
that while employees might experience stress due to technological changes and innovations in their
work, it does not directly affect their ability to bounce back and maintain resilience. However,
Technostress does have a significant direct effect on Employee Burnout, indicating that technological
challenges and demands in the workplace can contribute to burnout among employees.
The role of Resilience is particularly noteworthy in this study. Resilience acts as a partial mediator
between COVID-19 Stress and Employee Burnout. This suggests that individuals with higher levels of
resilience may be better equipped to handle the stress brought on by the pandemic, leading to lower
burnout levels. Building resilience among employees should be a priority for organizations, as it can
potentially buffer the negative effects of stressors and contribute to better well-being and performance.
Page 464
SEMIT 2023
6. Implications and Future Research Directions

This study provides empirical proof of the relationship between COVID-19 Stress, Technostress,
Resilience, and Employee Burnout, offering practical implications for organizations to support their
workforce effectively. By exploring the role of resilience as a mediator factor, this study provides
insight into how an employee’s ability to respond to such challenges faced at the workplace and how it
might affect his/her performance and relationship with his/her work. Understanding these specific
stressors and resilience factors can inform policymakers and managers to tailor interventions and
support systems to enhance employee well-being, job satisfaction, and productivity. The research is
timely and can contribute to the ongoing efforts to understand and address the challenges met by
employees in these sectors during challenging times. The research can offer insights into how
organizations can effectively manage technological changes to support employee well-being and
productivity. As the world is also changing and there is an obvious shift towards technological advances,
this research can shed light on the importance of managing technostress not only to maintain employees’
well-being but also to keep them and their organizations agile and competitive in such a complex and
rapid environment. From a theoretical perspective, the results will contribute to existing literature on
employee well-being, technostress, COVID-19 stress, and Organizational Behavior in general, offering
insights into the unique challenges faced by white-collar workers in the real estate, tourism, and
education sectors by incorporating multiple variables that can affect employee’s wellbeing. Also,
building on the JD-R model, suggest that one of the demands and resources one can have at his/her job
are psychological and in this research, these demands manifest by COVID-19 Stress and Technostress,
however, resilience manifest in job resource which can help understand how can this relationship occur,
align with the research model’s predictions and strengthens the validity of the conclusion.
Future research could build on the JD-R model to find additional resources that may mediate the impact
of Technostress on Employee Burnout, such as genetic factors and IT knowledge, and organizational
support to better understand individuals' coping mechanisms with technological stress. Especially that
Technostress had a direct impact on Employee Burnout and resilience did not play a role in mediating
this impact. Other theoretical frameworks adopted from the strategic management field can also be
incorporated to reflect on the managerial role in this relationship. Also, it can be held in multiple settings
to allow for more cross-cultural research. Future research can also conduct comparison research
between Locals and Foreigners residing in Turkey to investigate how some other factors can play a role
in this relationship.
7. Conclusion
This research sheds light on the complex relationship between COVID-19 Stress, Technostress,
Resilience, and Employee Burnout. It builds its theoretical background on the JD-R model to explore
how a resource like resilience can help mediate COVID-19 Stress and Technostress's impact on
Employee Burnout in the real estate, tourism, and education sectors in Istanbul. It highlights the
relevance of this model in explaining the dynamics of COVID-19 Stress, Technostress, Resilience, and
Employee Burnout in contemporary workplaces. The findings highlight that COVID-19 stress and
technostress act as significant job demands, imposing additional pressures on employees. However,
resilience emerges as a crucial job resource, playing a mediating role in mitigating the impact of only
COVID-19 stress on burnout. It contributes to the understanding of employees’ well-being in the face
of challenges like the rapid technological environment and the COVID-19 pandemic. In conclusion, the
study’s insight has important theoretical and practical implications and can lead to better and sustainable
work environments and improved employees’ well-being, performance, and overall health.
References
1. Abdi, H., & Williams, L. J. (2010). Principal component analysis. Wiley Interdisciplinary
Reviews: Computational Statistics, 2(4), 433-459.
2. Alqahtani, A., & Alghamdi, M. (2021). The impact of technostress on psychological well-being:
The moderating role of dispositional resilience. Human Resource Management Review, 31(1),
77–86.
Page 465
SEMIT 2023
3. Alter, O., Brown, P. O., & Botstein, D. (2000). Singular value decomposition for genome-wide
expression data processing and modeling. Proceedings of the National Academy of Sciences,
97(18), 10101-10106.
4. Azami, M., Zohar, D., & Luria, G. (2020). Technostress and well-being: The role of coping
strategies and social support. Computers in Human Behavior, 111, 106568.
5. Bakker, A. B., Hakanen, J. J., Demerouti, E., Xanthopoulou, D., & Schaufeli, W. B. (2020).
Burnout and work engagement during the COVID-19 pandemic. Burnout Research, 100169.
6. Bishop, C. M. (2006). Pattern recognition and machine learning. springer.
7. Brod, C. (1984). Technostress: The Human Cost of The Computer Revolution. Boston, Addison-
Wesley.
8. Chori, D. I. M., Al-Marwaziyyah, K., & Hasanati, N. (2021). What is the Predictor Variable of
Burnout? Burnout Literature Review. American Research Journal of Humanities & Social Science
(ARJHSS), 2(1), 108–113.
9. Demerouti, E., Mostert, K., & Bakker, A. B. (2010). Burnout and work engagement: A thorough
investigation of the independency of both constructs. Journal of Occupational Health Psychology,
15(3), 209–222.
10. Diller, H. (2016). Life is tough so you gotta be rough: How resilience impacts employees' attitude
towards ICT use (No. B-24-16). Passauer Diskussionspapiere-Betriebswirtschaftliche Reihe.
11. Du, L., & Liu, J. (2020). Technostress and well-being: The moderating role of mindfulness.
Human Resource Management Review, 30(4), 553–563.
12. Dunteman, G. H. (1989). Principal components analysis. Sage Publications.
13. Ennis, L. A. (2005). The evolution of technostress. Computers in Libraries, 25(8), 10–12.
14. Ercan, H. (2017). The Relationship between Resilience and the Big Five Personality Traits in
Emerging Adulthood. Eurasian Journal of Educational Research, 70, 83–103.
15. Feng, X., Chen, X., Ma, L., Zhao, J., & Li, Z. (2020). The impact of COVID-19 on resilience: A
study of Chinese employees. Stress and Health, 36(4), 773–782.
16. Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2010). Multivariate data
analysis (6th ed.). Pearson.
17. Hayes (2017). Introduction to Mediation, Moderation and Conditional Process Analysis. New
York, THE GUILFORD PRESS.
18. Hotelling, H. (1933). Analysis of a complex of statistical variables into principal components.
Journal of Educational Psychology, 24(6), 417-441.
19. Hull, J. C. (2015). Risk management and financial institutions (4th ed.). Wiley.
20. Jolliffe, I. T. (2002). Principal component analysis (2nd ed.). Springer.
21. Kang, L., Li, Y., Hu, S., et al. (2020). The Mental Health of Medical Workers in Wuhan, China
Dealing With The 2019 Novel Coronavirus. The Lancet Psychiatry, 7(3).
22. Kim, H., & Kim, S. (2021). Technostress and job performance: The mediating role of work-life
balance. Journal of Business Research, 118, 314–324.
23. Kim, J., & Lee, J. (2021). The effect of COVID-19 on psychological resilience: A moderated
mediation model. Stress and Health, 37(1), 29–38.
24. Lee, Y., & Kim, M. (2017). Technostress and its effect on job performance: A moderated
mediation model. International Journal of Human Resource Management, 28(7), 1229–1246.
25. Li, L., & Wang, X. (2020). Technostress inhibitors and creators and their impacts on university
teachers’ work performance in higher education. Cognition, Technology & Work, 24.
Page 466
SEMIT 2023
26. Li, Y., & Wang, G. (2020). The impact of techno-stressors on users’ adoption of digital services:
An empirical investigation. International Journal of Information Management, 51, 102019.
27. Lin, S., Chen, Y., & Lin, H. (2020). The impact of COVID-19 on mental health and well-being:
A systematic review. Journal of Occupational Health Psychology, 25(4), 372–394.
28. Liu, L., Meade, L. M., & Cao, J. (2005). A visualization tool for supply chain performance
measurement using the balanced scorecard and data envelopment analysis. Computers &
Industrial Engineering, 49(4), 530-544.
29. Mahapatra, M., & Pati, S. P. (2018). Technostress creators and burnout: A job demands-resources
perspective. In Proceedings of the 2018 ACM SIGMIS conference on computers and people
research, Buffalo-Niagara Falls NY USA, 18–20 June, 70–77.
30. Maslach, C., Schaufeli, W. B., & Leiter, M. P. (2001). Job Burnout. Annual Review of
Psychology, 52, 397–422.
31. Nimrod, G. (2018). Technostress: Measuring a new threat to well-being in later life. Aging &
Mental Health, 22(8), 1086–1093.
32. O'Brien, K., & Jones, D. (2016). Technostress and work engagement: The mediating role of work-
life balance. Human Resource Management Review, 26(4), 379–391.
33. Pearson, K. (1901). On lines and planes of closest fit to systems of points in space. Philosophical
Magazine, 2(11), 559-572.
34. Ragu-Nathan, T. S., Tarafdar, M., Ragu-Nathan, B. S., & Tu, Q. (2008). The Consequences of
Technostress for End Users in Organizations: Conceptual Development and Empirical Validation.
Information Systems Research, 19(4), 417–433.
35. Rosen, L. D., Whaling, K., Rab, S., Carrier, L., Cheever, N. A., Rokkum, J., & Carrier, L. M.
(2013). Technostress: The human cost of technology. In Technostress (pp. 1–5). IGI Global.
36. Sarros, J. C., Cooper, B. K., & Santora, J. C. (2002). Building a climate for innovation through
transformational leadership and organizational culture. European Journal of Innovation
Management, 5(1), 6-18.
37. Schneider, T., & Rauschenbach, P. (2016). Technostress in the digital age: Conceptualization,
antecedents, and consequences. Human Resource Management Review, 26(4), 363–378.
38. Şen, E., & Tarabah, N. E. H. (2020a). Döngüsel Ekonomi, Sürdürülebilirlik Ve İnovasyon. In
Sayın F. (Ed.), Döngüsel Ekonomi Makro Ve Mikro İncelemeler (pp. 67–85). Nobel Yayın,
Ankara.
39. Şen, E., & Tarabah, N. E. H. (2020b). COVID-19 ve Dijital Devrim Bağlamında Stratejik Zeka
ve Paydaş Yönetimi. In Şen, E., Hidiroğlo, D., & Yilmaz, O. (Eds.), COVID-19 Pandemisinde
Yönetim ve Ekonomi (pp. 439–458). Gazikitabevi, Ankara.
40. Şen, E., & Tarabah, N. E. H. (2020c). Girişimcilik, Dijitalleşme ve Örgütsel Çeviklik. In Karadal,
H., Halis, M., & Mert, G. (Eds.), Girişimcilik & Liderlik Güncel Gelişmeler (pp. 83–101). ATİ
Akademi Titiz Yayınlar, Istanbul.
41. Şen, E., & Tarabah, N. E. H. (2020d). Knowledge Management and Corporate Governance within
COVID-19 Period. In Mert, G., Şen, E., & Yılmaz, O. (Eds.), Data, Information and Knowledge
Management (pp. 541–556). Nobel Bilimsel Eserler, Istanbul.
42. Şen, E., & Tarabah, N. E. H. (2021). Neuroscience, Mental Health and Senism in The Post-
COVID-19 Era. In Şimşek, H. & Mečiar, M. (Eds.), The Social and Economic Impact of Covid–
19: Rapid Transformation of the 21st Century Society (pp. 97–118). IJOPEC Publication Limited,
London.
Page 467
SEMIT 2023
43. Şener, U., & Abrouk Elkedra, A. (2020). The Mediating Role Of Social Media And Customer
Engagement In The Impact Of Digital Content Marketing On Brand Awareness. International
Research Journal of Marketing & Economics, 7(11), 1–11.
44. Shen, W., Chi, N., & Wang, D. (2014). Technostress: A new challenge for human resource
management. Journal of Applied Psychology, 99(3), 501.
45. Shih, H., Chen, C., & Ho, Y. (2018). Technostress and well-being: Examining the mediating role
of coping strategies. Journal of Applied Psychology, 103(8), 935.
46. Shirom, A., Shamai, S., & Melamed, S. (2020). COVID-19 and burnout syndrome among
employees: A large-scale study. Journal of Occupational Health Psychology, 25(4), 415–424.
47. Smith, B. W., Dalen, J., Wiggins, K., Tooley, E., Christopher, P., & Bernard, J. (2008). The brief
resilience scale: Assessing the ability to bounce back. International Journal of Behavioral
Medicine, 15(3), 194–200.
48. Tarabah, N. E. H. (2021). The Impact of Technostress and COVID-19 Stress on Employee
Burnout Among Employees in Turkey Under the Mediating Role of Resilience. Master's thesis,
Istanbul Aydın University, Turkey.
49. Taylor, S., Landry, C. A., Paluszek, M. M., Fergus, T. A., Mckay, D., & Asmundson, G. J. G.
(2020). Development and Initial Validation of The COVID Stress Scales. Journal of Anxiety
Disorders, 72, 1–7.
50. Ukah, O. E., & Şener, U. (2021). Mediation of Online Brand Community on the Relationship of
Social Media and Purchase Intention. International Journal of Scientific Research and Engineering
Development, 4(6), 183–192.
51. WHO. (2019). Burn-out an "occupational phenomenon": International Classification of Diseases.
Retrieved from https://www.who.int/news/item/28-05-2019-burn-out-an-occupational-
phenomenon-international-classification-of-diseases (accessed 05 March 2022)
52. WHO. (2020). WHO Characterizes COVID-19 as a Pandemic. Retrieved from
https://www.who.int/emergencies/diseases/novelcoronavirus2019/events-as-they-happen
(accessed 05 March 2022)
53. Wold, S., Esbensen, K., & Geladi, P. (1987). Principal component analysis. Chemometrics and
Intelligent Laboratory Systems, 2(1-3), 37-52.
54. Wu, C., & Chi, N. (2020). COVID-19 and burnout among healthcare workers: A moderated
mediation model. Journal of Occupational Health Psychology, 25(4), 425–434.
55. Yıldırım, M., & Solmaz, F. (2020). COVID-19 burnout, COVID-19 and resilience: Initial
psychometric properties of COVID-19 Burnout Scale. Death Studies, 1–9.
56. Zhang, Y., & Chen, X. (2020). Resilience in the face of COVID-19: A study of Chinese
employees. Journal of Occupational Health Psychology, 25(4), 435–448.
57. Zohar, D., & Luria, G. (2018). The role of technostress in the relationship between the use of
communication technologies and work outcomes. Computers in Human Behavior, 85, 1–9.
Page 468
SEMIT 2023
EmpiricalAnalysis of Profit-Intended Vehicle FleetMaintenance Schemes in Developing

1
B.O OZERO, S.O.ENIBE2 and *P.A. Ako (formerly Ozor).2,3
1
Shell Petroleum Development Company of Nigeria Limited, Shell Industrial Area, Rumuobiakani,
Port Harcourt Nigeria
2
Nsukka, Nigeria)
3
Department of Quality and Operations Management
Faculty of Engineering and Built Environment, University of Johannesburg, P. O. Box 534 Auckland
Park, South Africa
(B.O. Ozero: ozero4real@gmail.com ): (SO Enibe: Samuel.enibe@unn.edu.ng)
*Corresponding Author: pozor@uj.ac.za; paul.ozor@unn.edu.ng
Abstract
Economic uncertainties, galloping inflation, and the overarching demand for customer safety as well as
enterprise profitability underscore the importance of proper analyses to evaluate the maintenance
decisions of service-producing systems. Whilst the maintenance planning of profit-intended equipment
can be done on the principles of minimum cost, appropriate models that strike a good balance between
reliability and cost can lead to the best maintenance policy for such systems. In this work, the existing
maintenance procedure for a typical commercial vehicle fleet Company in Nigeria, called T-tours, for
confidentiality is studied. The Company’s prevalent operation processes and control, maintenance costs,
the effect of climatic variations, and inflation were obtained through document analysis, observation, and
use of interviews with the opinion leaders. The resulting data provided the foundation for an empirical
investigation, cost model formulation, and application of appropriate quantitative techniques that
minimize the operation costs as well as determine the useful life span of the vehicles. The results show
that a commercial vehicle, typical of the Toyota Hiace bus, operated by the studied company should be
replaced after about 5 years. The result also shows that within this period, the vehicle would have fully
recouped the capital invested in it.
Keywords: Empirical Analysis, Vehicle Fleet Maintenance, Minimum Cost, quantitative models, Useful
life
Nomenclature
A = Optimal Cost variable

C = Purchase price of a vehicle
c p = Average fuel price [Naira/Litre]
d = Depreciation value or discount rate per naira during one year

i = Interest rate
M = Annual utilization intensity or Mileage of the vehicle in km
MC = Maintenance, repair and other miscellaneous costs
n = Number of years of exploitation of a vehicle
R n = Running cost of vehicle at the beginning of the nth year
r = Annual interest rate
Page 469
SEMIT 2023
S = Salvage value of vehicle

Z p = Average fuel consumption of a vehicle in consecutive years of exploitation [Litre/Km]
1. Introduction
Vehicle fleet Companies has experienced sharp decline of vehicle re-sale values, and increase of vehicle
insurance premiums in some developing Countries. Recent economic climate requires that the industry
aim at achieving maximum customer satisfaction while minimizing associated costs. There exist ongoing
technological progresses that stress the significance of properly performing economic analyses to
evaluate maintenance policies of equipment subject to increasing failure rate (due to aging) or usage
(Ozor et al, 2017). The time long theory in economic analysis for equipment maintenance programmes is
the depreciation over time due to wear and age (Karsak and Tolga, 1997). A service producing system
(SPS), such as the vehicle, becomes obsolete before its anticipated life expires. The motor vehicle is an
example of SPS used in transportation industry. Today’s economic climate requires that the transportation
industry aim at achieving maximum customer satisfaction while minimizing associated maintenance
costs. This means finding ways to maximize vehicle reliability and up-time. Although impressive
progress has been made in maintaining vehicles in developed countries in an effective manner,
maintenance of vehicles in Nigeria still leaves much to be desired due to factors such as size, cost,
complexity, breakdowns, unscheduled maintenance, poor road conditions, and competition (Eti, Ogaji
and Probert, 2004).
Most SPS get older with time and their failure rate keep increasing. This can lead to frequent failures and
breakdowns. In particular, the vehicle system can undergo minor failures as well as catastrophic ones
(Badia and Berrade, 2007). When the former take place, it returns to the right operating state with slight
repairs and at low costs, even if the failure causes the vehicle to stop working. Consider for instance a
part failure that presents some difficulties to work due to lack of lubrication, or a battery that does not
properly supply power. The problem is easily solved after testing the oil or the battery level. The
catastrophic failures are those that cause the vehicle to stop working properly and require major repairs
and high costs to fix them. Often, a replacement of the whole unit or a perfect repair that restores the part
to an as-good as- new condition must be carried out. There is also a difficulty in differentiating the failure
cause and effect. Each component or part of the vehicle has a certain inherent reliability and is exposed
to external or internal failure events. These components form a reliability network that defines the total
reliability of the vehicle system. There may be an almost infinite number of dynamic, static, operational,
logistical, electrical, mechanical and chemicalconnections that define the interdependencies of the vehicle
network. The failure event ofone component may cause failure effects that look as failure events to other
components.
The failure event may therefore propagate through the vehicle system causing a flow of failures for the
following components. These failures eventually increase the maintenance and running costs
sympathetically. The cost of maintenance is defined as costs that include lost opportunities in uptime,
rate, yield, and quality due to no operating or unsatisfactorily operating vehicle in addition to costs
involved with the vehicle-related degradation of the safety of people, property, and the environment
(Dhillon, 2002). The author asserts that cost of maintaining equipment often varies from 2 to 20 times
the acquisition cost. Also, the breakdown of the vehicle does not permit it to be used for transportation
purposes which is a loss (downtime cost) to the company. Hence the vehicle should be maintained by
scheduling a maintenance action to inspect the vehicle for faults and to remove all failed parts. A time
may eventually be reached when it is no longer economical to use the vehicle as the maintenance and
other costs now outweigh the cumulated profits from theuse of the vehicle and hence a replacement of
the vehicle becomes the best course of action.
According to Yoshinori & Gregory (2005), one factor that affects the vehicle companies profitability is
how they manage their vehicle replacement cycles (i.e., what age of vehicle to purchase and for how long
to use it). Theoretically, for every possible situation, there is a unique replacement cycle that maximizes
Page 470
SEMIT 2023
the vehicle life-cycle profit. In general, there are two fundamental replacement strategies (cases):
replacement on failure and preventive replacement (Redmer, 1999). The first category of replacement
strategies can be applied to systems that deteriorate with time or not, assuming that in both cases we can
finally expect a failure. The second strategy can be applied when the following conditions are met, namely
the total cost associated with a failure replacement is greater than the total cost associated with a
preventive maintenance. The hazard rate of the equipment is increasing with the time of utilization. In
case of both strategies, the general problem is how to determine the optimal time of utilization to
replacement or optimal cumulative usage of vehicle to replacement.
Many factors influence maintenance costs, including asset condition (i.e., age, type, and status), operator
expertise and experience, company policy, type of service, skills of maintenance personnel, operational
environment, equipment specification, uncertainty of Government financial subsidies, vehicle type,
business models, climate and regulatory controls (Xiao, et al. 2023). There is good evidence of research
efforts calculated to tackle vehicle fleet management and maintenance optimization in literature as
presented in the next section, but not to the extent of precluding investigations into new challenges that
may have been previously neglected, but has suddenly dominated the polity in a particular region or
Country. Such challenge could be instantaneous or long-lasting. Whereas impressive vehicle fleet
restoration efforts is evident in developed countries, the practice still face challenges of size, cost, bad
road network, unanticipated breakdowns leading to unscheduled maintenance in Nigeria. As noted in this
study, in addition to useful life expiration, prohibitive cost of preventive maintenance and increasing
hazard rate can justify a replacement action in a profit intended vehicle (Ozor and Mbohwa, 2018). The
general problem presented is how to determine the optimal time of utilization to replacement or optimal
cumulative usage of vehicle to replacement when the environmental realities makes fixed age
replacement difficult in Nigeria. A real life case of Vehicle Fleet Company is investigated in this study.
The Company is called T-Tours for the purpose of confidentiality. This study is an attempt to contribute
in the solution to the stated managerial problem.
2. Optimal Replacement Models

The determination of integrated replacement policies has received considerable attention in the past
couple of decades. One of the first contributions to traditional equipment replacement was Preinreich
(1940). The presentation viewed each reinvestment in equipment as an identical link in the replacement
chain. Terborgh (1949) emphasized the difference between deterioration and obsolescence, and was the
first to introduce the subject of obsolescence into equipment replacement study. In his pioneering work,
obsolescence was defined as not a characteristic of the machine itself but as a matter of the relative
capabilities of new machines, which he named as challengers. Constant salvage values and linear rate of
obsolescence were assumed in Terborgh’s model. Bellman (1955) was the first to view the replacement
problem as a multi-stage decision process and applied dynamic programming to find an optimal solution.
Dreyfus (1960) improved Bellman’s model and applied the solution procedure to an example problem.
Grinyer (1973) extended the earlier algebraic economic life models by substituting negative exponential
functions in place of linear functions.
Christer and Goodbody (1980) and Oakford et al. (1984) were among the pioneers to consider inflation
in replacement problems. The authors analyzed the problem of replacing members of a fleet of forklift
trucks under conditions of inflation and economic uncertainty as well as optimal sequence of
replacements when there exists multiple alternative challengers. However, the various approaches to
inflation were short of differential rates, which should ordinarily constitute a result closer to reality.
Inflation and tax effects were also neglected in a later article by Huirne et al. (1993), which provided a
stochastic dynamic programming formulation to determine the optimal replacement policy in swine
breeding herds. Optimal replacement models can be classified by the length of planning horizon for the
problem, the degree of effectiveness of each maintenance action, the optimization criteria, and nature of
the deterioration process and state space that characterizes the component (Valdez-Flores et al, 1989). A
key distinguishing feature among these models is the degree to which a component is repaired. A
particular advantage of perfect repair models is that they are able to take advantage of renewal theory, as
presented in Barlow and Proschan (1981).
Page 471
SEMIT 2023
A combination of imperfect, minimal and perfect repair can attend a replacement maintenance, depending
on whether a component, unit or whole system is involved. Imperfect repair is incorporated into models
in which maintenance does not renew a component to its original state. The associated mathematics of
imperfect repair is more involving, as standard renewal theory does not apply. In the literature, research
outputs presenting results for minimal and imperfect repair are properly documented (Bruns, 2000,
Biswas et al. 2003, Cha et al .2004). They make comparisons of the steady-state availability, based on
imperfect repair policies. Much earlier, Kijima and Nakagawa (1992) present replacement policies for a
shock model with imperfect preventive maintenance. Under a block replacement policy, maintenance
actions are performed at fixed and deterministic time intervals. Under age replacements preventive
maintenance policy, maintenance is initiated when a component has accumulated a certain amount of
operational time. Research in the area of age replacement policies is primarily concerned with the optimal
operational time a component is allowed to accumulate before initiating a maintenance action.
Numerous cost models have been presented to determine the optimal replacement age, most notably by
Ascher and Feingold (1984), Aven and Bergman (1986) and Nachlas (1989). Block replacement is
concerned with implementing a policy of maintenance at regularly spaced intervals without regard to the
age of component. Block replacement problems are popular, partially due to the fact that they are often
easier to implement mathematically and operationally. Barlow and Proschan (1981) proved that a block
replacement policy results in fewer failures than an age replacement policy, even though block
replacement requires more replacements and consequently incurs higher cost. Wortman et al. (1994)
showed that deterministic block replacement is preferable to random replacement. The literature discusses
many preventive maintenance variations of block replacement, with the purpose of presenting policies
that decrease the cost of frequent replacement or simplify the mathematical expressions incurred by block
replacement. Generally, block replacement models suffer from the potential to prematurely replace
components while age replacement policies require knowledge of the cumulative operational time for the
component under consideration. For instance, Ozor and Mbohwa (2018) compare the best replacement
model for a service producing system and posit that the age replacement model gave better results.
Zuckerman (1980) presents an optimal inspection interval and replacement rule with consideration for
the associated costs when deterioration is due to a Poisson shock process. Obviously, inspection can
reduce the frequency of failures as incident failures can be picked early and arrested. However,
management of vehicle fleet size and routes can become so large that inspection costs may get prohibitive,
warranting special scientific approaches. Interested readers can consult Militao and Tirachini (2021),
Zhang and Wang (2023) as well as Ahmed and Yousefikhoshbakht (2023). The authors discuss optimal
models for such and similar situations.
Cassady et al .(2004) presented an availability optimization model that applies to a general class of two-
state repairable components by determining a set of availability importance measures. Through five
examples, they demonstrated that focusing on the reduction of component failures provides greater
benefit than increasing the speed of equipment repair. The authors then established a set of three
optimization models that capture the trade-offs between improving availability performance and the
investments required to achieve that improvement. There is a Decision Making Trial and Evaluation
Laboratory (DEMATEL) model and Analytic Network Process (ANP) model due to Vujanovic et al.
(2012). The authors developed these models to evaluate nine identified indices that were considered
critical to vehicle fleet management in the Republic of Serbia. The aim was to achieve efficient energy
utilization through ensuring good environment for the vehicles, which in the long run leads to reduced
cost of operating vehicle fleet. Other presentations based on reducing cost of vehicle fleet operation and
subsequent replacement by providing good environment can be found in Li et al. (2022), and Toni et al.
(2023). The researchers introduce a new function called innovative Dashboard Functions (DF) to
optimize predictive maintenance at Bus garages and describe procedure for achieving sustainability in
Vehicle fleet operations.
Evolutionary Algorithms has recently been generated to detect failure nucleation zones and optimize
maintenance of vehicle fleet used for transportation and other purposes. The modeling procedure have
the advantage of particularizing failures to component level and reducing cost by ensuring that parts are
neatly identified and replaced, short of the costly alternative of replacing whole system or even replacing
other components thought to fail soon. Wang et al (2019) introduce a proof of concept real-world vehicle
Page 472
SEMIT 2023
fleet maintenance scheduling optimization problem, and solved it with a dedicated multi-objective
evolutionary algorithm framework (MOEA). In particular, the MOEA proposed, involve a grouping
strategy, three vector chromosome as well as corresponding genetic operators. It was reported to behave
better than other deterministic optimization models employed for the same situations.Wang et al. (2021)
noted that systems can be de-compartmentalized into its constituent components, making it possible for
failures to occur in a particular or diverse component(s).To model the situation, the authors constructed
a diversity indicator Based MOEA (DI-MOEA) with the objective of firstly, automatically generating the
region favoured for incipient failure in the optimization phase, and subsequently focusing on
implementing the solution set in the preferred region. Next, the vehicle fleet maintenance scheduling
optimization (VFMSO) problem is formulated, and an adapted MOEA proposed to optimize maintenance
decisions based on the candidate distribution of historical data of the concerned component of vehicles
in the fleet.
From the foregoing literature review, vehicle fleet maintenance can be considered as a vivacious problem
due to the changes it promises on the profitability of the operating companies. In addition, it has been
shown that traditional based replacement thought to restore systems to “as-good-as-new”, cost-based,
mathematical programming-based, algebraic and exponential-based, deterministic-based, analytical
hierarchical Processes-based and evolutionary algorithm are largely utilized among many methods for
evaluating vehicle fleet maintenance optimization. Among these methods, deterministic-based, cost-
based, analytical hierarchical processes-based and evolutionary algorithm were found to be more
successful than traditional and mathematical programming analysis. However, due to the reasons of
difficulty in application of the existing methods to the vehicle fleet maintenance in Nigeria where
unprecedented inflation has assumed astronomical dimension, this study propose an easy to apply cost
analytic model applicable to the case study.
3. Materials and Methods

In this section, the procedure for developing the profit-intended vehicle fleet maintenance (replacement)
model is explained in detail. Vehicle fleet belong to service producing systems that fail due to age and
usage, and enormous literature on its replacement strategies, including the ones already cited above, are
well documented. The choice of parameters in this study was carefully decided because of the
environmental constraints placed on many models with respect to their applicability to the Nigerian
situation. In particular, some of the existing models parameters cannot be determined with ease in Nigeria
due to reasons bordering on unavailability of competent instruments and proficient personnel in the
vehicle fleet companies, who can collect the required data.The running costs of the system can be modeled
into an equation in advanced countries by using appropriate mathematical tools which enable the
computation of results to be simpler (appropriate records are kept for analysis and environmental factors
like road condition are good and almost constant). In contrast with the Nigeria case, the absence of the
conditions in advanced countries leaves one with discrete running costs data, which fluctuate due to
inherent environmental conditions. It is obvious that the problem formulation should not involve variables
that cannot be easily measured if the solution procedure will strictly apply to the Nigerian situation. Since
every fault, wear and maintenance action on the vehicle will ultimately result in costs to the vehicle
owner, it therefore seems reasonable that the problem should be formulated on the basis of costs.
The next problem was to identify a variable whose change easily offsets the running costs of the vehicle.
It is understandable that the greater the distance travelled by the vehicle (mileage), the greater are its
chances of failure. In addition, the older a vehicle is the more prone to failure it is than a newer vehicle,
if other conditions are maintained. Therefore, in the choice of a variable that sympathizes with the running
costs of the vehicle, mileage and age of vehicle were considered as viable options. It is not easy to
compute maintenance costs as functions of mileage because in Nigeria, motor transportation companies’
only record maintenance costs and dates. Therefore, vehicle age was chosen over mileage for the reasons
mentioned above. In the developmental framework, the effect of astronomical and incessant changes in
the value of money with time (inflation), which has recently hit Nigeria, and which have not been
contemplated by any of the existing models was taken into consideration. The study presents a concise
mathematical formulation of the optimal vehicle replacement problem (considering all the associated
Page 473
SEMIT 2023
costs and inflation), and optimize the total cost function with respect to time to get an optimal replacement
time. The developed optimal replacement function is tested using real time cost values from a big vehicle
fleet company (T-Tours) with offices in various parts of the Country, to prescribe an estimated time of
replacement for the case study company.
3.1 Problem Formulation
The running costs of a vehicle are those cost involved in exploiting the vehicle, and they include fuel
costs, repair costs, servicing costs, insurance costs and general maintenance costs. Therefore the running
cost of the vehicle is given by: Rn = MC + Z p .c p .M
As the value of money changes with time due to inflation of the national currency, Naira, the
value of money now obviously is worth more than the same value at some future date. Hence for the
vehicle investment problem, the present worth of the money to be spent in a few years must be considered.
Let the rate of inflation of money per year be r . Then one Naira invested at present will be equivalent to
(1 + r ) in a year or (1 + r ) 2 in two years or (1 + r )3 in three years and (1 + r) n in n year’s time. In
other words, making a payment of one Naira now is equivalent to paying (1 + r ) − n in n years time. The
quantity (1 + r ) − n is called the present worth or present value of one Naira spent n years from now.
Present value of one Naira to be spent in n years= (1 + r ) − n = d n ,
where d = (1 + i ) −1 . With the above definitions, and assuming that all conditions of road, driver and
environment are constant, the present worth of all future costs of purchasing and running the vehicle with
a policy of replacing it after every n years is then given by:
Dn = {(C − S + R1 ) + dR2 + d 2 R3 + ... + d n−1 Rn } {for 1 to 𝑛 years}

+ {𝑑𝑛 (𝐶 − 𝑆 + 𝑅1 ) + 𝑑𝑛+1 𝑅2 + 𝑑1+2 𝑅2 + ⋯ + 𝑑2𝑛−1 𝑅𝑛 } {for 𝑛 + 1 to 2𝑛 years}
+ {𝑑2𝑛 (𝐶 − 𝑆 + 𝑅1 ) + 𝑑 2𝑛+2 𝑅2 + 𝑑1+2 𝑅2 + ⋯ + 𝑑 3𝑛−1 𝑅𝑛 } {for 2𝑛 + 1 to 3𝑛 years}
𝑛−1 𝑛 2𝑛
+… +𝑑 + 𝑅𝑛 + (1 + 𝑑 + 𝑑 + ⋯)
= 〈[{𝐶 − 𝑆 + 𝑅1 } + 𝑑𝑅2 + ⋯ 𝑑𝑛−1 𝑅𝑛 ]1 + 𝑑𝑛 + 𝑑2𝑛 + ⋯ 〉
1
= [𝐶 − 𝑆 + ∑𝑛𝑖=1 𝑑𝑖−1 𝑅𝑖 ] [ ] (1)
1−𝑑 𝑛
 1 
Where  n
= [1 + d n + d 2 n + ...] (sum to infinity of a series)
1 − d 
Therefore,
 n



C − S + 
i =1
d i −1Ri  = [(C − S + R1 ) + dR2 + ...d n−1Rn ]

This value Dn which is the value of equation (1) is the present sum of money required to settle all future
costs of purchasing and running the vehicle with a replacement policy of every n years. The research
intends to look for the optimal year n which minimizes the value of Dn and this will obviously be the
best period to replace the vehicle.
3.2 Problem Solution
For Dn to be optimal, the following inequality must hold:
Dn +1  Dn  Dn −1
Page 474
SEMIT 2023
That is the present sum of money needed to purchase and run the vehicle is optimal in a particular year if
it is smaller than that for the succeeding and preceding years.
From this the two inequalities Dn +1 − Dn  0  Dn − Dn +1  0
This implies that ( Dn − Dn −1 )  0  ( Dn +1 − Dn ) (2)
Let 𝑃(𝑛) = 𝐶 − 𝑆 + ∑𝑛𝑗=1 𝑑𝑖−1 𝑅𝑖

Therefore, from equation (1)
P (n)
Dn =
1− d n
And putting for n + 1 for n in equation above,
P (n + 1)
Dn +1 =
1 − d n +1
Therefore:
𝑃(𝑛 + 1) 𝑃(𝑛) 𝑃(𝑛 + 1)(1 − 𝑑𝑛 ) − 𝑃(𝑛)1 − 𝑑𝑛+1
𝐷𝑛+1 − 𝐷𝑛 = − =
1 − 𝑑𝑛+1 1 − 𝑑𝑛+1 (1 − 𝑑𝑛 )(1 − 𝑑𝑛+1 )
A few more algebraic analysis gives the value of 𝐷𝑛+1 − 𝐷𝑛 presented in equation (3),
𝑑 𝑛 (1−𝑑) 1−𝑑𝑛
𝐷𝑛+1 − 𝐷𝑛 = { 𝑅 − 𝑃(𝑛)} (3)
(1−𝑑 𝑛 )(1−𝑑 𝑛+1 ) 1−𝑑 𝑛+1
Since d  1 , and so 1 − d n  0 , therefore Dn +1 − Dn is always positive and has the same sign as the
quantity in bracket in equation (3)
Similarly putting n − 1 for n in equation (3), gives
𝑑𝑛 (1 − 𝑑) 1 − 𝑑𝑛
𝐷𝑛+1 − 𝐷𝑛 = { 𝑅 − 𝑃(𝑛 − 1)}
(1 − 𝑑𝑛 )(1 − 𝑑 𝑛+1 ) 1 − 𝑑 𝑛
Which upon further analysis yields equation (4) as:
𝑑 𝑛−1 (1−𝑑) 1−𝑑𝑛
𝐷𝑛+1 − 𝐷𝑛 = (1−𝑑𝑛 )(1−𝑑𝑛+1 ) { 1−𝑑 𝑅𝑛 − 𝑃(𝑛)} (4)
From the inequality in (2), ( Dn − Dn −1 )  0  ( Dn +1 − Dn )
That is to say,
1 − d n  1 − d n 
 R − P ( n )   0   Rn +1 − P(n)
 1− d  1− d
n
 
Which results in
1− d n 1− d n
Rn  P(n)  Rn +1 (5)
1− d 1− d
Since
1− d n
1 + d + d 2 + ... + d n −1 = = ∑𝑛𝑖=1 𝑑𝑖−1
1− d
Page 475
SEMIT 2023
 
 P ( n) 
Then the inequality in (5) becomes Rn   n   Rn +1
 d i −1 
 
i =1

The following constraints have to be taken into account:

1. Technical durability of a vehicle, e.g. maximal mileage of the vehicle M max should not be exceeded:
n
M
i =0
 M max
2. Decision variable, i.e. age of a vehicle at the moment of replacement must be positive, integer values
only:
n   \ 0
3. Maximal age of a vehicle at the moment of replacement cannot be greater than the length of the
planning horizon H: nmax  H
Since all the above data are readily available, the optimal replacement time that is the year
that satisfies the above inequality can be computed

Table 1 gives the values of running costs for an example Toyota Bus operated by T- Tours, while Table
2 shows the computation of relevant terms of the developed optimal equation.
Table 1. Computation of Running costs, Source: T-Tours
Year 𝒁𝒑 𝑴 𝑪𝒑 𝒁𝒑 . 𝑪𝒑. 𝑴 𝑴𝑪 𝑹𝒏
1 0.167 153300 50 1280055 180000 1460055
2 0.167 153300 50 1280055 209000 1489055
3 0.18 153300 50 1379700 224000 1603700
4 0.18 153300 65 1793610 300000 2093610
5 0.2 153300 65 1992900 315000 2307900
𝒁𝒑 ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮
13 0.2 153300 65 1992900 400000 2392900
14 0.2 153300 65 1992900 400000 2392900
Page 476
SEMIT 2023
Table 2. Computation of Running costs for i=1%

Year 𝒁𝒑 𝑴 𝑪𝒑 𝒁𝒑 . 𝑪𝒑. 𝑴 𝑴𝑪 𝑹𝒏 𝑑 𝑛−1 ∑ 𝑑 𝑛−1 𝑅𝑛 𝑑 𝑛−1 ∑ 𝑅 𝑑 𝑛−1 𝐶 − 𝑆 𝐽 𝐾
𝑛
1 0.167 1533 50 1280 1800 1460 1 1 1460 1460 2200 3660 3660
2 0.167 1533 50 1280 2090 1489 0.990 1.99 1474 2934 2200 5134 2580
3 0.18 1533 50 1380 2240 1604 0.980 2.971 1572 4565 2200 6706 2258
4 0.18 1533 65 1794 3000 2094 0.971 3.941 2032 6539 2200 8739 2217
5 0.2 1533 65 1993 3150 2308 0.961 4.902 2218 8756 2200 1096 2235
6 0.2 1533 65 1994 3500 2343 0.951 5.853 2229 1099 2200 1319 2253
7 0.2 1533 65 1994 3600 2353 0.942 6.795 2217 1320 2200 1540 2267
8 0.2 1533 65 1994 3600 2353 0.933 7.728 2195 1540 2200 1760 2277
9 0.2 1533 65 1994 4000 2393 0.923 8.652 2210 1761 2200 1981 2290
10 0.2 1533 65 1994 4000 2393 0.914 9.566 2188 1980 2200 2199 2299
4500000 𝐶−𝑆+∑𝑛
𝑗=1 𝑑
𝑖−1 𝑅
𝑖
𝐽 = 𝐶 − 𝑆 + ∑𝑛𝑗=1 𝑑𝑖−1 𝑅𝑖 ,𝐾 = ∑𝑛 𝑖=1
𝑖=1 𝑑
𝐶−𝑆+∑𝑛
𝑗=1 𝑑
𝑖−1 𝑅
𝑖
Writing A(n)= ∑ 𝑑 𝑖−1
for convenience, A(n) and n were computed for various Salvage values,
4000000
ranging from N500,000 to N1,000,000.The graphs are shown in figures 1-7. The results of the
computation indicate that it is optimal for the company to replace its buses plying the Warri Onitsha
route every five years since the minimum of A(n) occur when 𝑛 = 5 years
Below3500000
graphs were plotted for A(n) in the vertical axis and n-number of years in the horizontal axis for
4500000
varying salvage values (S) at interest rates (i) for 1 to 7 percent

4000000
3000000
3500000
S=N500,000
A(n)
2500000
3000000
4500000 S=N600,000
Series1
Series2 S=N700,000
2500000 Series3
Series4 S=N800,000
Series5
2000000
4000000 Series6 S=N900,000
2000000 n
0 0 2 2 4 4
6 8 6 10 8 12 10 12 S=1,000,000
n
1500000
Figure 1: Graph of A(n) versus n for i=1% at varying salvage values, S
1500000
4500000
3500000
1000000
4000000
500000
1000000
3000000
3500000
0
A(n)
S=N500,000
500000
3000000
2500000 S=N600,000
Series1
Series2 S=N700,000
2500000 Series3
Series4
S=N800,000
0
2000000 Series5 S=N900,000
n
Series6
2000000 0 2 4 6 8 10 12 S=1,000,000
0 2 4 6 8 10 12
n
1500000
1500000 Figure 2: Graph of A(n) versus n for i=2% at varying salvage values, S
1000000
1000000 Page 477

500000
0
3500000
SEMIT 2023
4500000
3000000
4000000
45000003500000 S=N500,000
2500000 S=N600,000
A(n)
3000000 S=N700,000
Series1
4000000 Series2
S=N800,000
2500000 Series3
2000000 Series4
S=N900,000
n
0 2 4 6 8 10 12
Series5
S=1,000,000
Series6
2000000
3500000 0 2 4 6 8 10 12
n
1500000
1500000
4500000
30000001000000
4000000
1000000
500000
3500000 S=N500,000
4500000
2500000 S=N600,000
A(n)
0
500000 3000000 S=N700,000
Series1 S=N800,000
Series2
4000000
20000002500000 Series3 S=N900,000
n Series4
0 0 2 4 6 8 10 12
Series5 S=1,000,000
Series6
2000000
0 2 4 6 8 10 12
n
3500000
1500000
1500000
4500000
1000000
4500000
3000000
1000000
4000000
500000
3500000
S=N500,000
4000000
2500000
A(n)
500000 S=N600,000
0
3000000
S=N700,000
Series1
S=N800,000
Series2
2500000 Series3
3500000
2000000
0 S=N900,000
Series4
n
Series5
0 2 4 6 8 10 12 S=1,000,000
Series6
2000000
0 2 4 6 8 10 12
n
3000000
1500000
4500000
1500000
4000000
1000000 S=N500,000
2500000
10000003500000 S=N600,000
500000
S=N700,000
A(n)
3000000 S=N800,000
0
2000000
500000 S=N900,000
Series1
Series2
n
0
2500000 2 4 6 8 10 12 S=1,000,000
Series3
Series4
Series5
Series6
2000000
1500000
0 0 2 4 6 8 10 12
n

1500000
10000001000000 Page 478

500000
500000
3500000
SEMIT 2023 4500000
3000000
4000000
S=N500,000
3500000
2500000 S=N600,000
A(n) S=N700,000
3000000
S=N800,000
Series1
2000000 Series2 S=N900,000
2500000 n
Series3
0 2 4 6 8 10 12 Series4 S=1,000,000
Series5
Series6
2000000
1500000 0 2 4
n
6 8 10 12
1500000
1000000 1000000
Conclusion
The study succeeded
500000in presenting a new vehicle replacement procedure that is easily applicable to the
Case study
500000and similar Vehicle Fleet Companies. The cost based modelling procedure adopted for the
research interestingly0 lead to significant cost reduction, among other improvements. The procedural steps
have been carefully described and results presented in Tabular and mostly graphical images to enhance
understanding. Although the initial models and results came from historical data from a real life Toyota
Bus, operated0 by T- Tours, the methodology can be scaled to other Vehicle Fleet companies with the
same challenges, as described in this work. The maintenance policies decided upon the basis of the
outcome of the study has the capability of saving the Company from prohibitive cost of frequent
preventive maintenance from vehicles that have spent their useful life. Hence, the five year replacement
period should be vigorously pursued by the Management of T- tours for optimum reliability and
profitability.
References
Ahmed, Z.H. and Yousefikhoshbakht, M. (2023),An improved tabular search algorithm for solving
heterogeneous fixed fleet open vehicle routing problem with time windows, Alexandria Engineering
Journal 64, 349–363
Aitan M. Militao, A.M. and Tirachini, A.(2021), Optimal fleet size for a shared demand-responsive
transport system with human-driven vs automated vehicles: A total cost minimization approach,
Transportation Research Part A 151 52–80
Ascher, H. and Feingold, H. (1984). Repairable Systems Reliability. Marcel Dekker, New York.
Aven, T. and Bergman, B. (1986). Optimal replacement times – a general set-up. Journal of Applied
Probability, 23, 432-442.
Badia F.G. &Berrade M.D. (2007). Minimal Repair Model for Systems under both Unrevealed Minor
Failures and Unrevealed Catastrophic Failures. International Mathematical Forum, 2(17), 819-834
Barlow, R. and Hunter, L. (1960). Optimum preventive maintenance with repair. Operations Research”
8, 90-100.
Barlow, R. and Proschan, F. (1981). Statistical Theory or Reliability and Life Testing: Probability
Models. New York, Holt, Rinehart, and Winston
Bellman R. (1955), Equipment replacement policy. Journal of the Society for Industrial and Applied
Mathematics 3 (3) 133-136.
Biswas, A., Sarkar, J., and Sarkar, S. (2003). Availability of a periodically inspected system, maintained
under an imperfect repair policy. IEEE Transaction on Reliability. 52, 311- 318.
Page 479
SEMIT 2023
Bruns, P. (2000). Optimal maintenance strategies for systems with partial repair options and without
assuming bounded costs. Technical report, Faculty of Mathematical Sciences, University of Twente.
Cassady, C. R., Pohl, E. A., and Jin, S. (2004). Managing availability improvement efforts with
importance measures and optimization. IMA Journal of Management Mathematics. 15, 161-174.
Cha, J. H., Lee, S., and Mi, J. (2004). Comparison of Steady system availability with imperfect repair.
Applied Stochastic Models in Business and Industry, 20, 27- 36.
Christer A.H., Goodbody W (1980). Equipment replacement in an unsteady economy. Journal of the
Operational Research Society 31 (6) 497-506.
Nachlas, J. A. and Cassady, C. R. (1999). Preventive maintenance study: A key component in engineering
education to enhance industrial productivity and competitiveness. European Journal of Engineering
Education. 24, 299-309.
Dhillon B.S. (2002). Engineering Maintenance: A Modern Approach. Florida, CRC Press.
Dreyfus S.E. (1960). A generalized equipment replacement model, Journal of the Society for Industrial
and Applied Mathematics 8 (3) 425-435.
Eti M.C., Ogaji S.O.T. & Probert S.D. (2004). Reliability of the Afam Electric Power Generating Station,
Nigeria. Applied Energy, 77(3), 309-315
Grinyer P.H. (1973), The effects of technological change on the economic life of capital equipment. AIIE
Transactions. 5 (3) 203-213.
Huirne R.B.M., Dijkhuizen A.A., Van Beek P. & Hendriks H.B. (1993). Stochastic dynamic
programming to support sow replacement decisions. European Journal of Operational Research 67, 161-
171.
Karsak E.E. &Tolga E (1997). An overhaul-replacement model for equipment subject to technological
change in an inflation-prone economy. International Journal of Production Economics. 56. pp. 291-301
Kijikma, M. and Nakagawa, T. (1992). Replacement policies of a shock model with imperfect preventive
maintenance. European Journal of Operational Research. 57, 100-110.
Li, K., Acha, S., Sunny, N., Shah, N. (2022), Strategic transport fleet analysis of heavy goods vehicle
technology for net-zero targets, Energy Policy 168 112988
Oakford R.V., Lohmann J.R. (1984), A. Salazar, A dynamic replacement economy decision model. IIE
Transactions on Operations Research 16 (1) 65-72.
Ozor, P.A. and Mbohwa, C. (2018) Availability Based Comparison of Fixed Age and Life Cycle Cost
Replacement, Proceedings of the International Conference on Industrial Engineering and Operations
Management Washington DC, USA, September 27-29, pp. 2006 – 2013
http://ieomsociety.org/dc2018/proceedings/
Ozor, P.A. Onyegegbu, S.O. and Agunwamba, J.C. (2017), Modelling composite performance variable
of deteriorating systems using empirical evidence and artificial neural network’, International Journal of
Reliability and Safety, Vol. 11, Nos. 1/2, pp. 23 – 49, https://doi.org/10.1504/IJRS.2017.088546
Preinreich, G. A. D. (1940) The Theory of Industrial Replacement, Scandinavian Actarial Journal,
V.1939, Is. 1. https://doi.org/10.1080/03461238.1939.10419246
Redmer A. (1999) ‘Vehicle replacement planning in freight transportation companies’ Advanced AI
methods in Operations Research
Terborgh G. (1949) Dynamic Equipment Policy, Machinery and Allied Products Institute, Washington
DC
Toni, A., Corazza M.V., Vasari, D.(2023), Improving the environmental performance of bus fleets in
Europe, Transportation Research Procedia 69 147–154
Page 480
SEMIT 2023
Valdez-Flores, C. and Feldman, R. (1989). A survey of preventive maintenance models for stochastically
deteriorating single-unit systems. Naval Research Logistics. 36, 419-446.
Vujanovic, D., Momcilovic, V., Bojovic, N., Papic, V. (2012), Evaluation of vehicle fleet maintenance
management indicators by application of DEMATEL and ANP. Expert Systems with Applications 39
10552–10563
Wang, Y., Limmer, S.,Olhofer, M., Emmerich, M., Back, T., (2021) Automatic preference based multi-
objective evolutionary algorithm on vehicle fleet maintenance scheduling optimization, Swarm and
Evolutionary Computation 65 100933
Wang, Y., Limmer, S., Olhofer, M., Emmerich, M.T., Bäck, T., Vehicle fleet maintenance scheduling
optimization by multi-objective evolutionary algorithms, in: 2019 IEEE Congress on Evolutionary
Computation (CEC), IEEE, 2019, pp. 442–449.
Wortman, M. A. and Klutke, G. A., and Ayhan, H. (1994). A maintenance strategy for systems subjected
to deterioration governed by random shocks. IEEE Transactions on Reliability. 43, 439-445.
Xiao, L., Zhang, J., Wang, C., Han, R. (2023), Optimal fleet replacement management under cap-and-
trade system with government subsidy uncertainty, Multimodal Transportation 2, 100077
Yoshinori S. & Gregory P. (2005). A vehicle replacement policy for motor carriers in an unsteady
economy. European Journal of Operatios Research, Transportation Research Part A. 39, 463-480
Zuckerman, D. (1980). Inspection and replacement policies. Journal of Applied Probability. 17, 168-177.
Page 481
SEMIT 2023

Future Perspective
Parametric Study of Shear Strength UHPC Tapered Beam

Aqeel A. Jbarah 1[0009-0005-1123-2418] and Nasser Hakeem TU'MA 2[0000-0003-1956-8539]
1
2
Abstract
This study introduces a verification of the structural confidence of the recently developed tapered beam
of new trends of tapered using a numerical analysis by ABAQUS program. The comparisons of numerical
results with the existing of previous experimental results, in scope of ultimate shear strength, load-
displacement relationships and other structural aspects. They showed a very good agreement between
than converges. Proposed rates ranged from 0.96 to 1.13. Parametric study that related to the developed
Ultra High-Performance Concrete beams, is included with many structural parameters such that crown
length, simulations of the tapered beam by multiple semi-scale factor, steel reinforcement strength,
provided steel ratio (longitudinal and transversal). It noted that raising up the longitudinal reinforcement
ratio induces an increase in the yield and ultimate strength of beams. Shear testing on steel concrete beams
with Steel fiber ratio 2% by volume yielded positive findings in terms of the influence of the shear
reinforcement ratio on shear strength. The greater the transverse reinforcement ratio, the more cracks
occurred with a smaller openings width was noticed. Also, the shear cracking under greater load increase
when the shear reinforcement was increased. When the scale factor 2.4 was increased, an increase in
shear strength was observed at 348%. The crown's length is increased from 0 to 200 mm, the shear
strength is increased by 10%.
Keywords: Shear strength, Scale Factor, Reinforcement Ratio, Ultra High-performance concrete,
Crown Length.
1. Introduction
Reinforced concrete beam of non-prismatic is an optimal beam and has special importance. The tapered
configuration is a specific mechanism could be utilized in structural engineering to get optimum material
investment[1]. Tapered concrete beam is a type of non-prismatic concrete beam members which varying
depths along the axis use in middle rise buildings, bridges, structural portal-frames. It is possible to design
these tapered concrete beam members to resist all the required external loads [2-4]. The midsection of the
tapered beam has a deeper section to resistant a significant flexural and shear force. In addition, there are
longitudinal openings in the member that will allow service facilities to pass through it and reduce
concrete materials, this is economic feasibility[2, 5, 6]. Because a major approach to evaluating the shear
strength of ultra-high performance concrete beams is to separately evaluate the shear contribution of the
concrete and stirrups parts, most research on this topic has been conducted on ultra-high performance
concrete tapered beams without stirrups first, which is also the main objective of this paper. Because to
the crown slope, the amount of concrete used was lowered. There has been no previous study on ultra-
high performance tapered beams [7-10]. The determination of the shear strength of beams is contingent
upon a myriad of factors that must be taken into account. These factors encompass a wide range of
considerations, such as the contributions made by the shear strength of concrete, which takes into
consideration variations in beam depth as well as the presence of existing openings. Additionally, the
shear steel reinforcements must be considered, as they have a direct impact on the overall shear
Page 482
SEMIT 2023
strength[11]. Furthermore, one must also take into account the effects of diagonal shear reinforcement
and inclined tensile steel reinforcement, both of which can significantly influence the shear strength of
the beams. The influence of the compression chord cannot be overlooked either, as it also plays a
significant role in determining the shear strength. Furthermore, it is worth noting that there are certain
approaches that can be employed to enhance the shear strength of beams. One such approach is the
utilization of different mixing techniques, such as the two-stage mixing approach and the treatment of
recycled aggregate concrete. These methods have the potential to significantly improve the shear strength
of the beams and should be carefully considered when designing structures[12]. Moreover, the shear
strength of steel fiber-reinforced concrete beams is influenced by several parameters that must be
carefully analyzed. These parameters include the steel fiber content, the shear-span-to-depth ratio, the
reinforcement ratio, the effective depth, and the concrete compressive strength. Each of these parameters
has the potential to impact the overall shear strength of the beams and should be taken into account when
conducting analyses. This study included four rounds of specimen tests to determine the shear resistance
of ultra-high performance concrete tapered beams without stirrups. Furthermore, Engineers [13] use this
type of beam based on an empirical and cautious design strategy, which is inaccurate and inefficient. To
achieve an acceptable design, a straightforward assessment technique with high accuracy and
compatibility for tapered beams must be proposed.
2. Research Significance
This numerical study presents cross-section variable models, containing longitudinal openings to know
the shear failure behavior of ultra-high performing concrete(Improved shear strength and reduced
weight of concrete mixture materials as well as use of longitudinal openings as service facilities) and
included many parameters such that crown length, simulations of the tapered beam by a multi-scale
factor for the geometry of beam, steel reinforcement strength, provided steel ratio (longitudinal and
transversal).
3. Description of Finite Element Model

3.1. Geometry of Model
Modern engineering programs can simulate experimental models and obtain compatible models. The
ABAQUS Finite Element (FE) program was used to simulate the non-prismatic beam variable section
shown in the Figure 1. The length of 1.9 meters and a variable depth beginning at 0.18 meters up to a
fair package 0.4 meters containing 50 mm diameter longitudinal holes. The meshes used for the
concrete part are tetrahedral mesh of C3D4 type, which is a three- dimensional 4-node mesh with a
grain size of 5 cm in all directions.
Fig. 1. Geometry and details of tested beams
Page 483
SEMIT 2023
3.2. Concrete Modeling

Concrete is a quasi-brittle material with different compression and tension behavior; hence, developing
a nonlinear material model for concrete that accurately captures all of its behavior is a difficult task.
The compression, part of the stress-strain curve for concrete, extends linearly for up to about 30% of the
maximum compressive strength. Above this point, the stress increases gradually and nonlinearly up to
the maximum compressive strength. The stress-strain curve's tension part is linearly elastic up to the
maximum tensile strength. After this, the concrete begins to crack and its strength gradually reduces
until it is zero [14]. Usually, the tensile strength of concrete is equivalent to 10% of its compressive
strength. The tensile strength of concrete is assumed to be equal to the rupture strength of concrete [15].
The tensile strengths for the tested Tapered-beam model were According to the recommendations of the
Japanese Society of Civil Engineering[13]. In ABAQUS, the Concrete Damage Plasticity model is
implemented by parameters shown in Table 1.
Fig. 2a. Element Types of UHPCBs Model.
Fig. 2b. Assembling of beam parts.
Table 1. Input Parameters of Concrete Damage Plasticity Model.

Parameter Value
Compressive Strength (MPa) 155
Concrete Tensile (MPa) 8
Modulus of elasticity ft (GPa) 45
Poisson Ratio (𝜈) 0.18
The Ratio of Biaxial Compressive Strength to Uniaxial Compressive
1.07
Strength (𝜎bo/𝜎co)
The Shape of The Failure Surface (Kc) 2/3
Dilation Angle (𝜓) 36°
Eccentricity (є) 0.1
Viscosity Parameter (v) 0.005
Page 484
SEMIT 2023
3.3. Steel Modelling

Reinforcement and wire mesh are designed using a bi-line elasto-plastic model that induces a steel
elasticity modulus 200 GPa and Poisson ratio 0.3 The Table 2. shows the physical properties of the
rebar used in simulation, and the meshes used for the reinforcements are beam element mesh with a
grain size of 5 cm. As shown in Figure 2a.
Table 2. Properties of Steel Bars.
Yield Ultimate
Bar size Strength
Strength
(mm)
(MPa) (MPa)
8 423 526
25 481 672
3.4. Experimental – Numerical Comparative Analysis

This study relies on comparing models with a previous study where comparing experimental - numerical
results showed good results [16, 17]. Only nine of the tapered beams was modeled as well as three groups
were compared. The first group (TB-1, TB-2, TB-5) included three different models of longitudinal
openings distribution in different zones of the beam in the prismatic, non-prismatic zone and in both of
them (prismatic and non-prismatic zones). The second group (TB-6, TB-8, TB-10) included concrete
beams strengthen by NS-GFRP with a different slope vertically and at a 30, 45-degree angle. The third
group (TB-7, TB-14, TB-15) involved different cases of shear span to effective depth ratio 2.3, 2.7, 2.9;
respectively Figure 4. is below shown the load-displacement of the computed results. The purpose of
verifying shear strength, compared with experimental results to check the input data and starting with an
independent parametric study. Proposed rates ranged from 0.96 to 1.13 and shown in the Table 3. While
Figure 3. shows the observed failure mode and Strain distribution for experimental and numerical of TB
10.
Table 3. Comparison Between the Numerical and Experimental Results.
Group Beam Shear Strength (kN)

No. ID EXP FEA EXP/FEA
TB 1 416 405 1.03
st
1 TB 2 362 365 0.99
TB 5 462 453 1.02
TB 6 486 499 0.97
nd
2 TB 8 446 456 0.98
TB10 432 402 1.07
TB 7 370 386 0.96
3rd TB14 441 451 0.98
TB15 487 432 1.13
Page 485
SEMIT 2023
Fig. 3. Failure Mode and Strain Distribution for Experimental and Numerical of TB 10.
500
400
Load (kN)
300
200
Experimental_TB1
100
Abaqus_TB1
0
0 3 6 9 12 15
Displacement (mm)
a. TB 1
400
300
Load (kN)
200
100 Experimental_TB2
Abaqus_TB2
0
0 3 6 9 12 15 18
Displacement (mm)
b. TB2
500
400
Load (kN)
300
200
Abaqus_TB5
0
0 3 6 9 12 15 18 21
Displacement (mm)
c. TB5
Page 486
SEMIT 2023
600
500
Load (kN)
400
300
200
Experimental_TB6
100
Abaqus_TB6
0
0 3 6 9 12 15 18 21
Displacement (mm)
d. TB6
500
400
Load (kN)
300
200
Experimental_TB7
100
Abaqus_TB7
0
0 3 6 9 12 15
Displacement (mm)
e. TB7
500
400
Load (kN)
300
200
Experimental_TB8
100
Abaqus_TB8
0
0 3 6 9 12 15 18 21
Displacement (mm)
f. TB8
500
400
Load (kN)
300
200
Experimental_TB10
100
Abaqus_TB10
0
0 3 6 9 12 15 18 21
Displacement (mm)
g. TB10
Page 487
SEMIT 2023
500
400
Load (kN)
300
200
Abaqus_TB14
0
0 3 6 9 12 15 18 21 24
Displacement (mm)
h. TB14
600
500
Load (kN)
400
300
200
Experimental_TB15
100 Abaqus_TB15
0
0 3 6 9 12 15 18 21 24
Displacement (mm)
i. TB15
Fig. 4. Comparison of the Numerical and Experimental Load-Displacement Curve.
3.5. Design for Flexural and Shear Strengths

To avoid shear failure, the ultimate flexural loads must be greater than the ultimate shear strength. Steps
have been adopted in the design of concrete component, fibre components, dowel action part and
effective depth [18, 19]. All tapered beams are designed by a derivative equation the yielding of main
reinforcement and the compression force, which is contain two parts. First one is uniform compressive
0.65 × 𝑓𝑐′ and the second in the same intensity. The tensile contain two parts also. One of them
represents the bar force and the second from the neutral axis to extreme tensile fibres 0.4 × √𝑓′𝑐,
which is the pressure of the tension zone. The concrete component and the fibre components are
designed depending on the Japanese Society of Civil Engineering [13].
3.6. Parametric Study
Nineteen Beams were modelling and run it until the failure under two-point loading. The main considered
variables were the Scale factor, crown length and reinforcement ratio (longitudinal and transvers). As a
result, the eighteen tapered beams were classified into three groups. All beams were the compression
reinforcement 2𝜑8. Details of the specimens shown in Table 4.
Page 488
SEMIT 2023
Table 4. Details of The Specimens of UHPC Tapered Beams.

Crown Reinforcement
Beam Scale
Length
ID Factor Longitudinal Transvers
(mm)
SF-1.0 1.0
SF-1.3 1.3
SF-1.6 1.6
1.0 0.027 𝜑8@845
SF-1.8 1.8
SF-2.1 2.1
SF-2.4 2.4
TL-1.0 1.0
TL-200 200
TL-300 1.0 300 0.027 𝜑8@845
TL-400 400
TL-500 500
LR-
0.022
2.24
LR-
1.0 1.0 0.027 𝜑8@845
2.73
LR-
0.030
3.01
TR-
𝜑8@845
845
TR-
1.0 1.0 0.027 𝜑8@200
200
TR-
𝜑8@120
120

4.1. Scale Factor
The change in the dimensions of the tapered beam from the length of the beam 1900mm, the width 150mm
and the height at the supported 180mm and in the middle of the beam 405mm is gradually changed to
five stages until it reaches dimensions close to semi-scale. Figure 1. the geometrical detail of tapered
beam (SF-1.0) while Table 5. depicts the related results analysis. Results show an increase in shear
strength and deflection compared to Model SF-1.0 by 348% and 135%, respectively. This is from the
representation of the volume of the concrete with dimensions that seem semi-scale where the concrete
receiving his works as resistance the greater compression force.
Page 489
SEMIT 2023
2500
2000 SF-2.4
SF-2.1
Load (kN)
1500
SF-1.8
1000 SF-1.6
SF-1.3
500
SF-1.0
0
0 7 14 21 28 35 42 49
Displacement (mm)
Fig.5. Load-Displacement for Scale Factor Effect.

Table 5. Effect of Scale Factor on Tapered Beam.
First
Shear Ultimate
Beam Scale Crack
Strength Deflection
ID Factor Load
(kN) (mm)
(kN)
SF-1.0 1.0 172.0 465.4 17.6
SF-1.3 1.3 183.0 691.5 20.3
SF-1.6 1.6 284.5 994.1 26.4
SF-1.8 1.8 346.8 1280.7 30.0
SF-2.1 2.1 402.9 1687.1 37.5
SF-2.4 2.4 499.5 2085.3 41.4
4.2. Crown Length

The Tapered Length region within overall tapered modes is investigated numerical-ly. The maintenance
of section strength with uniformity region reduction means, higher strength rating (ultimate strength /
beam overall weight) and mean reduction in the attainted dead load. Wide range of extended regions
distance as variable are considered, the adopted numerical modeling details are summarized in Figure 6.
The obtained results show that, as uniformity region increased from 0 to 500 mm, the ultimate shear
strength gets the increasing 13% improved from 465.4 kN to 526.7 kN. Table 6. exhibits results analysis
that relates to a tapered length investigation. These observations could be clearly monitored by the
introduced load -deflection curves that shown in Figure 7. and related ultimate strength-deflection.
Page 490
SEMIT 2023
Fig. 6. Specimens' Modelling Details.

Table 6. Effect of Crown Length on Tapered Beam.
First
Crown Shear Ultimate
Beam Crack
Length Strength Deflection
ID Load
(mm) (kN) (mm)
(kN)
TL-1.0 1.0 172.0 465.4 17.6
TL-200 200 173.6 511.7 15.30
TL-300 300 175.5 509.7 15.12
TL-400 400 190.1 524.0 15.10
TL-500 500 204.7 526.7 15.06
Page 491
SEMIT 2023
600
500
400
Load (kN)
300 TL-1.0
TL-200
200
TL-300
100 TL-400
0
TL-500
0 3 6 9 12 15 18 21
Displacement (mm)
Fig. 7. Load-Displacement for Crown Length Effect.

4.3. Reinforcement Ratio
4.3.1. Longitudinal Reinforcement
The effect of longitudinal steel ratio within ductility limits, on section flexural response is considered
and attainted in specimens (LR-2.24, LR-2.73, LR-3.01) The related specimens' modelling details are
shown in Table 4. The numerical results are summarized in Table 7. The section strength is improved
from 501 kN to 532 kN as steel ratio change from ρ=0.022 to ρ= 0.030 which are within ductility limit,
the over reinforcement section (ρ= 0.039), turns the section to be over reinforcement with brittle failure
mode. Figure 4-28 shows the observed failure modes in scope of crack patterns total Strain distribution.
Results showed that the ultimate shear strengths increase with increasing longitudinal reinforcement
ratio. The increase in longitudinal reinforcement ratio from LR-2.73 to LR-3.01 results in a 14%
increase in ultimate shear strength.
600
500
400
Load (kN)
300 LR-3.01
200 LR-2.73
100
LR-2.24
0
0 3 6 9 12 15 18 21
Displacement (mm)
Fig. 8. Load-Displacement for Longitudinal Reinforcement Ratio.
Table 7. Effect of Longitudinal Reinforcement on Tapered Beam.
First
Shear Ultimate
Beam Crack
𝝆𝒙𝟏𝟎−𝟐 Strength Deflection
ID Load
(kN) (mm)
(kN)
LR-2.24 2.24 162.6 501.5 19
LR-2.73 2.73 172.0 465.4 17.6
LR-3.01 3.01 170.7 531.7 18.7
*𝜌: Longitudinal Reinforcement Ratio
4.3.2. Transvers Reinforcement
The study was to verify the amount of transverse reinforcement using ultra high performance concrete
with the addition of 2% of fiber steel. The addition of transverse reinforcement in three phases with the
same diameter as 8mm bar and at different distances every 850mm, 200mm and 120mm as shown in
Page 492
SEMIT 2023
table 4. results show that the increased amount of reinforcement resulted in a significant increase in the
load capacity by 28% and 44% respectively compared to the standard model TB-TR845 above. There is
a significant effect on the mid-span deflection as the increase was observed by 79% and 148% on taking
over compared to the model TB-TR845 the results are clear in Figure 9. As the pattern of Failure Turns
Into flexural-shear failure mode.
800
700
600
Load (kN) 500
400
300 TR-845
200 TR-200
100
TR-120
0
0 7 14 21 28 35 42 49
Displacement (mm)
Fig. 9. Load-Displacement for Transvers Reinforcement.

Table 8. Effect of Transvers Reinforcement on Tapered Beam.
First
Transvers Shear Ultimate
Beam Crack
Reinforce- Strength Deflection
ID Load
ment (kN) (mm)
(kN)
TR-845 𝜑8@845 172.0 465.4 17.6
TR-200 𝜑8@200 178.3 596.0 31.5
TR-120 𝜑8@120 187.8 668.4 43.7
a. TR-845
b. TR-200
c. TR-120
Fig. 10. Failure modes (crack patterns).
5. Conclusions
The present numerical study is focused on the behavior of reinforced concrete Tapered beams. Based on
the results obtained from the FEM by using ABAQUS program. The comparison of nine numerical
Page 493
SEMIT 2023
models with experimental results in the previous study accurately predicted the structural behaviors of
UHPC beams models and showed a very good compatibility ranging from 0.96 to 1.13. Many factors
have a significant effect on the general behavior of the beam at failure and these effects can be
summarized as follows:
1. When the scale factor 2.4 was increased, an increase in shear strength was observed at 348%. This
is from the representation of the volume of the concrete with dimensions that seem semi-scale
where the concrete receiving his works as resistance the greater compression force.
2. The crown's length is increased from 0 to 200 mm, the shear strength is increased by 10%. After
that increase in crown length from 200-500 mm there was a slight increase.
3. The ultimate shear strengths increase with increasing longitudinal reinforcement ratio. This is may
be due to the increase of the dowel action component and therefore increased shear strength.
4. The greater the transverse reinforcement ratio, the more cracks occurred with a smaller opening’s
width. The increased shear reinforcement ratio increased the shear strength of the tested beams,
resulting in shear cracking under greater load. When the shear reinforcement was increased, the
usual Combined shear-flexural failure mode developed.
References
1. Ibrahim, S.K. and M.M. Rad. Numerical Plastic Analysis of Non-Prismatic Reinforced Concrete
Beams Strengthened by Carbon Fiber Reinforced Polymers. in Proc. of the 13th fib International
PhD Symposium in Civil Engineering. 2020. Marne-la-Vallée Paris France.
2. Al-Ahmed, A.H.A., A.H. Al-Zuhairi, and A.M. Hasan. Behavior of reinforced concrete tapered
beams. in Structures. 2022. Elsevier. https://doi.org/10.1016/j.istruc.2022.01.080.
3. Nimnim, H.T., A.T. Jassim, and A.A. Abdul Mohsen, Structural behavior of reinforced reactive
powder concrete tapered beams. Practice Periodical on Structural Design and Construction, 2019.
24(2): p. 04019002. https://doi.org/10.1061/(asce)sc.1943-5576.0000416.
4. Toshniwal, S., Shear Analysis of Non-Prismatic Concrete Beams. 2019.
5. Balaji, G. and R. Vetturayasudharsanan, Experimental investigation on flexural behaviour of RC
hollow beams. Materials today: proceedings, 2020. 21: p. 351-356.
https://doi.org/10.1016/j.matpr.2019.05.461.
6. Aziz, A.H., H.F. Hassan, and F.M.A. Razzaq, Experimental study on shear behavior of reinforced
self-compacted concrete tapered beams. Civ. Environ. Res, 2016. 8(8): p. 11-22.
7. Tena-Colunga, A., H.I. Archundia-Aranda, and Ó.M. González-Cuevas, Behavior of reinforced
concrete haunched beams subjected to static shear loading. Engineering Structures, 2008. 30(2): p.
478-492.
8. Hou, C., K. Matsumoto, and J. Niwa, Shear failure mechanism of reinforced concrete haunched
beams. Journal of JSCE, 2015. 3(1): p. 230-245.
9. Tayeh, B.A., L.K. Askar, and B.A. Bakar, Ultra-High-Performance Concrete (UHPC)-Applications
Worldwide: A State-of-the-Art Review. Journal of Engineering Research and Technology, 2023.
10(1).
10. Lateef, H.A., R.M. Laftah, and N.A. Jasim, Study on the Shear Failure of Reinforced Concrete
Beams Using Extended Finite Element Method (XFEM). Basrah Journal for Engineering Sciences,
2021. 21(3): p. 55-65.
11. Jajodia, R. and S. Gadve, Shear strength of plain concrete. Strain, 2023. 59(3): p. e12436.
12. Rajesh, C. and G.R. Kumar. A new shear test setup for determining shear strength of normal and
high strength concrete. in Structures. 2023. Elsevier.
Page 494
SEMIT 2023
13. JSCE, F., Recommendations for design and construction of ultra-high strength fiber reinforced
concrete structures (Draft). Japan Society of Civil Engineers Tokyo, Japan, 2004.
14. Lee, J. and G.L. Fenves, Plastic-damage model for cyclic loading of concrete structures. Journal of
engineering mechanics, 1998. 124(8): p. 892-900.
15. Lubliner, J., et al., A plastic-damage model for concrete. International Journal of solids and
structures, 1989. 25(3): p. 299-326.
16. Tu’ma, N.H., M.R. Aziz, and H.J.J. Barry, Residual Tensile Stress Estimation for Shear Strength of
UHPC Nonprismatic Beams. Civil and Environmental Engineering, 2021. 17(1): p. 164-177.
17. Tu’ma, N.H. and A.I. Said, Yield Line Model for Estimation Shear Strength UHPC Non-prismatic
Beams, in Geotechnical Engineering and Sustainable Construction: Sustainable Geotechnical
Engineering. 2022, Springer. p. 523-530.
18. Hakeem, N., Whitney Block Estimation for Flexural Strength of UHPC Beams. Journal of University
of Babylon for Engineering Sciences, 2022. 30(4): p. 21-35.
19. Said, A.I. and N.H. Tu’ma. Numerical modeling for flexural behavior of UHPC beams reinforced
with steel and sand-coated CFRP bars. in IOP Conference Series: Earth and Environmental
Science. 2021. IOP Publishing.
Page 495
SEMIT 2023
The Present and Future Scope of Virtual Reality in the Construction Industry
Aditya Singh1, *(0000-0001-9347-5627)
1
School of Civil Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, Tamil Nadu, India
*Email: aditya7777777k@gmail.com
Abstract
Virtual Reality is developing and spreading all over the world at a rapid rate, its applications are
being increasing with time. In the recent years, the presence of virtual reality can be observed in
the construction industry as well. With the technological advancements in the recent decades, the
traditional way of construction is being transformed slowly and steadily, due to which the
construction industry is no longer separate from advanced technologies like virtual reality. In this
paper, the focus on virtual reality was given and its significance in the construction sector at
present. Further, the potential of virtual reality to change as well as transform the construction
industry in the future, is covered in the paper. Some research gaps in the current studies in addition
to graphical analysis based on the obtainable data from different sources were performed to support
the study.
Keywords: Virtual Reality, Construction 4.0, Quality Control, Construction Safety, Sustainability
1 Introduction
Virtual Reality in short is often called as VR and it is a simulated three dimensional environ which allows
the user to not only reconnoitre but also the ability to intermingle with a given virtual surrounding in a
manner which comes closer to reality as how the senses of the user perceived it. With the assistance of a
software as well as computer hardware, it is possible to create such a virtual environment, granting that
the users are required to wear some devices like goggles or helmets, in order to have the ability to interact
with that particular created VR environment. If a user is able to immerse himself or herself too deeply in
a virtual reality environment provided that the nearby physical surrounding of that user is blocked out,
then that user can hang his or her faith as well as even believe that a bizarre VR environment is a real one
[24]. In other words, for a VR user to have an immersive feel of a given virtual world, VR technology
provides a simulated experience to that user, which can hire three dimensional near eye displays in
addition to pose tracking, to make it work. Mostly, Virtual Reality is applied in the entertainment industry,
especially the gaming industry; also in the education sector while training the user in military or medical
and healthcare industry, in addition to in businesses as well, where virtual meetings are organized when
required. Even particularly designed rooms having a number of large screens can provide the same VR
experience which is normally attained with the assistance of Virtual Reality headsets comprising of head
mounted display along with small screen to be placed in front of the VR user. Then, with the assistance
of haptic technology force feedback as well as additional sensory perception and input might be permitted,
apart from the classical inclusion of video as well as auditory feedback in Virtual Reality. On a careful
look into a brief history, from the middle of 15th century, the term virtual meant something which is being
present in essence or has effect, nonetheless not present in reality or factually present, which shows that
the concept of Virtual Reality was up to certain extent was already present in the recorded history. From
the year 1959 onwards, it was being used in the computer sense, where it meant that the software is
helping a non-existent thing in realism to appear. However, it is important to note that Antonin Artaud
who was a French avant-grade playwright, in the latter half of the fourth decade of the 20st century,
labelled the illusory nature of objects as well as characters as “la réalité virtuelle” which means Virtual
Reality in the English language, which was used in regards to theatre in his collection of essays, later on
translated in English a year before 1959, which shows the earliest usage of the word Virtual Reality in
any published form based on the records. However, it can be found that Myron Krueger created the term
Artificial Reality for the same in the 8th decade of the 20th century; but by the early ninth decade of the
20th century, Damien Broderick brought back the term Virtual Reality in the limelight after mentioning
in his science fiction novel. Further, the 1st business oriented VR hardware was designed by Jaron Lanier
Page 496
SEMIT 2023
and his company selling the products just began the real competition in the business world. By the early
last decade of the 20th century, Lawnmower Man was the movie which also added the popularity of VR
technology among the masses and then the development of VR technology significantly increased later
on in the succeeding decades [23]. At present, people generally first think of video games after hearing
the word Virtual Reality, and that is because the gaming industry utilised the VR technology very
efficiently so far unlike other industries. This led to the success of popular video games like Minecraft
VR, Beat Saber, in addition to Skyrim VR. However, it is also important to not forget about other areas
where promising growth of Virtual Reality is visible in the recent years, which are following [22]:
Education: Virtual Reality could give educational institutions novel methods to help not only medical,
aerospace engineering and defence students to learn as well as understand in an effective manner various
complex things, but also it could help construction management in addition to construction technology
engineering students to understand the concepts in a better way. This also increases their involvement
and activity level during their lectures. This might help them to understand and realize the terrain of
faraway inaccessible places in a way better manner.
Training: With Virtual Reality, it is easy to train piolets in commercial planes as well as in fighter planes
and helicopter. The same can be done to train doctors, surgeons, police officers, firefighters, and so on.
Similarly, labourers, engineers, and project staffs could also be trained in a proper way, which might
increase their awareness, alertness, sharpness and insight while working on a real project.
Retail: VR technology is useful in making the right decisions while selecting decorative items, clothes,
spectacles, and other daily fashion items, which will improve the retail business greatly and reduce the
chances of wrong selection considerably in the first time. However, only minor things can be seen for
people working in a project, like selecting the right size of safety equipment, work equipment and upto
some extent the quantity of materials. Still, this needs further research of how much VR technology can
be of a significant use in this case.
Healthcare: The medical and healthcare industry is getting some benefits and the quality of medical
professionals are expected to improve further. This directly doesn’t help in any project but the quality of
medical professionals being enhanced also increases the survival as well as recovery of labourers,
engineers, and other project staffs who often becomes a victim of mishaps during the course of a project.
Entertainment: Mostly Virtual Reality is benefiting the gaming industry but the remaining entertainment
industry has also noticeable chance to grow further, which was observed during the pandemic. Just like
the above one, it doesn’t have any direct effect on any project, but engineers, labourers, and other project
staffs who often have to relocate to different places based on the project location easily develop work
related stress and that affects their work. So, with a balanced way of some VR entertainment it might
reduce their work related stress and make them relax in unfamiliar locations after their work hours, which
will make them fresh before their next working schedule.
Real Estate: VR technology is expected to improve the real estate and make it more profitable. For
instance, touring HVAC systems become easier for building engineers, detailed plan of houses or
buildings become easier to understand in three dimensional view which also decreases the chance of
faults as well as improving any part if needed, and also observing the remodelling of the houses by owners
whenever required at their own comfort.
The industry of Virtual Reality has a long way to go before apprehending its vision of a complete
immersive environ which can permit VR users to involve numerous sensations in such a way which
approaches almost close enough to the real world. Nevertheless, VR technology has already reached some
major milestones, which includes giving genuine sensory engagement as well as demonstrates its
potential for a number of businesses in various industries.
1.1 What is the need and role of Virtual Reality in the development of the construction industry?
Virtual Reality is an advanced technology and one of the revolutionary technologies in the current
century. There is no doubt the gaming industry has been showing significant development with the
application of Virtual Reality but overall the entertainment industry has been improved considerably with
the appearance of VR. Further, medical and healthcare industry also shows that the Virtual Reality has a
Page 497
SEMIT 2023
good potential in enhancing the industry more in the future. Then, Psychologist also found that Virtual
Reality will help in the advancement of the human psychology and in improving mental health. There are
other areas also which are showing signs that VR technology might help in their transformation in the
future. Now, with the coming of advanced technologies in other industries, it is natural to affect the
construction industry slowly and steadily. Plus, advanced technologies like Virtual Reality can help in
developing and leading the construction industry towards Construction 4.0. Moreover, Virtual Reality is
compatible with other advanced technologies like AI, BIM, Robotics, Automation, and so on, which can
further help in the advancement of the construction industry considerably in the future [29][30].
1.2 Objectives
There are a several objectives of this paper, which are as follows:
• To understand the concept of Virtual Reality in the general sense and the difference between
Virtual Reality as well as in Augmented Reality.
• To understand the gaps in the current research in applying Virtual Reality in the construction
industry.
• To understand the current as well as future scope of Virtual Reality and how much beneficial it
is for the construction industry.
2 Motivation
The author performed this study as Virtual Reality technology is growing at an alarming rate over the
recent decade and its strong presence can be seen in various industries like entertainment which include
advanced games, medical and healthcare industry and so on. The author being a civil engineer felt that it
is important to understand how an advanced technology like Virtual Reality can be important and play
any major role in the development of the construction industry through this study.
3 Literature Review
Bakhoum et al. (2023) studied virtual reality and they worked on examining the impact of executing it in
the construction industry in the country Egypt. They developed Oculus Quest 2 as well as Fuzor VDC to
elaborate the use of technology, and in their study it also included four dimensional simulation as well as
meeting in VR. Further, they performed a survey to get the desired results [1]. Kang et al. (2023)
performed a quantitative approach in order to understand the distance estimation’s reliability in a given
VR space, particularly intended for construction management [2]. da Silva et al. (2022) studied virtual
reality and they worked on analysis of robotic construction by creating a simulation with the assistance
of it, in order to detect construction problems, forecast cost as well as duration of construction process,
through their study [3]. Luo et al. (2023) studied the design of an immersive VR based safety training in
the construction sector and they examined the impact of interaction fidelity as well as scenario on such
training experiences [4]. Hong et al. (2023) studied virtual reality and they worked on an analysis based
on it to perform on the exposure of the construction noise’s effects especially on work productivity of
masonry [5]. Rokooei et al. (2023) studied VR and they worked on applying it in the safety training in
the construction work [6]. Zhang et al. (2023) studied modular construction and they worked on VR
improved multirole collaboration particularly in the training of crane lifting intended to aid their studied
direction [7]. Bao et al. (2022) studied cross platform VR and they worked using it especially for the
safety training in the case of real time construction scenario with the assistance of foundation classes of
immersive web as well as industry [8]. Zhang et al. (2022) studied VR technology and they worked on
an extended technology acceptance model using it to support the construction safety training through
their study [9]. Han et al. (2022) performed an assessment on the basis of process as well as outcome to
differentiate VR driven construction safety training and conventional one [10]. Han et al. (2021) worked
in the direction of a complete construction design review with the support of VR technology and they
made a framework which was intended for semiautomatic one to detect completely the obstructed objects
in three dimensional models [11]. Dallasega et al. (2020) worked on Lean Construction Management and
they empower it using AR, BIM as well as VR, in order to make project simulated game, to obtain desired
Page 498
SEMIT 2023
study outcomes [12]. Kim et al. (2021) performed a VR experiment in order to forecast the carelessness
of workers to be struck by dangers through the observation of bio signals while performing a given
construction task [13]. Zhang and Pan (2021) studied VR technology and they worked on using it to aid
in the layout planning of an interactive tower crane, which was specifically intended for high rise modular
integrated construction through their study [14]. Delgado et al. (2020) studied VR as well as AR
technologies and they worked on a research programme which was intended for such technologies to be
used in engineering, architecture in addition to construction, to obtain the desired study results [15]. Jeon
and Cai (2021) studied VR as well as wearable electroencephalogram, and they worked on categorizing
hazards associated to perceptions in the construction projects with their assistance [16]. Getuli et al.
(2020) studied BIM and they worked on immersive VR on its basis, which was intended for the planning
of construction workspace [17]. Dinis et al. (2020) worked on linking laser scanning as well as VR, in
order to enhance project communication in the construction, engineering in addition to architecture
industries [18]. Shi et al. (2020) performed an experiment which was a pipe maintenance one using VR,
and they further worked on finding out the impact of engineering information formats on implementation
in addition to learning of a given construction operation through their study [19]. Li et al. (2023) worked
on enhancing bridge teaching communication by using immersive VR as a tool to perform their study
[20].
4 Research Gaps
After going through around 60 published papers, only the ones which were published in the recent years
were further considered. Based on the recent scientific and research papers or review papers, it was found
that most of the research on use of virtual reality in the construction industry in the recent years was
particularly performed in China and also from developed countries. There was lack of sufficient research
done in the recent years in regards to the use of virtual reality in the construction industry in India. More
research on the use virtual reality in other areas like medical and health care, entertainment, human
psychology as well as mental health and so on, in comparison to the construction industry.
5 Main Focus of the Paper Along with Issues and Problems

The paper focussed on explaining the concept of Virtual Reality and how it could benefit the construction
industry, in some detail. This paper also talked about the differences between Augmented Reality and
Virtual Reality, as well as how Virtual Reality is compatible with other advanced technologies, which
together can improve the construction industry considerably. However, there are a number of problems
and issues with Virtual Reality while using the technology in the construction industry, which can’t be
ignored. There is a significant difference between the real scenario in any construction project and what
could be observed through Virtual Reality. At times, the difference between the two is high enough which
might confuse the civil engineers or make their training through Virtual Reality of not much use. Then,
cost for the setup of Virtual Reality on a large scale is high enough to make them expensive, which might
make the contractor hesitate to use it. Many civil engineers who are used to the traditional ways of
construction practices, won’t be able to get used to Virtual Reality, as they lack any prior knowledge
about VR technology and it might be difficult for them to use sophisticated VR technology suddenly.
6 Main Types of Virtual Reality

The Virtual Reality systems differ considerably from one system to the other, based on the type of
technology utilized and their purpose, though they normally can be part of three main types, which are as
follows [22]:
• Non Immersive
• Semi Immersive
• Fully Immersive
Page 499
SEMIT 2023
Non Immersive
Non Immersive is a type of Virtual Reality, which denotes a three dimensional simulated environment
which can be accessed with the assistance of a computer screen. Based on the type of program used, this
type of virtual environment can also produce sound. In this case, the user has fractional control over this
type of virtual environment with the aid of a mouse, keyboard or other devices. Nonetheless this type of
virtual environment can only interact with the user indirectly. For understanding purpose, a video game
where a website can allow a user to help in designing the décor of a room, can serve as a good instance
of a Non Immersive Virtual Reality system.
Semi Immersive
Semi Immersive is another type of Virtual Reality, but compared to the earlier one, it offers an experience
which is partially virtual in nature, which can also be accessed with the assistance of a headset or some
specific type of glasses or computer screen. Main aim of this type of VR system is on the visual three
dimensional aspect of virtual reality, so it doesn’t execute any physical movement like the Fully
Immersive VR system does. A flight simulator which is used to train pilots in the military as well as in
the airlines, is a good instance of this type of VR system.
Fully Immersive
Fully Immersive is another major type of Virtual Reality, but in comparison to the other VR types, it
gives the highest level of Virtual Reality experience, where the user is entirely immersed in a given
simulated three dimensional world. It executes not only sound and sight, but also touch in some cases.
Some experiments had been even performed where smell was also added. To completely interact with
the virtual world, the users are require to wear special equipment like gloves, goggles, and helmets.
Equipment like stationary bicycles or treadmills can also be incorporated in the virtual environment, to
give the user an experience where the user is moving through the three dimensional space. This type of
Virtual Reality technology is still in the early or initial stage, nevertheless it had made inroads up to some
extent in the healthcare industry, but significant contribution in the gaming industry specifically. Also, it
is attracting other industries towards itself slowly and steadily.
7 Difference between Augmented Reality and Virtual Reality

AR and VR are many times used together and noncore people might not be able to differentiate both of
them. However, there is a major difference between them which are as follows [21][24]:
• In the case of Augmented Reality, a real world setting is used. On the other hand Virtual Reality
is entirely virtual.
• Smartphone can be used to access Augmented Reality environment, while device like a headset
is required in the case of Virtual Reality.
• Users in Augmented Reality has the ability to control their respective presence in the real world,
but in the case of Virtual Reality, the system control the users.
• Augmented Reality has the ability to improve both the real world as well as the virtual world.
However, a fictional reality can only be improved in the case of a Virtual Reality.
8 Methodology
In this section, application of Virtual Reality is shown in the construction industry, where numerous
benefits of Virtual Reality can be observed during the course of a construction project.
Page 500
SEMIT 2023
Figure 1: Flowchart showing application of VR technology in construction projects [25][26]
9 Result and Discussion

In this section, data from different sources were collected to perform graphical analysis.
Table 1: AR and VR Market Revenue of Top Countries [27]

AR & VR Market Revenue of Major Countries in 2023 In USD
China 6,606,000,000
Japan 1,880,000,000
USA 8,568,000,000
Table 2: Market Growth of VR in India [27]

Market Revenue in India In USD
VR Hardware in 2017 19,150,000
VR Software in 2017 6,490,000
VR Hardware in 2027 273,000,000
VR Software in 2027 121,400,000
Table 3: VR Market Dominance of USA in the World [28]

VR Users Comparison Total VR Users
USA 65,900,000
World 171,000,000
Table 4: Worldwide Market Growth of VR [28]

Global VR Market Growth In USD
Global Value of VR Market in 2016 2,020,000,000
Global Value of VR Market in 2027 26,900,000,000
Page 501
SEMIT 2023
8,568,000,000
10,000,000,000
6,606,000,000
8,000,000,000
6,000,000,000
4,000,000,000 1,880,000,000
2,000,000,000
0
In USD
China Japan USA
Figure 2: AR and VR Market Revenue of Top Countries [27]
121,400,000
273,000,000
In USD 6,490,000
19,150,000
0 100,000,000 200,000,000 300,000,000
VR Software in 2027 VR Hardware in 2027

VR Software in 2017 VR Hardware in 2017
Figure 3: Market Growth of VR in India [27]
180,000,000
160,000,000 171,000,000
140,000,000
120,000,000
100,000,000
65,900,000
80,000,000
60,000,000
40,000,000
20,000,000
0
TOTAL VR USERS
USA World
Figure 4: VR Market Dominance of USA in the World [28]
Page 502
SEMIT 2023
Global Value of VR Market in 2016 Global Value of VR Market in 2027
30,000,000,000 26,900,000,000
25,000,000,000
20,000,000,000
15,000,000,000
10,000,000,000
5,000,000,000 2,020,000,000
0
In USD
Figure 5: Worldwide Market Growth of VR [28]
9.1 Advantages of Virtual Reality

In this subsection, the various advantages of Virtual Reality and how it could it make a positive impact
in the construction industry would be discussed, based on the analyses of the above graphs:
• According to Fig. 2, USA is having AR and VR market revenue of over $8.56 Billion USD,
whereas Japan has over $1.8 Billion USD in the year 2023. This also means that advanced
countries are using technologies like Virtual Reality on a large scale. Then, China which is a
major developing country also has a strong AR and VR market revenue of over $6.6 Billion USD
in the same year, second only to USA. There is no doubt that China over the decades developed
at an astonishing rate and use of advanced technologies might have played a major role in its
development. It is safe to say that the developing countries like India and so on in the future could
invest more in advanced technologies like VR to increase their rate of development.
• According to Fig. 3, the Virtual Reality Hardware had a market revenue over $19 Million USD
in the year 2017 in India, but it is expected that its market revenue will reach $273 Million USD,
which means that around 14.26 times the market revenue is expected to increase in the span of
10 years only. Similarly, the market revenue of Virtual Reality Software in the year 2017 in India
was over $6.4 Million USD, but by the year 2027, it is expected to reach over $121 Million USD,
which means that its market revenue in the country might increase by 18.71 times in the span of
10 years. It is safe to say that with time, Virtual Reality Software and Hardware will have more
demand in the future. Since, both Virtual Reality Software and Hardware will be highly profitable
in India in the coming future, so it will be a smart move to invest more in Virtual Reality Software
and Hardware, in addition to it will be good to apply Virtual Reality technology in the
construction sector as well.
• Based on Fig. 4, there are one hundred seventy one Million Virtual Reality users in the world, on
the other hand close to sixty six Million users of Virtual Reality from USA. This means that over
38.5% users of Virtual Reality in the world are alone from America, which also shows the
dominance of America in the world of Virtual Reality. USA being an advanced and developed
country is using advanced technologies like Virtual Reality way more than other countries, also
shows how Virtual Reality which is an advanced technology is helping in the further development
of USA. It is safe to say, that the developing and underdeveloped countries also need to use
advanced technologies like Virtual Reality more in the future in order to reach the current point
of development of advanced and developed countries.
• Based on Fig. 5, the worldwide market value of Virtual Reality was over $2 Billion USD in the
year 2016, but by the year 2027, it is expected that its worldwide market value will reach close
to $27 Billion USD. This means that the market value of Virtual Reality on a global level, is
expected to increase by 13.32 times in the span of 11 years, which shows a significant positive
Page 503
SEMIT 2023
increase in its market value. So, it is safe to say that more investment can be done on Virtual
Reality on an international level, as the global demand of Virtual Reality is increasing at a
tremendous rate. Also, it will be a wise option to expand the application of Virtual Reality to
other areas also like construction industry, apart from the main ones.
10 Conclusion
The paper explained the concept of Virtual Reality in the general sense and a brief history of VR
technology. It also discussed the way Virtual Reality is having the potential to affect other areas and how
projects are also affected through them. Then, this paper talked about the way VR technology can help in
the development of the construction industry and some gaps in the current research. Some problems and
issues were also discussed in the paper in regards to the application of Virtual Reality in the construction
industry and how it could be applied in a construction project to improve it in some aspects. Further,
some graphical analyses were performed on data available from different sources, to support the study.
Since, both Virtual Reality Software and Hardware will be highly profitable in India in the coming future,
so it will be a smart move to invest more in Virtual Reality Software and Hardware, in addition to it will
be good to apply Virtual Reality technology in the construction sector as well. Also, the developing and
underdeveloped countries also need to use advanced technologies like Virtual Reality more in the future
in order to reach the current point of development of advanced and developed countries. Some more
research in the future is required to be done to reduce the current limitations of Virtual Reality while
applying in the construction industry.
Acknowledgement
This research received no specific grant from any funding agency in the commercial, public, or not for
profit sectors.
References
1. Bakhoum, E.S., Younis, A.A., Aboulata, H.K., and Bekhit, A.R.: Impact assessment of implementing
virtual reality in the Egyptian construction industry. Ain Shams Engineering Journal, Volume 14,
Issue 6, June 2023, 102184.
2. Kang, J.H., Yadav, N., Ramadoss, S., and Yeon, J.: Reliability of distance estimation in virtual reality
space: A quantitative approach for construction management. Computers in Human Behavior,
Volume 145, August 2023, 107773.
3. da Silva, N.P., Eloy, S., and Resende, R.: Robotic construction analysis: simulation with virtual
reality. Cell Press, Volume 8, Issue 10, October 2022, e11039.
4. Luo, Y., Ahn, S., Abbas, A., Seo, J.O., Cha, S.H., and Kim, J.I.: Investigating the impact of scenario
and interaction fidelity on training experience when designing immersive virtual reality-based
construction safety training. Developments in the Built Environment, Volume 16, December 2023,
100223.
5. Hong, J., Choi, J., Lee, J.H., Cho, S.H., Hong, T., Han, S.U., Park, H.S., and Lee, D-E.: Virtual
reality-based analysis of the effect of construction noise exposure on masonry work productivity.
Automation in Construction, Volume 150, June 2023, 104844.
6. Rokooei, S., Shojaei, A., Alvanchi, A., Azad, R., and Didehvar, N.: Virtual reality application for
construction safety training. Safety Science, Volume 157, January 2023, 105925.
7. Zhang, Z., Wong, M.O., and Pan, W.: Virtual reality enhanced multi-role collaboration in crane-lift
training for modular construction. Automation in Construction, Volume 150, June 2023, 104848.
8. Bao, L., Tran, S.V-T., Nguyen, T.L., Pham, H.C., Lee, D., and Park, C.: Cross-platform virtual reality
for real-time construction safety training using immersive web and industry foundation classes.
Automation in Construction, Volume 143, November 2022, 104565.
Page 504
SEMIT 2023
9. Zhang, M., Shu, L., Luo, X., Yuan, M., and Zheng, X.: Virtual reality technology in construction
safety training: Extended technology acceptance model. Automation in Construction, Volume 135,
March 2022, 104113.
10. Han, Y., Yang, J., Diao, Y., Jin, R., Guo, B., and Adamu, Z.: Process and Outcome-based Evaluation
between Virtual Reality-driven and Traditional Construction Safety Training. Advanced Engineering
Informatics, Volume 52, April 2022, 101634.
11. Han, B., Ma, J.W., and Leite, F.: A framework for semi-automatically identifying fully occluded
objects in 3D models: Towards comprehensive construction design review in virtual reality.
Advanced Engineering Informatics, Volume 50, October 2021, 101398.
12. Dallasega, P., Revolti, A., Sauer, P.C., Schulze, F., Rauch, E.: BIM, Augmented and Virtual Reality
empowering Lean Construction Management: a project simulation game. Procedia Manufacturing,
Volume 45, 2020, Pages 49-54.
13. Kim, N., Kim, J., and Ahn, C.R.: Predicting workers’ inattentiveness to struck-by hazards by
monitoring biosignals during a construction task: A virtual reality experiment. Advanced Engineering
Informatics, Volume 49, August 2021, 101359.
14. Zhang, Z., and Pan, W.: Virtual reality supported interactive tower crane layout planning for high-
rise modular integrated construction. Automation in Construction, Volume 130, October 2021,
103854.
15. Delgado, M.D., Oyedele, L., Dermian, P., and Beach, T.: A research agenda for augmented and
virtual reality in architecture, engineering and construction. Advanced Engineering Informatics,
Volume 45, August 2020, 101122.
16. Jeon, J.H., and Cai, H.: Classification of construction hazard-related perceptions using: Wearable
electroencephalogram and virtual reality. Automation in Construction, Volume 132, December 2021,
103975.
17. Getuli, V., Capone, P., Bruttini, A., and Isaac, S.: BIM-based immersive Virtual Reality for
construction workspace planning: A safety-oriented approach. Automation in Construction, Volume
114, June 2020, 103160.
18. Dinis, F.M., Sanhudo, L., Martins, J.P., and Ramos, N.M.M.: Improving project communication in
the architecture, engineering and construction industry: Coupling virtual reality and laser scanning.
Journal of Building Engineering, Volume 30, July 2020, 101287.
19. Shi, Y., Du, J., and Worthy, D.A.: The impact of engineering information formats on learning and
execution of construction operations: A virtual reality pipe maintenance experiment. Automation in
Construction, Volume 119, November 2020, 103367.
20. Li, W., Zhu, J., Dang, P., Wu, J., Zhang, J., Fu, L., Zhu, Q.: Immersive virtual reality as a tool to
improve bridge teaching communication. Expert Systems with Applications, Volume 217, 1 May
2023, 119502.
21. Singh, Aditya. "8 Augmented Reality and its use in the field of civil engineering". Handbook of
Augmented and Virtual Reality, edited by Sumit Badotra, Sarvesh Tanwar, Ajay Rana, Nidhi
Sindhwani and Ramani Kannan, Berlin, Boston: De Gruyter, 2023, pp. 117-136.
https://doi.org/10.1515/9783110785234-008.
22. Tech Target, https://www.techtarget.com/whatis/definition/virtual-reality, last accessed on
12/09/2023.
23. Wikipedia, https://en.wikipedia.org/wiki/Virtual_reality, last accessed on 12/09/2023.
24. Tulane University School of Professional Advancement, https://sopa.tulane.edu/blog/whats-
difference-between-ar-and-
vr#:~:text=AR%20uses%20a%20real%2Dworld,only%20enhances%20a%20fictional%20reality,
last accessed on 12/09/2023.
25. Neuroject, https://neuroject.com/virtual-reality-in-
construction/#:~:text=By%20leveraging%20VR%2C%20architects%20and,costly%20errors%20pri
or%20to%20construction, last accessed on 12/09/2023.
26. Program-Ace,https://program-ace.com/blog/vr-in-
engineering/#:~:text=Virtual%20reality%20in%20civil%20and%20structural%20engineering,-
Civil%20engineering%20is&text=In%20a%20similar%20vein%2C%20structural,objects%20in%2
0a%20virtual%20environment, last accessed on 12/09/2023.
Page 505
SEMIT 2023
27. Statista, https://www.statista.com/outlook/amo/ar-

vr/india#:~:text=Revenue%20in%20the%20AR%20%26%20VR,US%24211.30m%20in%202023,
last accessed on 12/09/2023.
28. Zippia, https://www.zippia.com/advice/virtual-reality-
statistics/#:~:text=Virtual%20Reality%20Statistics%20by%20Users,of%20VR%2FAR%20device
%20users, last accessed on 12/09/2023.
29. Singh, A. (2023). Mechatronics Engineering in the Modern World and Its Use in the Construction
Industry. In M. Mellal (Ed.), Trends, Paradigms, and Advances in Mechatronics Engineering (pp. 64-
92). IGI Global. https://doi.org/10.4018/978-1-6684-5887-7.ch004.
30. Singh, A. (2023). The Present and Future Role of Artificial Intelligence in the Promotion of
Sustainable Practices in the Construction Sector. In: Kahraman, C., Sari, I.U., Oztaysi, B., Cebi, S.,
Cevik Onar, S., Tolga, A.Ç. (eds) Intelligent and Fuzzy Systems. INFUS 2023. Lecture Notes in
Networks and Systems, vol 759. Springer, Cham. https://doi.org/10.1007/978-3-031-39777-6_2.
Page 506
SEMIT 2023
Mixed Reality and Its Future Scope in the Construction Sector

Aditya Singh1, *(0000-0001-9347-5627)
1
School of Civil Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, Tamil Nadu, India
*Email: aditya7777777k@gmail.com
Abstract
Mixed Reality is another advanced technology which has become popular in the recent times.
Through this study, how Mixed Reality can impact and transform the construction sector, are
covered. The paper also talked about the scope of Mixed Reality in the future briefly. Numerous
scientific and research papers are considered to find out the research gaps, in application of Mixed
Reality in the context of construction. Additionally, this paper is going to help in understanding
the way Mixed Reality technology can help in the advancement of construction sector towards
Construction 4.0. Further, data from a number of different sources are collected to perform
graphical analyses in order to support the study.
Keywords: Mixed Reality, Construction Sector, Sustainability, Construction Projects,
Construction 4.0.
1 Introduction
Mixed Reality is considered to be the succeeding wave in computing after following the footprints of
mainframes, Personal Computers, as well as Smartphones. It is observed that the MR technology is
becoming popular among not only for businesses but also for consumers. Mixed Reality frees people
from the experience of a screen bound system, as it gives an option for people to have instinctual
interaction with a given data present in the living spaces as well as with other acquaintances or friends.
With the assistance of devices which are handheld, it is possible to experience MR technology, as
considering the case of 100s of millions of online explorers in the whole world who have first-hand
experience with it. Mainly the typical MR solutions at present can be found with the assistance of
applications present in Smartphones particularly on the social media platform. Many humans doesn’t
realize that the Instagram which is a popular social media platform offers Augmented Reality filters,
which gives the experiences of MR technology. However, Windows MR upgrades all such experiences
of the users to a completely new level, as it allow humans to be represented by spectacular holographic
ones, the actual place around them as well as higher committed holographic three dimensional models,
which enriches the Mixed Reality experience of people. To understand MR technology, it can be said that
it is a mixture of digital as well as physical worlds, which solves the intuitive in addition to natural
interactions between computer, three dimensional humans as well as environment. Mixed Reality is
normally on the grounds of progress and development of graphical processing, input systems, computer
vision, cloud computing as well as display technologies. Initially, Paul Milgram along with Fumio
Kishino worked on a paper where for the 1st time Mixed Reality as a term was used in the first half of the
last decade of the 20th century, in their paper named “A Taxonomy of Mixed reality Visual Displays”.
Through their study they discovered not only virtuality continuum’s concept but also they classified
visual displays. Afterwards, MR technology’s application went outside the displays to also comprise
human understanding which included eye tracking, hand tracking, as well as input of speech; three
dimensional asset collaboration in MR spaces; spatial sound; environmental understanding which
included anchors as well as spatial mapping; and positioning as well as location in both virtual as well as
physical spaces. Some people see Mixed Reality as an innovative idea which until now hasn’t found its
correct place in the building industry. However, considering the case where Mixed Reality can be
accepted in the construction industry early, can bring numerous benefits in the industry. With the
assistance of Mixed Reality, the efficiency of work can be increased significantly as well as flow of info
can further be enhanced. Since MR technologies impersonates the real life, so they are instinctive,
nevertheless deprived of the physical world’s restrictions. Novel circumstances can be completely made
with the aid of MR technology as it integrates both virtual as well as real worlds, where the digital in
Page 507
SEMIT 2023
addition to physical elements are able to interact at the same time. Since it utilizes immersive technology
to integrate both real as well as virtual settings, it is the reason why the term Mixed Reality is used, due
to which the users are permitted to perceive the present space put over with holographic data from a three
dimensional model. With the assistance of MR technology, it is possible that the manner construction
companies interact as well as receive info is altering now. Due to quick feedback, crews can now
understand the particular task to be completed, which is possible because of the lately declared sol where
both software as well as hardware are combined in order to enable the collaboration of both offsite in
addition to onsite teams.
1.1 MR Technology used in Construction
The term Mixed Reality is mentioned to a technical development which is indistinct between reality and
fiction. MR technology assist the people in commendably understanding not only digital data as well as
physical objects but also spatial relationships. MR technology comprises of instances like AR as well as
different mixed configurations. Designers, architects, service providers, project managers, in addition to
engineers are offered unparalleled capacity by MR technology in order to help in making an imaginary
imitation of the things they had prepared to build, which improves the quality of the real thing later on.
This allows them to be able to walk around the imaginary structure and realize the appearance of the same
structure when it is constructed later on in reality. This gives them the similar vibes of magic and excites
people to another level. Since, Mixed Reality offers a virtual tour of the supposed to be constructed
structure in the real world beforehand, so it permits the concerned people like engineers to detect flaws
in the design beforehand, and this saves time as well as cost considerably. In the construction sector,
Microsoft HoloLens is making MR technology to be popular, as it is a self-contained holographic
computer as well as a wearable, which offers high quality experience of Mixed Reality. Sometimes people
incorrectly think that Mixed Reality as Augmented Reality where the latter not only complements but
also expands the physical realm with the imaginary features such as data; or as Virtual Reality, which
entirely created of the digital world. However, Mixed Reality integrates the physical as well as digital
worlds where the interaction between both the worlds take place in real time as well as the two worlds
exists together at the same time. In the simple instance, it can be said that an individual is present in a
real life setting and he or she can interact with imaginary characters who are roaming around him or her,
like they really exist in the real world. This same principle is used in the construction sector to improve
the design and quality of projects. For instance, an individual could make real world environs in which
the individual could place digital beams, which performs similar to real beams nonetheless the digital
beams are considerably way easier to edit in addition to move them. In numerous manners, the workflows
in the construction sector can be aided by MR technology. With virtual models, people who are working
on them doesn’t lose their info. Currently, people can use a number of applications as well as tools at
every level of the workflow, where any form of the transfer could consequence in the gap of the data. It
is possible to create extremely detailed three dimensional models, then change them to two dimensional
ones which is intended to be utilized on the construction site, as well as further work on the envisioned
three dimensional building by means of the abridgement. A constant amount of understanding could be
maintained with the assistance of MR technology during the entire course of the projects. The capability
to examine with digital duplicates of real structures is highly valuable while working with the real
structures. Before assembly testing could be performed in the field to identify detect faults, like at the
time of making onsite changes. In comparison to three dimensional models on a flat screen, Mixed Reality
is considerably highly reliable, as it gives a correct three dimensional experience which shows
dimensions, scale, in addition to proportions. Since, it entirely imitates real life events, so with Mixed
Reality it is possible to achieve astonishing precision. Also during the early stages, it moreover makes
sure as well as enables remote collaboration. When working with Mixed Reality, utilization of
headphones is a classic method, but this technology can change quickly as well as its uses will be
enhanced with time. To make Mixed Reality instinctive, HoloLens technology is equipped with the
classic hard helmets. With time, the technology is becoming less onerous. In the future, it is expected that
the lightweight items will be used to give rich Mixed Reality experience. Nevertheless, an individual will
already possess all the needed info without the requirement to carry any other external device or laptop.
By giving a faultless alignment of holographic data on the workplace or job site in the case of projects
through the devices which are hard hat attached ones, MR technology improves the construction sector.
Page 508
SEMIT 2023
The project coordination as well as way clear-cut collaboration is permitted by MR technology’s aid, as
workers involved in the project are able to view the proposed models to be put over in the physical world.
In-field uses can be done effortlessly as created by the predefined views, as a result of giving immediate
access to immersive visualizations of the proposed three dimensional which is designed. This also helps
workers involved in projects to compare the plans with the help of MR technology in order to complete
their respective work.
1.2 Objectives
There are several objectives of this study, which are as follows:
• To understand Mixed Reality in the general sense and its compatibility with other advanced
technologies.
• To understand the ways MR technology can improve the construction sector.
• To find out the gaps in the current research based on the usage of Mixed Reality in the
construction projects.
• To perform graphical analyses to understand the future scope of Mixed Reality especially in the
construction sector.
1.3 What is the need of Mixed Reality in the construction sector?
Mixed Reality is an advanced technology which has shown high potential to transform the future of
mankind. Gaming industry which is a part of the Entertainment industry and some other fields have
started incorporating Mixed Reality as well as extracted benefits from the technology. However, Mixed
Reality also has the potential to transform the construction sector, but currently it is still almost non-
existent in the construction sector. Since, MR technology is highly compatible with other advanced
technologies like AI, AR, Robotics and so on, it might play a major role in improving the construction
sector in quality as well as design aspects particularly in the future, which will also help in leading the
construction sector towards Construction 4.0 [25][26][27].
2 Motivation
The author felt that in the last decade Mixed Reality gained considerable amount of popularity and gaming
industry has gained the most benefits from Mixed Reality. Also in the recent years, Mixed Reality is
slowly and steadily being applied in other fields as well and being a civil engineer, the author thought to
perform a study to understand the scope of MR technology in the construction sector and the potential
benefits which the sector can gained from it.
3 Literature Review
Ogunseiju et al. (2022) studied MR technology in order to understand sensing technology, which they
could incorporate in construction. They examined the use of designed MR environment which was
intended for preparing or training construction engineering students as well as making them competent
enough to organize the given sensing technologies in a construction project, through their study. They
stated that their study results showed features like easy access to resources as well as resources, animation
quality, correctness of construction activities which are represented, all are essential while designing an
effective MR learning environment [1].
Wu et al. (2022) studied workforce safety in construction projects and they worked on visual warning
intended for it, on the basis of real time MR technology. They stated their manuscript incorporates
advanced technologies like Deep Learning, Digital Twin, as well as Mixed Reality together, to form a
novel developed visual warning system which was a real time one that can allow workers in construction
projects to proactively find out not only the status of their safety but also to prevent mishaps. They also
performed tests under 3 quasi on site situations, as well as they proved its feasibility in relation to
represent info on potential hazards visually to the concerned users, in addition to coordinating large scale
construction activities [2].
Page 509
SEMIT 2023
Dai et al. (2021) studied construction site’s safety communication and they worked on improving it with
the assistance of MR technology, as well as considered the viewpoint of an industry to perform their
study. They stated that the present practices are based on physical interaction of people as well as talk,
call through telephone or smartphone, in addition to video conferencing that doesn’t enable immediate
access to not only info, interaction visually, or perception which is based on a particular context, regarding
the given situation, which are in fact important for efficient communication according to the demand of
the construction workplaces in the modern times. They stated that their results indicated the MR
technology’s potential for communication, visualization, as well as remote collaboration in the case of
issues which are regards to the safety concerns in the construction projects [3].
Li et al. (2023) studied underground pipelines, and they worked on developing them using the MR
approach, particularly in the experiments associated to digital mechanics. They stated that the
conventional transmission of info between computers as well as people has absence of not only interaction
but also understanding in the experiments related to digital twins. They used BIM in the design phase to
combine info storage, models, comprising of material info as well as physical geometry. They also made
an analytical algorithm which was intended for communication of sensors, Socket to improve the ways
corresponding to computer to human interaction, with the assistance of MR Toolkits as well as Game
Engines, and programming languages like C++ or C#, to accomplish their study [4].
Babalola et al. (2023) studied immersive technologies and they worked on applying them for health as
well as safety management, through their methodical review paper, especially in the construction sector.
They performed an assessment of one hundred seventeen applicable manuscripts which they gathered
from 3 databases which were all online ones, comprising of Web of Science, Scopus in addition to
Engineering Village, to proceed with their study. They stated that the past studies aimed on applying
numerous immersive technologies intended for safety training, visualization as well as detection of
hazards, perception of risk, safety design, in addition to evaluation in many construction activities. They
also found many limitations of the past studies and lack of research done on applying it for health hazards
[5].
Karaaslan et al. (2022) studied smart bridge and they worked on examining it with the aid of MR
technology especially for the upcoming smart cities. They then examined a novel methodology intended
for structural examination with the assistance of not only real time ML but also MR technology in order
to speed up an examiner’s specific tasks like measurement, identification, easier accessibility to defect
locations, in addition to examination of defects. They also worked on a methodology which can work as
a conditional examination of defects in concrete with the assistance of MR technology to obtain the
desired study outcomes [6].
Al-Sabbag et al. (2022) studied MR technology and they worked on permitting machine – human
collaboration with its assistance especially in the examination of infrastructures. They used holographic
display of a Mixed Reality headset as well as exact tracing of head in order to permit examiners to not
only localize but also visualize info like defects in a structure, in the real situations that were collected by
a given robot as well as took the assistance of an offsite computational server to process it, to accomplish
their study [7].
Kido et al. (2021) studied MR technology and they used it to measure the future landscapes with the
semantic segmentation through deep learning technology. They stated that the living environment of the
people must be preserved while promoting engineering, architecture as well as construction projects in a
balanced way with the natural environs. They found out that the absence of standard visualization
approach intended for the upcoming landscapes which still doesn’t exist, there with the presence of MR
technology, it offers to improve the landscape design field, as it can help in overlaying virtual content
onto an actual scene. However, they noticed a major problem in Mixed Reality technology is that
obstruction that happens when working with the virtual objects in the case of incomprehensible real
objects, which must be reduced in the forefront [8].
Wang et al. (2023) studied lightweight timber structure’s construction and they worked on it using human
cyber physical system which was intended for post digital design. They focussed on the problem which
talks about the humanization particularly in the post digital architectural practices to improve the
Page 510
SEMIT 2023
flexibility of the workflow. They realized the intricate structural geometry via a blend of human directed
robotic fabrication, computational form finding approach, as well as MR assisted assembly in addition to
analysis, to obtain desired study outcomes [9].
Prabhakaran et al. (2022) studied immersive technology and they worked on comprehending its
challenges while applying it in the construction as well as architecture industry, through their methodical
review paper. Through the severe rejection in addition to inclusion criteria of their own, they selected
fifty one qualified published papers from the year 2010 to the year 2019, for their final evaluation. They
stated that their results showed 9 serious challenges which they also ranked in their study, to further help
the future researchers working on the similar area [10].
Hong et al. (2021) studied outdoor urban environs which was amplified with natural sounds and they
worked on their soundscape examination with the assistance of a MR method. They further performed in
situ experiment with the assistance of a MR head mounted display, and they selected 2 outdoor locations
which were near to an expressway, to accomplish their study. They stated that their study results revealed
that both natural sounds considerably decreased the perceived loudness of traffic as well as improved the
overall soundscape quality [11].
Hussein et al. (2020) studied MR technology and they worked on applying it in the direction of
Construction 4.0, through their review study. They stated that effective MR solutions allows the user to
instinctively as well as efficiently navigate the augmented or virtual space while generating, extracting,
interacting with, as well as altering the data. They found out that the necessary data is accessed instantly
and it results with an improved performance of the user while completing their tasks as well as their
decision making ability, with the support of Mixed Reality [12].
Rossi-Schwarzenbeck and Vannelli (2023) studied MR, AR as well as VR, and they worked on using
them as Communication and Verification Tools especially in File to Factory Process as well as in Design
which is a digitized one, intended for temporary housing in the case of Cold Formed Steel building
constituents. Their study represented a research project related to their study topic and this began from
early design ideas as well as leading towards the building components being disassembled in addition to
their reutilization. Their study results included the application itself, as well as especially the challenges
in addition to advantages of the utilization of the chosen tools in a given designed project to enhance the
imminent applications [13].
Shen and Hsu (2023) studied MR interface and they worked on developing it as well as synchronous
robot fabrication which was intended for collaborative construction through their paper. They found that
there is a problem of low birth rate as well as shortage of workers in Taiwan which is negatively affecting
the construction field. They proposed ROCOS system in order to combine robot fabrication as well as
MR interface intended for collaborative construction between the robotic arm and the designer, to obtain
the desired results [14].
Assila et al. (2023) studied MR, AR, in addition to VR, and they worked on combining them with BIM,
through their methodological review paper. To accomplish their study, they used four different electronic
databases to search for qualified publications, which were from IEEE Xplore, Google Scholar, Science
direct, as well as ACM Digital Library, where they found two hundred and thirty nine publications and
they further selected only twenty eight publications for their analysis [15].
El Bedewy et al. (2023) studied MR technologies and they used them to make architectural models in the
direction of a novel Science, Technology, Engineering, Arts as well as Mathematics practice. They used
broad spectrum of technologies to assist their proposed architectural modelling which was modelled in a
number of ways, to obtain the desired study outcomes [16].
4 Research Gap
Out of 200 published papers, only the ones which were published in the last few years were further
considered. Based on the recent scientific and research papers or review papers, it was found out that
there is an absence of research done in the recent years in India in using MR in the construction sector.
Page 511
SEMIT 2023
China was one of the leading countries in the last few years, in regards to applying MR in the construction
sector. There was insufficient research done on utilizing MR technology in the construction sector in the
past few years.
5 Main focus of the Paper Along with Issues and Problems

The paper highlighted the Mixed Reality’s presence and how could it affect the construction sector to
improve its quality. Also the trend of Mixed Reality and its applications in the construction projects were
understood through the paper, but there are some issues and problems which shouldn’t be ignored. Most
of the people working in the construction sector are lacking awareness about Mixed Reality and they are
having insufficient knowledge about MR technology, which also doesn’t allow them to think about using
Mixed Reality in projects. Then, the cost and maintenance of this sophisticated technology is high. Often,
the construction projects take place in difficult terrains and people face extreme weather conditions which
might make them unable to use MR technology at the site or nearby it. Then, it is difficult to train workers
with Mixed Reality who are clueless and completely unfamiliar with it, particularly on a large scale.
Normally, the civil engineering graduates doesn’t have any knowledge about MR technology unless they
are involved in the gaming industry, so that also hinders the application of Mixed Reality in the
construction sector.
6 Mixed Reality Vs Other Immersive Technologies

In the case of Augmented Reality, interactive objectives are supposed to be layered on top of a given
physical environ but it lacks the capability to control or change the augmented objects. Then, Augmented
Reality helps the digital info to be overlaid onto the physical world. Whereas in the case of Virtual Reality
the physical world is kept hidden as well as people are entirely immersed in given digital world and the
digital simulation is being made of a physical environ. On the other hand, Mixed Reality combines the
ability of Augmented Reality as well as Virtual Reality to bring both the digital as well as physical worlds
together, in order to create an environ where digital in addition to physical objects could not only interact
but also coexist in the real time.
7 Methodology
In this section, how Mixed Reality can be applied in different ways to enhance the construction projects,
is presentenced through the following figure.
Figure 1: Flowchart representing MR applications in construction
Page 512
SEMIT 2023
8 Result and Discussion

In this section, data from various sources are collected to perform graphical analysis to support the study.
Table 1: Expected MR Apps Delivery
MR Apps Delivery In (%)
Smartphones 75
Others 25
Table 2: Market Growth of MR

Market Growth of MR In USD
Year 2020 553,270,000
Year 2026 5,811,090,000
Table 3: Expected Maximum Impact of MR in Different Fields

Expected Maximum Impact of MR in Different Fields In (%)
Education 28
Training of Workforce 24
Table 4: World Market Share of MR by Different Regions By 2035

World Market Share of MR By Different Regions By In (%)
2035
North America 32
Asia Pacific Region 27
Others 41
Smartphones Others
75
25
In (%)
Figure 2: Expected MR Apps Delivery
Page 513
SEMIT 2023
5,811,090,000
IN USD
553,270,000
0 2,000,000,000 4,000,000,000 6,000,000,000
Year 2026 Year 2020
Figure 3: Market Growth of MR
Training of Workforce Education
24
In (%)
28
Figure 4: Expected Maximum Impact of MR in Different Fields
41
50
32
40 27
30
20
10
0
In (%)
North America Asia Pacific Region Others
Figure 5: World Market Share of MR by Different Regions By 2035
8.1 Advantages of Mixed Reality in the Construction Sector

This subsection
• According to Fig. 2, Mixed Reality Apps delivery is 75% through Smartphones and others is only
25%. It shows that Smartphones are delivering MR technology thrice the times in comparison to
the others. Since, Smartphones are easy to carry and most the people are easily able to afford it,
so it is safe to say that it can be carried to the site and used in projects in the future. In the future,
Page 514
SEMIT 2023
with more advancement of technology, it will be easy to have rich experience of Mixed Reality
without carrying any additional supporting devices.
• According to Fig. 3, the market value of Mixed Reality was over $553 Million USD in the year
2020, but by the year 2026, it is expected that the market value of Mixed Reality will reach over
$5.8 Billion USD. This means that the market value of MR technology is expected to increase
slightly above 10.5 times within the span of 6 years, which shows a significant demand and
market growth of Mixed Reality. It is safe to say that Mixed Reality is having a bright future and
it will be wise to invest in it in the future. Similarly, it will be a smart move to use MR technology
in the construction sector.
• Based on Fig. 4, the expected maximum impact of Mixed Reality in the field of education is 28%,
which means that it can also help in the education of civil engineering students in universities
and degree colleges to make them understand difficult engineering subjects and concepts in an
easy way. Similarly, the expected maximum impact of Mixed Reality in the field of training of
workforce is 24%, which means that civil engineers, workers and other project staffs can be
trained effectively with the assistance of MR technology, which will make their training more
interactive and efficient. This will definitely help in the construction projects significantly in the
future.
• According to Fig. 5, by the year 2035, it is expected that the North American continent will have
a market share of 32% in the world in the case of Mixed Reality, followed by Asia Pacific Region
with 27%, then others having 41%, in which the European continent will have a major share in
others and will be closely behind the Asia Pacific Region. It can be said the advanced countries
are using MR technology way more than others to develop themselves further. So, developing
and underdeveloped countries also need to use such advanced technologies to reach similar
heights in the future. Of course, the major countries in the Asia Pacific Region’s market is
expected to grow considerably in the case of Mixed Reality, which is also a good sign and hint
for future investment in MR technology by the major countries in the said region.
9 Conclusion
This paper explained the concept of Mixed Reality in the general sense and it’s importance in the
construction sector. The manuscript also talked about the potential of Mixed Reality to transform the
construction sector and it’s compatibility with other advanced technologies which can lead the
construction sector towards Construction 4.0. A number of scientific and research papers were considered
to find out the gaps in the recent research. It also covered the various ways Mixed Reality can be
incorporated in construction projects to improve them further. Then, data from different sources were
collected to perform graphical analyses to understand the current trend and future scope of MR
technology in the construction sector, to support the study. In the future, with more advancement of
technology, it will be easy to have rich experience of Mixed Reality without carrying any additional
supporting devices. Mixed Reality can also aid in the education of civil engineering students in
universities and degree colleges to make them understand difficult engineering subjects and concepts in
an easy way. Civil engineers, workers and other project staffs can be trained effectively with the
assistance of MR technology, which will make their training more interactive and efficient. This will
definitely help in the construction projects significantly in the future. In the future more research is
required to be done to apply MR technology in the construction sector to reduce the current problems and
issues.
Acknowledgement
This research received no specific grant from any funding agency in the commercial, public, or not for
profit sectors.
Page 515
SEMIT 2023
References
1. Ogunseiju, O.R., Gonsalves, N., Aknmu, A.A., Bairaktarova, D., Bowman, D.A., and Jazizadeh, F.:
Mixed reality environment for learning sensing technology applications in Construction: A usability
study. Advanced Engineering Informatics, Volume 53, August 2022, 101637.
2. Wu, S., Hou, L., Zhang, G.(K.)., Chen, H.: Real-time mixed reality-based visual warning for
construction workforce safety. Automation in Construction, Volume 139, July 2022, 104252.
3. Dai, F., Olorunfemi, A., Peng, W., Cao, D., and Luo, X.: Can mixed reality enhance safety
communication on construction sites? An industry perspective. Safety Science, Volume 133, January
2021, 105009.
4. Li, W., Wang, Y., Yang, H., Ye, Z., Li, P., Liu, Y.A, and Wang, L.: Development of a mixed reality
method for underground pipelines in digital mechanics experiments. Tunnelling and Underground
Space Technology, Volume 132, February 2023, 104833.
5. Babalola, A., Manu, P., Cheung, C., Yunusa-Kaltungo, A., and Bartolo, P.: A systematic review of
the application of immersive technologies for safety and health management in the construction
sector. Journal of Safety Research, Volume 85, June 2023, Pages 66-85.
6. Karaaslan, E., Zakaria, M., and Catbas, F.N.: Mixed reality-assisted smart bridge inspection for future
smart cities. The Rise of Smart Cities, Advanced Structural Sensing and Monitoring Systems, 2022,
Pages 261-280.
7. Al-Sabbag, Z.A., Yeum, C.M., and Narasimhan, S.: Enabling human–machine collaboration in
infrastructure inspections through mixed reality. Advanced Engineering Informatics, Volume 53,
August 2022, 101709.
8. Kido, D., Fukuda, T., and Yabuki, N.: Assessing future landscapes using enhanced mixed reality with
semantic segmentation by deep learning. Advanced Engineering Informatics, Volume 48, April 2021,
101281.
9. Wang, S., Lin, D., and Sun, L.: Human-cyber-physical system for post-digital design and construction
of lightweight timber structures. Automation in Construction, Volume 154, October 2023, 105033.
10. Prabhakaran, A., Mahamadu, A-M., and Mahdjoubi, L.: Understanding the challenges of immersive
technology use in the architecture and construction industry: A systematic review. Automation in
Construction, Volume 137, May 2022, 104228.
11. Hong, J.Y., Lam, B., Ong, Z-T., Ooi, K., Gan, W-S., Kang, J., Yeong, S., Lee, I., Tan, S-T.: A mixed-
reality approach to soundscape assessment of outdoor urban environments augmented with natural
sounds. Building and Environment, Volume 194, May 2021, 107688.
12. Hussien, A., Waraich, A., and Paes, D.: A review of mixed-reality applications in Construction 4.0.
In Sawhney, A., Riley, M., and Irizarry, J. (Eds) Construction 4.0 An Innovation Platform for the
Built Environment, 2020. A review of mixed-reality applications in Construction 4.0 | 7 | Const
(taylorfrancis.com), https://www.taylorfrancis.com/chapters/edit/10.1201/9780429398100-
7/review-mixed-reality-applications-construction-4-0-aseel-hussien-atif-waraich-daniel-
paes?context=ubx&refId=7d2d2495-e574-4e9e-81df-2a40a3ef0f04.
13. Rossi-Schwarzenbeck, M., Vannelli, G. (2024). Virtual, Augmented and Mixed Reality as
Communication and Verification Tools in a Digitized Design and File-To-Factory Process for
Temporary Housing in CFS. In: Barberio, M., Colella, M., Figliola, A., Battisti, A. (eds) Architecture
and Design for Industry 4.0. Lecture Notes in Mechanical Engineering. Springer, Cham.
https://doi.org/10.1007/978-3-031-36922-3_24.
14. Shen, YT., Hsu, JS. (2023). The Development of Mix-Reality Interface and Synchronous Robot
Fabrication for the Collaborative Construction. In: Chen, J.Y.C., Fragomeni, G. (eds) Virtual,
Augmented and Mixed Reality. HCII 2023. Lecture Notes in Computer Science, vol 14027. Springer,
Cham. https://doi.org/10.1007/978-3-031-35634-6_26.
15. Assila, A., Dhouib, A., Monla, Z., Zghal, M. (2022). Integration of Augmented, Virtual and Mixed
Reality with Building Information Modeling: A Systematic Review. In: Chen, J.Y.C., Fragomeni, G.
(eds) Virtual, Augmented and Mixed Reality: Design and Development. HCII 2022. Lecture Notes
in Computer Science, vol 13317. Springer, Cham. https://doi.org/10.1007/978-3-031-05939-1_1.
16. El Bedewy, S., Haas, B., Lavicza, Z. (2023). Architectural Models Created with Mixed Reality
Technologies Towards a New STEAM Practice. In: Cai, Y., Mangina, E., Goei, S.L. (eds) Mixed
Page 516
SEMIT 2023
Reality for Education. Gaming Media and Social Effects. Springer, Singapore.
https://doi.org/10.1007/978-981-99-4958-8_3.
17. Microsoft Learn, https://learn.microsoft.com/en-us/windows/mixed-reality/discover/mixed-reality,
Last Accessed 2023/10/07.
18. Construction Placements, https://www.constructionplacements.com/mixed-reality-in-construction/,
Last Accessed 2023/10/07.
19. Geospatial World, https://www.geospatialworld.net/blogs/mixed-reality-in-construction/, Last
Accessed 2023/10/07.
20. Geospatial World, https://www.geospatialworld.net/blogs/mixed-reality-in-
construction/#:~:text=The%20global%20mixed%20reality%20industry,benefits%20this%20amazin
g%20technology%20offers, Last Accessed 2023/10/07.
21. XR Today, https://www.xrtoday.com/mixed-reality/mixed-reality-statistics-to-know-in-2021/, Last
Accessed 2023/10/07.
22. Statista, https://www.statista.com/topics/11185/mixed-reality/#topFacts, Last Accessed 2023/10/07.
23. Research Nester, https://www.researchnester.com/reports/mixed-reality-
market/5066#:~:text=Mixed%20Reality%20Market%20size%20is,was%20over%20USD%202%20
Billion, Last Accessed 2023/10/07.
24. Mix Your Reality, https://www.mixyourreality.com/articles/15-statistics-every-executive-should-
know-about-ar-mr-vr, Last Accessed 2023/10/07.
25. Singh, Aditya. "8 Augmented Reality and its use in the field of civil engineering". Handbook of
Augmented and Virtual Reality, edited by Sumit Badotra, Sarvesh Tanwar, Ajay Rana, Nidhi
Sindhwani and Ramani Kannan, Berlin, Boston: De Gruyter, 2023, pp. 117-136.
https://doi.org/10.1515/9783110785234-008.
26. Singh, A. (2023). Mechatronics Engineering in the Modern World and Its Use in the Construction
Industry. In M. Mellal (Ed.), Trends, Paradigms, and Advances in Mechatronics Engineering (pp. 64-
92). IGI Global. https://doi.org/10.4018/978-1-6684-5887-7.ch004.
27. Singh, A. (2023). The Present and Future Role of Artificial Intelligence in the Promotion of
Sustainable Practices in the Construction Sector. In: Kahraman, C., Sari, I.U., Oztaysi, B., Cebi, S.,
Cevik Onar, S., Tolga, A.Ç. (eds) Intelligent and Fuzzy Systems. INFUS 2023. Lecture Notes in
Networks and Systems, vol 759. Springer, Cham. https://doi.org/10.1007/978-3-031-39777-6_2.
Page 517
SEMIT 2023
Identification of Green Retrofitting Procurement and Permitting Processes in High-rise Office

Buildings in Jakarta Based on PerMen PUPR No.21 Year 2021 and GBC Indonesia that Affects
Project Time Performance
Benedict Mario Gilbert F.S1*, Yusuf Latief1, Bernadette Detty Kussumardianadewi1

1
Abstract
The concept of green buildings or green buildings is highly needed in the present era, considering the
drastic decline in global climate conditions. However, its development is greatly hindered because the
implementation of green building concepts is primarily focused on new constructions, while almost two-
thirds of the world's buildings are already built. Therefore, the objective of this research is to enhance
the Performance of Green Retrofitting Implementation Time, which is expected to accelerate the growth
of existing green buildings in Indonesia by identifying the Procurement and Permitting Processes of
Green Retrofitting in High-Rise Office Buildings in Jakarta based on PERMEN PU NO 21 TAHUN 2021
(Regulation of the Minister of Public Works and Housing Number 21 of 2021) and Green Building
Council Indonesia (GBCI) through work breakdown structure (WBS) along with their high-risk activities
and improvement strategies using a preventive approach that Influences the Performance of Green
Retrofitting implementation time. This research analysis involves 15 respondents' data on risks in the
procurement and permitting processes, which will be processed using a qualitative risk approach. Based
on the results of literature studies, expert validation, and respondent questionnaires, a total of 83 activities
in the procurement and permitting processes of green retrofitting were identified, 214 risk indicators that
affect the performance of green retrofitting schedule, along with 56 prevention strategies for the high-
risk situation, aiming to accelerate the growth of existing green buildings in Indonesia.
Keywords: Green Retrofitting, High Rise Office Building, WBS, Procurement, Permitting, Schedule
Performance
1. Introduction
Several studies in recent years have estimated the impacts of global and regional climate change across
various sectors. These changes pose challenges for all countries worldwide [1]. Based on the literature
review, current climate change has reached a range of 0.5-0.8°C over the past century [2]. Scientists are
striving to keep the global average temperature increase below 2°C from pre-industrial levels and aiming
to limit the temperature rise to 1.5°C above pre-industrial levels. This is expected to significantly reduce
the risks and impacts of climate change [3].
Countries worldwide are making efforts to address the impacts of climate change, including through
agreements such as the Paris Climate Agreement. One key point in this agreement is the achievement of
net zero emissions. The Paris Agreement was initiated during the 21st Conference of the Parties (COP
21) to the United Nations Framework Convention on Climate Change (UNFCCC) held in Paris on
December 12, 2015. The event was attended by 195 countries, including Indonesia [4].
Gases such as CO2, methane, nitrous oxide, CFCs, and other elements are the main contributors to
climate change and the greenhouse effect. Among these gases, CO2 is the largest contributor, accounting
for approximately 50% of the global greenhouse effect, with billions of tons being emitted into the
Earth's atmosphere each year. Other gases, such as CFCs, CH4 (methane), O3, and NOx, contribute in
smaller proportions, approximately 20%, 15%, 8%, and 7% respectively [5]. In 2018, CO2 emissions
reached a record high of around 33.1 billion tons, further exacerbating global warming [6].
The construction/building industry has a significant contribution to CO2 emissions, which play a role in
global warming. It is estimated that around 33% of CO2 gas emissions worldwide come from buildings
[7]. According to the Global Share of Buildings and Construction Final Energy and Emissions 2019 data,
carbon emissions generated by the building industry reached 38% of the total global carbon emissions.
The construction sector itself contributes to emissions in several categories, with 27% coming from the
Page 518
SEMIT 2023
operational phase of existing buildings, 6% from the construction of new buildings, and 7% from other
sectors within the construction industry [8]. This indicates that the construction/building industry plays
a crucial role in environmental degradation and contributes significantly to CO2 emissions.
In response to that, it is necessary to implement the concept of green building. Green building is a concept
for 'sustainable buildings' and has specific requirements, including location, planning and design systems,
renovation, and operation, which adhere to energy-saving principles and must have a positive impact on
the environment, economy, and society [32]. Green building refers to structures and processes that are
responsible for the environment and efficient resource use throughout the building's life cycle: from site
selection to design, construction, operation, maintenance, renovation, and demolition. This practice
extends and complements classic building design concerns about economics, utility durability, and
comfort [32]. According to Regulation of the Minister of Public Works and Housing No. 21 of 2021, a
green building is a building that meets the Technical Standards for Building Construction and has a
significant measurable performance in energy, water, and other resource savings through the application
of Green Building principles (BGH) in accordance with its function and classification at each stage of its
implementation.
However, looking at the projected number of buildings already constructed by 2040, existing buildings
are estimated to make up two-thirds of the buildings in that year [8]. The demolition of non-green
buildings not only involves a significant waste of resources and energy but also leads to secondary
pollution and ecological damage. On the other hand, if non-green buildings continue to be used, their
negative impact on the environment will persist. Therefore, retrofitting existing buildings with green
solutions (Green Retrofitting) is a more resource-efficient and sustainable approach compared to
developing new green buildings.
To achieve success and ensure proper implementation in green retrofitting activities, it is important to
have adequate procurement and permitting processes. In this regard, a comprehensive reference is needed
that covers all stages of the building lifecycle, including procurement and permitting stages. The
development of a Work Breakdown Structure (WBS) and risk management are also crucial in this
context. WBS helps identify and organize the work activities involved in the project, while risk
management helps identify high-risk stages of work and offers appropriate solutions and responses to
mitigate those risks.
The Work Breakdown Structure (WBS) is a crucial component in project planning. It starts with a
hierarchical arrangement of tasks and levels that help identify the flow of the project within the designated
and established timeline [9]. With a well-defined WBS and effective risk management, the
implementation of green retrofitting activities can be carried out in a focused and efficient manner,
ensuring the successful achievement of project goals.
2.1 Concept of Green Retrofitting for High-rise Office Buildings
The definition of green retrofitting according to Regulation of the Minister number 21 of 2021 is the
effort to adjust the performance of a utilized building to meet the requirements of Green Building Criteria
(Bangunan Gedung Hijau
- BGH) [10]. Meanwhile, according to the U.S. Green Building Council (USGBC), green
retrofitting is any type of improvement to an existing building, either fully or partially occupied,
aimed at enhancing energy and environmental
performance, reducing water usage, and improving comfort and indoor environmental quality in terms
of natural light, air quality, and noise, all done in a financially beneficial manner for the owner [11].
The aspects considered in the implementation of green retrofitting are categorized into 6 criteria:
Appropriate Site Development (ASD), Energy Efficiency & Refrigerant (EER), Water Conservation
Page 519
SEMIT 2023
(WAC), Material Resources & Cycle (MRC), Indoor Air Health & Comfort (IHC), and Building &
Environment Management, as outlined by the Green Building Council Indonesia [12].
Regulations related to the assessment of green building performance/certifications currently include

assessments conducted by the government through Regulation of the Minister of Public Works and
Housing No. 21 of 2021, and non-government assessments by GBCI through Greenship, as well as
EDGE, developed by the International Finance Corporation (IFC).
GREENSHIP Existing Building (EB) is a building certification system designed for buildings that have
been constructed and operational for at least one year. The implementation of green building practices in
GREENSHIP Existing Building focuses on operational and building maintenance management [12].
GREENSHIP itself is a rating system product issued by the Green Building Council Indonesia, taking
into consideration the local conditions, natural characteristics, and regulations and standards applicable
in Indonesia. The certifications provided include New Building (NB), Existing Building (EB), Interior
Space (IS), Homes, Neighborhood (NH), and Net Zero Healthy (NZH).
Regulation of the Minister of Public Works and Housing (Peraturan Menteri PUPR) No. 21 of 2021 is a
regulation concerning the Assessment of Green Building Performance issued by the government and
enacted on March 31, 2021. The green building performance assessed in this regulation is divided
into two categories: new buildings and existing buildings (retrofitting). For existing buildings, the
stages are divided into two phases: utilization and disposal phases [10].
2.2 Procurement Process for Green Retrofitting in Indonesia
According to Presidential Regulation, Number 12 of 2021 concerning Government Procurement
of Goods/Services, procurement of goods/services through Providers is a way to obtain goods/services
provided by Business Entities. Business Entities refer to companies or individuals engaged in specific
fields of business or activities. The Government Goods/Services Providers, hereinafter referred to as
Providers, are Business Entities that provide goods/services based on contracts [13].
The only government institution responsible for developing and formulating policies for government
procurement of goods/services is the Government Goods/Services Procurement Policy Institution
(LKPP), while the implementation of government goods and services is carried out by the Electronic
Procurement Services (LPSE) [13].
E-procurement is an auction system for government procurement of goods/services that utilizes internet-
based technology, information, and communication to ensure effectiveness, efficiency, transparency, and
accountability [14]. According to Andrianto [15], e-procurement is defined as the digitalization process
of government procurement tender/auction assisted by the Internet. The features of e-procurement
include E-auction, E-Tender, and E-catalogue. E-auction is an online auction system between
goods/services providers and users. E-Tender is the procedure for selecting goods/services providers,
open to all registered providers in LPSE. E-catalog is an electronic list developed by LKPP.
The procurement process for green retrofitting activities only applies during the retrofitting phase for
private buildings, and the planning is carried out by the building management. On the other hand, for
government building renovation procurement, it applies to both the planning and implementation phases.
2.3 Permitting Process for Green Retrofitting in Indonesia
Permitting is a process where the government unilaterally grants approval to individuals or communities
to legally perform certain acts or activities. In essence, this process serves as a regulatory instrument
by the state to control the conduct of its citizens in activities that should not violate the law or harm
others [16]. Based on Ministerial Regulation Number 21 of 2021, the licensing process applicable to
green retrofitting activities involves PBG (Building Use Approval) and SLF (Certificate of Functionality)
during the utilization phase of green retrofitting work. Meanwhile, for the disposal phase, the required
permit is RTB (Technical Demolition Plan). According to GREENSHIP Existing Building (EB) by
GBCI, the necessary permits for green retrofitting are PBG, SLF, and UKL/UPL documents issued by
BAPPEDAL (Environmental Impact Control Agency) [12].
Page 520
SEMIT 2023
Building Use Approval (PBG) is a permit granted to the owner of a building to construct, modify, expand,
reduce, and/or maintain the building in accordance with the technical standards for building construction
[17]. The related regulations are governed by Government Regulation Number 16 of 2021. PBG is a
reporting permit that can be processed during the construction process [18].
Certificate of Functionality (SLF) is proof that a building has been tested for its safety and functionality.
By possessing an SLF, a building is officially recognized and is expected to provide a sense of safety and
comfort to its occupants. The regulations related to SLF are governed by the Minister of Public Works
and Housing Regulation Number 19/PRT/M/2018 on the Implementation of Building Construction
Permits (IMB) and Certificate of Functionality, and Minister of Public Works and Housing Regulation
Number 27/PRT/M/2018 on the Certificate of Functionality for Buildings [19].
Environmental Impact Analysis (AMDAL), Environmental Management Efforts, and Environmental
Monitoring Efforts (UKL-UPL) are Environmental Documents that must be prepared by businesses
whose activities have a significant impact on the environment. These documents will outline the process
of developing business infrastructure (such as buildings, factory installations, etc.), soil conditions or
geological aspects, types of potential environmental impacts (including liquid waste, solid waste, gas,
and noise), as well as how the business entity intends to manage and monitor its activities to minimize
potential environmental damage risks [20]. The regulations concerning UKL-UPL documents are
outlined in the Minister of Environment Regulation Number 16 of 2012 on Guidelines for the Preparation
of Environmental Documents [20].
The Technical Demolition Plan, abbreviated as RTB, is a planning document issued by the owner of a
building to the government as a declaration that the said building will be demolished, and there is a plan
for the demolition of the building [18]. The plan for the demolition of the building must comply with
technical standards as stipulated by laws and regulations. To determine whether the demolition plan meets
these technical standards or not, a consultation process involving experts with the relevant expertise and
skills related to building construction is necessary.
2.4 Work Breakdown Structure of Procurement and Permitting for Green Retrofitting
Work Breakdown Structure (WBS) is an extremely effective tool used in project management. It serves
as the foundation for effective project planning, execution, control, status tracking, and reporting. All the
work contained within the WBS must be identified, estimated, scheduled, and budgeted. The WBS is a
structure and code that integrates and links all project work along with the required resources (scope,
schedule, and cost) [21].
A project WBS involves grouping project work elements oriented towards products that organize and
define the total project scope. It is a multi-level framework that graphically represents elements indicating
the work that needs to be accomplished in logical relationships. Each lower level represents a more
detailed definition/division of project components. It is a structure and code that integrates and links all
project work and is used throughout the project life cycle to identify, establish, and track specific work
scope. The WBS is created with enough detail so that each control account has a unique WBS element
[21].
According to the PMBOK 6th Edition, the main benefit of creating a Work Breakdown Structure (WBS)
is to provide a framework for what needs to be delivered in the project. The process of creating a WBS
is only done once or at designated points in the project. Additionally, the WBS represents a
hierarchical decomposition of the total work scope to be performed by the project team to achieve
project objectives and deliver the required outcomes. The WBS organizes and defines the total project
scope and represents the work specified in the currently approved project scope statement [23].
Page 521
SEMIT 2023
Figure 1. Work Breakdown Structure of Procurement and Permitting for Green Retrofitting
Source: Self Administered
In this research, the Work Breakdown Structure (WBS) for the Procurement and Permitting Process of
Green Retrofitting Activity aims to define activities within a project related to the procurement process
and permitting process at the utilization and demolition stages. The development of the WBS structure
for the Procurement and Permitting Process of Green Retrofitting Activity is based on the requirements
of GREENSHIP Existing Building (EB) and on Peraturan Menteri PUPR No. 21 Year 2021 (Regulation
of the Minister of Public Works and Housing Number 21 of 2021). The following details the levels in the
WBS for the Procurement and Permitting Process of Green Retrofitting Activity:
Level 1: Green Retrofitting
Level 2: Work Phases
Level 3: Work Phase Processes
Level 4: Process Methods
Level 5: Process Support
Level 6: Work Activities
Page 522
SEMIT 2023
2.5 Risk Assessment
According to the PMBOK 6th Edition, a risk event is something that has been identified beforehand that
may or may not occur. If it happens, it can have either a positive or negative impact on the project.
Uncertainty is the lack of knowledge about an event that reduces confidence in the conclusions drawn
from data. Risk management is the process of identifying, evaluating, and planning responses to events,
both positive and negative, that may occur during the project [22].
Risk management is an approach taken toward the risks of a project, which, involves understanding,
identifying, and evaluating the project's risks. Subsequently, it considers what actions to take in response
to the potential impacts and the possibility of transferring or reducing the risks. Risk management
encompasses a series of activities related to risk, including planning, assessment, handling, and
monitoring of risks [23]
The purpose of risk management is to recognize risks within a project and develop strategies to reduce
or even avoid them, while also seeking ways to maximize existing opportunities [24]. According to the
PMBOK 6th Edition, risk management processes are divided into 7 processes, namely: Risk Management
Planning, Risk Identification, Qualitative Risk Analysis, Quantitative Risk Analysis, Risk Response
Planning, Risk Response Implementation, and Risk Monitoring. The PMBOK 6th Edition is widely used
as a reference for risk management planning, especially in construction projects [22].
The risk assessment in the green retrofit project was ranked for 19 risks using the risk criticality index
(RC), which is the result of the likelihood of occurrence and effect indexes (MI). Revealing that the top
5 risk rankings are as follows: "post-retrofit tenants’ cooperation risk," "regulatory risk," "market risk,"
"financial risk," and "pre-retrofit tenants’ cooperation risk." [25]. The highest risk in the rating was the
"post-retrofit renters' cooperation risk," which indicated that tenants were likely to be uncooperative once
the green retrofit was finished. Achieving the most effective use of energy-saving equipment and
maximizing all potential benefits would not be possible without the tenants' involvement. Bon-Gang
Hwang et al emphasized that the top 5 risks in LO (Likelihood of occurrence) and MI (magnitude of
impact) values are “Complex procedures to obtain approvals”, “Overlooked high initial cost”, “Unclear
requirements of owners”, “Employment constraint”, and “Lack of availability of green materials and
equipment”. The top position risk (Complex procedures to obtain approvals) was because green
residential building projects always involve some particular green features, which would result in
lengthier planning approval and permit procedures [26].
2.6 Project Time Performance
Time performance is the comparison of the actual project implementation period to the estimated project
implementation period. Time performance can be used as a measure of slippage in a construction
schedule. The project implementation begins on the agreed-upon date specified in the contract
agreement and ends when all the work is completed. This ratio or comparison serves as the standard
measure of project performance, allowing the estimation of project performance through statistics [27].
Time performance is crucial in construction implementation and is one of the primary factors for the
success of a construction project. Both design-build and traditional systems of construction projects
generally have specific implementation plans and schedules, determining when the project should start,
when it should be completed, how it will be executed, and how resources will be allocated [28].
Good time performance seems to be one of the indicators of project success. Poor time performance can
lead to various undesirable issues. Time performance is considered good when a project is completed
according to the agreed-upon schedule by all parties [29]. In the procurement process, time performance
control is crucial and requires leadership, strategy, and management systems. The time needed for
procurement and permits significantly impacts the project's implementation and completion time [30].
Project owners must have accurate scheduling, recording of work completion status, and contractual
authority to accelerate project completion (project time efficiency) [31].
Page 523
SEMIT 2023
To answer the objectives of this research, the researcher used a qualitative research method. Data
validation was carried out using a structured questionnaire and interviews with 5 experts who are
proficient in the procurement and permitting of buildings for the green retrofitting activity. Data
collection was conducted using a standardized questionnaire administered to 15 experienced
respondents who have experience in implementing procurement and permitting. The following figure
shows how this research was conducted in order:
Page 524
SEMIT 2023
Figure 2 Research Flowchart (Source: Self-Administered)
4. Conceptual Framework
Page 525
SEMIT 2023
Conceptual framework this research represents the Relationships Model between Variables in the
Relationship between Green Retrofitting WBS, Procurement Activity, Permitting Activity, Risk, and Risk
Strategies, and Project Time Efficiency of High-rise Office Buildings with The Accuracy Level based
can show likes:
Figure 3. Conceptual Framework of the relationship between green retrofitting and project time
(Source: Self-Administered)
When doing categorical analysis, factor variables must be transformed into a collection of indicator
variables. Finding the risks that have an impact on the scheduled time performance of green retrofitting
office high-rise buildings based on procurement and permitting activity is the next stage after achieving
the WBS standard. Below is an explanation of the risk variables used in this study:
Page 526
SEMIT 2023
Table 1 Research Indicator Variables
VARIABLE INDICATOR Refere

nce
X 1 Planning Stage [13]
X
1 2 Procurement Preparation Stage [13]
Z PROCUREM X
1
. 3 Procurement Implementation Stage [13]
X ENT STAGE X
1
. 4 Monitoring Stage [13]
1 X
1
. 1 Environmental Impact Assessment [20]
X 2
. and Environmental
X
1 PERMITTIN . Management Efforts (AMDAL
2X G STAGE UKL-UPL)
1 X 2 Building Use Approval (PBG) [18]
X
2 3 Certificate of Functionality (SLF) [19]
X
.
2 4 Technical Demolition Plan (RTB) [18]
2
.
X 1 Input [22]
X WBS X
.
3 2 Techniques and Equipment [22]
3 X
3
. 3 Output [22]
X
3
. 1 Risk Identification [22]
X RISK X
4
. 2 Risk Analysis [22]
4 X
4
. 3 Risk Strategies [22]
PROJECT 4
.
Y TIME
EFFICIENC Y
. Schedule Performance [22][3
Y 1]
5. Discussion & Conclusion:

The activities in the procurement and permitting process of green retrofitting have been organized using
the Work Breakdown Structure (WBS) tools based on GREENSHIP Existing Building Ver 1.1 by GBCI,
which includes regulations and guidelines related to environmental permits, such as: Environmental
Impact Assessment (AMDAL) Permit: Based on Regulation No. 16 of 2012 on Environmental Impact
Assessment Document Preparation for AMDAL UKL-UPL. Building Permit (PBG) Requirements:
Based on PP No. 16 of 2021 for building permit requirements. Technical Demolition Plan (RTB): Based
on the Technical Regulation of the Minister of Public Works and Housing No. 21 of 2021. Certificate of
Occupancy Permit: Reference is made to Regulation of the Minister of Public Works and Housing No.
19/PRT/M/2018 regarding the Implementation of Building Permit (IMB) and Certificate of Occupancy
for the execution of green retrofitting activities. Regarding procurement, it is based on Regulation of the
Ministry of Public Works No. 21 of 2021, which pertains to the implementation of procurement for green
retrofitting activities. There are a total of 27 work packages in the procurement phase, with 17 of them
related to procurement activities and the remaining 10 being associated with the permitting phase.
From the 27 work packages related to procurement and permitting activities for green retrofitting, it has
been determined that there are a total of 83 activities. These activities are based on GREENSHIP Existing
Building Ver 1.1 by GBCI for permitting in the implementation of green retrofitting activities and
Regulation of the Ministry of Public Works No. 21 of 2021 for procurement. Each of these activities can
be used to identify their objectives and targets, which are useful for identifying potential risks or risk
factors that may occur due to the non-fulfillment of activity goals. This risk identification is crucial for
the successful execution of green retrofitting projects, as it allows for proactive risk management and
mitigation strategies.
From the 27 work packages related to procurement and permitting, based on the findings from the
previous sections and in alignment with the research objectives, a total of 214 risks/factors have been
identified in the procurement and permitting phases of green retrofitting activities. Out of these, 84 are
associated with the procurement phase, and the remaining 130 risks/factors are related to the permitting
phase of green retrofitting activities.mitting activities for green retrofitting, it has been determined that
Page 527
SEMIT 2023
there are a total of 83 activities. These activities are based on GREENSHIP Existing Building Ver 1.1 by
GBCI for permitting in the implementation of green retrofitting activities and Regulation of the Ministry
of Public Works No. 21 of 2021 for procurement. Each of these activities can be used to identify their
objectives and targets, which are useful for identifying potential risks or risk factors that may occur due
to the non-fulfillment of activity goals. This risk identification is crucial for the successful execution of
green retrofitting projects, as it allows for proactive risk management and mitigation strategies.
Based on the risks/risk factors identified in the stages of procurement and permitting for green retrofitting
activities, it was found that out of the total risks, 56 were categorized as high risks, 72 were categorized
as medium risks, and 86 were categorized as low risks. The total of 83 activities in the procurement and
permitting processes of green retrofitting were identified, 158 risk and 56 high-risk indicators that affect
the performance of green retrofitting schedule from each of the procurement and permitting activities,
along with 56 prevention strategies for the high-risk situation.
The results of research by developing WBS on the procurement and permitting stage of Retrofitting are
guided by GBC Indonesia and PUPR Ministerial Regulation Number 21 of 2021 (Regulation of the
Minister of Public Works and Housing Number 21 of 2021) along with their derivative regulations, and
developing the risk strategies for the high-risk activity expected to significantly increase the project time
efficiency of green retrofitting in Indonesia.
References
1. Arnell, N. W., Lowe, J. A., Challinor, A. J., & Osborn, T. J. (2019). Global and regional impacts of
climate change at different levels of global temperature increase. Climatic Change, 155(3), 377–391.
https://doi.org/10.1007/s10584-019-02464-z
2. Lubis, R. F., Delinom, R., & Geoteknologi, P. (n.d.). PERUBAHAN IKLIM DAN PEMANASAN
GLOBAL DI INDONESIA; DAMPAKNYA TERHADAP KONDISI BAWAH PERMUKAAN STUDI
KASUS: DKI JAKARTA.
3. Skytt, T., Nielsen, S. N., & Jonsson, B. G. (2020). Global warming potential and absolute global
temperature change potential from carbon dioxide and methane fluxes as indicators of regional
sustainability – A case study of Jämtland, Sweden. Ecological Indicators, 110.
https://doi.org/10.1016/j.ecolind.2019.105831
4. BPHN. (n.d.). na_ruu_ratifikasi_paris_agreement (2). Na_ruu_ratifikasi_paris_agreement.
Retrieved July 23, 2023, from https://bphn.go.id/data/documents/na_ruu_ratifikasi_paris_agreement.pdf
5. Eko, W., Peneliti Bidang, C., Ozon, P., & Udara, P. (n.d.). PENGARUH PEMANASAN GLOBAL
TERHADAP LINGKUNGAN BUMI.
6. Izumi, F. (2020). Climate Emergency Declaration. Climate Emergency Declaration.
7. UGM (n.d) The building industry contributes to the process of environmental degradation
8. Architecture 2030. (n.d.). Actions For Zero Carbon Buildings Existing Buildings.
9. Devi, Tr., & Reddy, Vs. (n.d.). Work Breakdown Structure of the Project. In International Journal
of Engineering Research and Applications (IJERA) (Vol. 2). www.ijera.com
10. Regulation of the Minister of State for the Environment Number 16 of 2012 concerning
Performance Evaluation - Green Building
11. The U.S. Green Building Council. (n.d.). Building Retrofits.
12. Green Building Council Indonesia. (n.d.). GREENSHIP Rating Tools Existing Building.
13. Presidential Regulation Number 12 of 2021 concerning Government Procurement of Goods/Services
14. Sutedi, Adrian., (2012) Legal aspects of the procurement of goods and services and their various
issues
15. Andrianto, Nico., (2007) Transparency and public accountability through e-government
Page 528
SEMIT 2023
16. Priyanta, M. (2019). Regulasi Perizinan Mendirikan Bangunan dalam Mendukung Kemudahan
Berusaha Menuju Bangsa yang Adil dan Makmur. https://doi.org/10.24843/JMHU.2019.v08.i03
17. Simbawa, Anggreini Anglia., (2022) Procedure for obtaining building construction approval
18. Government Regulation Number 16 of 2021 concerning building
19. Minister of Public Works and Housing Regulation Number 27/PRT/M/2018 on the Certificate of
Functionality for Buildings
20. Minister of Environment Regulation Number 16 of 2012 on Guidelines for the Preparation
of Environmental Documents
21. Burghate, M. (2018). Work Breakdown Structure: Simplifying Project Management. In
International Journal of Commerce and Management Studies (IJCAMS) (Vol. 3, Issue 2).
www.ijcams.com
22. A Guide to the Project Management Body of Knowledge (PMBOK 6 ® Guide)
23. Kerzner, Project Management: A Systems Approach to Planning, Scheduling, and Controlling,
2017; Kerzner, Strategic Planning for Project Management Using a Project Management Maturity
Model, 2001 24. Wideman, Risk Management. A guide to managing risk opportunities, 1992
25. Zhao, X., Hwang, B.G., and See, Y.L. (2015). Green Retrofit Projects: Risk Assessment and
Mitigation. E (Eds) Procs 31st Annual ARCOM Conference, 7-9 September 2015, Lincoln, UK,
Association of Researchers in Construction Management, 125-134.
26. Hwang, B. G., Shan, M., Phua, H., & Chi, S. (2017). An exploratory analysis of risks in green
residential building construction projects: The case of Singapore. Sustainability (Switzerland), 9(7).
https://doi.org/10.3390/su9071116 [27] Bacon, R. W., & Besant-Jones, J. E. (1998). Estimating
construction costs and schedules Experience with power generation projects in developing countries.
In Energy Policy (Vol. 26, Issue 4).
28. (Kerzner, Project Management: A Systems Approach to Planning, Scheduling, and Controlling, 2017)
29. Yean, F., Ling, Y., Swee, ;, Chan, L., Chong, ; Edwin, & Ee, L. P. (n.d.). Predicting Performance
of Design-Build and Design-Bid-Build Projects. https://doi.org/10.1061/ASCE0733-93642004130:175
30. Lusi indah, Wijayanti., Development of quality management system procurement consultant
services to improve procurement performance in Ministry of Transportation (2017) 31. Huston, Charles
L. Management of Project Procurement. (1996)
32. Sudarwani, M. M. (2021). The implementation of green architecture and green building is an effort
to achievesustainable architecture.
Page 529
SEMIT 2023
Performance Investigation of Heat Sinks with Different Perforations Diameters

Rana A. J. AL-Luhaibi a, *, Assis. prof. Dr. Ibrahim T. Nazzalb
a
Mechanical Department/ Engineering College /Tikrit University/ Tikrit, Iraq
Email: rana.a.jasim43787@st.tu.edu.iq, ORCID: https://orcid.org/0009-0008-5346-3072
b
Mechanical Department/ Engineering College/Tikrit University /Tikrit, Iraq
Email: dribrahimthamer@tu.edu.iq; ORCID: https://orcid.org/0000-0002-8745-275X
Abstract
In this paper, the perforated heat sink was studied and compared with the non-perforated heat sink
experimentally. Four rectangular heat sinks were investigated to present this comparison; all heat sinks
used in this study have the same dimensions. Three of them are perforated and the fourth heat sink is
solid. The perforated heat sinks have 12 circular holes in each fin distributed in two rows (each row has
six holes) with different holes’ diameter: first heat sink 3 mm the diameter (PHS-Ø3mm), the second heat
sink is 3.5 mm called (PHS-Ø3.5mm), and the third one its hole diameter is 4 mm (PHS-Ø4mm). The
experiment was carried out using forced convection device in the heat transfer laboratory. After obtaining,
analyzing and discussing the results, it was found that the perforated heat sink is better at dissipating heat
than the non-perforated one. Also, the PHS-Ø3mm has the highest heat transfer enhancement (14.21%).
For the weight reduction, this paper proved that perforating the fins reduces the weight of the heat sink,
and the larger the perforation diameter, reduce more mass of heat sink. The highest reduction achieved
by PHS-Ø4mm.
Key words: Forced convection, heat sink, Perforated heat sink, perforation diameter.
Introduction
The rapid development of electronic devices in recent times has led to an increase in heat
generation, posing a significant challenge in effectively cooling these devices. Plate fin heat sinks, known
for their simple manufacturing process, are commonly used to dissipate the generated heat. In the pursuit
of improving their performance. Various studies have explored different approaches to enhance heat
transfer in heat sinks. Some researchers have focused on modifying the fin geometry [1], while others
have introduced vortex generation within the fins [2]. Furthermore, the use of perforated fins has garnered
attention of searchers, and many studies had been conducted in this field. Researchers such as Sara et al.
[3] conducted experiments on the pressure drop and heat transfer characteristics of perforated flat plates.
Their findings demonstrated that the use of perforated fins resulted in increased heat dissipation compared
to solid flat plates. The study conducted by Shaeri et al. [4] (2009) utilized numerical methods to examine
the fluid flow and convective heat transfer of rectangular fins with square holes positioned on their lateral
surfaces. The researchers concluded that these newly designed perforated fins exhibited superior heat
transfer capabilities and significantly reduced weight compared to solid fins. In a similar vein, Venkitaraj
and Siddikh [5] (2016) conducted a numerical analysis to evaluate the enhancement in convection heat
transfer achieved by an array of fins. They investigated different types of perforations, such as circular,
elliptical, square, and triangular, to assess their performance and impact on the fin's material mass. The
findings revealed that the circular perforated fin array outperformed other types in terms of heat transfer.
Sonawane and Palande [6] (2016) explored the potential for improving heat transfer in fin heat sinks
through the incorporation of various perforation designs. In comparison to solid pin-fin heat sinks, the
study highlighted that perforated pin-fin heat sinks offered higher rates of heat transmission. Furthermore,
Shaeri et al. [7] (2017) conducted an experimental investigation to compare the hydraulic and thermal
performance of laterally perforated-finned heat sinks. By employing square cross-sectional perforations
of different sizes evenly distributed along the heat sinks' length, the researchers observed that the
perforated fins demonstrated higher heat dissipation compared to unperforated fins. Chingulpitak et al.
[8] (2019) presented a study focusing on the thermal performance of laterally perforated plate-fin heat
sinks with varying numbers and diameters of circular perforations. To determine suitable design
Page 530
SEMIT 2023
parameters, the study recommended considering the thermal performance factor, which accounts for the
Nusselt number and friction factor. The results indicated that the perforated heat sinks exhibited a 10.6%
improvement in thermal performance and a 28% reduction in mass compared to solid heat sinks. Lastly,
Huang et al.[9] (publication year not provided) investigated the thermal performance of pin-fin heat sinks
with perforations and splitters. Their study highlighted a significant enhancement in thermal performance
attributed to the presence of perforations and splitters.
In conclusion, the introduction of holes in the fins of plate-fin heat sinks has emerged as a
promising approach to enhance their thermal performance. The sizes of perforations play a crucial role
in improving the heat sink's efficiency. While there have been numerous studies on the impact of
perforated fins on heat sink performance, further research in this area remains necessary.
Experimental setup
In order to achieve the goal of this research, a device utilizing forced convection in a horizontal
configuration is utilized, as illustrated in Figure [1].
Figure 1. Forced convection device-horizontal duct

1- Air blower. 2- Fan. 3- Throtle opening. 4- Test section. 5- Electrical heater. 6-Total head tube.
7- Test section. 8- Thermometer. 9- Air inlet. 10- Digital thermometer. 11- Control panel. 12-
Inclined manometer. 13- Watch the timer.
The configuration of the experiment is visually depicted in Figure [2]. It comprises essential
components such as a wind tunnel, a heating system, heat sinks, a blower, and measurement
instruments to assess temperature and flow velocity. The wind tunnel system itself includes a
blower, a flow straightener and a test section.
Figure 2. Schematic diagram of forced convection device
Page 531
SEMIT 2023
A continuous airflow steadily traverses the heated heat sink, resulting in the efficient dissipation of heat.
The duct's inlet is specifically designed as a curvy shape resembling a bell mouth to ensure a seamless
entry of air into the duct, eliminating the formation of separate air layers. The blower, situated at the end
of the wind tunnel, possesses a maximum flow rate of 29.5 m3/min and effectively draws the required air
through the test section from the immediate vicinity of the device. This air is subsequently directed out
through a vertical pipe, featuring a throttle opening measuring 125 mm in diameter. Furthermore, the
wind tunnel's exit is equipped with a movable gate that can be adjusted to regulate the airflow by opening
or closing the pipe, thereby controlling the quantity of air passing through the tunnel. Beyond the test
section, the duct incorporates a flow straightener, consisting of a honeycomb structure, which ensures
that the airflow remains laminar and well-organized. Constructed from iron with a thickness of 1 mm, the
flow straightener is a horizontal square duct measuring 125 x 125 mm and has a length of 1300 mm. As
depicted in Figure [3], the air continuously flows across the heat sink located within the wind tunnel,
allowing for the continuous removal of heat.
Figure 3. Heat sink inside the test section of the wind tunnel
Within the test section, all the heat sink samples were positioned. These heat sinks are fabricated from an
aluminum alloy (1006) with a thermal conductivity rating of (237) W/m.K. The heat sink base possesses
specific dimensions: 12 mm for width (Ly), 100 mm for length (Lx), and 55 mm for thickness(t). This
research project encompasses four variations of heat sink models, each incorporating different perforated
configurations, as detailed in Table [1].
Table 1. Types of the heat sinks
Heat sink type Perforation number Surface area (mm2) Weight (gr)
Solid Non 51420 339.7
PHS-3Ø 12 52324 315.2
PHS-3.5Ø 12 52211 314.3
PHS-4Ø 12 52022 307.14
Page 532
SEMIT 2023
Figure [4] demonstrates the samples that used to test in this paper, perforated and non-perforated heat
sinks.
ac bd
Figure 4. Types of heat sinks (a is non-perforated), (b is PHS-Ø3mm), (c is PHS-Ø3.5mm), (d is PHS-

Ø4mm)
The choice was made to use an electrical heater(AC), in order to maintain a steady heat flux surface for
heat sinks. An AC variac was utilized to adjust the power supplied to the heater. This transformer
controlled the heating rate by delivering a specific voltage to the heater. For temperature measurement at
the base of the heat sink, three thermocouple holes were positioned on the bottom. To ensure complete
insulation, the entire body of the heater was enveloped in Teflon. The low thermal conductivity of acrylic
allowed for disregarding heat loss from the sides of the test section. The findings indicated that the
temperatures of the acrylic sides were approximately equivalent to the surrounding temperature due to
the insulation provided by the heat sink samples. Within the data acquisition unit, a 16-channel
multiplexer is utilized to record data from different thermocouples. The Anbai AT4516 data logger is
equipped with a sixteen-channel selector to establish connections with the thermocouple set. K-type
thermocouples are employed for temperature measurement of the heat sink and the cooling air. All
thermocouples used in this research underwent calibration. Three K-type thermocouples were integrated
into the heat sink to measure its base temperature. Furthermore, eleven thermocouples were attached to
the heat sink's surface, as depicted in Figure [5].
Figure 5. Thermocouples location on the heat sink

By estimating and comparing the heat dissipated to the air-fluid with the power input of the heater, the
researchers were able to determine the heat transfer to the surroundings. The analysis revealed that the
heat loss to the surroundings accounted for less than 7.3% of the total power dissipation across all tested
samples. Various air flow velocities were forced to cool the heated heat sinks (1.5, 2, 2.5, 3, 3.5, 4)m/s.
Page 533
SEMIT 2023
To measure these velocities of the airflow entering the wind tunnel, an anemometer from BENETECH
was used, providing an accuracy of + 5% for velocities ranging from 0 to 30m/s. The investigation
focused on assessing the performance of all heat sinks, illustrated in Figure [4], All the samples had
identical dimensions, as outlined in Table [2]. They were constructed with a flat plate and extruded
rectangular fins measuring 2.5 mm in thickness. The spaces or channels between the fins measured 3500
mm2, and each heat sink contained seven channels with a width of 5 mm. These perforations were
introduced to evaluate the thermal performance of the heat sink in terms of heat dissipation and cooling
rate, while also facilitating a comparison with other samples.
Table 2. The dimensions of the heat sink
Dimensions Value (mm)
Heat sink width (W) 55
Heat sink length 100
Heat sink base thickness 12
Fin height 28
Fin thickness 2.5
The thickness of the interface material 1
Experimental procedures
After setting up the forced convection device and connecting all the main parts to each other, the first
sample of heat sinks is placed inside the wind tunnel and operation is started. At first, the temperature of
the heater is raised by Variac (i.e. raising the voltage to 20V), the air velocity is fixed at 1.5 m/s. The first
reading is taken, then increasing the air velocity 2 m/s to take the reading, and so on for all velocities (1.5,
2, 2.5, 3, 3.5, 4) m/s, then the voltage is raised to 40 volts and the first steps -changing air velocities- is
repeated until all thermal loads are completed (20, 40, 60, 80, 100) V. All of these steps are repeated each
time the heat sink is replaced. After recording all readings and testing all samples. Results are obtained,
analyzed, discussed and presented.
Data reduction
The performance of the heat sink is influenced by factors like the heat transfer coefficient, Nusselt
number, fin efficiency, and the heat transfer rate, all of which comply with Newton's second law for
cooling.:-
𝑄𝑐𝑜𝑛𝑣. = ℎ𝐴𝑠 (𝑇𝑠 − 𝑇∞ ) = 𝑚̇𝑎 𝐶𝑝 (𝑇𝑜𝑢𝑡 − 𝑇𝑖𝑛 ) ………………………(1)
The equation below can be utilized to compute the rate at which convective heat is transferred from the
heat sink surface to the ambient air: -
𝑄𝑐𝑜𝑛𝑣. = 𝑄𝑒𝑙𝑒𝑐 − 𝑄𝑙𝑜𝑠𝑠 .………………………(2)
Where 𝑄𝑒𝑙𝑒𝑐 is the electrical heat input that can be extracted as using the following equation:-
𝑄𝑒𝑙𝑒𝑐 = 𝐼𝑉 .……………………….(3)
The average heat transfer coefficient can be evaluated by:-
𝑄𝑐𝑜𝑛𝑣.
ℎ= 𝑇 +𝑇 ..………………………(4)
𝐴𝑠 (𝑇𝑠 − 𝑜𝑢𝑡 𝑖𝑛 )
2
When considering a flat plate fin, the area AS in equation [1] corresponds to the combined surface area
of the fin that interacts with the fluid passing through the duct. This value is obtained by summing up the
projected area and the total surface area contribution from the blocks
For flat plate fin:
Page 534
SEMIT 2023
AT = WL + 2Nf H[L+t] + 2B[L+W] ..………………………(5)

When dealing with a perforated fin, the area AS referred to in equation [1] takes into account both the
external surfaces of the fin and the internal surfaces of the perforations.
π
AT = WL + 2Nf H [L+t] + 2B [L+W] + Np Nf dt ̶ Np Nf 4 𝑑2 …………………………(6)
For cross-cut fin:
AT = WL + 2Nf H [L+t] + 2B [L+W] + Nf CH ̶ 2 Nf tH .….….….….….….….….(7)
Hence average convection heat transfer coefficient hav can be finding:
𝑄𝑁
hav = Tout + Tin ..…..……………………(8)
AT [Ts−( )]
2
Now, the thermal resistance is calculated as

1
Rth = havAT
…...…………………….(9)
The dimensionless values are computed as follows:

The Nusselt number, Nu is defined as:
ℎ𝑎𝑣 𝐷ℎ
Nu = Ka
…..…………………….(10)
The average velocity is determined by:
𝑉
Vav = …..…………………….(11)
𝐴−𝐴𝑓𝑟𝑜𝑛𝑡
The duct Reynolds number, Re is defined as:

𝜌𝑎 𝑉𝑎𝑣 𝐷ℎ
Re = 𝜇𝑎
…..…………………….(12)
The hydraulic diameter of a rectangular section of the wind tunnel is defined as:
𝐴𝑤
Dh = 4 𝑃𝑤 …..…………………….(13)
Aw denotes the cross-sectional area of the wind tunnel, while Pw represents its perimeter.
Evaluation of Fin Performance: To evaluate the effectiveness of fins, it is important to determine their fin
performance. This is achieved by comparing the actual heat transfer rate from the fin arrays (Qf) to the
heat transfer rate of plate-fin arrays (Qfp) and calculating their ratio:
𝑄𝑓
ɛfin = 𝑄𝑓𝑝
…..…………………….(14)
To evaluate the percentage enhancement in fin performance, one can utilize the provided equation:
𝑄𝑓−𝑄𝑓𝑝
%ɛfin = …..…………………….(15)
𝑄𝑓𝑝
When performing calculations, the thermophysical properties are acquired at the bulk mean temperature,
which serves as the representative average temperature:
𝑇𝑜𝑢𝑡+𝑇𝑖𝑛
Tmean = 2
…..…………………….(16)
The characteristics of the air are examined at ambient temperature, referred to as T∞. The heat sink
receives heat from the element heater positioned on its base. The average temperature of the heat source
is Ts. Once the total thermal resistance value is established, the overall heat dissipation can be computed.
Page 535
SEMIT 2023
Results and discussion

The present paper showed how the thermal performance of perforated heat sinks is affected by different
hole diameter. A comparison is drawn between three models of these heat sinks and a non-perforated heat
sink. as following.
a- Changing air velocity:
- Heat transfer (Q)
As the purpose of manufacturing heat sinks is to cool the hot bodies and remove the excess heat, the
amount of heat dissipated during testing the heat sink helps in choosing the best type. Figure [6] shows
the comparison between perforated and non-perforated heat sink, also shows the influence of perforation
diameter on the amount of heat dissipated(Q). It is noticeable that perforated heat sinks have a higher
value of Q than solid, where the solid fins read lowest value, as the perforation enlarge the surface area
for exchanging the heat. For perforated heat sinks, PHS-Ø 3mm has the highest rate of Q, which enhanced
Q by 14.21%, while other types improved the Q by (6.57%, 10.01%) for (PHS-Ø4mm, PHS-Ø3.5mm)
respectively at Re 20000. The size of the holes might explain this tendency. Holes increase the area for
exchanging heat with the air; bigger holes in the fin imply cold air stays for a longer time due to the vortex
that forms at the hole area, which prolongs heat exchange between the fins and the surrounding, resulting
in a faster cooling.
Non- perforated PHS-Ø 4mm Non- perforated PHS-8P
PHS-Ø3.5mm PHS-Ø3mm PHS-Ø 4mm PHS-Ø3mm

60 60
50 50
40 40
Q
30 30
20 20
10 10
10000 15000 20000 25000 30000 35000 11642 15518 19391 23265 27133 30998
Re Re
Figure 6. The relation between heat transfer rate and Re

- Nusslet number (Nu) and heat transfer coefficient (h)
It cleared in figure [7 and 8] that Nu increase with Re increasing for all types of heat sink, as Nu is a
function of Re. Also it cleared that non-perforated heat sink had the lowest value of Nu. Because of the
perforating fins, turbulence increased, resulting in better heat transmission. Furthermore, using a
perforated heat sink rather than a solid one increases the surface area. In the current study, perforated heat
sinks (PHS-Ø4mm, PHS-Ø3.5mm, PHS-Ø3mm) increased Nu by (22.3%, 38.61%, 47.94%) at Re 20000,
respectively. For h, the enhancement ratios were (18.20%, 30.51%, 34.3%) for (PHS-Ø4mm, PHS-Ø3.5mm,
PHS-Ø3mm) respectively, at same point of Re. PHS-Ø3mm has the greatest Nu and h that mean this type
cooling faster than others.
Page 536
SEMIT 2023
Non- perforated PHS-Ø 4mm Non- perforated PHS-Ø 4mm
PHS-Ø3.5mm PHS-Ø3mm PHS-Ø3.5mm PHS-Ø3mm

630 140
540 120
450 100
360 80
Nu
60
h
270
180 40
90 20
0 0
10000 15000 20000 25000 30000 35000 10000 15000 20000 25000 30000 35000
Re Re
Figure 7 and 8. The variation of Nu and h with Re
- Fin efficiency ƞf
The efficiency decreases as Re increases, as seen in figure [9]. As previously noted, this is due to the
larger surface area provided by perforation. Furthermore, in comparison to non-perforated heat sink
efficiency, the PHS-Ø3mm heat sink has the highest efficiency, with an improvement ratio of (4.24%),
while the PHS-Ø3.5mm, PHS-Ø4mm have less (an improvement ratio of (2.62% and 3.91%) respectively
at Re 30000.
PHS-Ø3mm PHS-Ø3.5mm
Non- perforated PHS-Ø 4mm PHS-Ø 4mm Non- perforated
PHS-Ø3.5mm PHS-Ø3mm 100%
100%
98%
98%
96% 96%
94% 94%
ᶯf
ᶯf
92% 92%
90% 90%
88% 88%
86%
86%
10000 15000 20000 25000 30000 35000
11642 15518 19391 23265 27133 30998
Re
Re
Figure 9. The relation between fin efficiency and Re
Weight reduction
The weight of the heat sink is very important to know, for economic reasons; lighter heat sinks are
desirable. As shown in figure [10], increasing the diameter of the holes decreases the weight of the heat
sink. Surface area is inversely related to heat sink mass. This trend suggests that perforating the fins
entails removing sections of the fins, resulting in a reduction in the fins' mass. The figure shows that PHS-
Ø3mm, PHS-Ø3.5mm, PHS-Ø4mm reduced the weight of the solid heat sink by (7.21%, 7.47%, 9.58%)
respectively.
Page 537
SEMIT 2023
340
335
330
325
320
Weight (gr.)
315
310
305
300
295
290
PHS-Ø 3mm
PHS-Ø 4mm
Solid
PHS-Ø 3.5mm
Figure 10. Weight reduction of heat sink types
b- Changing thermal loads:

All previous results were at constant voltage (80 v) and varying air velocity; this part will demonstrate
the influence of altering thermal loads (at constant air speed 2.5 m/s) on heat sink thermal performance.
- Nusslet number and heat transfer coefficient:
According to figures [11, 12], both Nu and h rise with increasing thermal loads. Nu was raised
by (4.97, 10.92%, 13.47%) while using tested heat sinks PHS-Ø4mm, PHS-Ø3.5mm, PHS-Ø3mm ,
respectively, when compared to non-perforated heat sinks. The perforated heat sinks (PHS-Ø4mm,
PHS-Ø3.5mm, PHS-Ø3mm) provide these readings for h (4.93%, 10.55%, 13.74%). The maximum
value of h and Nu was found while testing the PHS-Ø3mm, and the lowest value was obtained
when testing the solid heat sink. The temperature differential between the hot body and its
surrounds increases as the thermal load increases, resulting in a rise in the convective heat transfer
coefficient. Then, because the Nusslet number represents the ratio of convective to conductive
heat transfer inside a fluid [10], raise Nu-.
Non- perforated PHS-Ø 4mm

PHS-Ø 3.5mm PHS-Ø 3mm
100
550
90
475
80
400 70
Nu
325 60
250 50
175 40
100 30
10 20 30 40 50 60 70 80 90 100 110 10 20 30 40 50 60 70 80 90 100 110
Loads Loads
Figure 11 and 12. The variation of Nu and h with loads
- Fin Performance ƞf
Figure [13] depicts the variation in fin performance caused by thermal load impact. The performance of
the fins degrades as heat loads grow. as perforated heat sinks (PHS-Ø4mm, PHS-Ø3.5mm, PHS-Ø3mm) were
Page 538
SEMIT 2023
tested at (100V), their fin performance improved by (0.58%, 1.16%, 1.42%)) as compared to solid heat
sink. In terms of total thermal efficiency, the heat sink (PHS-Ø3mm) surpassed the other varieties. Because
the perforated fins have more surface area that interacts the surrounding air than the non-perforated fins,
it dissipates more heat.

100
95
ƞf
90
85
10 20 30 40 50 60 70 80 90 100 110
Loads
Figure 13. Variation of fin performance with loads
CONCLUSIONS
This study discusses the influence of perforation size on the thermal performance of heat sinks. Three
perforated heat sinks were employed and compared to a non-perforated heat sink to get the following
conclusion:
•Perforation increases the convective heat transfer coefficient value by increasing the heat exchange area
between heat sinks and air flow through the duct.
• When compared to a solid heat sink, increasing the diameter of perforations (Dp) enhanced heat transfer
by (6.57%, 10.01%, 14.21%) for (PHS-Ø4mm, PHS-Ø3.5mm, PHS-Ø3mm) respectively at Re 20000.
• The larger holes, the higher the weight decrease; PHS-Ø4mm lowered the weight of the solid heat sink
by 9.58% but PHS-Ø3mm reduced the weight by 7.21% only.
Nomenclature
Symbol Definition Unit
Q Heat transfer W
hav Average heat transfer coefficient W/m2.K
T Temperature °C
Ts Surface temperature °C
Tꚙ Mean temperature °C
Tin Temperature at inlet area °C
Tout Temperature at outlet area °C
Tm Mean temperature °C
ΔT Temperature difference °C
W Duct cross section width mm
A Area mm2
Ac Cross sectional area mm2
As Surface area mm2
Asf Surface area of fin mm2
Ass Surface area of fin space mm2
Page 539
SEMIT 2023
Ap Perforation area mm2

Np Perforation number -
Nf Fins number -
Wf Fin width (fin thickness) mm
Lf Fin length mm
Hf Fin height mm
𝑄̇ conv. Convective heat transfer W
𝑄̇ cond. Conductive heat transfer W
𝑄̇ rad. Radiative heat transfer W
𝑄̇ elect. Electric heat transfer W
I Current A
V Voltage V
v Air velocity m/s
K Thermal conductivity W/m.K
Re Reynolds number -
Nu Nusslet number -
m Mass Kg
𝑚̇ Mass flow rate kg/s
Cp Specific heat at constant pressure J/kg.K
Pp Perforation perimeter mm
Dp Perforation diameter mm
Ps Perforation space -
Ah Hole area mm2
Ns Space number -
R Thermal resistance °c/W
Rconv. Convective resistance °c/W
Rcond Conductive resistance °c/W
dh Hydraulic diameter mm
Ad Cross-sectional area of duct mm2
Pd Perimeter of duct mm
Greek Symbols
Symbol Definition Unit
Ø Hole diameter (perforation) -
𝜈 Kinematic viscosity m2/s
ꝭair Density of air Kg/m3
Ƞ Thermal efficiency -
ɛf Fin effectiveness -
Θb Temperature difference -
Subscripts
Symbol Definition
Av Average
Base Heat sink base
Fin, Fins Heat sink fin
In Inlet
Out Outlet
M Mean
ꚙ Infinity
Re Reynolds number
Nu Nusslet number
Pr Prandtle number
Page 540
SEMIT 2023
References
[1] Sakkarin Ch. a,b, Somchai W., “A review of the effect of flow directions and behaviours on the
thermal performance of conventional heat sinks”, International Journal of Heat and Mass Transfer, 81
(2015) 10–18.
[2] Jin-Cherng Sh., Jhao-Siang J., “Heat Transfer Enhancement of Plate-Fin Heat Sinks with Di_erent
Types of Winglet Vortex Generators”, Energies, 2020, 13, 5219.
[3] Kim O.N. Sara, T. Pekdemir, S. Yapici, M. Yilmaz, “Heat-transfer enhancement in a channel flow
with perforated rectangular blocks”, J heat and fluid flow,22 (2001) 509-518.
[4] M.R. Shaeri, M. Yaghoubi, K. Jafarpur, “Heat transfer analysis of lateral perforated fin heat sinks”.
Applied Energy, 86 (2009) 2019–2029.
[5] Venkitaraj K.P., sanoj Kidikh,“Natural Convection Heat Transfer Enhancement from Rectangular Fin
Arrays with Divers Geometrical Perforations” 2016, 78-1-4673-9925-8.
[6] Rahul S., D.D. Palande, “Heat Transfer Enhancement by using perforation: A Review”, IRJET,
(2016) 2395-0072.
[7] Mohammed R. Shaeri, R. Bonner, “Laminar forced convection heat transfers from laterally
perforated-finned heat sinks”, App. Thermal Engineering 2017 (116) 406–418.
[8] Sakkarin C., Ho Seon, Lazarus G.A., Smochai W., “Fluid flow and heat transfer characteristics of
heat sinks with laterally perforated plate fins”, International Journal of Heat and Mass Transfer, 138
(2019) 293–303.
[9] Cheng-Hung Huang, Yaun-Rei Huang, “An optimum design problem in estimating the shape of
perforated pins and splitters in a plate-pin-fin heat sink”, International Journal of Thermal Science, 170
(2021) 107096.
[10] Cheng-Hung Huang, Yaun-Rei Huang, “An optimum design problem in estimating the shape of
perforated pins and splitters in a plate-pin-fin heat sink”, International Journal of Thermal Science, 170
(2021) 107096.
Page 541
SEMIT 2023
SP13: Advancing Innovation and Collaboration in Science, Engineering, and

Information Technology
Improve Channel Capacity of TDMA

Aqeela N. Abed
College of Medicine, University of al- Iraqia, Iraq
Abstract
This paper gives time division multiple access analyses (TDMA) procedures and their crossovers with
the multiplexing of time division (TDM) for execution of Worldwide Portable the satellite
communications. In the satellite communication frameworks, and particularly in satellite systems,
numerous clients are dynamic at the similar time. The synchronous communications issue between
numerous multipoint or single versatile disciple clients, in any case, may be unraveled by utilizing the
plans of multiple access technique. Subsequently the frameworks assets as the power of transmitting and
constrained of the bandwidth, to utilize the fitting of channels by a total cost and to make a distinctive
multiple access to these channels. This produces an issue of partition and summation for signals within
gathering parts and the transmission, separately. By choosing the current issue comprises within this
improvement of transmission in arrange to isolate signals from various clients unmistakably on the
gathering portion of orthogonal channels.
The present work is appropriated to study and calculate the system capacity of TDMA/FDMA for
multiple beam satellites by performing analysis among multiple accesses. The system capacity could be
increased by using the time division multiple access (TDMA) techniques by various different variables
which have affected the channel capacity. The number of mobile users has been increasing day by day
with no or little increase in RF spectrum allocation, TDMA isn't bandwidth limited.
The system of digital capacity TDMA is marked in the present work. The improvement of capacity would
be in order tomes from 5-10 of FM analog in the absence of adding any cell sites. For instance, the TIA
standard in North America displays about 50 Erlang/km2 with a distance of 3 Km between site to site.
and then, the TDMA standard allowable faster handoff mechanism.
Therefore, can conclude from the present work that the digital air interface of cellular TDMA which used
to build networks of individual communications.
Thereby, the technique of TDMA is the key to providing effectively the methods of channel allocation
and handover.
Keywords: TDMA, capacity, Multiple Access, Channel Capacity
1- Introduction
The technique of multiple access (CDMA, FDMA, and TDMA) displays theoretically a similar capacity
in a perfect medium, thus the typical environment encountered in the communications of cellular, better
capacities are provided than others with several techniques. The earlier analog cellular systems have
limited capacity that employs modulation of frequency and this became clear about 1987, and a digital
mechanism that displays more capacities were suggested to overcome limitations. Time division
multiple access (TDMA) and code division multiple access (CDMA), where, the main digital
Page 542
SEMIT 2023
transportation technique which was researched. through the research, it was found systems of CDMA
that display the magnitude of highest capacity when contrasted with other competing technologies of
digital such as TDMA and with technologies of analog as FM.
The satellites of mobiles under contemplation to use antennas have multiple beams and utilize frequency
reuse of the appropriated L- FDMA/TDMA is an orthogonal system. They are suffered from the
limitations of bandwidth and sensibility to the interference of inter-beam.
The capacity of mobile conventions as TDMA (Time Division Multiple Access), FDMA (Frequency
Division Multiple Access), and CDMA (Code Division Multiple Access) frameworks are compared
accepting culminate power control and the comes about are compared with the past comes about.
Strategies of calculating the activity capacity of such frameworks are talked about. CDMA can give up-
to 20 times more activity capacity than FDMA and 5.3 times more activity capacity than TDMA.
In any case, whereas the FDMA conspire includes moderately straightforward frequency tuning for
getting to, and basically, by giving free channel on-off operation, the conspire of TDMA that requires
communication notions that are moderately modern. The oblige numerous clients, time of TDMA
interims may be essentially brief, burst-type transmissions requiring time interims of all clients may be
legitimately and synchronized precisely, by requiring a few timing levels of control. Hence, the
equipment of high-speed specified for these operations is modern and moderately, in numerous
situations, is still yet beneath development.
2- Channel capacity
Therefore, the two partitions A and B are placed in the same spot beam footprint of an exceptional
situation that presents itself. In spite of A&B might be physically far away, they are topologically in the
same effect of the point and through the viewpoint of accessing. the uplink of channel1 and the downlink,
as allocated to caller A, the accessibility is merely easy to part B and each one may be transmitted to the
another on the similar channel 1. As if they were sitting down next to each another so they both ending
were at their position. the A caller talks to the frontal end and the B hears in the tail end, after then they
reciprocate ends, and then B talks into the frontal end, while A listens at the end of the tail, all their
happens in spite whom they are utilizing a sole channel has the same way.
Partitions A and B have an extremely particular relation in which the transmissions of their voice
supplement each other in the time. Thereby, can be shared without collisions with channel 1, they use
similar to channel 1, for both uplink and downlink and the spectral bands respective for them, they are
filling up the silences of others without interference, thus, continuous transmission is generated, they
entirely use one channel instead of tying up the two channels. the second channel released from one of
the two channels; the usable channels number are doubled. Therefore, doubled the capacity itself; this
doubling up is feasible of all the endpoints for intra-beam communication to every satellite at a similar
situation considering the one-way uplinks/downlinks. For intra-beam communication through satellite is
especially considerable with GEO satellites and this is status almost unique, whereas there are multiple
spots beams by the apply channel doubling to all beam links in the one-way [1].
As capacity happens inside a one beam, and inside every and each one beam, there is a request its general
important into outlook of the truth where many possible combinations remaining inside linkages of intra-
beam. what grant special significance to communication of intra-beam is that it may be anticipated that
most or at smallest of the whole communication will be intra-beam investigates have exhibited that most
traffic of telephone is regional, or at minimum local territorial. Further, GEO systems strategy of
operational might frequently be for getting the communication to the regional networks inside the itself
beam for intercontinental voice calls, this lead of reduce the delay to space-induced diffusion.[33]
Page 543
SEMIT 2023
Capacity is clarified herein firstly to the channels frequency because its uncomplicated to express in terms
of channels frequency as the reality that a model of FDMA is significant in its own. capacity is applied
equally fine to other access/modulation methods special to CDMA and TDMA. The channel concept
defines as the system specified to the utilizer through the call if the frequency of channel as in FDMA as
a sampling slot of time as in TDMA, or apportionment of power (code labeled) as in CDMA. This entity
is underused when the channels of two one-way are allocated to two callers–paired in a connection of a
call and benefited maximally, thereby the two portions take part the doubling channel of one-way.
3- Channel Doubling
Doubling of channel supplies broadened by liberating channels for extra services with complicated of its
transfer dispensation from the satellite to the ground station. Therefore, through providing profit the
efficiency of the spectrum and capacity of the system doubling channel possibility to double the capacity
from point to point of fax traffic or voice. NPV effect upon may be greater than 1$ billion.
So, the flexibility of pricing: the cost must be lower cost frame for local (intra-beam) apposite the long
distance, economy and standard degree of serving, margin of high profit on economy service grades, and
beam overlap of planned can expand the intra-beam coverage. Complexity of processing may be
transferred to terrestrial resources, therefore, power saving and the request on the satellite and acquiring
in capacity there also. Doubling of channels might even lead the mix itself of desired service, and
economic tradeoffs on the basis. The channel doubling full value is carried out when the two-way of
paired-portions for communication and can both utilize the similar channel of one-way and contact with
alternating broadcasts. Along sides fax and voice communication, therefore each communication
including data, video operating, and voice in this mode of interactive can lower the need from two
channels of one-way to one channel participated in one-way. With n participants of intra-beam
teleconference call, there is a saving in the n-fold and interpreting to n-fold extension of capacity. In a
teleconference wider range, there are multiple capacities of extending inside every beam containing more
than one participant. Therefore, the multicast application is a perfect enactment of channel doubling in a
few times Inside every beam, all participators sharing with uplink one channel of single common and
downlink one of a single common. In each one moment is talking by using one participator with one
uplink channel the uplink channels of every channel are utilized one by one at time, when the participator
is talking in that beam. So, all the downlink channels are in continual utilize of the n participator lower
than a special commercial beam, therefore, the one uplink/downlink channel couple is required as a
comparison with n, further, the gain it is feasible because of all other participator are silencing when one
of them is talking [1,20].
4- Multiple Access
Techniques of multiple access are allowed a medium of communication to participate between different
users, which exemplify one of the most topics that challenge the digital communication. In the expression
of the huge users that can accommodate the channel which given, the channel bandwidth is N twice of
the individual bandwidth of user signals, these techniques are accommodated of N users without any
mutual interferences, but this limiting number of users there isn't a single additional can be supported
beyond through these techniques.
In the present work has been reviewed the limits of the capacity for the existing techniques of multiple
access, without any interference, some new is introduced to describe the concepts that allow
accommodating N users, also additional accommodating from users at the expense of SNR penalty.
[33,34]
Page 544
SEMIT 2023
The satellite is a communication joint which connects all sorts of users into the network can be
interconnected as flexible as possible. In a similar, time two key bandwidth of resources and power of
spacecraft may be used professionally. In a several implementations, must be essential that a satellite at
the same time accessing with hundreds of users making accessing problems of more multiplex.
Furthermore, complexity is included while elements as demand for treating a traffic mix (e.g. voice and
data) the variations of traffic and the necessity to incorporate traffic outgrowth are beheld. Ideally, the
access of multiple projects may be able to improve the following factors [3]:
1- Power of satellite radiated
2- Spectrum of RF
3- Connectivity
4- Adaptability to network growth and traffic
5- handled different traffic sorts
6- Economics
7- ground station Complexity
8- Secrecy for several application
A single technique can't improve wholly these factors, therefore, the trade-off analysis utilizing the viable
conditions that are essentially supplied, and the selection of accessed schema isn't clear. For example, if
the application is provisioning a huge number of low-cost mobile terminals for communication, the
schema of accessing would be durable and simple to allow the utilization of low-cost receivers of mobile.
At a similar time, the confirmed flexibility degree is needed to authorize the spectrum share among huge
numbers of mobiles that are added to the network and to accommodate them. By comparing with one
application, whereas a comparatively few large ground stations carrying intense traffic want to be
interconnected. In this state, the schema of accessing is multiplex and the main improvement criterion
should be optimal utilization of the obtainable bandwidth and power of satellite rather than the need for
simple ground stations [3].
The terms multiplexing access and multiple are indicated the sharing of fixed communication resource
(CR), there is a fluffy difference between multiplexing access and multiple with multiplexing users'
requirements or plans for CR sharing are fixed or most slowly changing. The resources allocation is
assigned before and the sharing is usually a process that happens inside the confines of a local site (e.g. a
circuit board) multiple access, however, usually includes the remote sharing of a resource such as a
satellite. by changing dynamically multiple access schemas, a system controller may become aware of
every user's CR needs the transfer information that has the amount of time required to constitute sets and
overhead of upper limit on the efficiency of the CR utilization [6].
multiple access Protocol or algorithms of multiple access (MAA) is a base of a user informs how to use
code, time, functions, and frequency to communicate through the satellite of other users. The multiple
access system is a hardware grouping and software which supports the MAA of the universal target of a
multiple access system to provide communications service in a timely, orderly, and efficient way [6].
figure (1) illustrates some architecture choices of basics for the multiple access of satellite system, the
legend indicates the symbols used for an earth station with and without an MAA controller, figure (1-a)
illustrates the case where one earth center station is designed as a controller or a major. This ground
station center owns an MAA computer and replies to the service requirements of entirely other users,
throughout the user's transmission, the satellite and back down to the controller were required, the
response of the controller unit requires extra transmission throughout the satellite, and then, there are
two downlink and up transmissions entailed for every service task. Figure (1-b) illustrates the case where
the MAA control is scattered among all the station of earth, therefore, there is no solitary controller, every
Page 545
SEMIT 2023
station of earth utilizes a similar algorithm and they ever have correspondent information regarding
assignments and access demands, thus, for every service assignment, there is one round trip only is
demanded, figure (1-c) demonstrates the case whereas the MAA controller inside the satellite. One round
trip is demanded for every service assignment is gained from service demand moves from user to satellite
and then responding from the satellite and the follow will be instantly [6].
(a) (b)
………… ……….
Ground station without MAA

Controller
(C)
Ground station with MAA
controller
Satellite with no controller
………. Satellite with controller
Figure (1) satellite multiple access architecture. (a) single earth station control. (b) Distribution
earth station control. (c) Satellite control.
5- TDMA
Time-domain multiple access offers access to the obtainable transponder spectrum on a time-participated
orthogonally basis in the easy example, one carrier fully occupying the transponder at a time. Every
section sends off a data burst in its exclusively assigned time period, the interval of the burst can be varied
to fit the station's instantaneous traffic need, at once the station's burst is competing, the other station gets
its assigned slot of time.
TDMA is characterized by being inveterately digital by appropriately utilizing the maximum power
capability of a transponder and being bandwidth efficient but it results in a more costly and complex
station of the earth than FDMA, because they want to preserve exact network timing and operates at burst
rates of 60-120 Mbits/sec. this extra cost is especially apparent in small traffic stations, in figure (2)
sharing of communication resources (CR) is achieved by appropriating bands of frequency, in figure (3)
the similar CR is participated by assigning every of M signals or users the occupancy of full spectral for
the system in a short period of the time called a slot of time.
The unutilized time regions between assignments of slot called the time of guard, permit for similar time
unpredictability between signals in adjacent slots of time and thereby act as zones of buffer to reduce the
interference, figure (4) is an illustration of a model TDMA application time of satellite is divided into
periods called frameworks, every framework is divided into assignable user slot of time, the framework
structure reiterations until a fixed assignment of TDMA shapes one or more slots that sometimes appear
through every framework time, every framework time station transports its data into bursts, the timed
until to attain the satellite coincident with its nominated slots of time, the bursts of the transponder of the
satellite were received, and the retransmission from the downlink together with the bursts from other
Page 546
SEMIT 2023
stations. the receiving station demultiplexed and detects suitable bursts with feeds the information to the
contemplated user [22,32].
Time division multiple access (TDMA) is characterized by access to the channel through a slot of time,
this has confident advantages: -
- At every immediate channel amplifies of satellite repeater only a solitary transporter which inhabits all
of the bandwidth of repeater channel, there are no intermodulation products and the benefits of carrier
from the channel saturation power.
- Transmission through for a large number of accesses remains high.
- There isn't want to control the conducting the station power.
- All station transmits and receive on the similar frequency whatsoever the origin.
TDMA disadvantage
- The need for synchronization which infers multifaceted procedures and the saving of two central
stations, fortunately these steps may be computer-driven and automated.
- The need to bandwidth and increases power resulting of high burst bit average compared to continuous
access for instance as with FDMA.
by considering a station – to –station link, the standard objective is assigned in terms of probability for
error, to determine the imposed value and the required value of the ratio (E/No). for the overall link ratio
(C/No) is determined by the relationship: C/No = ENo /R
It can see that the ratio (C/No) is relational to (R) for the capacity (Rb) the dimension of the station should
be determined in bandwidth and power to transmit a bit rate (R), where it is value high when the cycle of
duty (TB /TF) is lower. The station transmits fit bit rate (Rb) in FDMA and thus, the required (C/No) is
smaller. TDMA disadvantage is partly recompensated by the higher power supplied by the repeater of
satellite, the channel on the downlink compared with the overall case of FDMA, therefore TDMA implicit
more costly equipment at stations of the earth broadcast. Thereby, the equipment cost is however
compensated by better utilization of the space segment due to the higher transmission throughput in the
large number case of accesses, further digital processing leads to operational simplicity.[33,34]
Frequency
f5
f4 Guard band
f3
f2 Guard band
f1
f0
Time
Figure (2) Frequency division multiplexing
Page 547
SEMIT 2023
Frequency
Frequency band 3
Frequency band 2
Frequency band 1
Guard band
Guard band
Time
Figure (3) Time Division Multiplexing
Frame Frame
Guard band
Frequency band 2
Time Slot
Transmitted Burst
Of RF
Energy
Ground
Terminal
Page 548
SEMIT 2023
Figure (4) Typical TDMA Configuration (TDM/TDMA Network Architecture)
6- Throughput of TDMA
The throughput of TDMA transmission can be measured by the ratio of the channel of satellite repeater
capacity in solitary carrier running (only one access) and the capacity of the multiple access case for the
same channel, it assumed that the whole bandwidth I occupied in both cases. In single carrier running the
transfer capacity is R=B℘,
whereas (B) is the bandwidth of channel (Hz), and (℘) is the modulation spectral efficiency (bit/s. Hz).
With the capacity of multiple access is (R1 - ∑tj /TF ) where (∑tj) signifies to the sum of the times not
dedicated to the transmission of traffic ( burst headers plus guard times), thus, the throughput is:
ƞ = 1 – tj / TF
It expresses the ratio of the time devoted to transmission of traffic (which is the source of revenue for the
network operator) and the total channel utilization time (for which the network operator pays). The
throughput is bigger when the frame period TF is high and as soon as Σtj is small [20].
A throughput relies on the amount P of bursts in the edge, the number P be the bits in the heading and g
the equivalent period in bits of guard time, assuming that the frame comprises two reference bursts, this
gives: [30,32]
η = 1 – (P + 2) (P + g) / RTF
Whereas R is the bit rate of the frame (bit/s).
The throughput as the number function of accesses that is the numeral of station, N on the network relies
on the technique of traffic routing assumed it is down that:
- In the one carrier per link technique case P = N (N – 1)
- In the one carrier per transmitting technique of station routing P = N
Since throughput is lower when P is high the benefit of accepting the last technique is obvious, throughput
is directly involved in calculating the capacity of a network as the number function of telephone channels,
the character r be the signaling rate connected with one channel of telephone and the character n is
represents the telephone channels number, this gives:
Page 549
SEMIT 2023
n = η R/r
This expresses the number of telephone channels in the frame (bearer channels) the terrestrial telephone
channels number assigned to the network depends on the possible use of concentration by unit of digital
speech interpolation (DSI) and corresponding gain [20].
7- TDMA / FDMA Capacity

In the TDMA/FDMA state of transmission, the crest power of burst increased by the channel time
intervals number per frame of TDMA, so in a TDMA that has 8-channel per transmission of the carrier,
the burst power of the handset mobile increases by a value between 8 or 9 dB. The average of transmission
is increased for channels number owing to the inefficiencies of the frame such as a time of guard,
concurrence bits, etc. if Y is the number of channels per the carrier of the forward link [28]:
M (TDMA/FDMA forward clustered) =
𝑁 𝐵𝑐ℎ 𝑇 𝑁 𝑎 𝑎 𝑁
[ ( 𝐸° ) 𝑟 𝑌𝑥 − [( 𝐶 ) 𝐵𝑐ℎ𝐶 + (𝑁𝑃𝑅 ) 𝑐 + ( ) 𝐿 + ( 𝐶 ) 𝐴𝐷𝐽 + 1] 𝐴𝜑 + 1 ]
𝑑 𝑑𝑁𝑃𝑅
𝑌
𝑎𝑏 𝑁 𝑃𝑝𝑎2
[ 𝐵 [( 𝑃 ) 𝐵𝑐ℎ 𝑑𝐿 + 2 𝐴𝜑) ]
𝐿 −1
Whereas: N/P the overall power received in the L- band that received bandwidth of carrier, the channel
capacity of the reoccurrence linkage through Z channels for each carrier is:
M (TDMA/FDMA return collected) =
𝑁 𝐵𝑐ℎ 𝑇 𝑁 𝑁 𝑎
[ ( 𝐸° ) 𝑟 𝑍𝑥 − [( ) 𝐵𝑐ℎ𝐶 + ( ) 𝐴𝐷𝐽 + 1] 𝐴𝜑 − (
𝐶 𝐶 𝑁𝑃𝑅
)𝑐 +1 ]
𝑑
𝑍
𝑁 𝑃𝑝𝑎2
[𝑎 [( 𝑃 ) 𝐵𝑐ℎ 𝑑𝐿 + 2 𝐴𝜑 )]
𝐿 −1
8- Carrier Bandwidth of TDMA/FDMA and Bandwidth Limitation

In order to keep away from the operation of narrowband with acquisition problems and associated
drift the oscillator, the systems of mobile satellite utilize TDMA/FDMA, as example 32 channels TDMA
for the frontward link and 8 channels TDMA for the reoccurrence link.
The separation requirements and bandwidth for the TDMA/FDMA system counting frame efficiencies
(75%) should follow: [30,33]
Forward link of TDMA/FDMA, A channel rate = 6.4 (kbps) with 32 channels Rate 3/4:
Bitrate = 32 x 4.8 x (4/3) / 0.75 = 270 (kbps)
Bandwidth = 270 x 0.6 = 162 kHz
Carrier Parting (s) = 270 x 0.5 x 1.41 = 200 KHz
Reoccurrence link 8 channels
Bitrate = 8 x 6.4/0.75 = 68 kbps
Bandwidth (Bw) = 8 x 0.6 x 6.4/0.875 = 41 kHz
Carrier Separation (s) = 1.42 x 68/2 = 50 kHz
Through a 1° beam width for the site beam, the empty shape can be predictable at 1.35°, but the premiere
sidelobe at 1.75° through a width scan is 7°, one can capable to create approximately 140 to 150 spots
through a 7-beam group and frequency reprocess of 140/7 = 20 once
Page 550
SEMIT 2023
Utilizable L-band frequency of 4 MHz, the bandwidth – imperfect channel capacity is [28]:
[4𝑥20𝑥10]
M (band limited,TDMA) = [200𝑥10]
𝑥 32 = 12.800 circuits
9- Results and Discussion

FDMA / TDMA channel Capacity
FDMA / TDMA channel capacity for multiple beam satellite relationship can be stated in equation below,
which is obtained from figure ( ).
𝑁
( 𝐸 ° ) (𝐵𝑐ℎ 𝑇)
𝑏 𝑟 𝑁 𝑎 𝑎 𝑁
[ −[( ) +( ) +( ) +( ) +1]𝐴𝑠𝑏 +1]
𝑌𝑋 𝐶 𝐵𝑐ℎ𝜇𝐶 𝑁𝑃𝑅 𝐶 𝑑𝑁𝑃𝑅 𝐿 𝐶 𝐴𝐷𝐽
𝑌 …………….. (1)
𝑎𝑏 𝑁 𝑃2𝑝𝑎
[ [( 𝑃 ) + 2 𝐴𝜑]]
𝐵 𝐵𝑐ℎ𝑑𝐿 (𝐿 −1)
Program (1) implemented these equation and Graph (1) obtained the relation between channel capacity
(C ) versus noise to energy ratio (N/E) forward link.
PROGRAM (1):
t=[0:0.1:10];
x=[(t*2371.37/(32*0.1))-(1.0023+0.4/19.952+0.4/(7*19.952)+0.01+1)*1.023+1];
x1=[(t*2371.37 /(32*0.1))-(1.0023+0.4/19.952+0.4/(7*19.952)+0.01+1)*1.023+1];
x2=[(t*2371.37/(32*0.1))-(1.0023+0.4/19.952+0.4/(7*19.952)+0.01+1)*1.023+1];
y=0.4*141.253*(0.2+10*1.023/(64-1));
y1=0.4*141.253*(0.2+15*1.023/(64-1));
y2=0.4*141.253*(0.2+20*1.023/(64-1));
c=[x/y];
c1=[x1/y1];
c2=[x2/y2];
a=10*log10(t);
plot (a,c , a,c1 , a,c2)
xlabel('N/E(dB)')
ylabel ('Capacity (Bit/Sec)')
title ('figure Normalized TDMA channel capacity versus channel N/E (forward link) with change the
(P)')
grid
At constant value of average power factor and quantization level Pav2= 10 & L = 8
(C) is directly proportional to (N/E) with constant value of average power factor (P av) and number of
quantization level (L).
By changing the average power factor (pav) with constant (L) we get Graph (1) Pav2 = 15 & L = 8
Program (1) implemented these equations.
Page 551
SEMIT 2023
when there is an increase in the value of average power factor the capacity decreases from (12) Kbit/Sec
to (10) Kbit/Sec. but the capacity is directly proportional with the (N/E) ratio.
If Pav2 = 20 & L = 8 we get on Graph (1) program 1 implemented these equations,
by continuous increase in value of average power factor the capacity decrease from (10) Kbit/Sec to (8)
Kbit/Sec, with the quantization level at remaining constant value, but the capacity will be directly
proportional with the (N/E) ratio
By changing the number of quantization level (L) with (Pav) constant we get on Graph (2).
When:
Pav2 =10 & L = 16 program (2) implemented these equations,
PROGRAM (2):
t=[0:0.1:10];
x=[(t*2371.37/(32*0.1))-(1.0023+0.4/19.952+0.4/(7*19.952)+0.01+1)*1.023+1];
x1=[(t*2371.37/(32*0.1))-(1.0023+0.4/19.952+0.4/(7*19.952)+0.01+1)*1.023+1];
x2=[(t*2371.37/(32*0.1))-(1.0023+0.4/19.952+0.4/(7*19.952)+0.01+1)*1.023+1];
y=0.4*141.253*(0.2+10*1.023/(64-1));
y1=0.4*141.253*(0.2+10*1.023/(256-1));
y2=0.4*141.253*(0.2+10*1.023/(1024-1));
c=[x/y];
c1=[x1/y1];
c2=[x2/y2];
a=10*log10(t);
plot(a,c , a,c1 , a,c2)
xlabel('N/E(dB)')
ylabel('Capacity(Bit/Sec)')
title('figure Normalized TDMA channel capacity versus channel N/E (forward link) with change the
(L)')
grid
If we increase the quantization level the capacity will increase too from (12) Kit/Sec to (18) Kbit/Sec,
with constant value of average power factor, but the capacity will be directly proportional with the (N/E)
ratio.
If L = 32 & Pav2 = 10 we get on Graph (2), program (2) implemented these equation, at continuous
increasing quantization level will cause increase in capacity from (18) Kbit/Sec to (20) Kbit/Sec, with
constant value of average power factor, but the capacity is directly proportional with (N/E) ratio.
Page 552
SEMIT 2023
Graph (1)
Graph (2)
Page 553
SEMIT 2023
10- CONCLUSIONS
Satellite communication capacities for FDMA were studied for L/C – band system. Capacity in
FDMA has the narrowband obstructions is the similar for all clients in CDMA and TDMA with PN
sequences, all clients endure from signal – to – noise proportion (SNR) corruption that develops with
the interference power, and the channel completely unusable past a few limit esteems of the
interference power. So that the opposite to FDMA where no unused clients can be included without
adjustment of the framework hardware.
furthermore, the TDMA strategy is progressed to advance multiple access innovations that admit
further than one client to get the radiofrequency (RF) channels into the satellite and other broadcast
communications frameworks. A rule job of TDMA is that the signals ought to partition for
milliseconds-long parcels. At that point, distributing a channel of a single frequency for a brief time,
after that movable to other channels to grant it claim the interim. Just alike the FDMA method, the
TDMA plot contains times of watch that avoid any intervention between channels, and decline the
variables of transmission impedances. The TDMA method contains numerous focal points such as
great capacity with higher information ranges. Be that as it may, the TDMA conspire has numerous
drawbacks such as attenuation impairment.
11- SUGGESTION FOR FUTURE WORK

1- Hybrid combination between computer simulation and practical design must be used to
compute capacity, where the computer simulation alone is not feasible.
2- Multi user algorithms may be used to increase the channel capacity without any
interference.
TABLE (I) LINK PARAMETERS (FORWARD LINK)
PARAMETER 7BEAM/CLUSTER
EIRP at back off, EIRPb,ddBW, NPR=13 72.0
Nr beams, B/b (140)dB -21.5
Power control advantage, p, dB 2.5
EIRP into area coverage antenna, dBw 53.0
LFD (fading), dB -10.0
Ld (path loss) and atmospheric loss, dB -187.6
Gd, dB 0.0
Received total power, P, dBw -144.6
Boltzman constant KBI, dBw/KHz 228.6
T, dBk -25.0
W or Bch, dBHz -38.0
Nd, dBw -165.6
(N/P)d, dB (Noise to total power ratio) -21
Chiprate, bps 9.6*106
Bitrate, bps 4.8*103
Bandwidth, KHz 6.4
WT 1.332
Eb/No required, dB 5
(No/Eb) WT 0.4373
(N/C)u, dB 0.01
Asb 1.023
(N/C)fade, dB 0.10
NPR, dB 13
(N/C)sl 0.023
Page 554
SEMIT 2023
References
1. http://www.cow.soc.org/cit/privect/1997/oct/Braley.html/IEE communication magazine
2. Macmillan PressLTD (1995), M.Richharia, satellite Communication System Design
Principes,2nd Ed.
3. James Martin, (2000), Communication Satellite system.
4. Macmillan, (1985), Digital Communication and Spread Spectrum System, New York.
5. Hill.Inc., Mo Graw, (1995), Mobile cellular Telecommunications (Analog and Digital System).
6. Proakis, J.G., McGraw-Hill, (1983), Digital Communication, New York.
7. Zain ALabdeen Jawad kadhim, 2001, Wire Less Digital Exchange using CDMA Technique, Al-
Nahreen University.
8. Andrew J. Vitrebi, (1995), CDMA Principles of Spread Spectrum Communication, Addison-
Wesley Publishing Company.
9. Kun-Nyeong Changl and Dongwoo Kim. (2000), Optimal Time Slot and Channel Allocation for
CDMA/TDD Systems with Unbalanced Traffic, Division of Managemet, Silla University, Pusan,
Korea, School of Electrical Engineering and Computer Science, Hanyang University, Ansan.
10. Eric Lawrey, (1997), The Suitability of OFDM as a modulation technique For Wireless
telecommunication, With a CDMA Comparison, Thesis research at James Cook University,
October.
11. Graig Michael Teuscher, (1998), Low Power Receiver Design for portable RF Applications:
design and Implementation of an Adaptive Multiuser Detector for an Indoor, Wideband CDMA
Application, PhD Dissertation to the University of California, Berkeley.
12. Magda EL-Zarki, , Spring (1998), Chapter 2 is 95 CDMA, Tcom 510
13. A. Ahlen, M. Stemad and C. Tidestav, March 6 (2001), Research on Multi-user Detectors for
CDMA Based Cellular System, Sgnals and Systems Group, Uppsala University.
14. Othman Mohammed Daffa-Allah, November (1998), Interference Rejection DS-Spread
Spectrum CDMA, MSc.Thesis, Military College of Engineering, Iraq.
15. Peng Peng, April 24 (2000), A Critical Review of Packet Data Services in Wireless Personal
Communication System, Wireless Networks and Mobile Computing, Individual Project Report.
16. Steve Gardner, July (1999), Burst Modem Design Techniques Part I, Communication System
Design Magazine.
17. Fred Harris, July (2000), Carrier Synchronization Techniques for DSP-Based Modems,
Communication System Design (CSD) Magazine.
18. Luo Cheng, Xu Xibin, Yao Yan, (1999), Towards Realizing IS-95 in Software Radio, Tsinghua
University, Beijing, China.
19. http://www.Componedex.com/application-notes/satellite.htm.
20. G. Maral, (1999), Satelite Communications System, Techniques and Technology-Third Edition,
Geneva 20 Switzerland.
21. Walter L. Morgan Gary D. Gordon, John Wiley & Sons, (1989), Commuications Satellite,
Inc.Canda.
22. Bernd Sklar, Prrentic Hall, (1988), Digital Communication Fundamentals and Application.
23. http://www.ant.uni-bremen.de/research/cdma/
24. Robert C. Dixon, John Wiley & Sons, (1984), Sprad spectrum system, Second Edition, Copy
right Inc. Canada.
Page 555
SEMIT 2023
25. http://www.indosat.net.id/elektro/assi0199a.html.
26. http://www.ant.uni-bremen.de/research/capacity of CDMA system.
27. http://www.ant.uni-bremen.de/research/one approach to increase capacity in DS-CDMA
28. http://www.ASSI on line.com
29. http://www.Drucker.associates.com.
30. Ilcev D. S. (2011), “Global Mobile CNS”, Manual, CNS Systems, Durban, South Africa, 285.
31. Maini A.K. & Agrawal V., (2007), “Satellite Technolog - Principles and Applications”, John
Wiley, Chichester,696.
32. Ilcev, (2017), “Global Mobile Satellite Communications for Maritime, Land and Aeronautical
Applications - Volume2”, Springer, Boston, US, 686.
33. EMC, (2015), "TDMA vs. SCPC SatLink Technical Note", EMC Satcom Technologies, Lysaker,
Norway, 13.
34. Garvey G., (2013), "Multiple Access Techniques", Atlanta RF, LLC, Roswell, GA, 53.
Page 556
SEMIT 2023
Seismic Design of Underground Structures’ Vulnerability – Review

Monkam Ngameni Huguette Maeva, Haibing Cai,
School of Civil Engineering and Architecture
Anhui University of Science and Technology, Huainan, China.
Email: Huguettemaeva11@gmail.com
Abstract
This paper reviews underground structures' seismic susceptibility, concentrating on design and
mitigation. Modern infrastructure requires underground structures including tunnels, subway stations,
and storage facilities. Their underground placement makes seismic design and earthquake performance
difficult. Ground motion characterization, soil-structure interaction, structural response, and failure
mechanisms particular to underground structures are covered in this review of seismic susceptibility.
Geotechnical conditions, structural kinds, and construction methods affect their vulnerability, according
to the paper. Seismic stability design and analysis measure underground structures' dynamic response to
seismic loads. The objective is to avoid earthquake seismic forces from collapsing structures. Seismic
stability design includes advanced analytical and numerical methods like finite element analysis (FEA),
discrete element modeling (DEM), and boundary element method (BEM). It also examines modern
seismic vulnerability assessment methods and international design standards. This review also examines
seismic retrofitting and design innovations for underground structures. This paper synthesizes and
analyzes existing literature to understand underground structures' seismic vulnerability and encourage
seismic engineering research in this vital field.
Keywords: Stability Design: Infrastructure, Earthquake, modeling, engineering
1.0 Introduction
Civil engineering needs underground structures' seismic stability design and vulnerability analysis. Due
to their placement below the ground surface and earthquake motion, underground structures have a high
seismic vulnerability. Several studies have evaluated the seismic vulnerability of underground structures
and generated design standards. When seismically constructing underground structures, Ghaemian and
Kuo (2018) suggest taking site-specific ground motion, structural reaction, and failure mechanisms into
account. From conceptual to detailed structures, underground seismic design should be included in the
project's design. Mazzoni et al. (2019) recommended 3D seismic response analysis for underground
structures. 3D seismic analysis may be more realistic than 2D. The seismic design standards for
underground structures are found in ASCE Standard 61-19 (ASCE, 2019). The recommendations provide
a systematic approach to assessing seismic risks and designing earthquake-resistant underground
structures. Civil engineering relies on seismic stability design and analysis of underground structures, and
many studies have improved their seismic performance. The conceptual seismic design of underground
structures should take site-specific ground motion characteristics, the structure's reaction characteristics,
and probable failure mechanisms into account. It should also be included in the project's overall seismic
design, from conceptual to detailed structures. Seismic stability design and vulnerability analysis of
underground structures' vulnerability ensures occupant safety and structure functionality during and after
an earthquake, minimizes damage, incorporates the latest standards, conducts a site-specific seismic
hazard analysis, develops appropriate structural models, identifies potential failure modes and mitigation
measures, evaluates seismic design measures, and developsSeismic stability design and analysis
safeguard underground structures from earthquakes. These technologies analyze ground shaking threats,
optimize design parameters for optimal performance, measure structural seismic load capacity, and offer
retrofitting methods to improve existing structures. Seismic stability design and analysis reduce seismic
vulnerability in underground structures.
Page 557
SEMIT 2023
Figure 1: Earthquake Damage Structure

Figure 1 shows a structure devastated by an earthquake, a hypothetical seismic event of great scale and
destructive intensity. The figure shows how an earthquake could damage a building or infrastructure. The
image shows a structure under serious duress. Structural deformation, fissures, and partial collapse show
the building's vulnerability to the earthquake's strong shaking.
High-magnitude seismic events can destroy structures, as shown. It acts as a visual reminder of the
significance of robust seismic design and proper construction techniques to promote structural resilience.
Figure 2 emphasizes earthquake dangers and the need for good seismic design. To reduce damage and
save lives, structural strength, load-bearing capacity, and reinforcement are crucial.
This image may also inspire seismic engineering studies. It emphasizes the need to advance understanding
and deploy novel methods to strengthen structures against extreme seismic events.
Figure 2 shows an earthquake-damaged structure, showing the potential destruction of a major seismic
event. It emphasizes seismic design principles and continuous efforts to prevent earthquake damage to
buildings and infrastructure, creating safer and more resilient communities.
2.0 Literature Review

2.1esearch of scholars
Seismic design codes and recommendations identify seismic risk and develop earthquake-resistant
underground structures. New research and technologies update codes and guidelines. "Seismic Design of
Reinforced Concrete and Masonry Buildings" (ASCE, 2019) is a popular underground seismic design
code. This standard offers design loads, seismic hazard assessment, seismic design ideas, and seismic
details for underground structures. EN 1998-5 (Eurocode 8), "Design of structures for earthquake
resistance" (CEN, 2004), is also crucial. This code covers underground seismic design, seismic hazard
assessment, seismic detailing, and structural element design. Several organizations regulate specialized
underground construction. ITA (2014) seismically designs tunnels, caverns, and underground structures.
These rules assess, design, and detail underground structures' seismic dangers. ASCE Standard 61-19,
Eurocode 8, and ITA standards govern seismic design for underground structures. These codes and
guidelines cover seismic hazard assessment, seismic design, and seismic detailing for underground
structures. Underground construction's seismic resilience is well-studied. These studies examined
earthquake behavior, seismic design, and underground construction risk. Li et al. (2021) quantified a big
Chinese underground cavern's seismic performance. The study indicated that cavern size and structure
considerably affected seismic loads. Ghosh et al. (2020) studied India's underground construction seismic
risk. The study indicated that earthquake-prone underground building in seismic hazard zones needed
Page 558
SEMIT 2023
seismic design. Papadrakakis et al. (2018) used numerical simulations to study subway station earthquake
behavior. The study indicated that soil-structure interaction significantly impacts subway station
performance and that seismic isolation may lower structural seismic response. ASCE Standard 61-19
(ASCE, 2019) and Eurocode 8 (CEN, 2004) are two seismic design standards for underground structures
that have been studied. Seismic performance, susceptibility, design, codes, and guidelines have all been
examined in underground structure seismic stability studies.
On the seismic stability of underground structures, numerous research has been carried out. This study
examined seismic stability, underground structures' earthquake response, seismic design, and seismic
dangers. Finally, scholars have examined underground structures' seismic stability, including
performance, vulnerability, design, and codes and standards. On the seismic stability of underground
structures, numerous research has been carried out. This research examined seismic stability,
underground structures' earthquake response, seismic design, and seismic risks Seismic performance and
susceptibility, seismic design, seismic laws, and seismic standards for underground structures have all
been extensively studied.
2.2 Seismic Design for Underground Structures
Seismic risk is a problem in many places of the world since earthquakes can hurt and kill. FEA simulates
seismic loading in structures. This literature review discusses seismic hazard analysis and FEA. Seismic
hazard analysis describes seismicity—earthquakes' frequency and magnitude. Ground motion and site
characteristics predict earthquakes and their consequences on structures. FEA solves complex
engineering problems like the seismic loading behavior of structures. Break down a building and mimic
loading with math. Element behavior in FEA predicts structure behavior under different loads. FEA helps
assess seismic hazards. Baradaran et al. (2021) used FEA to study structures under high-intensity seismic
loads. FEA can predict extreme loading behavior for building structures to endure high-intensity seismic
events. Kaviani et al.'s 2019 FEA study studied steel structural seismic response. The study found failure
mechanisms and seismically hazard-resistant structures using FEA. Shao et al. (2020) used FEA to study
seismic loading dynamics. FEA can predict dam behavior under different loading conditions, which is
important for establishing the safety of these structures. FEA models seismic loading on structures. FEA
helps engineers build stronger, seismic-resistant structures. For underground structures, each seismic
analysis and assessment method has pros and cons. Normal method A FEM: This numerical technique
simplifies engineering problems. FEM's intricate geometry and boundary conditions are ideal for the
seismic investigation of underground structures. Kianoush et al. (2019) examined FEM fault-zone tunnel
seismic behavior. ELM simplifies seismic analysis with linear soil behavior and frequency-dependent
damping. ELM is used for preliminary seismic analysis of underground structures due to its simplicity
and processing efficiency. ELM assessed a Chinese metro station's seismic reaction by Jiang et al. (2020).
Acceleration time histories show seismic ground motion in RSM. Maximize a structure's ground motion
response spectrum. RSM is used for seismic analysis and the design of underground structures because it
takes ground motion frequency into account. Performance-Based Seismic Design (PBSD) meets seismic
goals. Seismic hazard evaluations, ground motion record selection, and structural performance objectives
are all included in the process. PBSD designs seismic underground structures like nuclear power reactors
and traffic tunnels. PBSD built a Chinese hydropower facility's deep underground cavern. For
underground structures, each seismic analysis and assessment method has pros and cons. Project
requirements and underground structural complexity determine the approach.
2.2 Flaws and Weaknesses In The Seismic Stability Design
Researchers and practitioners attempt to overcome the flaws and weaknesses in the seismic stability
design and analysis of underground structures.
1) No site-specific data: For seismic design and analysis of underground structures, site-specific
data is hard to get. Several researchers (e.g., Kausel, 1996) point out the need for good geological
and geotechnical data in accurately estimating underground structures' seismic sensitivity.
2) Simplified modeling assumptions: The complicated nature of earthquakes and the dynamic
behavior of underground structures frequently necessitate simplifications and assumptions during
the design and analysis process. These simplifications may underestimate the structure's
Page 559
SEMIT 2023
reactivity and vulnerability (Gazetas et al., 1996). The research includes better modeling and
more accurate assumptions.
3) Predicting ground motion during earthquakes is uncertain. The uncertainties in ground motion
prediction are exacerbated by limited seismic records and a lack of knowledge of seismic hazards
(Pecker et al., 2011). These uncertainties can impair the seismic design and analysis of
underground structures.
3.0 Methodology
3.1 Selection of Underground Structures for Analysis
Seismic risk, structural significance, and failure implications determine underground seismic study and
evaluation structures. Seismic structures: For roads, railroads, and subways, seismic assessment is
needed. These buildings could collapse in densely populated areas. Water Treatment Pipes, reservoirs,
and plants clean water. These structures must be built and assessed for seismic risks to function after an
earthquake. For underground storage tanks, including those for oil, gas, chemicals, and mines, seismic
investigation and assessment are needed. Chemical-laden buildings must be earthquake-proof.
Subterranean parking facilities must be seismically developed and assessed in heavily populated areas.
These systems' flaws hurt society and the economy.
Underground seismic stability design includes seismic hazard analysis, site characterization, numerical
modeling, seismic response analysis, and vulnerability assessment (Wang et al., 2020). Geological and
seismic data define subsurface structural seismic risk. This analysis identifies the underlying structure's
peak ground acceleration, velocity, and displacement. Site characterization involves soil, groundwater,
and rock mass data collection and analysis. This data creates a subterranean structural and soil numerical
model. Subsurface structure and soil finite element or finite difference models simulate. It simulates
earthquake reactivity, natural frequencies, mode forms, and damping ratios. Numerical ground motion
model inputs show seismic risk in seismic response analysis.
3.2. Importance of shock absorption structures
Seismic risk, structural significance, and failure implications dictate subsurface seismic study and
evaluation structures. Risk assessment and failure impact should guide selection. Seismic evaluation
requires subsurface structures. FEMA requires environmental, human, and economic risk assessments.
Subways, roads, and trains are seismically examined (Li and Wang, 2017). Water pipelines, reservoirs,
and treatment plants need seismic analysis and evaluation to function during and after earthquakes
(Ebrahimi and Roshan, 2017). Nuclear power plants need seismic monitoring and appraisal due to their
environmental and human implications. Underground storage facilities such as oil and gas tanks, chemical
tanks, and mines need seismic appraisal because of their dangers (Papadopoulos and Panagiotou, 2018).
Seismic-risk underground parking structures in densely populated areas must be planned and analyzed to
minimize catastrophic economic and social losses (Choi et al., 2017). Select subsurface structures for
seismic risk assessment using seismic evaluations. Tunnels, water treatment, nuclear power reactors,
underground storage, and parking garages are examined.
Figure 2: Ground Movement and Isolation Bearing
Page 560
SEMIT 2023
Figure 2 shows ground movement and isolation bearing in seismic design for structures. Earthquakes
cause ground vibrations that can exert enormous forces on buildings and infrastructure. These stresses
can cause structural damage, jeopardizing safety and functionality. In seismic design, isolation bearing
systems are used to reduce ground movements. The isolation bearing buffers the structure from the
rocking ground during an earthquake. The figure shows the superstructure, representing the building, and
the isolation bearing, between the superstructure and the ground. Arrows show lateral displacement and
rotation of the ground. The isolation bearing, usually comprised of elastic and damping materials, isolates
the superstructure from ground motion. Thus, forces imparted to the superstructure are greatly minimized,
reducing damage and enhancing seismic performance. The graphic shows the need for isolation bearing
design and installation for seismic protection. To enhance isolation system performance, consider bearing
stiffness, damping, and load-carrying capacity. Figure 1 emphasizes ground movement and isolation
bearings in seismic design. Engineers can protect lives and essential infrastructure from seismic
disturbances by strengthening structures using isolation systems.
4.0 Results and Analysis

4.1 Details of Underground Structures Analyzed
To determine seismic stability, a computer model of a subterranean structure and soil is subjected to
seismic loads representative of the region's seismic dangers. FEA, FDA, or BEA created the numerical
model. The numerical model requires structural geometry, material specifications, soil properties, and
seismic stresses. Structural geometry encompasses underground structure size, shape, internal supports,
bracing, and reinforcement. Materials have elastic modulus, Poisson's ratio, density, soil shear strength,
permeability, and porosity. PGA, PGV, and Sa load seismically. Seismic hazard analysis predicts future
earthquakes using geology, seismic wave propagation, and past earthquake data. Displacement, stress,
and strain evaluate the structure's seismic performance after applying seismic loads to the numerical
model. A computer model of an underground structure and its soil is subjected to seismic stresses
indicative of the region's seismic hazard to assess seismic stability. The analyses indicated seismic
retrofitting. Structural geometry, material properties, soil loads, and seismic loads are modeled. Analysis
may suggest retrofitting.
4.2 Finite Element Analysis Results
Finite Element Analysis (FEA) is a numerical method that is widely used to simulate the behavior of
structures under various loading conditions, including seismic loads. The results of FEA are typically
presented in the form of tables and graphs, which provide information on the behavior of the structure,
including its deformation, stresses, and strains. In conclusion, Finite Element Analysis (FEA) results are
typically presented in the form of tables and graphs, which provide information on the behavior of the
structure, including its deformation, stresses, and strains. These results were used to assess the
effectiveness of different strengthening measures and to guide the design of structures to withstand the
expected ground motions.
4.3 Seismic vulnerability assessment results
Table 1: Seismic Vulnerability Assessment Results for the Underground Structure
Damage State Sa (g) PGA (g) PGV (cm/s) Drift (%) Damage Probability
None 0.15 0.10 5.5 0.2 0.99
Slight 0.30 0.20 10.2 0.4 0.50
Moderate 0.60 0.40 21.5 0.8 0.10
Extensive 1.20 0.80 43.1 1.6 0.01
Complete 2.40 1.60 86.2 3.2 0.001
Page 561
SEMIT 2023
Fragility Curve for the Underground Structure

100
Parameters 80
60
40
20
0
None Slight Moderate Extensive Complete
Damage State
Sa (g) PGA (g) PGV (cm/s) Drift (%) Damage Probability
Graph 1: Fragility Curve for the Underground Structure
Table 1 shows underground structure fragility and seismic vulnerability data. The table shows damage
states for different ground motion levels based on Sa, PGA, PGV, and drift values. Damage probability
shows the likelihood of each damage state given a given ground motion level. The underground structure
fragility curve (Graph 1) shows the probability of a certain damage state for different ground motion
levels. Graph 1 shows the probability of each damage state for distinct Sa values, demonstrating its
likelihood for various ground motion levels. These studies examined the underground structure's seismic
vulnerability and found ways to mitigate earthquake damage or collapse.
4.4 Comparison of the results with existing codes and standards
General comparison of the analysis results with the requirements specified in some of the commonly used
codes and standards, such as ASCE 7-16 and the IBC, with the help of data gathered
Table 4: Comparison of Ground Motion Parameters

Ground Motion Parameter ASCE 7-16 Requirement Analysis Result
Peak Ground Acceleration 0.4g 0.35g
Spectral Acceleration 1.0g 0.9g
Duration 5 sec 3 sec
Comparison of Ground Motion Parameters

6
5
4
Values
3
2
1
0
Peak Ground Acceleration Spectral Acceleration Duration
Ground Motion Parameter
ASCE 7-16 Requirement (g) Analysis Result (g)
Graph 2: Comparison of Ground Motion Parameters
Page 562
SEMIT 2023
Table 4 and Graph 2 compared the analysis results to the codes and standards and assessed the
underground structure's seismic performance. The results of this comparison can help identify areas for
improvement or refinement in future analyses and provide insights into the adequacy of the seismic design
and analysis methods used in the Seismic Stability Design and Analysis of Underground Structures'
Vulnerability analysis.
4.5 Discussion
Seismic stability design and subsurface structural vulnerability analyses require sensitivity and
uncertainty analysis. These methods evaluate analytical uncertainty and input parameter sensitivity.
Sensitivity analysis alters ground motion parameters, soil conditions, and structure attributes and analyzes
the effects. This analysis identifies input parameters that affect underground structure seismic
performance and their effects. Factorial design, response surface techniques, and one factor were used in
a sensitivity analysis. Uncertainty assessment influences analysis findings. Ground motion, soil, structural
attributes, and modeling assumptions affect underground structure seismic stability design and
vulnerability analysis. Monte Carlo, Latin hypercube, and sensitivity-based approaches examined
uncertainty. They represent crucial input parameters that require more exact measurements or research
and give analysis results legitimacy. Seismic stability design and uncertainty assessment reduced
underground structures' seismic sensitivity. Identifying potential weaknesses or shortcomings, and
developing strategies to address them. Some areas for improvement in seismic stability design include:
• Improved site characterization
• Enhanced modeling techniques
• Robust design criteria
• Adequate detailing of reinforcement:
• Effective construction practices
These improvements increased seismic stability design and reduced earthquake damage. Seismic stability
design and vulnerability assessments help underground constructions survive earthquakes. Design and
analysis include seismic hazard analysis, finite element analysis, and vulnerability assessment. Engineers
may identify and mitigate weaknesses, making construction safer and more resilient. Despite restrictions
and norms, seismic stability design can be improved. Site characterization, modeling, rigorous design
requirements, reinforcement detailing, and construction techniques improve seismic stability.
Underground seismic stability design and vulnerability analysis are tough and require constant study,
innovation, and development. Engineers can develop underground structures that resist earthquakes,
protecting people and vital infrastructure.
5.0 Conclusion
The design and analysis of underground seismic structures are essential to ensuring their safety during
earthquakes. Underground structures are vulnerable to seismic occurrences due to their limited interiors
and complicated, unpredictable earthquakes. To address this susceptibility, architects and designers must
consider the site's geology, seismic risk, underground facility's structural design, and ground motion's
consequences. Seismic analysis and numerical modeling were used to evaluate underground structures'
seismic performance and identify weak places. Designers and engineers must apply the correct seismic
design standards and techniques to underground structures to ensure their structural integrity.
Underground structures must undergo routine maintenance and inspection to identify any damage or
possible flaws that might compromise their safety. Overall, seismic stability design and analysis are
essential for ensuring underground structures' safety and resilience. Underground structures were made
seismically robust by using design concepts and analytical techniques, saving lives and reducing
infrastructure damage.
Page 563
SEMIT 2023
References
American Society of Civil Engineers (ASCE). (2019). Seismic design of reinforced concrete and masonry
buildings. ASCE Standard 61-19.
Asadi, M., & Poursha, M. (2020). Seismic analysis of underground structures: A review. Tunnelling and
Underground Space Technology, 99, 103352.
Baradaran, M., Bagheri, A., & Rahmani, O. (2021). Seismic analysis of reinforced concrete structures
with high-intensity seismic loads. Advances in Structural Engineering, 24(8), 1457-1470.
Comité Européen de Normalisation (CEN). (2004). Eurocode 8: Design of structures for earthquake
resistance. Part 5: Foundations, retaining structures, and geotechnical aspects. Brussels, Belgium.
Chen, J., Hu, H., & Zhou, X. (2019). Seismic performance analysis of underground structures based on
the finite element method. Applied Sciences, 9(5), 902.
Choi, H.-J., Lee, J.-H., & Choi, S.-W. (2017). Seismic performance evaluation of underground parking
structures considering soil-structure interaction. Bulletin of Earthquake Engineering, 15(7), 3043-3063.
El Naggar, M. H., & Mahmoud, A. (2019). Seismic design of underground structures: A review.
Tunneling and Underground Space Technology, 93, 103099.
Ebrahimi, E., & Roshan, S. (2017). Seismic evaluation of water supply systems: A case study on a
reservoir. Geotechnical and Geological Engineering, 35(4), 1837-1845.
Federal Emergency Management Agency. (2006). FEMA 361, Pre-Standard and Commentary for the
Seismic Rehabilitation of Buildings. Washington, DC: FEMA.
Ghaemian, M., & Kuo, K. (2018). Seismic design and analysis of underground structures. Tunnelling and
Underground Space Technology, 73, 112-122.
Ghosh, A., Ghosh, P., & Chakraborty, S. (2020). Seismic vulnerability assessment of underground
structures in India. International Journal of Disaster Risk Reduction, 50, 101774.
Gazetas, G., Anastasopoulos, I., & Drosos, V. (1996). Seismic analysis of underground structures: State
of the art. Soil Dynamics and Earthquake Engineering, 15(7), 429-448.
International Tunnelling and Underground Space Association (ITA). (2014). Guidelines for the design
and construction of tunnels and shafts under seismic conditions. ITA Working Group No. 2. Paris, France.
Iervolino, I., Convertito, V., & Galasso, C. (2016). A comprehensive review of the main issues in seismic
hazard assessment. Geosciences, 6(4), 42.
Kamalian, M., Jafarzadeh, H., & Adib, A. (2017). A probabilistic seismic risk assessment of a nuclear
power plant considering epistemic uncertainties. Annals of Nuclear Energy, 109, 173-188.
Kausel, E. (1996). Seismic analysis of underground structures. Soil Dynamics and Earthquake
Engineering, 15(7), 415-427.
Kaviani, A., Vafaei, M., & Javaheri, R. (2019). Seismic response of a steel structure using finite element
analysis. Journal of Constructional Steel Research, 156, 427-440.
Kramer, S. L. (1996). Geotechnical earthquake engineering. Prentice Hall.
Li, Y., & Wang, Y. (2017). Seismic damage and reinforcement measures for subway tunnels in soft soil.
Journal of Performance of Constructed Facilities, 31(1), 04016034.
Li, Y., Lin, H., Zhang, X., & Li, J. (2021). Seismic performance evaluation of underground caverns
considering geological conditions. Soil Dynamics and Earthquake Engineering, 139, 106644.
Mazzoni, S., McKenna, F., Scott, M. H., & Fenves, G. L. (2019). Three-dimensional seismic analysis of
underground structures. Soil Dynamics and Earthquake Engineering, 117, 167-180.
Papadrakakis, M., Lagaros, N. D., & Fragiadakis, M. (2018). Seismic isolation in an underground
structures: a numerical study on a subway station. Bulletin of Earthquake Engineering, 16(6), 2591-2608.
Page 564
SEMIT 2023
Pecker, A., Del Gaudio, V., & Scotti, O. (2011). Seismic response of shallow tunnels: Uncertainties in
ground motion prediction. Bulletin of Earthquake Engineering, 9(6), 1865-188
Papadopoulos, I., & Panagiotou, G. (2018). Seismic assessment of underground storage tanks considering
soil–tank interaction. Journal of Loss Prevention in the Process Industries, 55, 52-59.
Shao, X., Guo, Y., & Liu, Y. (2020). Dynamic analysis of a concrete gravity dam subjected to seismic
loading. Engineering Structures, 209, 110288.
Wang, X., Yang, S., Cai, M., Chen, S., & Chen, J. (2020). Seismic stability analysis of underground
structures: A review. Tunnelling and Underground Space Technology, 100, 103449.
Page 565
SEMIT 2023
SP14: Uncertain Decision Making
Stochastic optimization methodology for production planning with uncertain demand and lead
time based on the digital twin
Dan Luo 1, Simon Thevenin1, and Alexandre Dolgui1
1
dan.luo2@imt-atlantique.fr
Abstract
Due to the manufacturing demand for mass customization and the widespread application of flexible
automated production equipment, manufacturers tend to shorten their production cycle. This results in a
loss of regularity in the production system, and it becomes difficult for manufacturers to control the
product and process quality. In the meantime, the upgrading of the production system puts forward higher
requirements for the decision-making of production planning. This means that production planning needs
to be as precise as possible. However, with the loss of regularity in the production process, it becomes
difficult to estimate the production capacity required to produce a given batch size. This paper considers
the lot-sizing problems for planning production with uncertainties. We propose a stochastic optimization
methodology for production planning with uncertain demand and lead time. The proposed method is
based on the simulation and Bayesian network forecast methods. Compared to the lot-sizing model with
uncertain demand, we integrate the lot-sizing model with the simulation model, domain model, and
Bayesian network model to forecast the lead time and improve the lot-sizing model with uncertain
demand and lead time.
Keywords: Stochastic optimization, machine learning, simulation, uncertain demand and lead
time.
1. Introduction
Nowadays, manufacturers are increasingly shortening production cycles and expand-ing product
assortments, which results in a volatile production environment characterized by customer demand
uncertainties, yields, production capacity, and lead times. To navigate these challenges effectively,
manufacturers require production planning tools capable of handling such uncertain parameters. Relying
on a single mathematical or simulation model with a sole objective is insufficient to offer efficient and
accurate decision support in complicated production scenarios. The advanced production environment
necessitates a comprehensive approach that considers various aspects and objectives. This need becomes
even more pronounced in the context of Industry 4.0, where research efforts increasingly focus on
integrating advanced technologies to effectively address complex and uncertain production environment.
Therefore, we propose a novel stochastic optimization methodology for production planning with
uncertain demand and lead time, which combines simulation and Bayesian network forecast methods. In
contrast to the model for lot sizing with uncertain demand, we have integrated the lot sizing model with
the simulation model, the domain model, and the Bayesian network model to predict the lead time and
improve the lot sizing model with uncertain demand and lead time. The production planning module
provides various production plans as input for the simulation model, from which we can determine
accurate lead time and planning adherence. Furthermore, the feedback data can be used to train the
Bayesian Network model to predict the probability distribution of lead time. In summary, our approach
pioneers the handling of complex production environments that are characterized by uncertainty. By
Page 566
SEMIT 2023
combining stochastic optimization methodology with simulation and Bayesian network methodologies,
we present a reliable solution for manufacturers to optimize production planning under uncertain demand
and lead time factors.
2. Literature review
The initial lot-sizing studies presupposed that all parameters were deterministic. However, in practice,
none of the planning parameters can be forecasted perfectly. The term "uncertainty" refers to the
discrepancy between the amount of information needed to complete a task and the amount of information
already available [1]. In the course of time, researchers have endeavored to formalize and model
uncertainty in manufacturing systems [2,3]. The production planning literature offers several approaches
and models to handle diverse forms of uncertainty. The primary three variables that are uncertain in
production planning are demand, lead time, and capacity.
Demand uncertainty is vital in production planning, particularly for manufacturers with prolonged
production lead times [4]. The various forms of demand uncertainty include order size and due date. For
instance, in semiconductor manufacturing, customers furnish a demand signal (an indication of their
ultimate orders) well in advance of the due date, and gradually adjust their orders as time progresses until
a firm order is finally obtained. Despite the significant difference between the demand signal and the firm
order, customers still insist on timely fulfillment of their orders [5].
The lead time is defined as the number of periods between placing an order and its arrival. The language
used is clear, objective, and unbiased, with a formal register and avoidance of contractions, informal
expressions, and unnecessary jargon. In production planning, two types of lead time are identified:
delivery lead time and production lead time. Quotations are clearly marked, and filler words avoided. The
former indicates the time taken by suppliers to deliver components, while the latter is the time from order
release to shipping date from the shop floor. Technical term abbreviations are explained the first time
they are used. The text adheres to conventional structure and balanced language, with precise word
choice, grammatical correctness, and consistent citation. Delivery lead time uncertainty is a common
issue in practice, often resulting from problems at the supplier production level or during transport [6].
Production lead time uncertainty occurs due to various factors, including inaccurate modeling of capacity
constraints when creating the production plan, machine breakdowns, and stochastic variations in
operation processing times [7]. To address this uncertainty, some studies propose modeling uncertain
lead times using discrete support probability distributions based on statistical data.
Very few studies have investigated the precise effects of each MRP parameter amidst demand and lead
time uncertainties, despite its usefulness in comprehending their impacts on the system's behavior in terms
of cost, service level, and stability [8].
3. Problem description and mathematical model

The typical production planning problem considered in real manufacturing systems/supply chains is a
multi-echelon multi-item capacitated lot-sizing problem (MMCLP). The objective of the MMCLP is to
determine the suggested production plan, including when to produce, how many items to produce, when
to buy materials, how many items to buy, and the amount of extra capacity required. These decisions are
made based on the demand, the bill of material (BOM), the production capacity, and the lead time. We
choose a generic enough model for MMCLP that would fit in most of the manufacturing industries,
detailed described in Luo et al [9]. In this model, we consider uncertain demand 𝐷𝑖𝑡 and uncertain lead
time 𝐿𝑖 , both can be represented with the probability distribution. The objective of the MMCLP is to
determine the suggested production plan. We define three decision variables. 𝑌𝑜𝑡 , a binary decision
variable, presents if a batch of operation 𝑜 is performed in period 𝑡. 𝑄𝑜𝑡 presents the quantity of operation
𝑜 to perform in period 𝑡. 𝑤𝑟𝑡 presents the amount of extra capacity required for resource 𝑟 in period 𝑡.
The objective function is the expected total cost.
Page 567
SEMIT 2023
𝑚𝑖𝑛 ∑ ∑(ℎ𝑖 𝐼𝑖𝑡 + 𝑏𝑖 𝐵𝑖𝑡 ) + ∑ ∑(𝑠𝑜 𝑌𝑜𝑡 + 𝑣𝑜 𝑄𝑜𝑡 ) + ∑ ∑ 𝑜𝑟 𝑤𝑟𝑡 (14)

𝑡∈𝑇 𝑖∈𝐼𝑒 𝑡∈𝑇 𝑜∈𝐼𝑐 𝑡∈𝑇 𝑟∈𝑅
𝐼𝑖𝑡−1 − 𝐵𝑖𝑡−1 + 𝑎𝑜𝑖 𝑄𝑜𝑡−𝐿𝑖 − 𝐼𝑖𝑡 + 𝐵𝑖𝑡 = 𝐷𝑖𝑡 𝑖 ∈ 𝐼𝑒 𝑜 ∈ 𝐼𝑐 𝑡 ∈ 𝑇 (15)

𝐼𝑖𝑡−1 + 𝑎𝑜𝑖 𝑄𝑜𝑡−𝐿𝑖 − ∑𝑜∈𝐼𝑐 𝑏𝑜𝑖 𝑄𝑜𝑡 − 𝐼𝑖𝑡 = 0 𝑖 𝑜 ∈ 𝐼𝑐 𝑡 ∈ 𝑇 (16)
𝑄𝑜𝑡 − 𝑀𝑌𝑜𝑡 ≤ 0 𝑜 ∈ 𝐼𝑐 𝑡 ∈ 𝑇 (17)
∑𝑜∈𝐼𝑐 𝑘𝑜𝑟 ∙ 𝑄𝑜𝑡 ≤ 𝐶𝑟 + 𝑤𝑟𝑡 𝑜 ∈ 𝐼𝑐 𝑡 ∈ 𝑇 (18)
𝑌𝑜𝑡 = {0 1}, 𝐼𝑖𝑡 ≥ 0, 𝐵𝑖𝑡 ≥ 0, 𝑄𝑜𝑡 ≥ 0 (19)
The objective function (1) is the expected total cost. Constraints (2) and (3) ensure the balance of flow
for all items in each period. Constraints (4) set the production quantities to zero in periods without
operations. Constraints (5) enforce limits on production capacity.
4. Methodology
The main idea of the digital-twin based methodology is to build the production planning tools, that can
automatically generate a production plan on a mid-term or long-term planning horizon. The production
planner module, consisting of a series of tools, can be used by a production manager to predict the
production demand and generate long-term and middle-term plans according to the existing demand and
demand forecast. Figure 1 presents the structure of the production planner module, the function and
relationship between these modules are introduced as follows:
Fig. 17 Digital-twin based production planning tools

Probabilistic forecast tool, described is used to predict the distribution for the uncertain parameters,
including the demand and lead time.
For demand, the forecasts in practice are always wrong, and the deviation of the actual value of the
parameters from the forecast may have a large impact on the quality of the production plan. Therefore,
rather than providing a point forecast, the proposed the probability learner estimates the probability
distribution of uncertain demand. The tools take as input historical data in the form of a time series, and
it outputs the probability distribution of the parameter. The user may choose between several approaches
to estimate the probability.
For uncertain lead time, we create a Bayesian network using the data from the domain model or
simulation model. The Bayesian network is built from the relations in the domain model, and we learn
the conditional probability with the pair copula. We compare two methods to optimize the lead time. One
is just integrating the simulation model with the production planning model to improve the calculation
method of the lead time.
Page 568
SEMIT 2023
Another one is integrating the simulation model and Bayesian network with the production planning
model to improve the calculation method of the lead time. The production planning module provides
different production plans as the input of the simulation model. By running the simulation model with
the production plans, we can get the real lead time and planning adherence. Furthermore, we can use the
feedback data to train the Bayesian Network model for predicting the probability distribution of lead time.
Production planning solver is the main tool that is used to generate mid-term or long-term production
plans. First, according to the determined data, we build the general mathematical model for production
planning. To consider uncertain demand with distribution, we use Quasi-Monte Carlo methods to
generate the scenario tree for uncertain demand. Then, according to the scenario tree, we can get the
customized mathematical model, which is the scenario-based mathematical model for the uncertain
demand. In the end, we build the basic heuristic algorithm of fix-and-optimize to get the solution of
production planning.
References
15. Galbraith, Jay. Designing complex organizations. Reading, Mass (1973).
16. Sethi, S. P., Yan, H., Zhang, H., & Zhang, Q. (2002). Optimal and hierarchical controls in dynamic
stochastic manufacturing systems: A survey. Manufacturing & Service Operations
Management, 4(2), 133-170.
17. Yano, Candace Arai, and Hau L. Lee. Lot sizing with random yields: A review. Operations
research 43.2 (1995): 311-334.
18. Aouam, T., Geryl, K., Kumar, K., & Brahimi, N. (2018). Production planning with order acceptance
and demand uncertainty. Computers & Operations Research, 91, 145-159.
19. Higle, Julia L., and Karl G. Kempf. Production planning under supply and demand uncertainty: A
stochastic programming approach. Stochastic Programming: The State of the Art In Honor of George
B. Dantzig (2011): 297-315.
20. Hnaien, Faicel, Alexandre Dolgui, and Desheng Dash Wu. Single-period inventory model for one-
level assembly system with stochastic lead times and demand. International Journal of Production
Research 54.1 (2016): 186-203.
21. Aghezzaf, El-Houssaine, Carles Sitompul, and Najib M. Najid. Models for robust tactical planning
in multi-stage production systems with uncertain demands. Computers & Operations Research 37.5
(2010): 880-889.
22. Dolgui, Alexandre, and Caroline Prodhon. Supply planning under uncertainties in MRP
environments: A state of the art. Annual reviews in control 31.2 (2007): 269-279.
23. Luo, Dan, Simon Thevenin, and Alexandre Dolgui. (2021, September). A Digital Twin-Driven
Methodology for Material Resource Planning Under Uncertainties. In IFIP International Conference
on Advances in Production Management Systems (pp. 321-329). Springer, Cham.
Page 569
SEMIT 2023
Sustainable Aviation Fuel Production and Aviation Sector

Ibrahim Temam Ibrahim, Ali Osman Kuşakcı, Amna Abdullah
Ibn e Haldun University, Turkey
Abstract
Sustainable aviation fuels (SAF) present a feasible solution to decarbonize modern aviation. Unlike
traditional jet fuels, SAFs are produced in a variety of ways, thereby choosing one of these processes
as a complicated Multi-Criteria Decision challenge that involves conflicting priorities. This study
evaluates SAF production processes using a multicriteria methodology, PROMETHEE-2. With SAF
technology in its nascent stage and limited data, several stakeholders in the aviation sector were
enlisted to assist in the collection of data and preferences. The suggested framework’s strength lies in
its adaptability to suit the subjective opinions of diverse stakeholders, selection of ranking system, and
robustness of outcomes. This research engaged stakeholders in a participative manner to rank 11 (A1
to A11) SAF production paths based on 24 parameters categorized into social, environmental,
economic, and technological evaluation criteria. Industry professionals were given a form to rate SAF
production methods according to a performance criterion. Data is validated using fuzzy TOPSIS and
fuzzy VIKOR and PROMETHEE-II to reduce professionals’ judgmental personal prejudice. Results
indicate the optimal feedstock for SAF production is the direct transition process of CO2 to SAF (A11)
in the gasification or Fischer-T synthesis group.
Keywords: Aviation, Green fuel, Sustainable Aviation Fuel, PROMETHEE-2.
1. Introduction
The air transportation sector has grown exponentially in recent years, contributing to substantial
economic benefits and increased employment worldwide (IATA, 2019). However, this boon also
brings about ecological concerns, primarily due to greenhouse gas emissions (GHG), of which carbon
dioxide (CO2) holds significant responsibility, accounting for approximately 53% of the overall
greenhouse gas effect. Even though the airline industry is responsible for just 2%-2.6% of annual
global CO2 emissions, the ramifications of these emissions disproportionately impact the environment
(ICAO, 2016; Kivits et al., 2010).
The International Air Transport Association (IATA) foresees that by 2035, global passenger counts
will grow substantially. As a result, by 2050, the airline sector may contribute between 4.6% and
20.2% to global fossil fuel CO2 emissions (O’Connell et al., 2019). Decarbonizing the aviation
industry hinges on technological innovations and operational advancements. Potential solutions
include refined aviation designs, propulsion systems, and considering the potential of electrifying
long-haul planes. However, electric planes might not be commercially viable until around 2050
(Schäfer et al., 2018).
Another promising avenue is developing and utilizing SAF. These fuels play a pivotal role in emission
control and reducing dependency on fossil fuels (The Royal Society, 2019). Given the lack of technical
preparedness in alternative options, liquid fuels will remain to serve a central role in flight operations.
Therefore, the utilization of SAF is viewed as a moderate strategic plan for controlling emissions and
lowering reliance on fossil fuels. Thus far, the production of SAF poses significant challenges;
ranging from technological hindrances, public perceptions, and environmental implications of
production and distribution, to economic barriers (Bann et al., 2017; Wang et al., 2019). Despite the
crucial role of SAF in the future of aviation, current literature indicates a lack of research focusing on
SAF production techniques (Dožić, 2019). Addressing this gap, the present study introduces a
participative model based on a multi-criteria decision framework.
Page 570
SEMIT 2023
This model incorporates insights from aviation industry experts regarding low-carbon aviation fuel
production. Merging a comprehensive literature review and professionals' perspectives, the study
identifies performance criteria across social, environmental, economic, and technological realms.
Moreover, expert opinions were harnessed to ascertain each criterion's relative importance and rank
alternative SAF production methods. Employing this data, the study simulated professional
preferences to prioritize SAF production pathways. Notably, the research also juxtaposed the multi-
criteria ranking results using various Multi-Criteria Decision Making (MCDM) techniques.
The remaining of this study is organized as follows. Section 2 provides a literature review on SAF
while Section 3 addresses briefly different production methods. Next, the research methodology is
introduced in Section 4. Section 5 presents the main findings of the research and provides an analysis
of the results. Finally, Section 6 culminates with conclusive remarks.
The literature review is structured according to research examining societal, environmental,
economic, and technological aspects. Chiaramonti, (2019) underlined the challenges and prospects
for sustainable fuel adoption while. Kivits et al., (2010) studied the significance of airports in aiding
cleaner aviation. Wang et al., (2019) examined the production impacts of airline fuels in a Brazilian
economic system and found favorable socioeconomic outcomes for employment and Gross domestic
product on balance. O’Connell et al., (2019) prioritized energy efficiency and reduced greenhouse
gas (GHG) emissions as the two important factors and found that some Sustainable aviation fuel
conversion methods are more energy-intensive than others, excluding those that utilize trash and
leftovers as feedstock.
Moreover, Staples et al., (2018) examined emission reductions resulting from the manufacture of SAF
using non-food feedstocks via several conversion methods across their whole life cycle. It revealed
that annual investments of approx. 12B USD will be required to reduce it by at least 50% by 2050.
Besides, Ganguly et al., (2018) performed a Life Cycle Assessment (LCA) on a wood-related
feedstock for SAF manufacturing. It was discovered that this type of SAF can reduce global warming
by 78% as compared to traditional aviation fuel. Seber et al., (2014) utilized LCA to determine the
GHG discharge reductions and manufacturing expenses related to Hydro-processed Esters and Fatty
Acids (HEFA). It was discovered that LCA of green fuel derived from yellow grease created fewer
GHG emissions than sustainable aviation fuel derived from tallow when compared with aviation fuel
derived from petroleum.
Furthermore, Bann et al., (2017) used a financial technique to determine the most suited SAF
production process from an economic standpoint. They employed stochastic modeling focused on a
net current value computation for six alternative approaches to defining the minimal selling price of
sustainable aviation fuels. In addition, it was found that SAF produced using vegetable oil-based
feedstocks is more comparable in price to standard aviation fuel than SAF derived from non-food
feedstocks. Trivedi et al., (2015) found that the gasification or F-T production method is more
financially feasible when compared to the hybrid method by focusing on energy investment return
analysis intended for Fischer Tropsch, Hydro processed ester, and fatty acids, & improved
fermentation procedures utilizing various feedstock.
Neuling & Kaltschmitt, (2018) found that the F-T procedure with switchgrass yielded the highest
energy ROI at 9.8% through a techno-economic assessment, including environmental factors for SAF
production paths, concluding that such operations use the mass and energy balance method.
Achieving technological, economic, and environmental objectives concurrently for a single method
is not possible. Li et al., (2018) found that the production parameters analyzed, the minimum selling
price (MSP) of sustainable aviation fuel ranged from $0.40 to $0.17 per liter, with feedstock
expenditure being the most sensitive factor. Michailos, (2018) also conducted a thorough production
design and cost evaluation of sustainable aviation fuel from sugar cane residual (bagasse). It was
discovered that one kilogram of dry bagasse may create 0.121 kilograms of sustainable aviation fuel
at a rate of $2.78 per liter.
Page 571
SEMIT 2023
Furthermore, Lokesh et al., (2015) explored the effects of sustainable aviation fuel hydrocarbon
composition, thermodynamic characteristics, & fuel combustion on airplane performances studied
three alternative SAFs and compared them to standard aviation fuel, including camelina, jatropha, as
well as microalgae. Hileman & Stratton, (2014) focused on many criteria to evaluate the viability of
various sources of fuels for the airline industry and showed that the fuels obtained from the Fischer
Tropsh & HEFA procedures were shown to be practical for supplementing the existing aviation fuel
supplies. However, hydrogen is deemed impracticable because of airplanes' existing engine
technology. None of the studies employs the multi-criteria decision-making conceptual approach,
which is gradually being introduced into the Life cycle assessment or techno-economic assessment
toolbox as a definitive endpoint approach, notwithstanding the issue reasoning reflecting a multi-
criteria dilemma.
This research intends to address this disparity by presenting an investors' participative Multi criteria
decision structure that addresses the opinions of airline-experienced professionals on lower-carbon
aviation energy production approaches.
3. Sustainable Aviation Fuel Production Pathways

3.1. Aviation Industry at Present
The aviation industry plays a significant role in contributing to global warming, accounting for
approximately 3.5% of global warming when considering non-CO2 emissions and around 2.5% of
global CO2 emissions in 2018 (Ritchie, 2020). Despite its environmental impact, aviation is perceived
as a challenging sector to decarbonize both technologically and economically. In 2019, emissions
from jet fuel combustion reached 1027 Metric tons of CO2, constituting 12% of the CO2 emissions
from the overall transportation industry (Atag.org, 2022; IEA, 2022).
“The UN Climate Change Conference” set a target to limit the average global temperature rise to
below 2˚C above pre-industrial levels, with an aspiration for 1.5˚C (United Nations., 2015).
Governments worldwide have been making efforts to develop strategies to achieve these goals, but
many existing initiatives have fallen short. The transportation sector, including aviation, lags behind
other industries in its decarbonization efforts, and the European Union has pointed out that it is not
on track to meet its climate goals. Criticism has been directed at the aviation sector's short-term carbon
reduction plans as insufficient in the face of the current climate crisis (Reuters., 2021).
ICAO (International Civil Aviation Organization) member nations endorsed the Carbon Offsetting
and Reduction Scheme for International Aviation (CORSIA) in 2016, a market-based framework
aimed at achieving net-zero growth in global aviation emissions from 2020 onwards. However, most
of these plans and initiatives were based on assumptions of significant corporate air traffic growth
and did not account for the sudden disruption caused by the COVID-19 pandemic, resulting in a
historic decline in global aviation passengers (International Civil Aviation Organization, 2021). While
aviation traffic has partially rebounded in 2021, it remains lower than pre-pandemic levels.
The pandemic's impact on aviation is considered temporary, and long-term projections suggest that
passenger traffic in 2050 may still be 8% lower than pre-pandemic estimates. Despite this, the need
for mitigation strategies remains, and some existing carbon reduction strategies have had to adapt to
the changing circumstances (International Civil Aviation Organization., 2022). Various strategies can
reduce the aviation sector's greenhouse gas (GHG) emissions, but their effectiveness varies.
Optimizing flight routes for minimal emissions, improving maintenance schedules and other
operational initiatives, and developing new weight management resources and related technologies
are all important contributors to emissions reduction. However, these efforts alone are insufficient
given the growing aviation traffic and emissions (Timperley, 2021).
While the current generation of aircraft is more energy-efficient than previous generations, the rate of
improvement is slowing, and it is unlikely to keep up with the expected growth in aviation traffic
(International Civil Aviation Organization., 2019). Additionally, emissions during the takeoff and
climb phases, as well as ignition and warm-up emissions, are significant. Some argue for halting
Page 572
SEMIT 2023
aviation industry expansion, but this is impractical given industry competition. To achieve meaningful
emissions reductions, commercial aviation must reduce its reliance on fossil fuels (Alkema, B., 2022;
Fortier et al., 2014).
Hydrogen-powered and battery-powered aircraft are potential alternatives for the future, but they
require substantial changes to aircraft design and face delays in new deliveries and high short-term
costs (Scheelhaase et al., 2019). For short and medium-range flights, "drop-in" SAF offer a viable
alternative to traditional jet kerosene. Various pathways to produce SAF have been licensed for
commercial use, with bioactive feedstocks being the most common. However, sustainability remains
a challenge for larger-scale operations, and power-to-liquid options may be a more suitable solution,
albeit at higher costs (aviationbenefits.org, 2022).
3.2. Green Fuels Used in the Aviation Industry
SAF refer to drop-in kerosene alternatives made from various feedstocks. The ICAO defines SAF as
alternative fuels that meet specific sustainability criteria (ICAO, 2022). Although SAF research
initially focused on economic and supply concerns, it has gained momentum due to environmental
factors. By 2019, over 45 airlines had flown more than a quarter million SAF flights, albeit
representing less than 0.1% of total jet fuel usage (IEA, 2022). One challenge for SAF adoption is the
existing restrictions on blending ratios. SAF can significantly reduce CO2 emissions and other
pollutants, but blending ratios are typically limited to 5-50%, depending on the production process.
Early SAF formulations did not meet the aromatic component requirements, but recent advancements
in SAF transfer techniques and O-ring technology may lead to regulatory changes. United Airlines
successfully operated an aircraft with 100% SAF in 2021 (Palmer, 2021).
3.2.1. Certification Procedure for Sustainable Aviation Fuels
SAF must meet the same characteristics as conventional jet fuel, ensuring that it is "drop-in ready."
The global standard for classifying kerosene fuel in commercial aviation is the American Society for
Testing and Materials (ASTM) D1655, which applies to both Jet A and A-1 fuels. Jet A-1 is preferred
for long-distance flights using arctic routes due to its low freezing point, but it comes at a higher
production cost (Chevron Products Company, 2004). ASTM D45054 established a three-stage, four-
tiered testing method, which is a mandatory procedure for qualifying and certification of new airline
turbine fuels and extracts, which is required for a novel aircraft fuel to be licensed for business use.
In summary, the aviation industry faces significant challenges in reducing its greenhouse gas
emissions, and various strategies are being explored. SAF represent a promising option, but there are
obstacles to their widespread adoption, including blending restrictions. Certification processes like
ASTM D4054 are essential to ensure the quality and safety of SAF for aviation use.
3.2.2. Biofuels Utilized in the Aviation Sector
SAF must meet stringent criteria and undergo certification processes before they can be used in
commercial aviation. ASTM D4054 outlines a comprehensive testing procedure that requires various
fuel quantities at each level. This process, lasting 3 to 5 years, imposes a minimum cost of $5 million
on prospective fuel manufacturers (Heyne et al., 2021). To accelerate research and development,
ASTM introduced a fast-track annex for D4054 in January 2020, although SAFs certified through this
method are limited to a blending ratio of 10% (US Dept. Of Energy, 2020).
3.3. Renewable Feedstocks for Biofuels
Biofuels can be produced from renewable feedstocks, which are carbon-based materials capable of
renewal. The production of biofuels from plants involves processes that reduce the carbon life-cycle
emissions, as some of the CO2 released during production is absorbed by the next crop generation. A
comparison of carbon life-cycle emissions between fossil-based aviation fuel and bio SAF is shown
in Figure 1. Some frequent feedstocks used for biofuels are listed below.
Page 573
SEMIT 2023
Figure 1. LCA of carbon for fossil aviation fuels (b) sustainable biofuels (VDOCUMENT, 2017)
Camelina: Camelina sativa is a non-edible fuel crop with a high fatty oil content, making it a valuable
raw material for biofuel production. The byproduct of oil extraction can also be used as supplemental
feed for livestock. Camelina is employed as a cover crop, contributing to soil quality improvement
and reducing erosion (World Economic Forum., 2020).
Jatropha: Jatropha Curcas is another potential source of lipid oil for biofuel production. It can thrive
in marginal or unproductive environments, making it a promising candidate for biofuel production.
However, challenges such as low seed production, vulnerability to diseases, and extraction
complexity need to be addressed (Moniruzzaman et al., 2017).
Halophyte: Halophytes are salt-tolerant plants that show potential as feedstocks for biofuels, both oil
and lignocellulosic. They can grow in harsh environments and have adaptability for use in fuel
production. Their growth may also contribute to desalinizing arable land (CAAFI, 2019).
Algae: Algae, often referred to as "algae fuels," are considered third-generation biofuels. They have
rapid growth rates, require less land, and excel in carbon sequestration. Algae can be farmed in
marginal sites and offer significant potential for biofuel production (Chisti, 2007).
Waste Oil: Various waste materials can be converted into SAFs, including residual cooking oil,
livestock fats, tall and fish oils, and other leftover lipids. Forest and agricultural residues, as well as
municipal solid waste, also have potential as eco-friendly fuel sources (O’Malley, J., Pavlenko, N.,
and Searle, 2021).
3.4. Accredited Paths for Biofuel Production
ASTM has validated eight technological systems or conversion methods for producing SAFs, which
are outlined in annexes to standards like D7566. Fischer-Tropsch (FT) synthesis is one such method,
with two variants: FT-SPK and FT-SPK/A. FT-SPK has a maximum allowable mix percentage of
50%. It involves a reactor using a cobalt/iron compound to catalyze a thermochemical reaction,
resulting in synthetic paraffinic kerosene (SPK) that is functionally equivalent to conventional
aviation fuel. FT-SPK/A includes benzene to increase the aromatic content of the final hydrocarbon,
improving compatibility with modern engines (CAAFI, n.d.; SkyNRG., 2020).
3.5. Challenges Faced in Biofuel Adoption
Several challenges hinder the development and widespread use of alternative fuels, including:
• Insufficient Maturity of Fuel Approaches: Many alternative fuel approaches are not fully
mature, limiting their practicality.
• Lack of Raw Material: Adequate quantities of raw materials for fuel production, particularly
for biofuels, can be scarce.
• Sustainability Concerns: The sustainability of feedstock generation, especially for biofuels,
raises environmental and ethical concerns.
• Lack of International Support: The absence of support from international bodies and
governments for transitioning from fossil fuels to alternative fuels hampers progress.
Page 574
SEMIT 2023
These challenges impact the cost-efficiency of SAFs and may hinder the aviation industry's efforts to
reduce emissions and achieve sustainability goals. While SAFs have made progress in recent years,
achieving ultimate technical ease on demand remains a challenge (Ricardo Energy and Environment.,
2020). In conclusion, biofuels hold promise as a sustainable alternative to traditional aviation fuels,
with various feedstocks and production methods under consideration. However, significant
challenges, such as feedstock availability and sustainability concerns, must be addressed to ensure the
successful adoption of these fuels in the aviation industry.
3.6. Approaches to Producing Sustainable Aviation Fuels (SAFs)
SAF can be produced through various methods, each with its unique technological, economic,
societal, and environmental characteristics. These methods can be broadly categorized into
biochemical and thermochemical routes, as depicted in Figure 2.
HEFA is a process that produces long-chain hydrocarbons through hydrogenation and isomerization
of oils, including animal fat, vegetable oil, residual grease, and algae oil. The resulting SAFs have
high energy content, good thermal stability, and lower tailpipe emissions. ASTM has approved HEFA
fuels for use in blends with regular aviation fuel, with a maximum blending limit of 50%. Several
airlines, such as KLM, Lufthansa, and Etihad, have conducted successful test flights using HEFA-
based mixtures (Dayton & Foust, 2020).
Advanced fermentation includes two processes.
Alcohols To Jet (ATJ): This process produces long-chain hydrocarbons similar to traditional Jet-A1
fuel from alcohols like ethanol, methanol, or n-butanol. The alcohol can be generated from sugar
fermentation, catalytic transformation of biomasses, or direct carbohydrate conversion into
hydrocarbons. The ATJ process consists of four steps: dehydration, oligomerization, hydrogenation,
and distillation. It has reached an advanced stage of development but involves higher infrastructure
costs compared to Fischer-Tropsch (F-T)(Geleynse et al., 2020). Direct Sugars to Hydro-Carbon
(DSHC) does not require alcohol generation. Concentrated carbohydrates are converted into
hydrocarbons using anaerobic fermentation. The process includes phase separation to produce
aviation fuel (Zhang, Wang, et al., 2020).
Figure 2 Process classification on SAF production paths (Shahabuddin et al., 2020)

Fischer-Tropsch (F-T) production involves the conversion of syngas (carbon monoxide and
hydrogen) from carbon-rich biomass into liquid fuel through a catalytic process. Various biomass
resources, such as wood wastes, poplar, willow, and agricultural wastes like corn stover and wheat
Page 575
SEMIT 2023
straws, can be used to create carbon-free aviation fuel. The F-T process requires high temperatures
and pressures, making it a costly method (Dayton & Foust, 2020).
Pyrolysis produces gas, charcoal, and bio-oil as its primary products. Bio-oil contains a wide range
of organic molecules but has a high oxygen concentration. To make it suitable for use as aviation fuel,
oxygen molecules must be removed through hydrotreating in the presence of high-pressure hydrogen.
After purification, a variety of hydrocarbons, including aviation fuels, can be obtained (Chen et al.,
2020).
HTL (Hydrothermal Liquefaction) employs high temperatures and pressurized water to convert
biomass into liquid fuel sources. Feedstocks for HTL include algae, composts, wastewater sludge,
and lignocellulosic materials like corn stover. The resulting bio-crude has a high oxygen
concentration, which is eliminated during hydrotreating. Distillation is then used to separate the
refined hydrocarbons into aviation fuel, diesel, and other byproducts (Castello et al., 2019).
Co-Valorization of Carbon Dioxide and Waste Biomass approach, waste biomass is gasified to
produce syngas rich in carbon monoxide (CO), while syngas rich in hydrogen (H2) is generated
through the co-electrolysis of CO2 and water. These gases serve as dual feedstocks for Fischer-
Tropsch (F-T) processes (Zhang, Fang, et al., 2020).
3.3. Policy and Support for SAFs
To encourage the transition from fossil fuels to renewable energy sources, policies and financial
assistance systems play a crucial role. Examples include newly authorized tax breaks in the United
States and direct project finance. Legislation, like the European Commission's Refuel EU Aviation
directive, sets individual objectives for sustainable aviation fuel integration in European Union
airports. Such policies set targets for the incorporation of biofuels and e-fuels in aviation fuel mixes,
providing a regulatory framework for SAF adoption (Dyk, 2021).
In conclusion, various methods are employed to produce sustainable aviation fuels, each with its
advantages and challenges. These approaches are essential for reducing greenhouse gas emissions and
promoting the use of renewable energy sources in the aviation industry.
In this research, the methodology used to evaluate the effectiveness of SAF production methods is
presented. The approach involves an MCDM framework designed to be adaptable and flexible to meet
the needs of various stakeholders in the airline sector. The initial stage comprises a PROMETHEE-
based criteria selection stage, while the next step evaluates the alternatives (see Figure 3).
Page 576
SEMIT 2023
Start
Inputs: Professional's ratings for each alternative & criteria weights
Select a fuzzy modeling approach & calculate an aggregate fuzzy rank and weight for each
criterion and for every alternative.
Calculate aggregate weighted normalized fuzzy rankings.
Fuzzy TOPSIS Fuzzy VIKOR
Calculate fuzzy FPIS & FNIS Defuzzify the weighted normal aggregate fuzzy
ranking to a crisp value
Computing the distances between every alternate in

Calculating the CPIS & CNIS
FPIS & FNIS
Calculating the nearest coefficient for every alternate Calculating the utility (S) & regret (R) measures
Ranking alternate relying on nearest coefficient values Calculate the Q index for every alternative
Ranking the alternate relying on Q values
Compare the rankings
End
Figure 3. The flowchart on the fuzzy multicriteria decision method

The initial phase involved identifying performance standards and measurements for SAF production
techniques. Fifteen different production paths for SAF were found in the literature review.
Additionally, a list of 45 performance indicators related to social, economic, environmental, and
technological aspects of sustainable transport fuels was compiled. To refine the criteria, a value tree
approach integrated into the Delphi technique was employed. Experts and stakeholders participated
in an online survey, and a consensus level was established. The final model included 24 criteria and
11 SAF production paths, which were evaluated based on various parameters. Criterion and codes for
final evaluation are illustrated in Table 1 and sustainable aviation fuel production paths evaluated in
this research are given in Table 2. The references that used the same or similar evaluation criteria are
listed in the Appendix (see Table A2).
Table 1. Final Evaluation of Criterion and Codes
Category Code Criterion Description
Social Socl1 Traceability Transparency of the whole
manufacturing procedure from the raw
materials to finished products.
Socl2 Economical This signifies the establishment of new
contribution businesses, industrial areas, rural
growth, etc.
Socl3 Food safety The effect of feedstock utilized to
produce green fuel on food availability.
Socl4 Social Public opinion towards SAF production
acceptance and utilization.
Environmental Envt1 Sustainability It signifies the continued availability of
on feedstock feedstock for green fuel production.
Page 577
SEMIT 2023
Envt2 Savings on CO2 discharge levels compared to jet

GHG fuels.
emissions
Envt3 The impact SAF production changes land use
caused by a directly and indirectly.
change in land
usage
Envt4 Water and Soil Impact of using fertilizers and
pollution pesticides to produce biomass.
Economic Econ1 Alternative Other potential feedstock applications
uses with besides SAF production like electricity
feedstock and biomethane.
Econ2 Profitability of Monetary benefits in generating a
feedstock particular feedstock.
Econ3 Minimal price Minimum selling price estimated for
for selling the SAF.
Econ4 Input energy Energy consumption during SAF
usage manufacture.
Econ5 Productivity of Incorporation of short crop rotations or
Land intensive farming methods.
Econ6 Operational Operational and maintenance costs
and related to the SAF production plant.
maintenance
costs
Econ7 Cost on Costs of acquiring raw materials
feedstock required for manufacturing of fuel.
Econ8 Cost of plant Construction-related costs of
capital production and associated services.
Technical Tech1 Blending Amount of authorized alternative fuel
limits that can be blended with standard jet
fuels.
Tech2 Compatibility Fuel characteristics, such as flashpoint,
with Standard viscosity, density, and energy content,
jet fuels closely resemble those of conventional
jet fuel.
Tech3 Local The availability of locally accessible
technical production technologies.
capability
Tech4 Integration The capability of a production method
process to be integrated with established jet-fuel
refinery infrastructures.
Tech5 Maturity of The current stage of development for a
process workflow is either initial, demo, or
technology commercialized.
Tech6 Method yield Quantity of sustainable aviation fuel
acquired via conversion method.
Tech7 Scalability on Capability for future expansion of
production sustainable aviation fuel processing
volume facilities
Tech8 Composition Quality of SAF batch
and quality of
feedstock
Page 578
SEMIT 2023
Table 2. Sustainable Aviation Fuel Production Paths Evaluated in this Research
Production method Feedstocks Code
Algae and microalgae A1

Hydro processed Esters
Used cooking oils and animal fats A2
and Fatty Acids (HEFA)
Oilseeds A3
Municipal solid wastes A4

Fischer-Tropsch
synthesis or gasification
Wood residues and agricultural wastes A5

Devolatilization or
Pyrolysis

Hydrothermal
liquefaction (HTL)
Advanced fermentation Wood residues and agricultural wastes A 10
CO2 co-valorization and Combination of industrial waste gases

A 11
biomass waste CO2 and wood residues
To understand the relative importance of criteria, interviews were conducted with 22 European
aviation fuel supply chain experts. The summary of the expert pool was given in Table A1. These
experts, each with more than five years of experience, ranked four significant criteria (social,
environmental, economic, and technological) concerning one another. This ranking system provided
a total score out of 100 for each criterion.
Airline industry professionals were surveyed to rate SAF production paths based on different criteria.
Due to the limited availability of data, a numerical scale was introduced, ranging from 1 to 9, with
corresponding descriptions like extremely poor, poor, average, good, and excellent. These values
were then converted to a score out of 100. The PROMETHEE-II approach (Brans & Vincke, 1985),
known for handling both qualitative and quantitative data, was chosen for the analysis. This approach
helps address conflicting objectives and trade-offs, making it suitable for decision problems involving
multiple criteria. PROMETHEE-II approach is especially useful when there are conflicting objectives
and trade-offs to consider, making it applicable to a wide range of decision problems. It can help
decision-makers better understand the relationships between criteria and alternatives, This means that
subjective judgments and expert opinions can be integrated into the model.
PROMETHEE-II can be conceptualized as a modeling and key paradigm for MCDA difficulties in
which the difficulty in deciding is modeled as a fully connected system where nodes are alternative
and the arc reflects preferences relations among node pairs or alternatives (𝑎, 𝑏), like 𝜋(𝑎, 𝑏).
The alternative or node’s strength 𝑎, which is called the net outranking flow, is written as ∅(𝑎) or
∅𝑛𝑒𝑡 (𝑎), and it is calculated by deducting outward outranking flow ∅+ (𝑎) from inward outranking
flow ∅+ (𝑎) i.e., ∅𝑛𝑒𝑡 (𝑎) = ∅+ (𝑎) − ∅− (𝑎). The answer to the question of rating a group of
alternatives according to the MCDM approach is provided by arranging alternatives as per their
strength in descending order as calculated through net outranking flows, ∅𝑠 . Following is a breakdown
of the method in detail.
Page 579
SEMIT 2023
Input: Weighting scheme and the decision matrix

In PROMETHEE-2, the data is integrated into a decision matrix as depicted in Eq. 1 below.
DM = [𝑐1 (𝑎1 ) ⋯ 𝑐𝑛 (𝑎1 ) ⋮ ⋱ ⋮ 𝑐1 (𝑎𝑚 ) ⋯ 𝑐𝑛 (𝑎𝑚 ) ] (1)
where the alternative set is denoted as A={𝑎1 , 𝑎2 , …, 𝑎𝑖 , …, 𝑎𝑚 }, and performance criteria set is
considered as C={𝑐1 , 𝑐2 , …, 𝑐𝑗 , …, 𝑐𝑛 }. The alternative 𝑎𝑖 ′𝑠 𝑝𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒 on criterion j is indicated
by 𝑐𝑗 (𝑎1 ) where; i = 1, ...., m, and j = 1, …., n. After that, we obtain the comparative importance
weight (𝑤𝑗 ), of nth performance criterion. 𝑤𝑗 = (𝑤1 , 𝑤2 , …, 𝑤𝑛 ), where j= 1, …, n, it includes non-
negative weights with ∑𝑛𝑗=1 𝑤𝑗 = 1.
Step 1: For every alternative pair (𝑎, 𝑏) on a criteria j, we calculate the difference of opinion in criterion
j performance concerning alternative 𝑎 and 𝑏, represented as 𝐷𝑗 (𝑎, 𝑏), (Eq. 2):
𝐷𝑗 (𝑎, 𝑏) = 𝑐𝑗 (𝑎) − 𝑐𝑗 (𝑏), 𝑗 = 1, … . , 𝑛∀𝑎, 𝑏 ∈ 𝐴 (2)
Step 2: To find every alternative pair (𝑎, 𝑏) for criterion j, the method calculates a local preference
index 𝑃𝑗 (𝑎, 𝑏), that considers the difference of opinion in performance of criterion j concerning
alternative 𝑎 and 𝑏 that is calculated previously using step 1. Brans and Vince (1985) highlighted the
different preference index functions. This study adopts the type 1 general preference as given in Eq. 3.
𝑃(𝐷) = {0, 𝐷 ≤ 0 1, 𝐷 > 0 (3)
Output: The concluding rating of each of the alternatives is determined based on the net ranking,
∅𝑛𝑒𝑡 (. ). That is, the higher the net ranking, the more desirable the alternative is.
To compare the results of the ranking, this study utilized two alternative MCDM methodologies:
TOPSIS and VIKOR. To account for the unpredictability of professional preferences, the research
adopts the fuzzy version of the TOPSIS and VIKOR approaches (Awasthi et al., 2018). A summary
of detailed implementation information for the formulations of fuzzy TOPSIS and VIKOR analysis
is discussed here.
𝑘
Inputs: The ranking done by the K professionals of 𝑚 alternatives with 𝑛 criteria is denoted with 𝑥𝑖𝑗 ;
where 𝑖 = 1, … , 𝑚, and 𝑗 = 1, … , 𝑛, and 𝑘 = 1, … , 𝐾 and every professional’s given weight for every
criterion is 𝑤𝑗𝑘 ; where 𝑗 = 1, … , 𝑛, 𝑎𝑛𝑑 𝑘 = 1, … , 𝐾.
Step 1: Choosing a proper modelling context and computing the aggregated fuzzy rankings for every
alternative and weights on every criterion. Utilizing fuzzy triangular numbers to develop the fuzzy
data is easier and more common, whereas this study relates to the rankings. Hence, for every
alternative "i" (i = 1, … , m) and criteria "j" (j = 1, … , n), the rankings by K professionals are
combined to singular triangle fuzzy rankings, assume x̃ij = (aij , bij , cij ), like:
𝑘 1 𝑘 𝑘
𝑎𝑖𝑗 = 𝑚𝑖𝑛{𝑥𝑖𝑗 }; 𝑏𝑖𝑗 = ∑𝑘 𝑥𝑖𝑗 ; 𝑐𝑖𝑗 = 𝑚𝑎𝑥 {𝑥𝑖𝑗 } (4)
𝑘 𝐾 𝑘
Likewise, compute the aggregated fuzzy weights for all criteria j (j = 1, … , n), assume w
̃ j,1 = wj,1 ,
wj,2 , wj,3 , like:
1
𝑤𝑖,1 = 𝑚𝑖𝑛{𝑤𝑗𝑘 }; 𝑤𝑗,2 = 𝐾
∑𝑘 𝑤𝑗𝑘 ; 𝑤𝑗,3 = 𝑚𝑎𝑥 {𝑤𝑗𝑘 } (5)
𝑘 𝑘
Summarize aggregate fuzzy rankings as a decision matrix and fuzzy weight as vector criterion.
Step 2: Computing weighted normal aggregate fuzzy rankings, assume ñij where i = 1, … , m and j =
1, … , n, like:
𝑎
𝑖𝑗 𝑖𝑗𝑏 𝑖𝑗 𝑐 𝑖𝑗 𝑖𝑗 𝑎 𝑖𝑗 𝑏 𝑐
𝑛̃𝑖𝑗 = 𝑤
̃𝑗 ⊗ (𝑚𝑎𝑥 , 𝑚𝑎𝑥 , 𝑚𝑎𝑥 ) = (𝑤𝑖,1 𝑚𝑎𝑥 , 𝑤𝑖,2 𝑚𝑎𝑥 , 𝑤𝑖,3 𝑚𝑎𝑥); where, 𝑖 = 1, … , 𝑚 and 𝑗 ∈ 𝐶 + (6)
𝑖 𝑖 𝑖 𝑖 𝑖 𝑖
𝑚𝑖𝑛 𝑚𝑖𝑛 𝑚𝑖𝑛 𝑚𝑖𝑛 𝑚𝑖𝑛 𝑚𝑖𝑛

̃𝑗 ⊗ ( 𝑎𝑖 ,
𝑛̃𝑖𝑗 = 𝑤 𝑖
𝑏𝑖𝑗
, 𝑖
𝑐𝑖𝑗
) = (𝑤𝑖,1 𝑖
𝑎𝑖𝑗
, 𝑤𝑖,2 𝑖
𝑏𝑖𝑗
, 𝑤𝑖,3 𝑖
𝑐𝑖𝑗
); where, 𝑖 = 1, … , 𝑚 and 𝑗 ∈ 𝐶 − (7)
𝑖𝑗
Page 580
SEMIT 2023
where C − (respective C + ) indicates the cost criterion set (respective benefit criterion) in which low
values (respective high values) are better. Based on the values obtained at this step, fuzzy TOPSIS
follows the steps below.
Step 3: Computing the Fuzzy positive and Negative ideal solutions (FPIS and FNIS) accordingly
assuming ñ+ and ñ− , like:
𝑛̃𝑗+= { 𝑚𝑖𝑛 𝑛̃𝑖𝑗 𝐼𝐹 𝑗 ∈ 𝐶 − 𝑚𝑎𝑥 𝑛̃𝑖𝑗 𝐼𝐹 𝑗 ∈ 𝐶 + ; 𝑗 = 1, … , 𝑛 (8)
𝑖=1,…,𝑚 𝑖=1,…,𝑚
𝑛̃𝑗−= { 𝑚𝑎𝑥 𝑛̃𝑖𝑗 𝐼𝐹 𝑗 ∈ 𝐶 − 𝑚𝑖𝑥𝑖=1,…,𝑚 𝑛̃𝑖𝑗 𝐼𝐹 𝑗 ∈ 𝐶 + ; 𝑗 = 1, … , 𝑛 (9)

𝑖=1,…,𝑚
Step 4: Computing the distances i.e., d(i, ñ+ ) and d(i, ñ− ) between every alternate i in FPIS (ñ+ )and
FNIS (ñ− ). Where, i = 1, … , m:
1
1 2 2
𝑑𝑖+ = {𝑛 ∑𝑗 (𝑛̃𝑖𝑗 − 𝑛̃𝑗+ ) } (10)
1
1 2 2
𝑑𝑖− = {𝑛 ∑𝑗 (𝑛̃𝑖𝑗 − 𝑛̃𝑗− ) } (11)
Step 5: Choosing the similar scores, which is also known to be the closest coefficient for every
alternate i (i = 1, … , m) as shown below:
𝑆𝑖− = 𝑑(𝑖, 𝑛̃− )/(𝑑(𝑖, 𝑛̃− ) + 𝑑(𝑖, 𝑛̃+ )) (12)
Step 6: Ranking the alternate in ascending order concerning their similarities in the ranks. Hence, the
effective alternate method is distanced from FNIS and nearest to the FPIS.
Alternatively, fuzzy VIKOR determines the optimal ranking based on the following procedure, where
Steps-3-6 are redefined as follows.
Step 3: Defuzzification of the weighted normal aggregate fuzzy rankings, i.e., ñij (i = 1, … , m; j =
1, … , n) as crisp value by considering nij , as shown below:
1 𝑖𝑗 𝑎 𝑖𝑗 𝑖𝑗 𝑏 𝑐
𝑛𝑖𝑗 =
6
(𝑤𝑗,1 𝑚𝑎𝑥 , 4𝑤𝑗,2 𝑚𝑎𝑥 , 𝑤𝑗,3 𝑚𝑎𝑥); 𝑤ℎ𝑒𝑟𝑒, 𝑖 = 1, … , 𝑚; 𝑎𝑛𝑑 𝑗 ∈ 𝐶 + (13)
𝑖 𝑖 𝑖
1 𝑚𝑖𝑛 𝑚𝑖𝑛 𝑚𝑖𝑛

𝑛𝑖𝑗 = (𝑤𝑗,1 𝑖
, 𝑤𝑗,2 𝑖
, 𝑤𝑗,3 𝑖
); 𝑤ℎ𝑒𝑟𝑒, 𝑖 = 1, … , 𝑚 𝑎𝑛𝑑 𝑗 ∈ 𝐶 − (14)
6 𝑎𝑖𝑗 𝑏𝑖𝑗 𝑐𝑖𝑗
Step 4: Computing the crisp positive ideal solution (CPIS) considering 𝑛+ Moreover, calculating the
rating as shown below:
𝑛𝑗+= { 𝑚𝑖𝑛 𝑛𝑖𝑗 𝐼𝐹 𝑗 ∈ 𝐶 − 𝑚𝑎𝑥 𝑛𝑖𝑗 𝐼𝐹 𝑗 ∈ 𝐶 + ; 𝑤ℎ𝑒𝑟𝑒, 𝑗 = 1, … , 𝑛 (15)
𝑖=1,…,𝑚 𝑖=1,…,𝑚
Step 5: Calculate the performance marks i.e., 𝑄𝑖 for every alternate 𝑖, 𝑤ℎ𝑒𝑟𝑒; (𝑖 = 1, … , 𝑚):𝑄𝑖 =
𝑆𝑖 − 𝑆 + 𝑅𝑖 − 𝑅 +
𝛼( ) + (1- 𝛼) ( );0 ≤𝛼≤1
𝑆 − −𝑆 + 𝑅 − −𝑅+
where,
𝑆𝑖 = ∑𝑚
𝑗=1 (𝑛𝑗+ − 𝑛𝑖𝑗 ); 𝑅𝑖 = 𝑚𝑎𝑥 {(𝑛𝑗+ − 𝑛𝑖𝑗 )} (16)
𝑗
𝑆 + = 𝑚𝑖𝑛𝑆𝑖 ; 𝑆 − = 𝑚𝑎𝑥 𝑆𝑖 ; 𝑅 + = 𝑚𝑖𝑛𝑅𝑖 ; 𝑅 − = 𝑚𝑎𝑥 𝑅𝑖 (17)

𝑖 𝑖 𝑖 𝑖
Consider α = 0.5
Step 6: Ranking the alternate methods according to their performance in descending order. Hence, the
superior alternate method holds the lowest value.
Page 581
SEMIT 2023
5. Research Findings and Analysis

This section discusses the weights assigned to criteria, ranks based on broader implications, and
PROMETHEE-II ratings. After gathering expert opinions, geometric means are used to derive criteria
weights. Economic and environmental criteria carry the most weight, at 31% and 28%, respectively.
Technological criteria account for 25%, while social criteria are the least significant at 16%. Local
weights highlight the importance of specific criteria within each category. For example, food security
and social acceptance are crucial in societal impact, with weights of 29.8% and 26.9%, respectively.
Among environmental concerns, reducing greenhouse gas emissions is a top priority at 36.6% (see
Figure 4).
Composition and quality of feedstock 2.60%

Scalability on production volume 3.00%
Process yield 3.40%
Technical
Maturity of process technology 3.40%

Integration process 3.60%
Local technical capability 2.20%
Compatibility with Standard jet fuels 4.60%
Blending limits 2.20%
Cost on plant capital 4.50%
Cost on feedstock 3.70%
Operational & maintenance costs 2.80%
Economical
Productivity of Land 2.50%

Input energy usage 3.70%
Minimum price for selling 5.00%
Profitability of feedstock 2.60%
Alternative uses with feedstock 3.30%
Environmental
Water and Soil pollution 5.40%

Impact cause by change in Land-usage 5.80%
Savings on GHG emissions 11.30%
Sustainability on feedstock 8.50%
Social acceptance 4.30%
Food safety
Social
4.80%
Economical contribution 3.80%
Traceability 3.20%
0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00%
Figure 4. Global weight of selected criteria Source
Principal Components Analysis (PCA) yielded a 2D GAIA plot (see Figure 5), showing relationships
and conflicts between criteria. This plot helped identify criteria with competing values, such as the
profitability of feedstock conflicting with minimum selling price and capital costs. Utilizing the GAIA
model, disagreement, commonalities, and interdependencies between criteria are evaluated. As the
length of a criterion vector increases, it becomes more discriminatory. In this instance, a 24-D
criterion is predicted, with 67.1 percent.
This shows that the information offered by the group of professionals is reliable. Criteria indicating
competing values can be seen in Figure 5. For instance, the profitability of feedstock (Econ2) and
productivity of land (Econ5) are incompatible with alternate feedstock usage (Econ1), Minimum
selling price (Econ3), and capital costs of the plant (Econ8). Identical preferences are indicated by
vectors pointing toward the same direction.
Page 582
SEMIT 2023
Sustainability of feedstock (Envt1) and Greenhouse gas emission saving (Envt2) from environmental
effect, whereas standard aviation fuel compatibility (Tech2), local technical capabilities (Tech3), and
integration process (Tech4) have similar views. The most important factors to consider are food
security (Socl3), the effect of land use changes (Envt3), the operating costs and maintenance (Econ6),
as well as the level of technological maturation in the procedure (Tech5). Tech5 is the only
differentiating criterion on the decision axis along the vector line (Red).
Figure 5. Graphical Representation of GAIA Plan for Selected Criteria
5.1. Evaluation Based on Impacting Criteria

SAF production methods have been evaluated based on stakeholders' perspectives, including social,
environmental, economic, and technical impacts. Initially, utilizing weight vectors of 0.200, 0.233,
0.289, and 0.269 for social criterion, the study ranked SAF production methods as illustrated in Figure
6.
Gasification and Fischer Tropsch synthesis methods (A4 and A11) ranked highest from a social
standpoint, as they use wood residue and municipal solid waste as feedstock, which is widely
accepted. From an environmental perspective, A11 demonstrated the best performance due to its use
of confined CO2, followed by A5 and A8. A6, A3, and A9 ranked poorly due to environmental
drawbacks associated with algae and oilseed production.
Economically, chemical mixing procedures (A4, A5, and A11) were top-rated, while A6 and A7 were
middle ranked. HEFA methods (A2 and A3) showed promise due to their lower costs compared to
other alternatives. From a technical viewpoint, A5 was the best due to its high SAF yield and
development status, while A1, A9, and A6 lagged.
However, regarding technological ability, A1 is the least desirable of the production routes, following
A9 and A6. Because of the elevated water and oxygen concentration in the biocrude oil generated,
A1 and A6 with algae or microalgae have a reduced sustainable aviation fuel yield. According to
research data, A6 has a superior production method than A9.
Page 583
SEMIT 2023
Social Economic
0.8 0.4
0.6 0.2
0.4
0.2 0
A10 A3 A9 A1 A7 A6 A11 A2 A5 A8 A4
0 -0.2
-0.2 A3 A6 A8 A7 A10 A5 A2 A1 A9 A4 A11
-0.4
-0.4
-0.6 -0.6
Environment Technical
0.6 1
0.4
0.5
0.2
0 0
A9 A6 A3 A2 A1 A7 A4 A10 A8 A5 A11 A1 A9 A6 A7 A2 A8 A3 A4 A10A11 A5
-0.2
-0.5
-0.4
-0.6 -1
Figure 6. Graphical Representation of GAIA Plan for Selected Criteria
5.2. SAF Production Methods and Global Ratings

PROMETHEE-II was used to evaluate SAF production methods based on 24 criteria. A11
consistently ranked the highest, followed by A4 and A5. These methods were favored for their cost-
effective use of feedstock like wood residue and municipal solid waste. Table 3 illustrates the inflow
and outflow in sustainable aviation fuel production methods while Figure 7 shows the ranking of the
alternatives.
Figure 7 shows that Method A6 has the lowest outranking flow of all Sustainable aviation fuel
production methods, with a value of 0.207, and A9 and A3 are the next two extremes, with values of
0.2045 and 0.111 accordingly. The A11 production method has a total outranking flow of 0.240,
greater than all other methods. The fact that the highest-ranked production methods, A4, A5, and A11,
are gasification or Fischer-T synthesis procedures demonstrates that professionals acknowledge that
they are the most appropriate methods for Sustainable aviation fuel production. Syngas, the primary
constituent of gasify or Fischer Tropsch synthesis, is produced from Municipal Solid wastes, wood
residues, and Agri wastes, all of which are inexpensive feedstock. This offers gasification or Fischer-
T synthesis-related production methods a strategic benefit by offsetting its higher capital costs and
smaller production scale in comparison to HEFA methods (Neuling and Kaltschmitt, 2018).
Besides gasification or Fischer-T synthesis, the study indicates that A10, which is advanced
fermenting, is the second-best method. In contrast to the existing research, which claims that advanced
fermenting (DSHC and ATJ) related methods are constrained in feedstock sustainability and
incapable of reaching the commercial development stage, the results presented here appear plausible.
Nevertheless, it might be challenging to scale up the advanced fermented production process. Figure
7. illustrates that A10 ranks comparatively higher because of its superior overall technical and
environmental attributes.
Although A8 (HTL methods with algae or microalgae) scored higher than A2 (HEFA using waste
cooking oils or animal fat), this production method’s possible capacity expansion is constrained by
the availability of inputs. Eventually, other options outperform A6 and A9 (production method
using pyrolysis and HTL, respectively). These outcomes are constant with relevant research on
biofuel generation technologies (Fiorese et al., 2013). According to their study of industry
Page 584
SEMIT 2023
professionals, gasification or Fischer-T synthesis and oil-based methods, including HEFA, Pyrolysis,
and HTL, were the most common production methods.
The optimal feedstock for SAF production is the direct transition process of CO2 to SAF (A11) in the
gasification or Fischer-T synthesis group. This might be the case since experts believe that CO2
extraction from industrial technologies is a more direct way than employing MSW (A4), wood
residues, and agricultural waste (A5). The reduction in Greenhouse gas emission by using landfill gas
to counterbalance the use of fossil fuels has been dropped even though converting MSW into
biofuel prevents Greenhouse gas emission that derives waste in landfills and incinerating processes
(Staples et al., 2018).
Table 3. Inflow and outflow in sustainable aviation fuel production methods
Production Code Feedstocks Inflo Outflow Net
method w Phi
Hydro processed A1 Algae and 0.442 0.518 -0.077
Esters and Fatty Microalgae
Acids (HEFA) A2 Animal fats or 0.487 0.465 0.023
Used cooking oils
A3 Oilseeds 0.423 0.534 -0.112
F-T synthesis or A4 Municipal solid 0.583 0.386 0.198
gasification wastes
A5 Wood residues and 0.552 0.423 0.130
Agri wastes
Devolatilization or A6 Algae and 0.375 0.582 -0.208
Pyrolysis Microalgae
A7 Wood residues and 0.446 0.517 -0.072
Agri wastes
Hydrothermal A8 Algae and 0.483 0.464 0.020
liquefaction (HTL) Microalgae
A9 Wood residues and 0.380 0.584 -0.205
Agri wastes
Advanced A 10 Wood residues and 0.509 0.447 0.063
fermentation Agri wastes
CO2 co- A 11 Combination of 0.605 0.365 0.25

valorization and Industrial waste
biomass waste gases CO2 and
wood residues
0.3
0.2
0.1
0
A6 A9 A3 Al A7 A8 A2 A10 A5 A4 A11
-0.1
-0.2
-0.3
Figure 7. MCDM Ranks Sustainable Aviation Fuel Production Methods
Page 585
SEMIT 2023
5.3. Comparison of MCDM Methodologies

The rank order produced with TOPSIS and VIKOR using fuzzy methodologies is shown in Table 4.
Here, A4 and A5 were consistently placed in the highest 3 by both techniques. However, A6 and A9
were consistently ranked in the lowest three. Nevertheless, due to variations in the theoretical basis
of both methods, the rank given to A3 and A11 in both methodologies are distinct.
Table 4 Fuzzy VIKOR and TOPSIS for Sustainable Fuel Production Method Ranking Source
6. Conclusions
The airline industry’s efforts to reduce CO2 emissions and find new sources of jet fuel have the chance
to benefit significantly when using sustainable fuels in aviation. SAF could be produced in various
methods, making choosing a particular route a challenging strategic choice. Utilizing TEA or LCA
research, the existing literature compares the effectiveness of one or more methods. This study
proposed a comprehensive assessment approach for the SAF selection problem.
6.1. Implications
This research used a stakeholder engagement strategy to collaborate with professionals to create a
comprehensive structure based on evaluation criteria including societal, environmental, economic,
and technological factors. To determine the most important factors for
investment/production/purchase/usage of SAF, this study created a questionnaire to obtain
stakeholders’ opinions. In-depth conversations with industry professionals confirmed the criterion.
Industry professionals were given another form to rate SAF production methods according to a
performance criterion. Data is validated using fuzzy TOPSIS and fuzzy VIKOR and PROMETHEE
to reduce professionals’ judgmental personal prejudice.
The analysis concluded that economic and environmental considerations outweigh technological and
social considerations. The focus on these two groups is logical given that SAFs are marketed as having
less harmful environmental consequences than normal Jet-A fuel yet, serve as a more expensive
alternative. Greenhouse gas reductions, sustainability on feedstocks, minimum selling price,
compatibility of jet fuel, and traceability of sustainable aviation fuel are the essential performance
factors for rating their production methods. A comparison of sustainable production methods based
solely on one criterion reveals that no single production method dominates the 24 assessment criteria.
In addition, various results are observed when examining each effect category separately in mono-
criteria rankings. HEFA-based output scores highest in economic effects. This conclusion should be
interpreted cautiously, as HEFA appears to be a viable opportunity among the remaining 3 impact
categories. According to global rankings, gasification or Fischer-T synthesis, advanced fermentation,
and hydrothermal liquefaction are the top conversion processes for SAFs. Although technologically
developed and frequently used commercially, HEFA-based procedures are less desirable. It is
noticeable that collected waste from gas, wood, and farms is superior feedstock when compared to
algae or microalgae and purpose-grown seeds that are non-edible for oil extraction.
Animal fats or used cooking oil varieties are additional enticing options, but their restricted
availability prevents them from increasing SAF production. The findings suggest
integrating/designing/modifying the waste gases and leftovers supply chain for SAF production. The
SAF production method analysis and ranking approach aid policymaking. A nation’s technical,
feedstock, and economic conditions must guide SAF production methods.
Page 586
SEMIT 2023
Some countries, like Brazil, make SAF from sugarcane, whereas the United Kingdom has found that
municipal solid waste and agricultural waste make the best feedstock for making SAF. To secure a
consistent supply of raw materials to produce SAF, feedstock suppliers might receive monetary
benefits. An improved feedstock supply chain will assist in minimizing the unpredictability in the
production of SAF and ensure commercial stability for both feedstock providers and the SAF
production companies. Technology-wise, strengthening SAF production paths and R&D capacities
will boost economic competitiveness and the bio-economy. These competencies would be sold to
other countries.
In addition, as indicated in the research, the plant capital investment of the SAF production method
is a critical factor. In this regard, it is recommended to simplify and underwrite debt or equity funding
by state governments. This scenario will increase the trust of private capitalists, allowing new
investors and establishing refineries providing biofuel for road transportation to incorporate SAF into
their business models. The results of this research may also be used to develop policy choices such as
production or buying quotas from a particular SAF production method, taxes on standard jet fuel
(creating a comparable environment for that SAF production method), including subsidies for
sustainable production of aviation fuels.
In addition, as indicated by Trejo-Pech et al., (2019) SAF production can be regarded as a catalyst for
creating national climate change policies. In conclusion, it is anticipated that policymakers and
decision-makers will consider the research results when investing or making a policy decision about
SAF production methods. These findings raise new research questions. For instance, it could be
intriguing to examine conflicts across all the societal, economic, technological, and environmental
criteria. It is also intended to integrate more synthetic methods for SAF production and a range-based
technique to investigate ranking distribution.
References
Ahmad, S., Nadeem, A., Akhanova, G., Houghton, T., & Muhammad-Sukki, F. (2017). Multi-criteria
evaluation of renewable and nuclear resources for electricity generation in Kazakhstan. Energy, 141,
1880–1891. https://doi.org/10.1016/j.energy.2017.11.102
Alkema, B., R. G. (2022). Aviation’s net-zero ambitions meet resistance in the run-up to COP26 -le.
runwaygirlnetwork.com/2021/10/22/aviations-net-zero-ambitions-meet-resistance-cop26/
Atag.org. (2022). Air Transport Action Group. (2020, September). Facts and figures.
www.atag.org/facts-figures.html
Atsonios, K., Kougioumtzis, M.-A., D. Panopoulos, K., & Kakaras, E. (2015). Alternative
thermochemical routes for aviation biofuels via alcohols synthesis: Process modeling, techno-
economic assessment and comparison. Applied Energy, 138, 346–366.
https://doi.org/10.1016/j.apenergy.2014.10.056
aviationbenefits.org. (2022). Air Transport Action Group. (2021). Waypoint 2050.tle.
https://aviationbenefits.org/environmental-efficiency/climate-action/waypoint-2050/
Awasthi, A., Govindan, K., & Gold, S. (2018). Multi-tier sustainable global supplier selection using a
fuzzy AHP-VIKOR based approach. International Journal of Production Economics, 195, 106–
117. https://doi.org/10.1016/j.ijpe.2017.10.013
Bann, S. J., Malina, R., Staples, M. D., Suresh, P., Pearlson, M., Tyner, W. E., Hileman, J. I., & Barrett,
S. (2017). The costs of production of alternative jet fuel: A harmonized stochastic assessment.
Bioresource Technology, 227, 179–187. https://doi.org/10.1016/j.biortech.2016.12.032
Baudry, G., Macharis, C., & Vallée, T. (2018). Can microalgae biodiesel contribute to achieve the
sustainability objectives in the transport sector in France by 2030? A comparison between first,
second and third generation biofuels though a range-based Multi-Actor Multi-Criteria Analysis.
Energy, 155, 1032–1046. https://doi.org/10.1016/j.energy.2018.05.038
Page 587
SEMIT 2023
Brans, J. P., & Vincke, P. (1985). Note—A Preference Ranking Organisation Method. Management
Science, 31(6), 647–656. https://doi.org/10.1287/mnsc.31.6.647
CAAFI. (n.d.). Fuel Qualification. (2022b). Retrieved December 13, 2022, from
https://www.caafi.org/focus_areas/feedstocks.html
CAAFI. (2019). Etihad Airways Flies from Abu Dhabi to Amsterdam on AJF Blend from Halophytes.
caafi.org/news/NewsItem.aspx?id=10442
Castello, D., Haider, M. S., & Rosendahl, L. A. (2019). Catalytic upgrading of hydrothermal liquefaction
biocrudes: Different challenges for different feedstocks. Renewable Energy, 141, 420–430.
https://doi.org/10.1016/j.renene.2019.04.003
Chen, Y.-K., Lin, C.-H., & Wang, W.-C. (2020). The conversion of biomass into renewable jet fuel.
Energy, 201, 117655. https://doi.org/10.1016/j.energy.2020.117655
Cheng, F., & Brewer, C. E. (2017). Producing jet fuel from biomass lignin: Potential pathways to alkyl-
benzenes and cycloalkanes. Renewable and Sustainable Energy Reviews, 72, 673–722.
https://doi.org/10.1016/j.rser.2017.01.030
Chevron Products Company. (2004). Aviation Fuels Technical Review.
Chiaramonti, D. (2019). Sustainable Aviation Fuels: the challenge of decarbonization. Energy Procedia,
158, 1202–1207. https://doi.org/10.1016/j.egypro.2019.01.308
Chisti, Y. (2007). Biodiesel from microalgae. Biotechnology Advances, 25(3), 294–306.
https://doi.org/10.1016/j.biotechadv.2007.02.001
Dayton, D. C., & Foust, T. D. (2020). Alternative Jet Fuels. In Analytical Methods for Biomass
Characterization and Conversion (pp. 147–165). Elsevier. https://doi.org/10.1016/B978-0-12-
815605-6.00010-X
de Jong, S., Hoefnagels, R., Faaij, A., Slade, R., Mawhood, R., & Junginger, M. (2015). The feasibility
of short‐term production strategies for renewable jet fuels – a comprehensive techno‐economic
comparison. Biofuels, Bioproducts and Biorefining, 9(6), 778–800.
https://doi.org/10.1002/bbb.1613
Diederichs, G. W., Ali Mandegari, M., Farzad, S., & Görgens, J. F. (2016). Techno-economic comparison
of biojet fuel production from lignocellulose, vegetable oil and sugar cane juice. Bioresource
Technology, 216, 331–339. https://doi.org/10.1016/j.biortech.2016.05.090
Dožić, S. (2019). Multi-criteria decision making methods: Application in the aviation industry. Journal
of Air Transport Management, 79, 101683. https://doi.org/10.1016/j.jairtraman.2019.101683
Dyk, S. van. E. C. R. (2021). ReFuelEU Aviation proposal details SAF blending obligation on fuel
suppliers. https://www.greenairnews.com/?p=1374
Efroymson, R. A., Dale, V. H., & Langholtz, M. H. (2017). Socioeconomic indicators for sustainable
design and commercial development of algal biofuel systems. GCB Bioenergy, 9(6), 1005–1023.
https://doi.org/10.1111/gcbb.12359
Fiorese, G., Catenacci, M., Verdolini, E., & Bosetti, V. (2013). Advanced biofuels: Future perspectives
from an expert elicitation survey. Energy Policy, 56, 293–311.
https://doi.org/10.1016/j.enpol.2012.12.061
Fortier, M.-O. P., Roberts, G. W., Stagg-Williams, S. M., & Sturm, B. S. M. (2014). Life cycle assessment
of bio-jet fuel from hydrothermal liquefaction of microalgae. Applied Energy, 122, 73–82.
Ganguly, I., Pierobon, F., Bowers, T. C., Huisenga, M., Johnston, G., & Eastin, I. L. (2018). ‘Woods-to-
Wake’ Life Cycle Assessment of residual woody biomass based jet-fuel using mild bisulfite
pretreatment. Biomass and Bioenergy, 108, 207–216.
https://doi.org/10.1016/j.biombioe.2017.10.041
Page 588
SEMIT 2023
Gegg, P., & Wells, V. (2017). UK Macro-Algae Biofuels: A Strategic Management Review and Future
Research Agenda. Journal of Marine Science and Engineering, 5(3), 32.
https://doi.org/10.3390/jmse5030032
Geleynse, S., Jiang, Z., Brandt, K., Garcia-Perez, M., Wolcott, M., & Zhang, X. (2020). Pulp mill
integration with alcohol-to-jet conversion technology. Fuel Processing Technology, 201, 106338.
https://doi.org/10.1016/j.fuproc.2020.106338
Heyne, J., Rauch, B., Le Clercq, P., & Colket, M. (2021). Sustainable aviation fuel prescreening tools
and procedures. Fuel, 290, 120004. https://doi.org/10.1016/j.fuel.2020.120004
Hileman, J. I., & Stratton, R. W. (2014). Alternative jet fuel feasibility. Transport Policy, 34, 52–62.
https://doi.org/10.1016/j.tranpol.2014.02.018
IATA. (2019). Sustainable Aviation Fuels Fact sheet.
https://www.iata.org/contentassets/ed476ad1a80f4ec7949204e0d9e34a7f/fact-sheet-alternative-
fuels.pdf.
ICAO. (2016). Environmental Report. www.icao.int/environmental-protection/Pages/env2016.aspx.
ICAO. (2022). Sustainable Aviation Fuels (SAF). www.icao.int/environmental-
protection/pages/SAF.aspx
IEA. (2022). Aviation – Analysis. https://www.iea.org/reports/aviation
International Civil Aviation Organization. (2019). Trends in Emissions that Affect Climate Change.
https://www.icao.int/environmental-protection/Pages/ClimateChange_Trends.aspx
International Civil Aviation Organization. (2022). COVID-19 impacts and 2022 CORSIA periodic
review. https://www.icao.int/environmental-protection/CORSIA/Pages/CORSIA-and-Covid-
19.aspx
International Civil Aviation Organization. (2021). 2021 global air passenger totals show improvement
from 2020, but still only half pre-pandemic levels. https://bit.ly/3ruI5p8
Kivits, R., Charles, M. B., & Ryan, N. (2010). A post-carbon aviation future: Airports and the transition
to a cleaner aviation sector. Futures, 42(3), 199–211. https://doi.org/10.1016/j.futures.2009.11.005
Klein, B. C., Chagas, M. F., Junqueira, T. L., Rezende, M. C. A. F., Cardoso, T. de F., Cavalett, O., &
Bonomi, A. (2018). Techno-economic and environmental assessment of renewable jet fuel
production in integrated Brazilian sugarcane biorefineries. Applied Energy, 209, 290–305.
Lanzini, P., Testa, F., & Iraldo, F. (2016). Factors affecting drivers’ willingness to pay for biofuels: the
case of Italy. Journal of Cleaner Production, 112, 2684–2692.
https://doi.org/10.1016/j.jclepro.2015.10.080
Li, X., Mupondwa, E., & Tabil, L. (2018). Technoeconomic analysis of biojet fuel production from
camelina at commercial scale: Case of Canadian Prairies. Bioresource Technology, 249, 196–205.
https://doi.org/10.1016/j.biortech.2017.09.183
Lokesh, K., Sethi, V., Nikolaidis, T., Goodger, E., & Nalianda, D. (2015). Life cycle greenhouse gas
analysis of biojet fuels with a technical investigation into their impact on jet engine performance.
Biomass and Bioenergy, 77, 26–44. https://doi.org/10.1016/j.biombioe.2015.03.005
Michailos, S. (2018). Process design, economic evaluation and life cycle assessment of jet fuel production
from sugar cane residue. Environmental Progress & Sustainable Energy, 37(3), 1227–1235.
https://doi.org/10.1002/ep.12840
Moniruzzaman, M., Yaakob, Z., Shahinuzzaman, M., Khatun, R., & Aminul Islam, A. K. M. (2017).
Jatropha Biofuel Industry: The Challenges. In Frontiers in Bioenergy and Biofuels. InTech.
https://doi.org/10.5772/64979
Page 589
SEMIT 2023
Moore, R. H., Thornhill, K. L., Weinzierl, B., Sauer, D., D’Ascoli, E., Kim, J., Lichtenstern, M., Scheibe,
M., Beaton, B., Beyersdorf, A. J., Barrick, J., Bulzan, D., Corr, C. A., Crosbie, E., Jurkat, T., Martin,
R., Riddick, D., Shook, M., Slover, G., … Anderson, B. E. (2017). Biofuel blending reduces particle
emissions from aircraft engines at cruise conditions. Nature, 543(7645), 411–415.
https://doi.org/10.1038/nature21420
Neuling, U., & Kaltschmitt, M. (2018). Techno-economic and environmental analysis of aviation
biofuels. Fuel Processing Technology, 171, 54–69. https://doi.org/10.1016/j.fuproc.2017.09.022
O’Connell, A., Kousoulidou, M., Lonza, L., & Weindorf, W. (2019). Considerations on GHG emissions
and energy balances of promising aviation biofuel pathways. Renewable and Sustainable Energy
Reviews, 101, 504–515. https://doi.org/10.1016/j.rser.2018.11.033
O’Malley, J., Pavlenko, N., and Searle, S. (2021). Estimating sustainable aviation fuel feedstock
availability to meet growing European Union demand.
Palmer, W. (2021). United Flies World’s First Passenger Flight On 100% Sustainable Aviation Fuel
Supplying One of Its Engines. https://www.ge.com/news/reports/united-flies-worlds-first-
passenger-flight-on-100-sustainable-aviation-fuel-supplying-one
Reuters. (2021). current targets on sustainable aviation fuel “pathetic.”
https://www.reuters.com/business/aerospace-defense/uks-johnson-says-current-targets-
sustainable-aviation-fuel-pathetic-2021-11-02/
Ribeiro, L. A., Pereira da Silva, P., Ribeiro, L., & Dotti, F. L. (2017). Modelling the impacts of policies
on advanced biofuel feedstocks diffusion. Journal of Cleaner Production, 142, 2471–2479.
https://doi.org/10.1016/j.jclepro.2016.11.027
Ricardo Energy and Environment. (2020). Targeted Aviation Advanced Biofuels Demonstration
Competition – Feasibility Study Final Report for Department for Transport, UK.
https://docslib.org/doc/10188308/targeted-aviation-advanced-biofuels-demonstration-competition-
feasibility-study-final-report-report-for-department-for-transport-uk
Ritchie, H. (2020). Climate change and flying: what share of global CO2 emissions come from aviation?
https://ourworldindata.org/co2-emissions-from-aviation
Schäfer, A. W., Barrett, S. R. H., Doyme, K., Dray, L. M., Gnadt, A. R., Self, R., O’Sullivan, A.,
Synodinos, A. P., & Torija, A. J. (2018). Technological, economic and environmental prospects of
all-electric aircraft. Nature Energy, 4(2), 160–166. https://doi.org/10.1038/s41560-018-0294-x
Scheelhaase, J., Maertens, S., & Grimme, W. (2019). Synthetic fuels in aviation – Current barriers and
potential political measures. Transportation Research Procedia, 43, 21–30.
https://doi.org/10.1016/j.trpro.2019.12.015
Schillo, R. S., Isabelle, D. A., & Shakiba, A. (2017). Linking advanced biofuels policies with stakeholder
interests: A method building on Quality Function Deployment. Energy Policy, 100, 126–137.
Seber, G., Malina, R., Pearlson, M. N., Olcay, H., Hileman, J. I., & Barrett, S. R. H. (2014).
Environmental and economic assessment of producing hydroprocessed jet and diesel fuel from
waste oils and tallow. Biomass and Bioenergy, 67, 108–118.
https://doi.org/10.1016/j.biombioe.2014.04.024
Sikarwar, V. S., Zhao, M., Fennell, P. S., Shah, N., & Anthony, E. J. (2017). Progress in biofuel
production from gasification. Progress in Energy and Combustion Science, 61, 189–248.
https://doi.org/10.1016/j.pecs.2017.04.001
SkyNRG. (2020). Technology | Sustainable Aviation Fuel. https://skynrg.com/sustainable-aviation-
fuel/technology/
Staples, M. D., Malina, R., Suresh, P., Hileman, J. I., & Barrett, S. R. H. (2018). Aviation CO2 emissions
reductions from the use of alternative jet fuels. Energy Policy, 114, 342–354.
Page 590
SEMIT 2023
The Royal Society. (2019). Sustainable synthetic carbon based fuels for transport.
https://royalsociety.org/-/media/policy/projects/synthetic-fuels/synthetic-fuels-briefing.pdf
Timperley, J. (2021). The Fastest Ways Aviation Could Cut Emissions.
https://www.bbc.com/future/article/20210525-how-aviation-is-reducing-its-climate-emissions
Trejo-Pech, C. O., Larson, J. A., English, B. C., & Yu, T. E. (2019). Cost and Profitability Analysis of a
Prospective Pennycress to Sustainable Aviation Fuel Supply Chain in Southern USA. Energies,
12(16), 3055. https://doi.org/10.3390/en12163055
Trivedi, P., Olcay, H., Staples, M. D., Withers, M. R., Malina, R., & Barrett, S. R. H. (2015). Energy
return on investment for alternative jet fuels. Applied Energy, 141, 167–174.
United Nations. (2015). The Paris Agreement. https://www.un.org/en/climatechange/paris-agreement
US Dept. Of energy. (2020). Sustainable Aviation Fuel Review of Technical Pathways.
https://www.energy.gov/sites/prod/files/2020/09/f78/beto-sust-aviation-fuel-sep-2020.pdf
Wang, Z., Pashaei Kamali, F., Osseweijer, P., & Posada, J. A. (2019). Socioeconomic effects of aviation
biofuel production in Brazil: A scenarios-based Input-Output analysis. Journal of Cleaner
Production, 230, 1036–1050. https://doi.org/10.1016/j.jclepro.2019.05.145
World Economic Forum. (2020). Clean Skies for Tomorrow: Sustainable Aviation Fuels as a Pathway to
Net-Zero Aviation. https://www.weforum.org/reports/clean-skies-for-tomorrow-sustainable-
aviation-fuels-as-a-pathway-to-net-zero-aviation
Zemanek, D., Champagne, P., & Mabee, W. (2020). Review of life‐cycle greenhouse‐gas emissions
assessments of hydroprocessed renewable fuel ( <scp>HEFA</scp> ) from oilseeds. Biofuels,
Bioproducts and Biorefining, 14(5), 935–949. https://doi.org/10.1002/bbb.2125
Zhang, H., Fang, Y., Wang, M., Appels, L., & Deng, Y. (2020). Prospects and perspectives foster
enhanced research on bio-aviation fuels. Journal of Environmental Management, 274, 111214.
https://doi.org/10.1016/j.jenvman.2020.111214
Zhang, H., Wang, L., Van herle, J., Maréchal, F., & Desideri, U. (2020). Techno-economic evaluation of
biomass-to-fuels with solid-oxide electrolyzer. Applied Energy, 270, 115113.
https://doi.org/10.1016/j.apenergy.2020.115113
Page 591
SEMIT 2023
Appendix
Table A1: Details of the Expert Pool

Participants Years of Expertise Departmet Job Role Company
Participant 1 5+ Supply Supply Chain Executive Shell Aviation
chain
Participant 2 5+ Supply Supply Chain Assistant BP Aviation
chain
Participant 3 10+ Supply Supply Chain Total
chain Operations Manager Energies
Participant 4 6 Supply Supply Chain Lufthansa
chain Coordinator Aviation Fuel
Participant 5 8 Supply Supply Chain Assistant Sky tanking
chain
Participant 6 5 Supply Supply Chain Air BP
chain Administrator
Participant 7 6 Supply Supply Chain Specialist World Fuel
chain Services
Participant 8 5 Supply Downstream Purchasing CEPSA
chain Assistant
Participant 9 12 Supply Supply Chain Manager ExxonMobil
chain Aviation
Participant 10 7 Supply Supply Chain Executive
chain Gazpromneft-
Aero.
Participant 11 6 Supply Supply Chain Analyst Shell Aviation
chain
Participant 12 9 Supply Supply Chain BP Aviation
chain Coordinator
Participant 13 14 Supply Supply Chain Assistant TotalEnergies
chain Manager
Participant 14 10+ Supply Supply Chain Manager Lufthansa
chain Aviation Fuel
Participant 15 8+ Supply Supply Chain Sky tanking
chain Operations Coordinator
Participant 16 6 Supply Supply Chain Specialist Air BP
chain
Participant 17 7 Supply Fuel Supply ExxonMobil
chain Coordinator Aviation
Participant 18 5 Supply Supply Chain ExxonMobil
chain Coordinator Aviation
Participant 19 17 Supply Supply Chain Manager Gazpromneft-
chain Aero
Participant 20 8+ Supply Downstream CEPSA
chain Procurement Manager
Participant 21 7+ Supply Supply Chain Air BP
chain Coordinator
Participant 22 12+ Supply Fuel Supply Executive World Fuel
chain Services
Page 592
SEMIT 2023
Table A2: References to the Evaluation Criterion

Category Selected Criterion Citation
Social Traceability (Lanzini et al., 2016).
Economical (Wang et al., 2019).
contribution
Food safety (Sikarwar et al., 2017) .
Social acceptance (Gegg & Wells, 2017).
Environ Sustainability on (Chiaramonti, 2019).
mental feedstock
Savings on GHG (Zemanek et al., 2020).
emissions
The impact caused by (Lanzini et al., 2016).
a change in land usage
Water and Soil (Efroymson et al., 2017).
pollution
Economi Alternative uses with (Hileman & Stratton, 2014).
c feedstock
Profitability of (Klein et al., 2018).
feedstock
Minimal price for (Ribeiro et al., 2017).
selling
Input energy usage (Baudry et al., 2018).
Productivity of Land (Li et al., 2018).
Operational and (Diederichs et al., 2016).
maintenance costs
Cost on feedstock (de Jong et al., 2015).
Cost of plant capital (Moore et al., 2017).
Technica Blending limits (Cheng & Brewer, 2017).
l Compatibility with (Hileman & Stratton, 2014).
Standard jet fuels
Local technical (Neuling & Kaltschmitt, 2018).
capability
Integration process (Ahmad et al., 2017).
Maturity of process (Bann et al., 2017).
technology
Method yield (Schillo et al., 2017).
Scalability on (Atsonios et al., 2015).
production volume
Composition and (Fiorese et al., 2013).
quality of feedstock
Page 593
SEMIT 2023
SP15: Industry 4.0, Supply Chain 4.0, and Logistics 4.0
A Draft Model Proposal for Total Quality Management (TQM) Based Industry 4.0 Maturity
Model
Kerem Elibal1[0000-0003-0328-9332] and Eren Özcelyan2[0000-0002-5213-6335]
1
BCS Metal Co., Gaziantep, Turkey
2
Industrial Engineering Department, Gaziantep University, Gaziantep, 27300, Turkey
Abstract
This paper aims to propose a draft model for Total Quality Management (TQM) based Industry 4.0
maturity model. “As the Quality Management Principles (QMPs) are globally accepted business
practices, an Industry 4.0 maturity model based on these principles should leverage the
organization’s efficiency and productivity within the context of Industry 4.0 technologies” is the
main argument of this study. Seven QMPs declared by International Standardization for
Organization (ISO) and recommended actions of these principles has been considered as factors;
total 32 factors and maturity statements of these factors has been constructed. This paper should
be assessed as a draft framework and should be developed in order to verify the proposed model
as a future research.
Keywords: Maturity, Industry 4.0, TQM.
1. Introduction
The design phase of a maturity model mainly focuses on the questions “what will be measured” and “why
will be measured”, in other words, determination of dimensions and sub-dimensions. Since the “Industry
4.0 maturity model” concept has been included into the literature at 2006, development phases, especially
design phases of maturity model for the 4th industrial revolution concept is a challenge for academicians
and practitioners [1]. This challenge depends on two main aspects; one of them is poor theoretical basis
of maturity model development [2] indicated that “data for appraisal frequently depends on the people
being asked (also referred to as key informant bias).” The second aspect is the lack of consensus on the
Industry 4.0 concept as in shown in Table 1.
Table 22. Some comments about Industry 4.0 concept.
• [3] emphasized that one of the challenges of the fourth industrial revolution is to identify a common
language, that makes it possible to find a common ground or at least a starting point for the
discussions.
• According to the [4] the research community did not reach unanimous consensus towards a
definition of Industry 4.0 and a number of authors have raised concern over this lack of agreement
• [5] declared that there are some number of concepts and technologies that are associated with
Industry 4.0. Some of them are even used synonymously for the term Industry 4.0. These include,
for example, the terms Cyber-Physical System, Cyber Physical Production System, Smart Factory
or Internet of Things. However, it is hardly possible to distinguish these terms from each other as
well as from the term Industry 4.0. This is due to the fact that all of these terms have so far been
ambiguously defined, which also applies to the term Industry 4.0
• Industry 4.0 is not so far adequately standardised, spread and lacks the clear boundaries [6].
• Despite the current discussion in research and practice, the different perceptions and
characterizations of Industry 4.0 show that there is currently no clear definition [7].
Page 594
SEMIT 2023
These two main challenges mentioned above; insufficient theoretical basics about maturity model
development and confusion about the Industry 4.0 concept arise a new question for Industry 4.0 maturity
developers; “how an Industry 4.0 maturity model can be developed?”. To answer this question, first of
all a general question which is “how a maturity model can be developed?” should be investigated. [2] and
[8] try to propose a systematic approach for maturity model development. According [2] and [8],
extensive literature review and expert opinions about the related concept are the main steps of maturity
model development.
[8] demonstrated a six phased stage. In the design and populate phase it has been indicated that what will
be measured and how will be measured should be identified. To answer the “what will be measured?”
question, extensive literature review about the concept should be done. Also, exploratory research
methods such as the Delphi technique, Nominal Group technique, case study interviews and focus groups
can be considered. [2] demonstrated that in the design phase theory-driven (via literature review) and/or
practitioner-based (via expert opinions) approaches should be applied to determine the domains which
will be assessed in the maturity model.
[1] have conducted a systematic literature review about Industry 4.0 maturity models and indicated that
current maturity models are designed according to the principles of [2] and [8] on a large scale. According
to [1], insufficiency of current Industry 4.0 maturity models can be summarized as below;
• Reviewed maturity models have different dimensions, which means researchers offered different
aspects to be measured in the concept of Industry 4.0 assessment. (According to [1] 44 different
researches offer 171 different dimensions). When it is considered that each dimension has other sub-
dimensions according to some publications, it becomes a complicated situation for practitioners to
decide which dimensions and sub-dimensions should be used to assess the maturity level of Industry
4.0
• A maturity model should also create an agile and effective roadmap. [1] indicated that current Industry
4.0 maturity models are not sufficient enough to meet this criterion. The question “which technology
and/or concept should be leveraged first” cannot be answered clearly by the results of assessment
models.
For this study, the proposed systematic literature review by [1] updated at the date of 06-2023 and it has
been seen that newly added 16 researches are still not enough to close the gaps mentioned above.
Findings above are the main motivation of this study. According to this paper, an Industry 4.0 maturity
model; i) should include dimensions simple and really worth to be measured in a quick and effective way,
ii) should give simple and agile ideas to leverage the maturity. Principles of Total Quality Management
[9] has been considered to serve this aim in the context of this study.
According to [10] as the TQM principles lead an organization to efficiency, it can be assumed that if the
maturity dimensions serve these principles then a maturity model is efficient, so an Industry 4.0 maturity
model should measure the aspects of an organization in the perspective of TQM principles.
Fig. 18. Seven TQM principles [9]
Page 595
SEMIT 2023
Also [9] indicates that “The Quality Management Principles (QMPs) can be used as a foundation to guide
an organization’s performance improvement. They were developed and updated by international experts
of ISO/TC 176, which is responsible for developing and maintaining ISO’s quality management
standards”
The structure of this paper is as follows. Section 2 gives information about the current literature for TQM
based Industry 4.0 maturity modelling. Section 3 describes the methodology and finally conclusion has
been done.
To understand the current situation of TQM and/or quality based Industry 4.0 maturity modeling, a
literature review has been conducted on Scopus and Web of Science Databases until June 2023 with the
keywords mentioned in Table 2.
Among 29 articles found from the conducted literature review, 14 of them are indicated as the studies
which develops Industry 4.0 maturity models in the concept of quality.
Table 23. Literature review search queries
'tqm' and 'industry 4.0' and 'maturity model'
'quality management' and 'industry 4.0' and 'maturity model'
'efqm' and 'industry 4.0' and 'maturity model'
'business excellence' and 'industry 4.0' and 'maturity model'
'tqm' and 'industry 4.0' and 'assessment model'
'quality management' and 'industry 4.0' and 'assessment model'
'efqm' and 'industry 4.0' and 'assessment model'
'business excellence' and 'industry 4.0' and 'assessment model'
'quality 4.0' and 'maturity'
'quality 4.0' and 'assessment'
'tqm' and 'industry 4.0' and 'readiness'
'quality management' and 'industry 4.0' and 'readiness'
'efqm' and 'industry 4.0' and 'readiness'
'business excellence' and 'industry 4.0' and 'readiness'
'quality 4.0' and 'readiness'
'quality 4.0' and 'measurement'
'tqm' and 'industry 4.0' and 'measurement'
'quality management' and 'industry 4.0' and 'measurement'
'efqm' and 'industry 4.0' and 'measurement'
'business excellence' and 'industry 4.0' and 'measurement'
[11] explores the potential of the ISO 9004:2018 quality management system (QMS) as a basis for
assessing the maturity level of quality in Industry 4.0. They propose a maturity model based on the five
elements of the ISO 9004:2018 QMS: Context and Identity, Resource Management, Process
Management, Performance Management, and Improvement, Learning, and Innovation. They linked 31
self-assessment elements to Industry 4.0 and claimed that ISO 9004:2018 self-assessment tool can also
be used to measure the maturity level of an organization.
Page 596
SEMIT 2023
Table 3 shows the other 13 researches assessment methodology. These researches proposed different
dimensions to measure the Industry 4.0 readiness in the context of quality management.
Table 3. Dimensions of reviewed literature

Publication Quality 4.0 dimensions
[12] Top management support-Quality 4.0 vision and strategy-Leadership for Quality-
Trainings and rewards -Knowledge and awareness of Quality 4.0-Organizational
culture -Customer readiness -Supplier readiness
[13] Strategic Leadership-Quality Culture-Customer centricity-Quality Management
System-Compliance-Competence-Analytic Thinking-Metric and data-driven
decision making-Advanced analytics-Data governance-Innovation-New age
technological tools
[14] Top management commitment and support, Leadership, Organizational culture,
Employee’s commitment, ISO QMS standart implementation
[15] Workforce data analyse skills-Data quality-Big data-Connectivity-Collobration-
Horizantel, vertical and end-to-end integration
[16] Big data management- Application of Industry 4.0 technologies-Integration-
Digitalization-Smart machinery and factory-Real time acts-Robotics
[17] Collaboration- Customer centredness- Financial set-up- Government set-up –
Infrastructures -Knowledge and awareness of Q4.0 -Leverage technology adoption-
Organisation culture- Q4.0 readiness- Q4.0 vision and strategy -Supplier
centredness- Top management support- Training and Q4.0 skills
[18] Leveraged technology adoption- Training on Quality 4.0 skills and Reward-
Infrastructures such as electricity, Internet connectivity- The government set up in
terms of policy, strategy and plans- Top management support- Quality 4.0 vision
and strategy- Collaboration and Leadership- Knowledge and awareness of Quality
4.0- Organisational culture toward Quality 4.0- Financial set-up- Customer
centredness- Supplier centredness
[19] Management commitment to technology and innovation-Customised Customer
focus (Voice of the customer)-Technology-driven employee involvement and
empowerment-Process and systems integration and management-Knowledge for
decision-making and future prediction-Root cause analysis of operations
disturbances and sustainable solution-Operational environment bilgileşim
[20] Strategic direction-People and Culture-Processes-Method and Tools
[21] Compliance-Competency-Leadership-Connectivity-Collobration-Culture-
Management System-Web-based applications-Scalability-Data-Analytics
[22] Sensing Technologies including QR codes, sensors, actuators -IoT-Big Data (BD)-
Cloud Computing (CC) -Machine Learning (ML) -Deep Learning (DL) -Artificial
Intelligence (AI) -Data Science (DS) -Blockchain (BCh) -Additive Manufacturing
(AM) -Extended Reality (ER) including augmented reality, virtual reality, mixed
reality
[23] Data, Analytics, Connectivity, Collobration, App. Development, Scalability,
Management Systems, Compliance, Culture, Leadership, Competence
[24] Top management support-Quality 4.0 vision and strategy-Leadership-Training and

reward-Knowledge and awareness of Quality 4.0-Organizational culture toward
Quality 4.0-Customer centredness-Supplier centredness
Page 597
SEMIT 2023
As a result, it has been seen that the findings mentioned in the first chapter are also available in studies
that try to measure Industry 4.0 maturity within the scope of quality management. Studies are; i) not
capable to create agile and effective roadmaps, ii) use different dimensions that may confuse practitioners.
The most important issue is, reviewed methods do not answer the questions “What should be done after
the assessment?” and/or “Which technology should be leveraged first?”. For example, according to the
Table 3, if a practitioner identifies the maturity level of issue 12 (“Improvement of Quality 4.0-related
processes performances”) “determined” as in explained in [11], what should be the next step to improve
the mentioned issue?
Another examples can be given for other studies mentioned in Table 3. In the study of [23] , there is a
survey question which measures the usage of “ wearable devices that generate and transmit employee
data”. If the usage of weareble devices in an organizaton does not exist or in low level, should that
organization supposed to be low in the context of Industry 4.0 maturity? In another words, is this this
technology really urgently needed for Industry 4.0 implementation?
To overcome this insufficiencies mentioned above, a TQM based Industry 4.0 maturity modelling has
been proposed as in explained in section 3.
3. Methodology
TQM is a managerial approach which leads organizations to use their all resources in an efficient way
with a continuous improvement approach to satisfy the customers’ expectations and satisfactions. It is an
evolving concept since the industrial revolution and the practices of the concept are changing or emerging
via many researchers or practitioners. [25] gives detailed information about the history, key concepts and
case studies about the TQM. Among many success factors and principles about TQM which are proposed
by both academicians and practitioners, seven ISO principles are the ones that are widely used. These are
customer focus, leadership, engagement of people, process approach, improvement, evidence-based
decision making and relationship management [9].
These seven principles and recommended actions by [9] for each principle has been considered
Industry 4.0 maturity assessment factors. Determined factors will be measured according to the maturity
statements mentioned in Table 4.
Tables below will be used as survey questions and from the participants it will be requested to select the
most appropriate level according to their situation. Maturity levels are defined as:
• Ad-Hoc: There is no or ad-hoc applications for the related action.
• Traditional: Traditional activities with no or limited digital technologies are executed
• Aware/Visible: Company is aware of digitalization and uses digital technologies in some extent
which also enables visible and shareable data
• Optimizing/Predictive: Actions are executed in digital environment, company is highly aware of
digitalization and predictive proceedings can be executed
• Advanced/Autonomous: Highly integrated and strategized Industry 4.0 actions exsists,
autonomous predictive actions can be executed.
Page 598
Table 4. Maturity statements (Limited view)
Factor Activity [9] Ad-Hoc Traditional Aware Optimizing Advanced/Autonomous
Code
We do not have a We collect potential …. ….. We can analyze big data
systematic customer customer information collected through various data
identification method. through activities such as sources (surveys, social media
• Identification of indirect Customer information is fair participation and listening, contact forms, chat
and direct customers who generally obtained from surveys. We store this bots, web cookies) with artificial
CF-1 receive value (product or coincidental hearings and information in platforms intelligence algorithms (text
service) from your encounters, and we such as Excel and Word mining, etc.) and recognize
organization communicate with and perform superficial indirect customers and/or
potential customers reviews as needed. segment existing customers.
through this information.
CF-2 • Determining customer We don’t have any We conduct surveys and …. …. We conduct predictive data
current and future systematic method. We interviews and save them analysis with artificial
expectations, communicate with in electronic media such intelligence- based systems to
customers through as Excel and Word. The understand customers' future
• Actively manage customer methods such as information collected is needs with big data obtained
relationships telephone calls and e-mail examined superficially. from various data sources
correspondence. (customer portals , surveys,
social media listening, past sales
records, chatbots , web cookies,
etc.).
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
LS-2 • Establishing and There is no action to Verbal meetings, posters, We have an artificial
Page
maintaining shared values, establish and sustain and signs are used to intelligence- based HRM system
equity and ethical models shared values, justice and establish and maintain that can analyze the
599
ethical models. sustainability of common values,

Code
for behavior at all levels of shared models of values, justice and ethical models and
the organization justice, and ethics. make autonomous training
recommendations on an
• Creating a culture of trust individual basis .
and honesty
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
EOP-2 • Promoting collaboration We do not have a We create project groups, We have an AI-based predictive
throughout the systematic method. share documents via team building system based on
organization Word, Excel via email or skills and previous projects.
meetings
• Facilitating the sharing of
knowledge and experience
and open discussion
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
PA-1 • Defining the goals of the We do not have a Process and target We have a data management
system and the processes systematic method. definitions are system that checks real-time
required to achieve these documented in electronic data, filters it and compares the
goals media such as Word and correct data with defined targets.
Excel.
• Ensuring the availability of
necessary information to
implement and improve
processes and monitor,
analyze and evaluate
overall system
performance
Page
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
600
…. …. …. …. …. …. ….
Code
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
IP-3 • Planning, implementing, We do not have a We document We use AR, VR, simulation
completing and systematic method, we improvement plans, techniques, AI-based inspection
monitoring, reviewing and announce improvement implementation and and inspection system to plan
inspecting the results of plans and results verbally. results in electronic media and implement improvement.
improvement projects such as Word and Excel.
• Integrating the
improvement concept into
the development of new or
modified products ,
services and processes
• Recognition and
acceptance of improvement
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
EDM-4 • Ensuring that individuals We do not have a We use digital training We use training simulations for
are competent to analyze systematic method. documents in Word, Excel data analysis with real-time data
and evaluate the data for basic data analysis.
needed
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
…. …. …. …. …. …. ….
RM-4 • Establishing development We do not have a We receive and review We use AI-based data analytics
and improvement activities systematic method. suggestions from our and guide our partners on
in cooperation with suppliers and partners recommendation opportunities
Page
suppliers, business and potential gains from these

partners and other improvements.
601
relevant parties
Code
• Encouraging and
recognizing improvements
and achievements made by
suppliers and business
partners
Page
602
SEMIT 2023
4. Conclusion and Future Researchs

In this study, a draft model for Industry 4.0 maturity modelling has been proposed. Conducted literature
review about Industry 4.0 maturity modelling showed us that current models do not serve any business
practices, fuzzy to understand for practitioners and have insufficiencies about creating agile and effective
roadmap.
To overcome these gaps, a TQM based Industry 4.0 maturity model has been proposed. As the Quality
Management Principles have aimed organizations to achive sustainable improvement and efficiency, then
an Industry 4.0 maturity model which aims these principles provide agile and understandable assessment
process for practitioners. Because the Quality Management Principles are globally wide business
practices, these principles are used as base for proposed maturity model. 32 factors have been determined
and maturity statements for these factors have been constructed.
This paper should be considered as a first step for TQM based Industry 4.0 maturity modelling. According
to the constructed statements for 32 factors, a survey study and Confirmatory factor analysis (CFA) and
Structural equation modelling (SEM) to verify the proposed model should be one of the important future
research. Also, expanding the statements for different sectors which uses different technologies should
be another future research of this paper.
References
24. Elibal, K., Özceylan, E : A systematic literature review for industry 4.0 maturity modeling: state-of-
the-art and future challenges. Kybernetes 50.11, 2957-2994 (2021).
25. Mettler, T.: Maturity assessment models: a design science research approach. International Journal
of Society Systems Science 3.1-2,81-98 (2011).
26. Kane, Gerald C., et al.: Coming of age digitally. MIT Sloan Management Review (2018).
27. Hofmann, C., et al.: Development of an agile development method based on Kanban for distributed
part-time teams and an introduction framework. Procedia Manufacturing 23,45-50 (2018).
28. Stefan, L., et al.: Concept for an evolutionary maturity based Industrie 4.0 migration model. Procedia
Cirp 72, 404-409 (2018).
29. Schuh, G., et al.: Industrie 4.0 Maturity Index: Die digitale Transformation von Unternehmen
gestalten. Herbert Utz Verlag, (2017).
30. Bauer, W., et al.: Industrie 4.0–Volkswirtschaftliches Potenzial für
Deutschland. Berlin/Stuttgart, (2014).
31. De Bruin, T., et al.: Understanding the main phases of developing a maturity assessment model.
Australasian Conference on Information Systems (ACIS). Australasian Chapter of the Association
for Information Systems (2005).
32. ISO, https://www.iso.org/files/live/sites/isoorg/files/store/en/PUB100080.pdf, last accessed
28/06/2023
33. Elibal, K, and Özceylan, E.: Comparing industry 4.0 maturity models in the perspective of TQM
principles using Fuzzy MCDM methods. Technological Forecasting and Social Change 175, 121379
(2022).
34. Glogovac, M, et al.: ISO 9004 maturity model for quality in industry 4.0. Total Quality Management
& Business Excellence 33.5-6, 529-547 (2022).
35. Antony, Ji, et al.: An exploration of organizational readiness factors for Quality 4.0: an
intercontinental study and future research directions. International Journal of Quality & Reliability
Management 40.2, 528-606 (2023).
Page 603
SEMIT 2023
36. Sureshchandar, G. S.: Quality 4.0–a measurement model using the confirmatory factor analysis
(CFA) approach. International Journal of Quality & Reliability Management 40.1, 280-303 (2023).
37. Zulfiqar, M., et al.: A readiness assessment of Quality 4.0 in packaging companies: an empirical
investigation. Total Quality Management & Business Excellence, 1-19 (2023).
38. Antony, J, et al.: Benefits, challenges, critical success factors and motivations of Quality 4.0–A
qualitative global study. Total Quality Management & Business Excellence, 1-20 (2022).
39. Maganga, D.P. and Taifa, I.: Quality 4.0 conceptualisation: an emerging quality management concept
for manufacturing industries. The TQM Journal (2022).
40. Maganga, D.P. and Taifa, I.: The readiness of manufacturing industries to transit to Quality
4.0. International Journal of Quality & Reliability Management (2022).
41. Maganga, D.P. and Taifa, I.: Quality 4.0 transition framework for Tanzanian manufacturing
industries. The TQM Journal (2022).
42. Mtotywa, M.: Developing a Quality 4.0 Maturity Index for Improved Business Operational Efficiency
and Performance. Quality Innovation Prosperity 26.2, 101-127 (2022).
43. Nenadál, J, et al.: Quality 4.0 maturity assessment in light of the current situation in the Czech
Republic. Sustainability 14.12, 7519 (2022).
44. Zulqarnain, A., et al.: Developing a quality 4.0 implementation framework and evaluating the
maturity levels of industries in developing countries. Sustainability 14.18 ,11298 (2022).
45. Sütőová, A., et al.: Learning needs determination for industry 4.0 maturity development in automotive
organisations in Slovakia. Quality Innovation Prosperity 24.3, 122-139 (2020).
46. Armani, C.G., et al.: Proposal and application of a framework to measure the degree of maturity in
Quality 4.0: A multiple case study. Advances in Mathematics for Industry 4.0. Academic Press, 131-
163 (2021).
47. Sony, M., et al.: Motivations, barriers and readiness factors for Quality 4.0 implementation: an
exploratory study." The TQM Journal 33.6, 1502-1515 (2021).
48. Kiran, Dr R.: Total quality management: Key concepts and case studies. Butterworth-Heinemann,
(2016).
Page 604
SEMIT 2023
Prioritizing Criteria for Selecting Product Groups in Smart Agriculture Practices
Melike Erdoğan1, Zekeriya Konurhan 2, Melih Yücesan3, Muhammet Gül4

1 Duzce University/Computer Engineering Department, Duzce, Turkey
melikeerdogan@duzce.edu.tr
2 Munzur University/Department of Geography, Tunceli, Turkey
zkonurhan@munzur.edu.tr
3 Munzur University/Department of Emergency Aid and Disaster Management, Tunceli, Turkey
melihyucesan@munzur.edu.tr
4 Istanbul University/Department of Transportation and Logistics, İstanbul, Turkey
muhammetgul@istanbul.edu.tr
Abstract
Smart agriculture refers to a more efficient, sustainable, and smart agricultural system by integrating
traditional agricultural methods with digital technologies. In this sense, it is important to determine the
main factors affecting the selection of products and product groups to use smart agriculture effectively.
For this purpose, a multi-criteria analysis was adopted to prioritize the factors required to identify critical
fruits and vegetables within the framework of smart agriculture. Bayesian-Best-Worst-Method (B-BWM)
based on pairwise comparison was utilized and the outcome of the analysis identifies the most and least
crucial criteria for selecting products or product groups that are most suitable for smart agriculture
practices in Turkey.
Keywords: Bayesian Best Worst Method, MCDM, Smart Agriculture
Introduction
Agriculture 4.0, an integral component of sustainable agriculture, involves the application of Industry
4.0's technological advancements within the agricultural domain [1].Turkey has a significant potential in
the agricultural sector; however, for countries like Turkey, blessed with diverse agricultural resources,
solely relying on these technologies for determining agricultural practices is impractical. This study
introduces a multi-criteria decision-making (MCDM) model designed with a set of criteria to identify the
most suitable products or product groups for smart agriculture practices in Turkey and prioritize them for
investment. The chosen model for this endeavor is B-BWM [2], an enhanced version of BWM tailored
for group decision-making. Initially, the emphasis is on determining the importance weights within a
hierarchy comprising nine criteria. In this context, the model aims to contribute significantly to
pinpointing which products or product groups are best suited for smart agriculture practices.
In the literature-review step, papers adopting bayesian approach and decision-making analysis on smart
agriculture investigated. As a result of many studies examined, it has been determined that there is no
study on determining the products & product groups that should be taken into consideration as a priority
when investing in smart agriculture applications for any country, including Turkey. At this point, it can
be claimed that problem addressed in this study has been considered for the first time in literature.
Proposed Methodology
The best-worst method (BWM) is a technique developed by Rezaei [3] which yields solution by solving
an optimization model and is mostly used to determine the criterion weights in a decision problem. In
recent years, it has been successfully applied to decision problems encountered in academic literature and
practical industrial environments. It is preferred because it involves fewer pairwise comparisons and
produces more consistent results than other pairwise comparison-based MCDM techniques such as AHP.
The first two methods suggested by Rezae [3],[4] are known as nonlinear and linear versions of BWM.
Page 605
SEMIT 2023
Both methods work based on the evaluation of a single decision maker. It does not produce a solution for
group decision. For this reason, it is extended as Bayesian BWM by [2]. In Bayesian BWM, data is
prepared similarly to the traditional BWM. Unlike BWM, Monte Carlo and JAGs are used in the solution.
For details, please refer to [2]. Some transformations of multinomial, Dirichlet, Gamma distributions, and
a Markov-chain Monte Carlo sampling are used in the algorithm. The mathematical models regarding
Bayesian BWM in this paper are solved by MATLAB.
Real Case Analysis

This study aims to identify which products or product groups are suitable for smart agriculture
applications in Turkey using MCDM approaches. The purpose of this study is identifying which products
or product groups are suitable for smart agriculture practices in Turkey through the use of MCDM
approaches. For this purpose, the relevant literature was first reviewed. Among the relevant criteria, those
that could be used in Turkey were selected. With the help of MCDM analysis, a decision was made as to
which of these criteria would be important. Since the data used in the evaluation phase contains
uncertainties and more than one expert is needed, the MCDM approach was integrated with Bayes
Network. Table 2 shows the relevant criteria.
Table 2. Evaluation Criteria
A1: Social Factors
A2: Economic conditions
A3: Topography conditions
A4: Climate conditions
A5: Supply chain
A6: Technology
A7: Environmental conditions
A8: Agricultural management
Results and Discussion

As a result of the Bayesian BWM decision model applied, in which the criteria that should be primarily
considered in smart agriculture practices have been determined, the highest weight value is “Economic
conditions” with 18.47%. It is followed by the “Technology” criterion with 18.42%. The criterion with
the lowest importance weight is “Social factors” with 6.05%. Other criteria have values between 925%
and 13.72%. It has also been confirmed from the credal ranking graphs of B-BWM that the first two
important criteria do not have a clear superiority over each other and are the most important criteria.
According to the results of a B-BWM, the most important criterion was determined as Economic
Conditions, while the least important criterion as Social Criteria. In future studies, it is aimed to compare
the results using different MCDM methods.
References
[1] A. Walter, R. Finger, R. Huber, and N. Buchmann, “Smart farming is key to developing sustainable
agriculture,” Proceedings of the National Academy of Sciences of the United States of America, vol. 114, no. 24.
National Academy of Sciences, pp. 6148–6150, 13-Jun-2017.
[2] M. Mohammadi and J. Rezaei, “Bayesian best-worst method: A probabilistic group decision making
model,” Omega, vol. 96, p. 102075, Oct. 2020.
[3] J. Rezaei, “Best-worst multi-criteria decision-making method,” Omega, vol. 53, pp. 49–57, Jun. 2015.
[4] J. Rezaei, “Best-worst multi-criteria decision-making method: Some properties and a linear model,”
Omega, vol. 64, pp. 126–130, Oct. 2016.
Page 606
SEMIT 2023
SP16: Marketing and E-marketing for Resources Management
Enhancing Repurchase Intentions of Online Music Platform Customers for the Middle East
Tanya Haqi*1, Uğur Şener2
1
Graduate student, Graduate School of Social Sciences, Istanbul Aydin University, ORCID ID: 0009-
0008-4739-0866, Istanbul, Turkey.
2 Assistant Professor, Business Management, Istanbul Aydin University, ORCID ID: 0000-0001-5524-
1707, Istanbul, Turkey.
Extended Abstract
This study investigates the concurrent mediation impact of site design and e-satisfaction on the link
between e-service quality and repurchase intention in the context of on-demand music platforms in Iraq.
Given that on-demand music services are still relatively new in Iraq, it was thought appropriate to look
at how customers choose the platform and what factors influence their decision-making; In order to stay
competitive in the current digital landscape, online music platforms, particularly those in the Middle
East, need to prioritize the quality of their e-services, site design, and e-satisfaction in order to increase
the likelihood that their customers would make more purchases (Khatib et.al, 2019).
In an attempt to disentangle the complex links influencing consumer behavior in this sector, the study
explores the mediating effects of e-satisfaction and website design. It's important because it adds to the
body of knowledge by examining the special parallel mediation relationship between e-satisfaction and
web design in the Iraqi setting, especially in the music sector. As local and international platforms
compete for market supremacy, the findings might provide insightful information for other Middle
Eastern nations.
Next, Principal Component Analysis (PCA) is employed to reduce the number of items according to
their contribution to the variables of the model. Cronbach alpha coefficients are also computed after item
reduction. (M. Ringnér (2008). A regression-based parallel mediation model with three mediators
(Hayes,2018) was performed following the completion of item reduction with PCA and reliability.
The correlation between all the variables was measured before the regression equations were performed,
and the results showed that all of the correlation values were positive including the repurchase intention
and price relationship, which is an interesting finding regarding previous literature and is similar to a
study by Graciola, A. P., Toni, D. D., De Lima, V. Z., & Milan, G. S. (2018).
The results of the regression equations demonstrated that all of the variables were related in some way,
with the exception of the fourth and final equation, which indicated that the repurchase intention was
represented by the variable (y). In contrast to the other equations, the e-SQ variable did not show
statistical significance, with a value of (0.403), but the remaining variables did.
And for direct, total and indirect effects, the total effect of eservice quality on repurchase intention is
positive and statistically significant, indicating that e service quality is associated with an increase in
repurchase intention. However, the direct effect of eservice quality on repurchase intentions is
inconclusive, suggesting that there is no clear direct relationship between them.
In the regression equation and indirect effect, the mediating role of website design is noteworthy, serving
as a mediator between e-service quality and repurchase intention. Comparable positive outcomes were
observed in an Indonesian online shopping setting, suggesting that website design affects e-satisfaction
through service quality (Farisa, A., Oliveira, T., & Rita, P. (2019). Additional research, including
(Tandon, U., Kiran, R., & Sah, A. N. (2017)), corroborates these conclusions by showing that e-
satisfaction, an indirect measure of website design quality, influences repurchase intention. Repurchase
Page 607
SEMIT 2023
intention is influenced by e-satisfaction, which is directly impacted by the quality of a website's design
(Tandon, U., Kiran, R., & Sah, A. N. (2017).
E-service quality positively influences e-satisfaction, mediating the connection to repurchase intention,
consistent with Zhang, X., & Prybutok, V. R. (2005). This aligns with Zhang's findings that e-service
quality positively affects e-satisfaction. Additionally, website design and price mediate the link between
e-service quality and e-satisfaction, echoing the significance of convenience and service quality in
satisfaction, as discussed by Zhang et al.2005 . Contrary to our hypothesis that the price have a big
impact on the repurchase intention, price shows minimal impact on e-satisfaction and repurchase
intention, in contrast to Zeithaml et al.'s findings .
This study contributes to understanding e-service quality, website design, price, and customer behavior
in online music. Findings suggest prioritizing improvements in e-service quality, website design, and
pricing for enhanced customer satisfaction and repurchase intention. Future research may delve into
factors like brand reputation and social influence in the online music industry.
Our findings suggest that online music platforms in Iraq and the larger Middle East should focus on
improving the quality of their e-services, web design, and e-satisfaction in line with earlier research in
order to increase customer repurchase intention.
Keywords: e-service quality, ESERVQUAL, website design, repurchase intention, on-demand music
platforms, parallel mediation, PC.
References
1. Khatib, S. M., Seong, L. C., Chin, W. S., & Tze, H. K. (2019). Factors of E-service Quality among
Malaysian Millennial Streaming Service Users. International Journal of Economics and
Management, 63-77.
2. Ringnér, M. (2008). What is principal component analysis? NATURE BIOTECHNOLOGY, 303-
304.
3. Hayes, A. F. (2018). Introduction to Mediation, Moderation, and Conditional Process Analysis, a
regression-based approach. New York: The Guilford Press.
4. Graciola, A. P., Toni, D. D., De Lima, V. Z., & Milan, G. S. (2018). Does price sensitivity and price
level influence store price image and repurchase intention in retail markets? Journal of Retailing and
Consumer Services, 201-213.
5. Farisa, A., Oliveira, T., & Rita, P. (2019). The impact of e-service quality and customer satisfaction
on customer behavior in online shopping. Elsevier Ltd., 2405-8440.
6. Tandon, U., Kiran, R., & Sah, A. N. (2017). Customer Satisfaction as Mediator Between Website
Service Quality and Repurchase Intention: An Emerging Economy Case. SERVICE SCIENCE, 106-
120.
7. Zhang, X., & Prybutok, V. R. (2005). A Consumer Perspective of E-Service Quality. IEEE
TRANSACTIONS ON ENGINEERING MANAGEMENT, 461-477.
8. Zeithaml, V. A., Parasuraman, A., & Malhotra, A. (2002). Service Quality Delivery Through Web
Sites: A Critical Review of Extant Knowledge. Journal of the Academy of Marketing Science, 362-
37.
Page 608
SEMIT 2023

Brand Credibility
Sepehr Mobayena, Amirhossein Raoof Ghotboddinib*, Farinaz Tanoorianc
a,b
c
EMBA, Faculty of Management & Accounting, Shahid Beheshti University, Tehran, Iran
* Corresponding author: amirraoof@mail.um.ac.ir
Abstract
Given that customers are the foundation of business survival, all businesses need to establish strong
relationships with their customers. Social media marketing activities can have a significant impact on
brand credibility and purchase intention by creating a two-way relationship and exchanging brand-
related information between customers and firms. This correlational-applied survey aimed to investigate
the effect of firm-generated content (FGC) on customer purchase intention through the mediating role
of brand credibility. The study population consisted of all construction companies in Mashhad (Iran),
among which 76 companies were selected as the sample. The reliability of the questionnaires was
assessed using Cronbach’s alpha and composite reliability methods, while the validity of the
questionnaires was examined using convergent validity and confirmatory factor analysis (CFA). The
results confirmed the reliability and validity of all items. Finally, the data were analyzed in Smart-PLS
and SPSS using the partial least squares structural equation modeling (PLS-SEM) method. The results
show that firm-generated content has significant positive effects on both brand credibility and customer
purchase intention. On the other hand, a significant positive relationship was found between brand
credibility and customer purchase intention. Therefore, the mediating role of brand credibility in the
relationship between firm-generated content and customer purchase intention was confirmed. The results
of this study can help companies and marketers better understand how firm-generated content influences
consumer purchase intention for industrial products.
Keywords: Firm-generated content, Brand credibility, Purchase intention, Social media marketing,
User-generated content.
Introduction
The recent development of social media allows consumers to communicate not only with other
consumers, but also directly with companies (Mangold & Faulds, 2009); as a result, consumer interaction
with a company's social media output has shifted from passive to active (Bruhn, Schoenmueller &
Schäfer, 2012). Therefore, a company's social media platforms have developed into a crucial component
of its operations and a vital tool for many customer services (Schivinski & Dabrowski, 2015). Social
media can bring a business a variety of advantages, including (1) an extra channel of communication and
a connection with its loyal customers; (2) a way to reduce costs and improve customer service; and (3) a
tool to boost sales, improve customer satisfaction, and enhance brand image (Sadek, Elwy & Eldallal,
2018).
On the other hand, in a market that is getting more competitive every day, several key factors, such as
advertising credibility, brand credibility, and consumer satisfaction, can affect consumers' buying
decisions (Sutia et al., 2023). Stronger bonds between customers and a brand are made possible by brand
credibility, which offers businesses many advantages in the competitive environment of organizations.
These advantages include high relationship equity, reduced perceived risk, and positive changes in
customer consumption and purchasing behavior (Chin, Isa & Alodin, 2020). Similarly, increased brand
credibility influences customers' brand choice intentions, influences perceptions of high quality, and
reduces information search costs. Furthermore, in highly competitive markets, authentic brands can help
build brand equity and increase purchase intent (Martin-Consuegra et al. 2018). Since the consumer-
brand relationship is controlled and manipulated by the company or brand, it is critical to fully understand
their impacts. As a result, marketers must comprehend how FGC influences consumer behavior and also
Page 609
SEMIT 2023
gain insights into how consumer perceptions and behavior are influenced by brand presentation (Poulis,
Rizomyliotis & Konstantoulaki, 2019).
Despite the rapid growth of social media popularity among consumers and firms, a few studies have
addressed the effect of FGC on consumer purchasing behavior and intention. In addition, since most
previous studies in this field have been conducted on industries such as clothing and mobile network
providers in European countries. In this regard, Schivinski and Dabrowski (2015) state that there is a need
to further investigate the effects of these variables in other countries and industries (Sadek et al., 2018).
To partially bridge this research gap, this study aims to investigate the effects of FGC on brand credibility
and purchase intention of customers of industrial products.
Literature review
Firm-generated content (FGC)
The adaptability of social media enables firms to produce content in a variety of formats and distribute it
across various platforms. Furthermore, these contents are used to present products on the Internet in order
to increase brand visibility. This procedure aims to establish a channel of communication between the
brand and the emerging market and, finally, attract new customers (Poulis et al., 2019). User-generated
content (UGC) and firm-generated content (FGC) are two types of social media marketing strategies that
have been developed to create and exchange brand information with target customers (Perera, Nayak &
Nguyen, 2022). It is important to first define the distinction between FGC and UGC; FGC is any type of
communication controlled by a brand through a marketer or a brand representative (Bruhn et al., 2012),
whereas UGC is a communication channel directly created by users, so it is not under the brand's control
(Schivinski & Dabrowski, 2016). According to Schivinski and Dabrowski (2016), FGC is a type of
advertising that is entirely created and controlled by the company. Furthermore, Mangold and Faulds
(2009) argue that FGC is a type of communication that a company carefully develops in order to convey
its values. It is noteworthy that FGC depends on consumer feedback in a variety of formats; accordingly,
consumers may “like” FGC or be asked to comment on or share them. Therefore, the brand's evaluation
may improve or degrade depending on whether consumers respond favorably or unfavorably to the
content (Ceballos et al., 2016; Kumar et al., 2016).
Brand credibility
Brand credibility refers to customers' perception of the brand's ability, motivation, and willingness to
provide accurate and truthful information and deliver what the brand promises to its customers (Bougoure
et al, 2016). It is also defined as the consumer's belief in the reliability of information provided by a
company's brand (Swait & Erdem, 2007). In this process, the customer needs to comprehend that a brand
has the ability (i.e. expertise) and willingness (i.e. trustworthiness) to consistently deliver on its promises
(Martin-Consuegra et al. 2018). A customer's perception of a brand's expertise, trustworthiness, and
attractiveness may influence his or her evaluation of the brand (Chin et al., 2020). According to Anridho
and Liao (2013), when a brand has a high level of perceived expertise, trustworthiness, and attractiveness,
consumers evaluate its products positively. This is because high credibility indicates higher product
quality, value, and worthiness, which leads to a positive consumer attitude. Consumers perceive less risk
and gather less information during the decision-making process as credible brands demonstrate their
product positioning. As a result, brand credibility is an important factor in shaping a company's overall
image (Martin-Consuegra et al., 2018).
Purchase intention
According to Kim and Koo (2012), purchase intention is defined as a combination of consumer interest
and the likelihood that they will buy a product. The intention to purchase a particular brand, product, or
service, on the other hand, necessitates a thorough assessment of all products, services, and brands
provided by rivals (Teng, Laroche, & Zhu, 2007). Purchase intention refers to the final stage of the
Page 610
SEMIT 2023
purchasing process in which consumers decide whether or not to buy a specific brand's product (Grewal,
Monroe & Krishnan, 1998). Many studies have shown that purchase intention is highly influenced by
preferences and attitudes toward a brand or a product (Irelli & Chaerudin, 2020). As a result, measurement
of purchase intention helps us to predict consumer behavior based on their attitudes. When companies
provide features that meet a consumer's need, it results in a purchase intention. Furthermore, purchase
intention is an attitudinal variable used to assess customers' future involvements in a brand; thus, purchase
intention can eventually lead to actual purchase behavior. Therefore, the higher the purchase intention,
the greater the consumer's willingness to purchase a product or service (Martin-Consuegra et al. 2018).
Effects of FGC on PI
The assumption is that consumers have the intention to purchase before the actual buying process.
Therefore, consumer purchase intention predicts the real purchase behavior of consumers (Makudza,
Mugarisanwa & Siziba, 2020). Poturak and Softic (2019) argued that a discernible relationship can be
identified between FGB and consumer purchase intention. Permatasari and Kuswadi (2018), thus
indicate that through social media marketing, consumer’s sequential process has been reduced as
consumers can skip other stages and develop an intention to buy which is more likely to lead to purchase
behavior. Poturak and Softic (2019) concluded that FGC is considered valid and instructive since it is
developed by trained professional employees. This was supported by Rad, Joyin & Moyada (2011), who
found that formal reviews which were posted by the company were predictive of consumer purchase
intention and action. Kumar et al. (2016) argued that FGB has a strong, positive effect on PI. They
believed that firms that directly generate content and communicate through social media affect
consumers’ purchasing. Thus, our first hypothesis is:
H1: FGC has a positive impact on PI.
Effects of FGC on BC
The elaboration likelihood model (ELM) was proposed concerning social media and electronic word of
mouth (eWOM), with a focus on how marketers can affect the attitudes of consumers toward a brand
through persuasive situations, such as online communities, weblogs, and online marketing content
(Perera, Nayak & Nguyen, 2022). The ELM suggests that social media users have a series of elaboration
approaches to process the existing messages by using cognitive thinking (central route) and/or by only
following simple clues (peripheral route) for decision-making (Goh & Chi 2017). When recipients of an
advertising message have both the motivation and ability to consider detailed information in a message,
persuasion occurs through the central route (Chang, Yu & Lu, 2015). It requires them to think hard about
issue-related arguments in a message and reflect on the relative competence and relevance of those
arguments (Cyr et al., 2018). In contrast, the peripheral route requires lower cognitive efforts, where
individuals either accept or reject a message without active thinking about the attributes of the issue of
the object of consideration (Liao and Huang, 2021). Instead of high cognitive effort, message recipients
rely on various cues that allow them to make quick decisions (Shi et al., 2018). This is due to the lack of
motivation or ability to cope with detailed information in a message. On the other hand, FGC is processed
through the peripheral route where individuals feel that they are emotionally connected to the brand
(Perera et al., 2022).
If the brand creates a meaningful image through FGC on social media such as the introduction of physical
facilities, firm's ranking in the market, activities with shareholders, and offering superb services, this can
represent the ability and expertise of the brand to offer superior performance (Wang, Kao, and
Ngamsiriudom, 2017). In that case, users perceive the sources of the FGC information as credible, non-
commercial, unbiased information. As such, individuals use some cues that the brand shared in FCGs to
form their attitudes and make the decision toward the brand (Perera et al., 2022). A significant brand
image shared by FGC on social media provides cues for the ability and expertise of the brand to offer
superior performance over other competitive brands (Wang et al., 2019). As a result, the second
hypothesis can be formulated as follows:
Page 611
SEMIT 2023
H2: FGC has a positive impact on BC.
Effects of BC on PI
It requires consumers to understand that a brand has both the capability (i.e., expertise) and willingness
(i.e., trustworthiness) to continuously fulfill its promises. As the trustworthiness and expertise of a brand
are based on the cumulative effects of all the previous marketing measures of the brand, it is not surprising
that BC reflects the stability of mixed marketing strategies through brand investments, such as
advertising. As credible brands represent their product situations, consumers perceive lower risks and
collect lower information during the decision-making process; consumers tend to obtain information to
minimize perceived risk before the purchase (Martin-Consuegra et al. 2018). To increase the purchase
intention, it is required to enhance consumer trust in a brand by offering credible information on the
product or direct product experience (Zajonc & Markus, 1982). The perception of credibility in an
information source can affect consumers’ satisfaction with the product and help the brand persuade PI
(Bahtar & Muda, 2016). Consumers need FGC since it contains important factors that can affect their PI,
such as expertise, relevance, risk reduction, positive value, usefulness, and credibility (Irelli & Chaerudin,
2020). Finally, it can be said that strong credibility cues conveyed by a brand to induce a positive attitude
in consumers have the capacity to affect PI (Maia Bairrada et al., 2021). Therefore, it can be assumed that
BC has a direct, positive effect on PI. Therefore, the proposed hypothesis is as follows:
H3: BC has a positive impact on PI.
Effects of FGC on PI mediated by BC

Consumers who realize that a brand is credible enough would consider it in their purchase decisions since
BC provides possible data for positioning in customers’ minds (Cuong, 2020). If messages or comments
generated by a brand have high credibility, users have a higher perception of the usefulness of such
messages or comments, affecting their brand attitude and purchase decision (Wu & Wang, 2011). FGC
helps brands retain and enhance relationships with their target consumers. At the same time, it can be
observed to have a strong, positive effect on sales and, therefore, profitability. Furthermore, this new
information source contributes to brand popularity (Poulis et al., 2019). Trustworthiness refers to
customers’ trust in information offered through different sources in terms of authenticity, sincerity, and
integrity (Wang & Scheinbaum, 2017). Furthermore, Thomas and Johnson (2017) showed that the
expertise source has direct and mediating effects on PI. In this regard Martin-Consuegra et al. (2018)
observed that BC has a significant effect on PI. Kareem and Venugopal (2023) also found that the attitude
toward BC has a positive effect on PI. So it can be assumed that when consumers believe that a company
can fulfill its promises (brand credibility), they have more favorable perceptions about the brand,
recommend it to others, and are probably more interested to buy the products of that brand. Consequently,
the hypothesis is as follows:
H4: BC mediates the relationship between FGC and PI.
Methodology
The statistical population consisted of construction firms in Mashhad, Iran, which could not be accurately
determined. However, based on field surveys, it is estimated that the number of relevant contracting
companies in this research, involved in ongoing construction and design projects, is approximately
between 70 to 100 companies. To determine the sample size, an initial sample of 30 units was collected
from the target research population and fitted using Smart-PLS software from the results of this fitting,
the degrees of freedom index and fitness index were obtained subsequently using these two indices, the
appropriate sample size was calculated to be 70 companies. Finally, after distributing and collecting
questionnaires to the representatives of each company, 76 questionnaires were collected. To collect the
data, a questionnaire approach was adopted. The three-item questionnaire of Bruhn et al. (2012) was used
to measure FGC, the six-item questionnaire of Baek and King (2011) was utilized for measuring BC, and
Page 612
SEMIT 2023
the three-item questionnaire of Grewal et al. (1998) was exploited to measure PI. Table 1. reports the
descriptive and demographic data.
Table 1. demographic data
Variable Category Frequency %
Age (in years) 18-29 13 17.10
30-34 21 27.63
35-39 16 21.05
40-44 15 19.73
Above 45 11 14.47
Work experience (in
Less than 4 18 23.68
years)
5-9 23 30.26
10-14 12 15.78
15-19 14 18.42
Above 20 9 11.84
Company’s field contract work 38 50
Executive / project
22 28.94
manager
Design and consulting 16 21.05
Time spending on
social 1-2 28 38.84
Media (h/day)
2-3 19 25
3-4 16 21.05
Above 4 13 17.10
Measurement model
The present study analyzed the measurement model to evaluate the reliability, composite reliability (CR),
and average variance extracted (AVE) of constructs. Cronbach’s alpha (CA) and CR were employed to
measure reliability. CA and CR were calculated to be 0.877 and 0.924 for FGC, 0.848 and 0.887 for BC,
and 0.810 and 0.887 for PI, respectively. According to Hair, Ringle and Sarstedt (2011) CA and CR are
required to be greater than 0.70; CA and CR were found to be in the acceptable ranges. The Fronell-
Larcker criterion and Heteroit-Monotrait (HTMT) ratio were assessed to test discriminant validity. The
HTMT ratio has recently been preferred over the Fronell-Larcker criterion (Baloch et al., 2017; Henseler,
Hubona & Ray, 2016). As shown in Table 3, the Fronell-Larcker tests shows higher correlations between
the variables. The HTMT ratio was below the threshold value of 0.090 (see Table 4). Furthermore,
convergent validity was evaluated to obtain AVE, with all AVE values exceeding the threshold of 0.50
(0.569, 0.803, and 0.725 for BC, FGC, and PI), as proposed by Henseler et al. (2016) (see Table 2). The
variance inflation factor (VIF) was also analyzed to evaluate multicollinearity in the data. Aiken, West
and Reno (1991) proposed that the VIF should be smaller than 10, and this study found the VIF values to
be in this range. This suggests that the data had no multicollinearity (see Table 2).
Page 613
SEMIT 2023
Table 2. Measurement model.

Item Outer
Construct Loading VIF CA CR AVE
Code Weights
Brand Credibility (BC) 0.848 0.887 0.569
BC1 0.697 0.213 1.81
BC2 0.670 0.155 1.57
BC3 0.825 0.223 3.23
BC4 0.819 0.254 2.85
BC5 0.693 0.199 1.67
BC6 0.805 0.271 2.10
Firm Generated Content (FGC) 0.877 0.924 0.803
FGC1 0.856 0.330 2.04
FGC2 0.919 0.394 2.82
FGC3 0.913 0.389 2.72
Purchase Intention (PI) 0.81 0.887 0.725
PI1 0.847 0.413 1.71
PI2 0.807 0.331 1.72
PI3 0.897 0.427 2.14
Table 3. Discriminant validity (latent variable correlation and square root of AVE).
BC FGC PI
BC 0.754
FGC 0.672 0.896
PI 0.731 0.663 0.851
Note: Brand Credibility (BC); Firm Generated Content (FGC); Purchase Intention (PI).
Table 4. HTMT (Heterotrait–Monotrait ratio).

BC FGC
FGC 0.878
PI 0.850 0.779
Note: Brand Credibility (BC); Firm Generated Content (FGC); Purchase Intention (PI).
Table 5. Saturated model results

Construct R2 Adj. R2 Q2
Brand Credibility 0.597 0.591 0.327
Purchase Intention 0.558 0.546 0.383
Note: predictive relevance (Q2); determination of coefficient (R2).
Structural model evaluation

SEM was evaluated using 5000 bootstraps in Smart PLS. The coefficient of determination (R2) should be
greater than 0.1 (1998). It was found that 59.8% of BC variance was explained by FGC, and 55.8% of PI
variance can be explained by FGC and BC. Furthermore, Q2 should be higher than 0. Thus, according to
Falk and Miller (1992) the results of the present study were significant, and the predictive relevance of
the model was obtained (see Table 5).
Page 614
SEMIT 2023
Structural equation modeling

The PLS-SEM findings show that (H1) Firm Generated Content orientation has positive and significant
effects on Purchase Intention (β = 0.246, t = 2.208, p = <0.05( . (H2) Firm Generated Content orientation
has positive and significant effects on Brand Credibility (β = 0.772, t = 19.526, p = <0.05). Moreover,
(H3) Brand Credibility orientation has positive and significant effects on Purchase Intention (β = 0.418,
t = 5.348, p = <0.05). Thus, we accepted the direct relationships of H1, H2 and H. Moreover, the results
show that (H4) the indirect effect of Brand Credibility between the relationship of Firm Generated
Content and Purchase Intention were positive and significant, with β = 0.418, t = 5.088, p = <0.05 (see
Table 6). The past literature suggests that the indirect relation particularly includes a third variable, which
acts as an intermediating variable in the relationships between dependent and independent variables.
Technically, the effects of the independent variable (X) on the dependent variable (Z) are intermediated
by a third variable (Y) [3]. Moreover, the direct effects of Firm Generated Content on Brand Credibility
(β = 0.772, t = 19.526), BC on PI (β = 0.541, t = 5.348) and Firm Generated Content on Purchase Intention
(β = 0.246, t = 2.208) were positive and significant, and the indirect effects of Brand Credibility between
the relationship of Firm Generated Content and Purchase Intention were significant with β = 0.418, t =
5.088, which shows partial mediation in the model.
Figure1. PLS-SEM showing t-statistic values
Page 615
SEMIT 2023
Figure 2. PLS-SEM showing Path coefficients
Table 6. Hypothesis constructs.

Effects Relationships Beta Mean (STDEV) t-Value Decision f2
H1 FGC → PI 0.246 0.238 0.111 2.208* yes 0.055
H2 FGC → BC 0.772 0.777 0.04 19.526* yes 1.478
H3 BC → PI 0.541 0.551 0.101 5.348* yes 0.267
H4 FGC → BC → PI 0.418 0.428 0.082 5.088* yes
Note: * p < 0.05, Brand Credibility (BC); Firm Generated Content (FGC); Purchase Intention (PI).
Discussion and Conclusion

This study mainly sought to demonstrate the direct and indirect effects of FGC on BC and PI for B2B
and industrial product customers. It was found that FGC shared on social media platforms has a
significant effect on PI. This is consistent with Balakrishnan et al. (2014), Poulis et al. (2018), Kareem
and Venugopal (2023) and Irelli and Chaerudin, (2020). So, considering the positive effects of FGC on
consumers’ decision-making process, this should be considered and implemented as a key attribute of
marketers’ marketing communication. Moreover, our findings are in accordance with Perera et al. (2022)
and Cheng and Khan (2017) that showed FGC has a direct impact on BC; thus, according to these studies,
firm-generated content created by brands is a powerful and influential type of marketing communication
that can affect consumers' intention to make a purchase.
Our research supports the directly impact of BC on PI (Martin-Consuegra et al. 2018; Cuong, 2020); that
means consumers tend to purchase products when they receive expertise and originality in the
advertisement content source. Finally, the findings of our research indicated the effect of FGC on PI with
the mediation of BC. This is consistent with Cheng and Khan (2017), who argued that when FGC seems
less commercial, consumers are more likely to purchase products. The findings imply that consumers
tend to purchase products when they receive expertise and originality in the advertisement content source.
Page 616
SEMIT 2023
Social media should not be used only as a tool since rival brands may also use it; it should be considered
as an opportunity that helps brands enhance their BC and increase PI among consumers.
Our findings suggest that brand marketers should prioritize trustworthiness information exchange and
BC in their advertising messages on social media when creating FGC. In addition, firms should design
social media content to affect consumers’ attitudes toward brands since the quality and credibility of their
messages are important factors that affect consumers’ behavior after they are exposed to such messages
(Chaiken, 1980). Therefore, it is recommended that marketing managers should not make promises in
their FGC on social media that cannot be fulfilled by the firm as it would harm BC and decrease customer
satisfaction and PI in the long term. Furthermore, brand managers should evaluate tools to induce
perceptions in consumers’ minds through FGC (Mangold and Faulds, 2009). This, in turn, proactively
triggers eWOM among consumers and potentially has a positive effect on their PI.
References
Aiken, L. S., West, S. G., & Reno, R. R. (1991). Multiple regression: Testing and interpreting
interactions. sage.
Akar, E., & Topçu, B. (2011). An examination of the factors influencing consumers' attitudes toward
social media marketing. Journal of internet commerce, 10(1), 35-67.
https://doi.org/10.1080/15332861.2011.558456
Anridho, N., & Liao, Y. K. (2013). The mediation roles of brand credibility and attitude on the
performance of cause-related marketing. International Journal of Social Sciences & Education, 4(1),
226-276.
Bahtar, A. Z., & Muda, M. (2016). The impact of User–Generated Content (UGC) on product reviews
towards online purchasing–A conceptual framework. Procedia Economics and Finance, 37, 337-
342. https://doi.org/10.1016/S2212-5671(16)30134-4
Bougoure, U. S., Russell-Bennett, R., Fazal-E-Hasan, S., & Mortimer, G. (2016). The impact of service
failure on brand credibility. Journal of Retailing and Consumer Services, 31, 62-71.
https://doi.org/10.1016/j.jretconser.2016.03.006
Bruhn, M., Schoenmueller, V., & Schäfer, D. B. (2012). Are social media replacing traditional media in
terms of brand equity creation?. Management research review, 35(9), 770-790.
https://doi.org/10.1108/01409171211255948
Ceballos, M., Crespo, Á. G., & Cousté, N. L. (2016). Impact of firm-created content on user-generated
content: using a new social media monitoring tool to explore Twitter. In Rediscovering the
Essentiality of Marketing: Proceedings of the 2015 Academy of Marketing Science (AMS) World
Marketing Congress (pp. 303-306). Springer International Publishing. https://doi.org/10.1007/978-3-
319-29877-1_61
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus
message cues in persuasion. Journal of personality and social psychology, 39(5), 752.
https://doi.org/10.1037/0022-3514.39.5.752
Chang, Y. T., Yu, H., & Lu, H. P. (2015). Persuasive messages, popularity cohesion, and message
diffusion in social media marketing. Journal of Business Research, 68(4), 777-782.
Cheng, E., & Khan, N. (2017, May 30). Game of Trust: Brand-generated content vs Consumer-generated
advertising - A qualitative study of Arla Foods. Master Thesis, Uppsala University, Business Studies,
Uppsala.
Chin, P. N., Isa, S. M., & Alodin, Y. (2020). The impact of endorser and brand credibility on consumers’
purchase intention: the mediating effect of attitude towards brand and brand credibility. Journal of
Marketing Communications, 26(8), 896-912. https://doi.org/10.1080/13527266.2019.1604561
Page 617
SEMIT 2023
Chin, W. W. (1998). The partial least squares approach to structural equation modeling. Modern methods
for business research, 295(2), 295-336.
Cuong, D. T. (2020). The impact of brand credibility and perceived value on customer satisfaction and
purchase intention at fashion market. power, 12(3), 691-700.
https://doi.org/10.5373/JARDCS/V12SP3/20201308
Cyr, D., Head, M., Lim, E., & Stibe, A. (2018). Using the elaboration likelihood model to examine online
persuasion through website design. Information & management, 55(7), 807-821.
https://doi.org/10.1016/j.im.2018.03.009
Erdem, T., & Swait, J. (2004). Brand credibility, brand consideration and choice. Journal of Consumer
Research, 31(1), 191–199. https://doi.org/10.1086/383434
Falk, R. F., & Miller, N. B. (1992). A primer for soft modeling. University of Akron Press.
Fornell, C., & Larcker, D. F. (1981). Structural equation models with unobservable variables and
measurement error: Algebra and statistics. https://doi.org/10.2307/3150980
Goh, D., & Chi, J. (2017). Central or peripheral? Information elaboration cues on childhood vaccination
in an online parenting forum. Computers in Human Behavior, 69, 181-188.
https://doi.org/10.1016/j.chb.2016.11.066
Grewal, D., Monroe, K. B., & Krishnan, R. (1998). The effects of price-comparison advertising on
buyers’ perceptions of acquisition value, transaction value, and behavioral intentions. Journal of
marketing, 62(2), 46-59. https://doi.org/10.2307/1252160
Hair, J. F., Ringle, C. M., & Sarstedt, M. (2011). PLS-SEM: Indeed a silver bullet. Journal of Marketing
theory and Practice, 19(2), 139-152. https://doi.org/10.2753/MTP1069-6679190202
Henseler, J., Hubona, G., & Ray, P. A. (2016). Using PLS path modeling in new technology research:
updated guidelines. Industrial management & data systems, 116(1), 2-20.
https://doi.org/10.1108/IMDS-09-2015-0382
Hyun Baek, T., & Whitehill King, K. (2011). Exploring the consequences of brand credibility in
services. Journal of Services Marketing, 25(4), 260-272.
https://doi.org/10.1108/08876041111143096
Irelli, R. S., & Chaerudin, R. (2020). Brand-generated content (BGC) and consumer-generated
advertising (CGA) on Instagram: the influence of perceptions on purchase intention. KnE Social
Sciences, 882-902. https://doi.org/10.18502/kss.v4i6.6649
Kareem, S. A., & Venugopal, P. (2023). Social Media Influencers’ Traits and Purchase Intention: A
Moderated Mediation Effect of Attitude Towards Brand Credibility and Brand Familiarity. FIIB
Business Review, 0(0). https://doi.org/10.1177/23197145231162257
Kim, A. J., & Ko, E. (2012). Do social media marketing activities enhance customer equity? An empirical
study of luxury fashion brand. Journal of Business Research, 65, 1480–1486.
Kumar, A., Bezawada, R., Rishika, R., Janakiraman, R., & Kannan, P. K. (2016). From social to sale:
The effects of firm-generated content in social media on customer behavior. Journal of
marketing, 80(1), 7-25. https://doi.org/10.1509/jm.14.0249
Liao, L., & Huang, T. (2021). The effect of different social media marketing channels and events on
movie box office: An elaboration likelihood model perspective. Information & Management, 58(7),
103481. https://doi.org/10.1016/j.im.2021.103481
Maia Bairrada, C., Fontes da Costa, J., Santos, R. M., & Coelho, A. (2021). Determinants and
consequences of brand credibility: a case study in the pharmaceutical industry. International Journal
of Pharmaceutical and Healthcare Marketing, 15(2), 282-297. https://doi.org/10.1108/IJPHM-05-
2020-0037
Page 618
SEMIT 2023
Makudza, F., Mugarisanwa, C., & Siziba, S. (2020). The effect of social media on consumer purchase
behaviour in the mobile telephony industry in Zimbabwe. Dutch Journal of Finance and
Management, 4(2), em0065. https://doi.org/10.29333/djfm/9299
Mangold, W. G., & Faulds, D. J. (2009). Social media: The new hybrid element of the promotion
mix. Business horizons, 52(4), 357-365. https://doi.org/10.1016/j.bushor.2009.03.002
Martin-Consuegra, D., Faraoni, M., Díaz, E., & Ranfagni, S. (2018). Exploring relationships among brand
credibility, purchase intention and social media for fashion brands: A conditional mediation
model. Journal of Global Fashion Marketing, 9(3), 237-251.
https://doi.org/10.1080/20932685.2018.1461020
Perera, C. H., Nayak, R., & Nguyen, L. T. V. (2022). The impact of social media marketing and brand
credibility on higher education institutes’ brand equity in emerging countries. Journal of Marketing
Communications, 1-26. https://doi.org/10.1080/13527266.2022.2086284
Permatasari, A. and Kuswadi, E. (2018). The impact of social media on consumers’ purchase intention:
A study of ecommerce sites in Jakarta, Indonesia. Review of Integrative Business and Economics
Research, 6(1), 321-335. https://doi.org/10.1504/IJTTC.2018.097422
Poturak, M. and Softić, S. (2019). Influence of social media content on consumer purchase intention:
mediation effect of brand equity. Eurasian Journal of Business and Economics, 23(12), 17-43.
https://doi.org/10.17015/ejbe.2019.023.02
Poulis, A., Rizomyliotis, I., & Konstantoulaki, K. (2019). Do firms still need to be social? Firm generated
content in social media. Information Technology & People, 32(2), 387-404.
https://doi.org/10.1108/ITP-03-2018-0134
Rad, R., Joyin, R. and Moyada, C. (2011). Modelling social commerce. Journal of Business Management,
11(2), 37-51.
Sadek, H., Elwy, S., & Eldallal, M. (2018). The impact of social media brand communication on
consumer-based brand equity dimensions through Facebook in fast moving consumer goods: The
case of Egypt. Journal of Business and Retail Management Research, 12(2).
https://doi.org/10.24052/JBRMR/V12IS02/TIOSMBCOCBBEDTFIFMCGTCOE
Schivinski, B., & Dabrowski, D. (2016). The effect of social media communication on consumer
perceptions of brands. Journal of Marketing Communications, 22(2), 189-214.
https://doi.org/10.1080/13527266.2013.871323
Schivinski, B., & Dabrowski, D. (2015). The impact of brand communication on brand equity through
Facebook. Journal of Research in Interactive Marketing, 9(1), 31-53.https://doi.org/10.1108/JRIM-
02-2014-0007
Sutia, S., Riadi, R., Tukirin, T., Pradipta, I & Fahlevi, M. (2023). Celebrity endorsement in social media
contexts: understanding the role of advertising credibility, brand credibility, and brand satisfaction.
International Journal of Data and Network Science, 7(1), 57-64.
https://doi.org/10.5267/j.ijdns.2022.12.006
Thomas, T., & Johnson, J. (2017). The impact of celebrity expertise on advertising effectiveness: The
mediating role of celebrity brand fit. Vision: The Journal of Business Perspective, 21(4), 367–374.
https://doi.org/10.1177/0972262917733174
Wang, S. W., Kao, G. H. Y., & Ngamsiriudom, W. (2017). Consumers' attitude of endorser credibility,
brand and intention with respect to celebrity endorsement of the airline sector. Journal of Air
Transport Management, 60, https://doi.org/10.1016/j.jairtraman.2016.12.007
Wang, S. W., & Scheinbaum, A. C. (2017). Enhancing brand credibility via celebrity endorsement.
Journal of Advertising Research, 58(1), 16–32. https://doi.org/10.2501/jar-2017-042
Page 619
SEMIT 2023
Wu, P. C. S., & Wang, Y. C. (2011). The influences of electronic Word-of-mouth message appeal and
message source credibility on brand attitude. Asia Pacific Journal of Marketing and Logistics, 23,
448–472. https://doi.org/10.1108/13555851111165020
Zajonc, R. B., & Markus, H. (1982). Affective and cognitive factors in preferences. Journal of consumer
research, 9(2), 123-131.
Page 620
SEMIT 2023

Management Strategy on Creativity of Core Members in Live Broadcasting Teams - The
Ding Haoqing 12
1
Innovation College, North-Chiang Mai University, Chiang Mai, 50230, Thailand
2
Guangxi Vocational College of Water Resources and Electric Power, China
Abstract: To study the influence of high-involvement human resource management (HRM) strategies on
the creativity of the core members of live teams, the article constructs a research model with “resource
security”, “typical model”, and “empowerment support” as independent variables and “creativity” as the
dependent variable in the organizational innovation climate. A total of 106 team anchor samples and 506
individual anchor samples were collected through questionnaire research, and the data were statistically
analyzed using SPSS 26.0 and Mplus 8.3. The exploratory factor analysis and validation factor analysis
reveal that resource security, typical demonstration, and empowerment support play mediating roles
between high-involvement HRM strategies and creativity, while anchor traits play moderating roles
between them. Finally, specific recommendations are made in the recruitment of anchors and the
formation of live teams based on the study’s results.
Keywords: Creativity, Organizational Innovation Climate, High-Involvement Human Resource
Management.
1 Introduction
With the rapid changes in mobile Internet technology, live industry laws and regulations continue to
improve. Additionally, from the beginning, the network anchor will only show talent as a “net red
anchor”, which is gradually vertically subdivided into different types of anchors, such as e-commerce
anchors, game anchors, travel anchors, and so on. As the core of the live broadcast mode of electric
business, the electric business anchor is highly valued. As an emerging marketing model, e-commerce
live broadcasting also successfully integrates the professionalism of expert anchors and the huge clientele
of e-commerce platforms, thus having created numerous sales myths in just a few years since its creation.
Therefore, it is crucial for the current, increasingly competitive environment of traditional e-commerce
enterprises to develop marketing strategies and improve their competitiveness. Simultaneously, however,
the uneven personal abilities of e-commerce anchors, as well as live content homogeneity caused by
unfair competition, has caused much trouble for consumers during live purchases, to a certain extent
affecting consumers’ overall impression of e-commerce anchors and negatively impacting e-commerce
enterprises’ continued live business.
Goldsmith believes that an opinion leader influences others in a variety of ways, such as by
recommending information to others, giving directions to others, telling others what they think is good,
and so on, all of which can directly impact the influenced in terms of product searches and purchases [1].
Flyn et al. argue that consumers trust the recommendations of opinion leaders because they believe that
the advice or information that opinion leaders provide is directly related to the products they need and
that opinion leaders often exist as consumers and do not represent businesses, so consumers are more
likely to trust the information they provide over advertisements that businesses place [2]. However,
Gefen’s proposal to build trust through communication and other means to gain mutual understanding is
more applicable to live e-commerce platforms [3]. Studies on trust and purchase intention also show that
consumers’ purchase intention can increase with trust in opinion leaders [4, 5]. In general, however, there
is a relative lack of in-depth explorations into e-commerce anchors themselves in the e-commerce live
streaming context of the academic community. Therefore, this paper conducts an in-depth study on
various dimensions of e-commerce anchors’ trait attributes, as well as discusses the intrinsic paths and
mechanisms involved in anchor trait participation in regulating anchor engagement and the organizational
innovation climate. Finally, this study researches the cross-level impact mechanism of high-engagement
human resource management (HRM) strategies on live creativity.
Page 621
SEMIT 2023
In the context of the knowledge economy, employee creativity has become a core element for
companies’ continuous innovation and achievement of a competitive advantage, which are crucial for
organizational survival and development [6]. Employee creativity refers to the novel and practical ideas
or thoughts that individuals propose to achieve task goals and is influenced by individual characteristics
such as emotions, motivation, and cognition, as well as environmental factors such as the team,
organization, and network [7]. Among them, the impact on employee creativity is gradually receiving
attention from academia and industries. However, studies have been conducted mainly from a team
leadership style [8] or team composition [9] perspective, which is limited. In practice, teams usually have
to face complex and changing market and economic environments and must make dynamic adjustments
in personnel in response to these changes. Therefore, it is important to focus on the impact of HRM
practices on employee creativity in terms of theoretical significance and practical value.
Numerous scholars have argued that high-engagement HRM practices have a positive effect on
creativity, such as Shung Jae Shin et al., who argue that high-engagement HRM practices have a
significant cross-level effect on individuals’ intrinsic work motivation and work creativity, with intrinsic
work motivation playing a mediating role [10]. Additionally, Muhammad Yasir et al. argue that three
dimensions of HRM practices, namely empowerment, motivation enhancement, and opportunity
enhancement, have a direct impact on employees’ functional flexibility and innovative work behavior
[11]. Maarten Renkema et al. argue that high-engagement HRM practices have a positive effect on
innovative behavior, with affective commitment and autonomy playing fully mediating roles in this
relationship.
In addition, team leaders impact high-engagement management. Olivier Doucet et al. identify three
statistically significant interactions between transformational leadership and high-engagement
management practices: information sharing and power sharing practices act as leadership enhancers,
while skill development practices act as leadership substitutes [12]. Furthermore, companies that value
human resources are more likely to have better team effectiveness, as Johngseok Bae et al. argue. They
find that companies that score higher in valuing HRM and viewing people as sources for a competitive
advantage are more likely to have a high-engagement HRM strategy, and these variables have a positive
effect on firm performance. Therefore, companies with a high-engagement HRM strategy have better
performance [13].
Previous studies have shown that the organizational innovation climate is an effective predictor of
employee creativity and organizational innovation, such as Michael L.A. and Hsu et al. Using a large
sample of 16 organizations and a cross-level analysis with employee psychological competencies as a
mediating variable, the study concludes that both the organizational innovation climate and employee
psychological capital have a significant effect on employee innovation behavior [14]. Moreover, previous
studies have shown that the work climate within an organization significantly impacts the level of
innovative productivity of the organization, such as Scott G. and Isaksen et al., who conclude that the
creative climate plays a key role in the relationship between leadership behavior and innovative
productivity [15]. Additionally, Rui Sun et al. conclude that the organizational innovation climate and the
external motivation of employees have a positive effect on employees’ innovation behavior [16].
In developed markets, traditional innovative behavior is directed at customers. Since these customers
have the greatest purchasing power, innovation in developed markets targets affluent customers first,
aiming to provide them with products with rich functionality. The business models and distribution
methods of traditional innovation are also based on the rich resources of the West. However, traditional
innovation requires costly product design and development, which increase the cost of producing the
product. Further, the majority of the world’s population, which has limited spending power, still has
simple needs for products, which is why the idea of plain innovation emerged. The concept of plain
innovation originated in the 1950s with the “appropriate technology” movement, which emphasized the
need to specify the technology needed based on the needs of local customers and to develop products that
fit local economic and resource conditions [17]. To match the implementation of plain innovation,
companies had to change their traditional organizational structure. Zeschky et al. argue that companies
must adapt their organizational structure and develop organizational capabilities to facilitate the
Page 622
SEMIT 2023
development of plain innovation [18]. Moreover, Sehgal et al. argue that cross-functional teams, non-
traditional supply chains, and top-down support from leaders in the organization are more conducive to
plain innovation. Among them, the implementation of cross-functional teamwork allows for a
comprehensive consideration of all cost reduction possibilities; non-traditional supply chains help to
strengthen cooperation between firms in the supply chain, leading them to jointly set cost targets and cost
structures to optimize costs across the system; and commitment from the top of the organization facilitates
the development of common goals and drives plain innovation [19]. Therefore, we conclude that the core
members of the live team can effectively promote the development of individual and team creativity,
while the organizational innovation atmosphere of the team can, in turn, promote the creativity of
individual employees.
While previous studies, whether focused on traditional innovation or plain innovation, have been
conducted microscopically, the creativity study in this paper combines both team (organizational
innovation climate) and individual (creativity) levels in a cross-level analysis, finally concluding that the
team level and individual level influence each other’s roles.
3. Research Model and Hypotheses

3.1 Theoretical Foundation——Human–environment Matching Theory
The creativity component model that Amabile proposed suggests that professional-related
knowledge skills, creativity-related knowledge skills, and motivation are the three elements that
individuals need to achieve creativity [20]. These three elements all require a “correlation” between
individual qualities and the task, i.e., a match. Further, person–environment match theory suggests that
individual and environmental factors influence individual behavioral outcomes through the individual’s
subjective perception of the person–environment match [21].
3.2 Research hypotheses
3.2.1 The mediating role of the organizational innovation climate
Innovative activities require a great deal of effort, time, and risk and deplete the psychological,
emotional, and material resources of the innovator. According to resource conservation theory, when
faced with stress, individuals tend to conserve existing resources if they do not have access to sufficient
resources. A good organizational innovation climate can, thus, help employees to meet their needs for
resources, such as financial and technical resources and leadership support, as well as the need for
knowledge and information exchange among team members, thus alleviating employees’ stress caused
by resource depletion and promoting employees’ innovative behavior [22].
First, the technical, financial, and equipment resources that the organization provides to support
innovation comprise an important material resource base for employees’ innovative behavior. Shanker et
al. argue that organizations that support innovation provide employees with more resources and
opportunities to innovate and allow them to try and test new ideas [23]. That is, the more security and
resources an individual perceives that they are receiving from the organization, the more motivated they
will be to engage in innovative activities.
Second, people learn from “role models”. If supervisors lead by example and actively serve as
models of innovation, their subordinates will follow their lead and be willing to innovate, generate ideas,
and apply these ideas practically [24, 25].
Finally, scholars have confirmed that the higher the employee’s sense of organizational support, the
higher their psychological level of empowerment and the more focused and enthusiastic they will be at
work. Additionally, the higher the level of recognition and encouragement from superiors, the higher the
level of employees’ perception of empowerment, which further motivates employees to perform better at
work [26]. The Corsun study concludes that supportive work relationships are positively related to the
prediction of employees’ psychological empowerment levels [27]. Through research and statistical
analysis, Spreitzer also proposes that when the management range of a company is large, such as when
Page 623
SEMIT 2023
organization members are allowed to participate in the decision-making of major corporate matters, the
organizational culture and atmosphere of democratic freedom and so on will effectively enhance the level
of employees’ empowerment perception [28]. Employees who have a strong sense of organizational
support believe that the organization is supportive and affirming of them, which enhances their
confidence and enthusiasm in their work. Furthermore, based on the principle of reciprocity, to repay the
organization’s affirmation, employees have a strong self-demand for their own competence and mastery
of work skills, they regard the organization’s interests as equally important to their own, and they tend to
give full freedom to their work autonomy and influence to impact their work results and bring better
returns for the company.
In general, effective and trusting teamwork facilitates in-depth communication and exchange among
members, as well as the efficient sharing of information and knowledge resources. Numerous studies
have found that a good organizational innovation climate helps employees to establish an active
information and resource sharing platform that provides them with access to key information and
knowledge resources for innovation. The organizational innovation climate, therefore, provides
employees with abundant material, emotional, and information resources; relieves the pressure of losing
and preserving resources; and enhances employees’ willingness to express creative ideas, thus promoting
innovative behavior.
Based on the above, the following hypotheses are proposed:
• Research hypothesis 1: Resource security plays a mediating role between high-engagement HRM
strategies and creativity.
• Research hypothesis 3: Empowerment support plays a mediating role between high-engagement
HRM strategies and creativity.
• Research hypothesis 3: Empowerment support plays a mediating role between high-engagement
HRM strategies and creativity.
3.2.2 The moderating effect of anchor traits
There are three dimensions of anchor traits, namely one’s external image, internal quality, and
professional ability. The meaning of one’s external image is the appearance and related characteristics
that the anchor displays during live work, such as their expression, appearance, physical appearance, hair
style, clothing, makeup, and so on. The anchor’s inner image, on the other hand, is their work attitude,
policy position, knowledge level, business abilities, personal charm, inner temperament, culture
cultivation, character perfection, and so on. An anchor’s creation and live broadcasting mainly
incorporate audible language and body language, so a good language expression ability is one of the most
important criteria to become an anchor. Therefore, a good grasp of standard Mandarin, a professional
explanation ability, good communication and interaction skills, proper promotion and recommendation,
and excellent quality control and display abilities are the professional abilities that an anchor must have.
During live broadcasting, in addition to the anchor’s active “performance”, the audience plays an
important role. That is, the main body of the network’s live broadcast is the anchor and the audience, with
the interaction mechanism built by the pop-up screen providing a form of interaction between the two
that completes their self-expression. The process can be summarized as editing and outputting pop-ups—
watching the pop-ups—interpreting the pop-ups—both parties interacting through the pop-up content.
In terms of the media image and language expression in the past, TV anchors conveyed a serious
impression to the audience, either with their clothing, facial expressions, or audible language expressions,
while network anchors today often break through the expressions of conventional news anchors during
live broadcasting, adding lively, humorous, romantic, cute, and exaggerated elements to the process of
discussing goods. They also pay more attention to colloquial and popular expressions, coupled with the
occasional use of today’s network buzzwords, attracting a large number of young audience members.
Some of the realistic factors of live broadcast technology may increase the sense of isolation of
small-screen viewing. For example, the realistic scenes in which different audiences are, the sense of
isolation, and even the interwoven scoring environment of the mood of the moment may become aspects
through which the audience may resist accepting the information that the anchor is imparting. To better
Page 624
SEMIT 2023
achieve the goal and effect of live selling activities, the above resistance needs to be broken as much as
possible in the process of live broadcasting to gain the advantage of self-information dissemination. As
between Zhu Guangquan and Li Jiaqi during their live broadcast of goods, their grasp of standard
Mandarin, professional explanation ability, topics about life, instant interaction with the audience, oral
expression in a casual talk show style, and so on all aim to break the communication resistance to
guarantee the realization of the ultimate benefits of the live target. In addition, in the mobile space
constructed by real space and virtual space, combined with their own psychological needs and emotional
expectations, they also call for more multi-faceted possibilities for anchors, and these possibilities suggest
that anchors can be given a more three-dimensional media image and wider social influence.
Based on these points, the following hypotheses are proposed:
• Research hypothesis 4: Anchor traits play a positive moderating role between high-involvement
HRM strategies and resource security
HRM strategies and typical demonstration
HRM strategies and empowerment support
Fig 1. Model of the effect of high-involvement human resource management strategies on creativity
(with the organizational innovation climate as a mediating variable)
Interpretation of Figure 1.
The research propositions are derived from the conceptual model and are shown as arrows connecting
the structures.
0. Conceptualization of the anchor traits and scale development

Since there is no reference scale for anchor traits, this paper developed its own anchor trait scale to
conduct the analysis. SPSS 26.0 and Mplus 8.3 were used for statistical analysis of the data. Mean and
standard deviation were used for the continuous variable. Additionally, frequency and percentages were
used to describe the categorical variable. Cronbach’s alpha coefficient was used as a reliability measure.
The validity of the scales was explored using exploratory factor analysis (EFA) and confirmatory factor
analysis (CFA), and composite reliability (CR) and average variance extracted (AVE) were used to test
the combined reliability and convergent validity of the study, respectively. Further, correlation analysis
was used to examine the interrelationships among the variables, and if the AVE open-square value was
greater than the correlation coefficient between the variable and the other variables, and the significance
level was 0.05 (p < 0.05), the difference was considered statistically significant.
0.4 Basic characteristics of the survey respondents
The basic characteristics of the survey respondents can be found in Table 1 below.
Page 625
SEMIT 2023
Table 1. Basic characteristics of the survey respondents

Variable Attribute Frequency Percentage
Gender Male 49 21.8
Female 176 78.2
Age 18–22 years old 3 1.3
22–27 years old 148 65.8
28–32 years old 33 14.7
32–40 years old 12 5.3
40 years old or above 29 12.9
Education High school and below 46 20.4
College 52 23.1
Undergraduate 82 36.4
Master’s 40 17.8
Doctorate or above 5 2.2
Occupation Student 7 3.1
Employee at an enterprise or public institution in the
78 34.7
workforce
Individual/Freelancer 131 58.2
Other 9 4.0
Time with Less than 6 months
76 33.8
goods
6 months–1 year 45 20.0
1–2 years 51 22.7
More than 2 years 53 23.6
0.4 Description of the variables

The mean, standard deviation, kurtosis, and skewness of each measurement question item were
statistically analyzed using SPSS 26, and the statistical results are shown in Table 2 below. As can be
seen from the table, first, the mean value of each variable measurement question item was approximately
5, which means that the survey respondents held a more agreeable view. Second, the maximum value of
each item of the survey data was 7, while the minimum value was 1, and the absolute values of the
skewness and kurtosis were less than 3. The sample data were normally distributed (it is usually
considered that when the absolute value of the skewness of the sample items is less than 3 and the absolute
value of the kurtosis is less than 10, the sample data can be regarded as normally distributed), which
satisfied the conditions for subsequent analysis.
Table 2. Description of the variables
Item n Minimum Maximum Mean Std. Deviation Skewness Kurtosis

EI1 225 1.00 7.00 5.24 1.61 −0.83 0.00
EI2 225 1.00 7.00 5.17 1.75 −0.84 −0.15
EI3 225 1.00 7.00 5.20 1.66 −0.72 −0.12
EI4 225 1.00 7.00 5.25 1.79 −1.07 0.28
EI5 225 1.00 7.00 4.99 1.63 −0.71 −0.02
IQ1 225 1.00 7.00 5.92 1.49 −1.59 2.09
IQ2 225 1.00 7.00 5.64 1.44 −1.10 0.59
IQ3 225 2.00 7.00 5.76 1.46 −1.12 0.36
IQ4 225 1.00 7.00 5.74 1.50 −1.33 1.34
Page 626
SEMIT 2023
PC1 225 1.00 7.00 5.70 1.54 −1.22 0.80

PC2 225 1.00 7.00 5.88 1.48 −1.32 1.02
PC3 225 1.00 7.00 5.73 1.37 −1.13 0.98
PC4 225 1.00 7.00 5.75 1.42 −1.30 1.23
PC5 225 1.00 7.00 5.71 1.61 −1.30 0.67
PC6 225 1.00 7.00 5.81 1.48 −1.32 1.14
4.3Reliability analysis
Reliability analysis was conducted on the variables included in the scale, and the Cronbach’s alpha
values measured for each variable and each of its measurement items are shown in Table 3. From this
table, it can be seen that the Cronbach’s alpha values of each dimension were above 0.8, which indicates
that all of these variables had good consistency and stability according to the determination rules
described earlier. In addition, the minimum values of the CITC were all greater than its minimum standard
acceptance value of 0.4, and the deleted alpha values of each measurement question item were smaller
than Cronbach’s alpha, indicating that all variables in the scale had high reliability and could reflect the
sample situation consistently.
First, the standardized path coefficients (factor loadings) were all > 0.5, indicating that the factor
loadings were all large; second, the AVE of each variable was greater than 0.50, suggesting that the latent
variables in this study had good convergent validity. The CR of each variable was also higher than 0.70,
which suggested that the latent variables in this study had good combined reliability. Therefore, the
validity of the scale was considered good from the perspective of validation factor analysis.
Table 3.Reliability analysis
Item Mean Std. Deviation CITC CAID Cronbach’s Alpha n of Items

EI1 5.24 1.610 0.708 0.859 0.882 5
EI2 5.17 1.748 0.704 0.859
EI3 5.20 1.657 0.689 0.863
EI4 5.25 1.786 0.761 0.846
EI5 4.99 1.630 0.722 0.855
IQ1 5.92 1.490 0.711 0.867 0.885 4
IQ2 5.64 1.442 0.748 0.853
IQ3 5.76 1.459 0.752 0.852
IQ4 5.74 1.496 0.788 0.837
PC1 5.70 1.540 0.716 0.888 0.902 6
PC2 5.88 1.483 0.760 0.881
PC3 5.73 1.370 0.685 0.892
PC4 5.75 1.424 0.763 0.881
PC5 5.71 1.607 0.757 0.882
PC6 5.81 1.483 0.721 0.887
0.4 Related analysis

Correlation analysis is used to test the correlation between variables based on their correlation
coefficients, and Pearson’s coefficient is the most commonly used correlation coefficient in social science
research. If the correlation coefficient passes the significance test, then it indicates that there is a
statistically positive or negative correlation between the variables; conversely, if the correlation
coefficient does not pass the significance test, then it indicates that there is no statistical correlation
Page 627
SEMIT 2023
between the variables. The correlation coefficients in this study were as follows: < 0.4 indicates a weak
correlation; 0.4–0.7 indicates a moderate correlation; > 0.7 indicates a high correlation; > 0 indicates a
positive correlation; < 0 indicates a negative correlation. The results of the related analysis can be seen
in Table 4 below.
Table 4 .Related analysis
EI IQ PC
EI 0.774
IQ .485** 0.812
PC .442** .516** 0.780
Note: *p < 0.05; **p < 0.01; diagonal line is the AVE open-square value
The results showed that there was a significant positive correlation between external image and
internal quality and professional competence (p < 0.01), with Pearson’s correlation coefficients of 0.485
and 0.442, respectively; there was also a significant positive correlation between internal quality and
professional competence (p < 0.01, Pearson’s correlation coefficient = 0.516).
5. Analysis
5.1 Sample characterization
Among the individual samples, 21.1% were male and 78.9% were female; 4.5% were aged 18–22
years, 68.4% were aged 22–27 years, 12.8% were aged 28–32 years, 7.1% were aged 32–40 years, and
7.1% were aged 40 years or above. Furthermore, 22.9% had a high school education or below, 26.1% had
a college qualification, 33.6% had a bachelor’s degree, 15.2% had a master’s degree, and 2.2% had a
doctoral degree or above. D and above accounted for 2.2%; the occupation of students accounted for 3%,
employees at enterprises and institutions in the workforce accounted for 33.4%, individuals/freelancers
accounted for 59.9%, and others accounted for 3.8%. Those with less than 6 months of experience
accounted for 30.8%, those with experience of 6 months–1 year accounted for 19.8%, those with 1–2
years of experience accounted for 22.3%, and those with more than 2 years of experience accounted for
27.1%. In addition, those with a family member or partner who supported their career as an anchor
accounted for 69%. Those who thought the future development of the live broadcasting industry
accounted for 92.9%, with no accounts for 7.1%. Those who thought that the live broadcasting of goods
is an indispensable shopping method accounted for 88.9%, with no accounts for 11.1%. Finally, those
who thought the system specifications of the live industry need improvement accounted for 100%, with
no accounts for 0%.
5.2 The roles of resource security, exemplary modeling, and empowerment support in mediating
between high-engagement human resource management strategies and creativity
Fig 2. Multilevel structural equation model diagram
Page 628
SEMIT 2023
Table 5.The roles of resource security, exemplary modeling, and empowerment support in mediating
between high-engagement human resource management strategies and creativity
RS TD LS CT
Intercept distance 1.877* 1.892* 1.638** −0.581
Intercept
(0.846) (0.845) (0.619) (1.117)
Individual layer Gender 0.018
Control variables (0.124)
Age −0.117
(0.073)
Education 0.092
(0.056)
Live time 0.129*
(0.052)
Team size 0.201 0.019 0.184 −0.125
Team layer (0.106) (0.149) (0.137) (0.087)
Established −0.056 0.085 0.182 −0.022
(0.180) (0.136) (0.141) (0.084)
Control variables 0.573*** 0.591*** 0.585*** 0.358*
HHS
(0.151) (0.156) (0.102) (0.178)
0.293**
RS
(0.085)
Independent variables 0.270**
TD
(0.100)
Mediating variables 0.218**
LS
(0.082)
1.267***
σ2
(0.187)
0.842*** 0.742*** 0.821*** 0.353***
τ00
(0.140) (0.142) (0.162) (0.081)
Since p < 0.001, γ = 0.591, hypothesis H3 holds. Further, there was a significant positive effect of
high-involvement HRM strategies on empowerment support (p < 0.001, γ = 0.585), so hypothesis H4
holds. There was also a significant positive effect of resource security on creativity (p < 0.01, γ = 0.293),
so hypothesis H5 holds. In addition, hypothesis H6 was confirmed because there was a significant positive
effect of typical demonstration on creativity (p < 0.01, γ = 0.270). Finally, hypothesis H7 was confirmed
because there was a significant positive effect of empowerment support on creativity (p < 0.01, γ = 0.218).
Further tests were conducted to examine the mediating effects of resource security, typical
demonstration, and empowerment support between high-engagement HRM strategies and creativity, and
the results of the analysis are presented in the following table.
The 95% confidence intervals of the total effect, direct effect, and total mediated effect were
[0.489,1.137], [0.009,0.708], and [0.243,0.666], respectively, and none of the 95% confidence intervals
contained 0, indicating that the total effect, direct effect, and total mediated effect were significantly
present, and their effect sizes were 0.813, 0.358, and 0.455, respectively. In addition, the total
intermediary effect accounted for 56.0% of the total effect.
For the mediating effect of resource security between high-involvement HRM strategies and
creativity, the 95% confidence interval was [0.015,0.321]; the confidence interval did not contain 0,
indicating that the mediating effect of resource security between high-involvement HRM strategies and
creativity was significant. Further, the value of the mediating effect was 0.168, so hypothesis H8 is valid.
For the mediating effect of typical demonstration between high-involvement HRM strategies and
creativity, the 95% confidence interval was [0.027,0.292]; the confidence interval did not contain 0,
indicating that typical demonstration mediated significantly between high-involvement HRM practices
and creativity, with a mediating effect size of 0.159. Thus, hypothesis H9 holds. For the mediating effect
Page 629
SEMIT 2023
of empowerment support between high-involvement HRM strategies and creativity, the 95% confidence
interval was [0.009,0.246]. This confidence interval also did not contain 0, indicating that empowerment
support mediated significantly between high-involvement HRM practices and creativity, with a mediating
effect size of 0.127; therefore, hypothesis H10 holds.
Table 6 . Intermediary effect size estimation
95% 95%
Estimate S.E.
LCI UCI
Total effect 0.813 0.165 0.489 1.137
Direct effect 0.358 0.178 0.009 0.708
Total indirect effect 0.455 0.108 0.243 0.666
HHS–RD–CT 0.168 0.078 0.015 0.321
HHS–TD–CT 0.159 0.068 0.027 0.292
HHS–LS–CT 0.127 0.061 0.009 0.246
5.3 The moderating effect of anchor traits between high-involvement human resource management
strategies and resource security, typical modeling, and empowerment support
After developing a multilevel moderating model, the results of the analysis for the moderating effect
of anchor traits between high-engagement HRM strategies and resource security, typical modeling, and
empowerment support are presented in Table 7 below.
Table 7. The moderating effect of anchor traits between high-involvement human resource management
strategies and resource security, typical modeling, and empowerment support
RS TD LS
Intercept 4.906*** 4.970*** 4.629***
Intercept
(0.328) (0.311) (0.265)
Team layer Team size 0.205 0.069 0.275
Control variables (0.106) (0.154) (0.145)
Established −0.100 0.028 0.139
(0.171) (0.134) (0.136)
0.499*** 0.506*** 0.534***
HHS
(0.114) (0.121) (0.107)
Independent variable 0.076 0.225 0.319*
ST
(0.108) (0.131) (0.151)
***
Moderating variable 0.448 0.404** 0.088
HHS*ST
(0.092) (0.142) (0.141)
Interaction term 0.749*** 0.65*** 0.757***
τ00
(0.133) (0.122) (0.132)
Anchor traits had a significant positive moderating effect between high-engagement HRM strategies
and resource security (p < 0.001, γ = 0.499); but there was a non-significant moderating variable effect
of anchor traits on resource security (p > 0.05). For the moderating effect, the interaction term (HHS ×
ST) had a significant positive effect on resource security (p < 0.001, γ = 0.448), so it was concluded that
there was a significant positive moderating effect of anchor traits between high-involvement HRM
strategies and resource security; thus, hypothesis H11 holds.
To more visually represent the moderating effect of anchor traits between high-involvement HRM
strategies and resource security, the moderating effect graph was produced as follows (see Figure 3): The
positive effect of high-involvement HRM strategies on resource security under low group anchor trait
Page 630
SEMIT 2023
moderation was lower than the effect of high-involvement HRM strategies on multiple resource security
under high anchor trait moderation (slope), indicating that with the increase of anchor traits, the effect of
high-involvement HR meta-management strategies on resource security increased between groups.
Resource security
High-involvement HRM strategy

Fig 3. The moderating effect of anchor traits on the relationship between high-involvement human
resource management strategies and resource security
Anchor traits had a significant positive moderating effect between high-engagement HRM strategies
and typical demonstration (p < 0.001, γ = 0.506). However, there was no significant moderating variable
effect of anchor traits on typical demonstration (p > 0.05). Moreover, for the moderating effect, the
interaction term (HHS × ST) had a significant positive effect on typical demonstration (p < 0.01, γ =
0.404), so it was concluded that there was a significant positive moderating effect of anchor traits between
high-involvement HRM strategies and typical demonstration. Therefore, hypothesis H12 was accepted.
For a visual representation of the moderating effect of anchor traits between high-involvement HRM
strategies and typical demonstration, a moderating effect graph was made as follows (see Figure 4): The
positive effect of high-involvement HRM strategies on typical demonstration under low group anchor
trait moderation was lower than the effect of high-involvement HRM strategies on multiple typical
demonstrations under high anchor trait moderation (slope), indicating that with the increase of anchor
traits, the effect of high-involvement HR meta-management strategies on typical demonstration increased
between groups.
Typical demonstration
High-involvement HRM strategy

Fig 4. The moderating effect of anchor traits between high-involvement HRM strategies and
typical demonstration
Page 631
SEMIT 2023
For the moderating effect of anchor traits between high-engagement HRM strategies and
empowerment support, the analysis results showed that the main effect was significant and positive (p <
0.001, γ = 0.534). Additionally, there was a significant positive moderating variable effect of anchor traits
on empowerment support (p < 0.05, γ = 0.319), but the interaction term (HHS × ST) did not have a
significant effect on empowerment support (p > 0.05). Therefore, it was concluded that the moderating
effect of anchor traits between high-involvement HRM strategies and empowerment support was not
significant, so hypothesis H13 was not valid.
5.4 Assumptions summary
For a summary of the assumptions made using the results of this study, see Table 8 below.
Table 8. Assumptions summary
Research hypothesis Results
H1 High-involvement HRM strategies positively affect creativity Established
H2 High-involvement HRM strategies positively affect resource security Established
H3 High-involvement HRM strategies positively affect typical demonstration Established
H4 High-involvement HRM strategies positively affect empowerment support Established
H5 Resource security positively influences creativity Established
H6 Modeling positively influences creativity Established
H7 Empowerment support positively influences creativity Established
H8 Resource security mediates between high-involvement HRM strategies and Established
creativity
H9 Exemplary modeling mediates between high-involvement HRM strategies and Established
creativity
H10 Empowerment support mediates between high-involvement HRM strategies Established
and creativity
H11 Anchor traits play a positive moderating role between high-involvement HRM Established
strategies and resource security
H12 Anchor traits play a positive moderating role between high-involvement HRM Established
strategies and exemplary modeling
H13 Anchor traits positively moderate the relationship between high-involvement Not supported
HRM strategies and empowerment support
H14 Anchor traits moderate the mediating role of resource security between high- Established
involvement HRM strategies and creativity
H15 There is a moderating role of anchor traits in mediating the role of typical Established
modeling between high-involvement HRM strategies and creativity
H16 There is a moderating role in mediating the role of anchor trait empowerment Not supported
support between high-engagement HRM strategies and creativity
6. Discussion and conclusion

In this paper, a total of 106 team anchor samples and 506 individual anchor samples were collected
through questionnaire research, and the data were statistically analyzed using SPSS 26.0 and Mplus 8.3.
The EFA and validated factor analysis concluded that high-involvement HRM strategies positively affect
creativity, further proving the view of scholars such as Ambile and Avey, who show that if employees
perceive anchor support for innovative behavior, they will be more likely to engage in new innovative
activities.
This study focused on the specific context of the organizational innovation climate, for the first time
verifying that resource security, exemplary modeling, and empowerment support play mediating roles
between high-involvement HRM strategies and creativity and that anchor traits moderate the mediating
role between resource security and exemplary modeling regarding high-involvement HRM strategies and
creativity. Thus, the relationship between high-involvement HRM strategies and team creativity was
identified, enriching the research literature in the field of live team creativity.
Page 632
SEMIT 2023
7. Limitations and future research

This study was the first to explain the positive impact of high-involvement HRM strategies on team
creativity from the theoretical perspective of the organizational innovation climate. Future research can
further explore whether there is an effect on employee creativity, for example, whether there is a reverse
effect of individual creativity on the team or team leader. Furthermore, this study only focused on the
moderating variable of anchor traits, so future research can explore the market effects of individuals and
teams on live consumption from the perspective of a team’s combined abilities.
1. Goldsmith RE, Flynn LR, Goldsmith EB. Innovative consumers and market mavens. Journal of
Marketing theory and practice. 2003;11(4):54-65.
2. Flynn LR, Goldsmith RE, Eastman JK. Opinion leaders and opinion seekers: Two new measurement
scales. Journal of the academy of marketing science. 1996;24(2):137-47.
3. Gefen D, Straub D. Managing user trust in B2C e-services. e-Service. 2003;2(2):7-24.
4. Gefen D. E-commerce: the role of familiarity and trust. Omega. 2000;28(6):725-37.
5. Koufaris M. Applying the technology acceptance model and flow theory to online consumer behavior.
Inf Syst Res. 2002;13(2):205-23.
6. Amabile TM, Conti R, Coon H, Lazenby J, Herron M. Assessing the work environment for creativity.
Academy of management journal. 1996;39(5):1154-84.
7. Anderson N, Potočnik K, Zhou J. Innovation and creativity in organizations: A state-of-the-science
review, prospective commentary, and guiding framework. Journal of management. 2014;40(5):1297-333.
8. Wang P, Rode JC. Transformational leadership and follower creativity: The moderating effects of
identification with leader and organizational climate. Human relations. 2010;63(8):1105-28.
9. Huang X, Hsieh J, He W. Expertise dissimilarity and creativity: The contingent roles of tacit and
explicit knowledge sharing. Journal of Applied Psychology. 2014;99(5):816.
10. Shin SJ, Jeong I, Bae J. Do high-involvement HRM practices matter for worker creativity? A cross-
level approach. The International Journal of Human Resource Management. 2018;29(2):260-85.
11. Yasir M, Majid A. High-involvement HRM practices and innovative work behavior among
production-line workers: mediating role of employee's functional flexibility. Employee Relations: The
International Journal. 2020.
12. Doucet O, Lapalme M-È, Simard G, Tremblay M. High involvement management practices as
leadership enhancers. International Journal of Manpower. 2015.
13. Bae J, Lawler JJ. Organizational and HRM strategies in Korea: Impact on firm performance in an
emerging economy. Academy of management journal. 2000;43(3):502-17.
14. Hsu ML, Chen FH. The cross‐level mediating effect of psychological capital on the organizational
innovation climate–employee innovative behavior relationship. The Journal of Creative Behavior.
2017;51(2):128-39.
15. Isaksen SG, Akkermans HJ. Creative climate: A leadership lever for innovation. The Journal of
Creative Behavior. 2011;45(3):161-87.
16. Sun R, Wang N. Empirical research on organizational climate for innovation, extrinsic motivation
and employee innovation in China. 2009 International Symposium on Information Engineering and
Electronic Commerce: IEEE; 2009. p. 764-7.
17. Bhatti YA, Ventresca M. The emerging market for frugal innovation: fad, fashion, or fit? Fashion, or
Fit. 2012.
18. Zeschky M, Widenmayer B, Gassmann O. Frugal innovation in emerging markets. Research-
Technology Management. 2011;54(4):38-45.
19. Sehgal V, Dehoff K, Panneer G. The Importance of Frugal Engineering. or. 2010.
20. Amabile TM. A model of creativity and innovation in organizations. Research in organizational
behavior. 1988;10(1):123-67.
21. Yu KYT. Affective influences in person–environment fit theory: Exploring the role of affect as both
cause and outcome of PE fit. Journal of Applied Psychology. 2009;94(5):1210.
Page 633
SEMIT 2023
22. Mahmood IP, Zhu H, Zajac EJ. Where can capabilities come from? Network ties and capability
acquisition in business groups. Strategic Management Journal. 2011;32(8):820-48.
23. Shanker R, Bhanugopan R, Van der Heijden BI, Farrell M. Organizational climate for innovation and
organizational performance: The mediating effect of innovative work behavior. Journal of vocational
behavior. 2017;100:67-77.
24. Amabile TM, Schatzel EA, Moneta GB, Kramer SJ. Leader behaviors and the work environment for
creativity: Perceived leader support. The Leadership Quarterly. 2004;15(1):5-32.
25. Avey JB, Patera JL, West BJ. The implications of positive psychological capital on employee
absenteeism. Journal of leadership & organizational studies. 2006;13(2):42-60.
26. Sparrowe RT. The effects of organizational culture and leader-member exchange on employee
empowerment in the hospitality industry. Hospitality Research Journal. 1994;18(3-1):95-109.
27. Corsun DL, Enz CA. Predicting psychological empowerment among service workers: The effect of
support-based relationships. Human relations. 1999;52(2):205-24.
28. Spreitzer GM. Psychological empowerment in the workplace: Dimensions, measurement, and
validation. Academy of management Journal. 1995;38(5):1442-65.
Page 634
SEMIT 2023
The Impact of Industrial Cluster on SMEs Performance: Evidence from the Wood Industry in
Ghana
Masud Ibrahim1*, Dorah Yeboah2
1
Department of Management Studies Education, Akenten Appiah-Menka University of Skills Training
and Entrepreneurial Development, Ghana
2
GIMPA Business School, Achimota, Ghana
*
mibrahim@aamusted.edu.gh
Abstract
Purpose: Small and medium enterprises undertake their activities in a dynamic environment that is
affected by various processes that impact on their business sustainability. This study assessed the effect
of industrial cluster on the performance of Small and Medium Enterprises (SMEs) in Ghana’s wood
industry. Design/Methodology: Using the structured questionnaire method, we surveyed 320 SME wood
manufacturers in three regions of Ghana. Results: Findings from this study revealed that the factors of
production, horizontal networking, innovation and degree of network were positively related to
competitive advantage. This study further reveals a positive and significant relationship between
competitive advantage and firm performance. In addition, the findings of this study revealed that; the
degree of network, horizontal networking, factors of production, and innovation indirectly influence firm
performance through competitive advantage. Implication: The managerial implications of this study are
also discussed. Originality/Value: This study contributes to extant literature on cluster innovation from
a developing economy context.
Keywords: Industrial cluster, Competitive advantage, SMEs, SMEs performance, Ghana
Introduction
SMEs are undoubtedly regarded as one of the strong pillars that propels countries economic growth
(Werekoh, 2022). This supports previous studies that not only show that most countries depend largely
on the performance of SMEs for their economic emancipation (Abbas et al., 2019; Hashim, Raza, and
Minai, 2018) but also nations can have enormous economic challenges when such firms’ fail to redeem
their tax compliance obligations (Swistak, 2015) According to Werekoh (2022) it is especially the case
with developing countries, of which Ghana is no exception. However, due to the nature of the business
environment as well as the unfavourable circumstances in recent times, SMEs have found themselves
wanting in meeting the expectations as the enablers of economic growth (Khan, 2022; Werekoh, 2022).
SMEs operate in a competitive as well as dynamic environment that is influenced by a number of forces
which affect their operations and threaten their survival. This suggests that to be able to survive in this
turbulent and uncertain business environment of weak economy, high exchange and inflation rate as well
as lower purchasing power of consumer (Khan, 2022), SMEs need to be entrepreneurial in finding new
methods and innovative strategies to compete in the marketplace (Epede and Wang, 2022). One method
of modern entrepreneurship that SMEs could adopt to ensure their survival and growth in the industry in
this competitive business era is cluster cooperation (Havierniková and Kordoš, 2019). Cluster cooperation
becomes imperative due to SME’s inability to raise the needed capital amidst scarce resources, limited
skills set and knowledge to be able to compete effectively. SME competitiveness is determined by their
ability to sustain their business as well as maintain their competitiveness. Thus, SME sustainability and
competitiveness would be achieved and maintained mainly through innovative ideas that could be
obtained by working in a network cluster to share resources, knowledge, ideas and skills transfer
(Lundequist and Power, 2002; Epede and Wang, 2022).
These networks and exchanges enable firms to operate geographically with other firms in that same
industry or business category, thereby building mutually beneficial relationships, and creating what is
termed clusters (Havierniková and Kordoš, 2019). In addition, networks and exchanges from cluster
cooperation among SME’s are noted to stimulate digital innovation, knowledge sharing and cost
Page 635
SEMIT 2023
reduction (Leckel, Veilleux, and Dana, 2020; Ben Arfi and Hikkerova, 2021). Clusters are business
entities that operate within a specified geographical location (Foghani, Mahadi and Omar, 2017). Patrick
(2017, p. 116) explains clusters as consisting of “firms and related economic actors and institutions that
draw productive advantage from their mutual proximity and connections”. There is evidence to show
that clusters enable individual firms to operate at a higher level of efficiency by sharing information on
markets customers and transferring knowledge among the participants to achieve their individual
objectives (Ben Arfi and Hikkerova, 2021; Foghani, Mahadi and Omar, 2017; Havierniková, and Kordoš,
2019; Leckel, Veilleux, and Dana, 2020). However, the ability to form, promote and sustain cluster
cooperation activities may vary from country to country (Chowdhury and Audretsch, 2021) considering
that firms’ operate within different contexts to achieve specific objectives. Thus, making it necessary to
understand the impact of clusters on a firms’ interactions in achieving set goals.
Cooperation among individual participants in the network creates what is referred to as spill over effect
that allows for interactions among participants, customers, suppliers and other key stakeholders in the
business value chain to enhance business operation and reduce cost of doing business (Balaz and Hamara,
2016; Ben Arfi and Hikkerova, 2021; Leckel, Veilleux, and Dana, 2020). It can therefore be argued that
through this type of business cooperation, SMEs for instance are able to learn from the network and
achieve some benefits which they would not have if they were on their own.
In their opinion, Adamowicz and Machla (2016) claim SMEs tend to be innovative due to their
cooperation with other firms in research and development. They added that, one other benefit offered
through cooperation due to the clustering is that, SMEs are able to connect with other firms for business
development both local and international (Havierniková and Kordos, 2019) However, notwithstanding
the importance of SMEs as contributing to the economic fortune of most world economies, it has been
argued that, little has been done in terms of supporting SMEs to achieve this lofty vision (Ali and Anwar,
2021; Mulligan et al., 2021). Developing countries for instance, have been noted to churn out high rates
of business start-ups mortality rates despite the high rate of start-up business rates compared with the
most developed economies (Acolatse, 2012; Nakalembe et al., 2023). This seeming lack of support for
SMEs business start-ups calls for action to help these start-ups sustain their businesses to help grow the
economies of these less developed economies.
A review of the literature shows that the concept of innovation or industrial cluster has received much
attention in research (Quaye and Mensah, 2017; Amoah et al., 2018; Patrick, 2017; Li et al., 2019;
Carpinetti, Galda´mez and Gerolamo, 2008; Frisillo, 2007; Karlsson, 2007; Pan et al., 2019; Liu and
Wang, 2022). Despite the extensive view on the importance of clustering in shaping industrial
development, specifically in least developed economies (Schmitz and Nadvi, 1999; Anshari and
Hamdan,2022) and also enhance the ability to innovate (Frisillo, 2007; Antunes and Tate, 2022), little is
known of the effect that clustering has on competitiveness and performance of firms especially in the
wood industry cluster in developing countries such as Ghana.
Additionally, previous research investigated the concept mainly from a regional perspective while
overlooking micro-firm characteristics. Again, most studies on cluster focus on interorganizational
relations among large firms whiles neglecting small and medium enterprises who are mostly in need of
such innovation. This study therefore is positioned to fill the gap in literature on lack of focus on cluster
studies focusing on SMEs especially in Africa. This study assesses the effect of industrial cluster on the
performance of SMEs in the wood industry in Ghana. Additionally, the study investigates the relationship
between industrial clustering and competitive advantage in the wood manufacturing sector in Ghana.
Materials and method

Relationship between industrial clusters and competitive advantage
Research conducted by Amoah et al. (2018) indicated that industrial clusters have significant effect on
competitive advantage. They conceptualized industrial clusters as involving factors of production (such
as land, labour and equipment), horizontal networking (information sharing and joint action), vertical
networking (suppliers, customers and public/private institutions), healthy competition and innovation.
Page 636
SEMIT 2023
Factors of production
Factors of production involves the necessary ingredients including resources needed for the performance
or production of goods and services. Accordingly, these factors of production include natural resources
(land, trees, mineral deposits etc.), capital (buildings, machines etc.), labour and enterprise (Kleynhans,
2006; Quaye and Mensah, 2017). This suggests that obtaining the right quality and quantity of these
factors of production is critical in producing the right quality output needed for firm’s business. However,
considering that the availability of these resources are scarce to small and medium firms, these firms
become driven by these needs. Consequently, seeking assistance from counterpart firms in the industry
for the necessary tools and skills for successful business process becomes imperative (Antunes and Tate,
2022). Ideally, industrial cluster has become an avenue where knowledge and skills transfer manifest for
the enhancement of business process and work progress (Krugman, 1998; Anshari and Hamdan,2022).
Previous studies suggest that the wood industry requires a lot of investments due to the high cost of
essential industry equipment (Agyekum et al., 2021; Asamoah et al., 2020; Quaye and Mensah, 2017)
For instance, these studies note that tools like industrial cutting, design machines, sandpapering
equipment and other working equipment are expensive. This suggests that most of the business owners
might not be able to afford them or those who have them might have to either borrow them from other
business owners or purchase them on credit. Hence, the need to create a system where businesses in the
cluster are able to share tools and equipment to reduce the high investment in the purchase of these
essential tools. We therefore hypothesise that:
H1: Access to factors of production in a cluster leads to competitive advantage
Horizontal Networking
Horizontal networking has been explained as the relationship and contacts businesses make with other
competing firms (Amoah-Mensah, 2015). Quaye and Mensah (2017) explained the type of relationship
that exist among firms in a cluster as information sharing regarding information on market opportunities,
finance opportunities and technologies. Rosenfield (1995) indicates that firms that network in a cluster
benefit from the many opportunities presented like new markets, business process integration, resources,
as well as shared knowledge that enhances competitiveness.. This aligns with Scholars (Terziovski, 2003;
Premaratne,2001; Boschma and Ter Wal,2007; Jiang et al., 2018; Haffer, 2021) findings that networking
positively influences SME performance. We therefore hypothesise that;
H2: Horizontal networking in industrial clusters enhances SMEs competitive advantage
Innovation/Innovativeness
Innovation has been defined variously by different scholars. Lundvall (1985, p.79) conceptualised
innovation as “an acquired knowledge and experience that could be an incremental technical change or
an increase in technical opportunities”. Innovation is defined as the “production or adoption,
assimilation, and exploitation of value-added novelty in economic social spheres; renewal and
enlargement of products, services, and markets; development of new methods of production, and
establishment of new management systems” (Yusheng and Ibrahim, 2020: 2). In addition, innovation has
been explained as anything that assist firms to identify and satisfy the needs of customers at a profit
(Škerlavaj, Song and Lee, 2010; Gopalakrishnan and Damanpour, 1997). Further, according to Hurley
and Hult (1998), innovation is an aspect of firm’s philosophy and openness toward new ideas which gives
the firm the ability to adopt or implement new ideas, processes, or products successfully. In essence,
these works suggest that innovation has become critical to not only firms’ success but also necessary for
sustainability and competition. However, it is worthy of mentioning that much as this study is interested
in a firms’ innovative practices, its main focus is the exchange of innovative ideas among clusters that
can render firms with opportunities for competition.
This supports previous studies that firms in a cluster easily develop and share knowledge rapidly, hence
help in developing inimitable innovations that enhances firm’s competitiveness (Quaye and Mensah,
2017; Antunes and Tate, 2022). Thus, becoming imperative to study innovative mechanisms that accrue
to cluster firms within the wood industry that oftentimes seems to be consistent in practice, unique and
informal because it is difficult for them to imitate innovation of other firms from different sectors
Page 637
SEMIT 2023
McCann and Folta (2011) affirmed that SMEs who participate in a geographical cluster can obtain novel
ideas due to the tendency of knowledge spill overs. Also, Fukugawa (2006) mentioned that sharing
information in an industrial cluster boosts participants propensity to be innovative (Quaye and Mensah,
2017). Lichtenthaler (2020) asserts that innovation is very important especially for business start-ups in
achieving competitive advantage. We hypothesise therefore that:
H3: Innovation in a cluster positively leads to competitive advantage
Relationship between firm performance and Competitive Advantage

Firms would achieve competitive advantage when they are able to provide better services, goods and
value for customers more than the competition (Bateman & Snell, 2007; Denga, Vajjhala and Rakshit,
2022). However, Barney (1991) argued that firms that possess rare and valuable resources which are
imitable would enjoy some level of competitive advantage and better performance. In addition, firms that
are able to create innovative-resource are more likely to gain competitive advantage and achieve higher
performance (Li et al., 2023; Zeebaree and Siron, 2017; Rosenbusch, Brinckmann and Bausch, 2011).
According to resource theory, the competitiveness of a firm depends on sustainable resources, difficult
to imitate, and replace that differentiates this enterprise from its competitors (Zaridis, 2009). This supports
the idea that improving the performance of a business depends on the availability and access to valuable
resources, rare, inimitable, non-substitutable (Barney, 1991). Consequently, these two assertions show
that, the strength or weakness of a company is based on its resources (financial capital and physical
capital), and that allows it to enjoy a competitive advantage (Wernerfelt, 1984; Denga, Vajjhala and
Rakshit, 2022; Li et al., 2023). We propose the hypothesis that:
H4: There is a positive and direct relationship between Competitive advantage and SMEs performance
Moderating effect of Degree of Network
Networking could be explained as collaborative relationships established by firms to achieve certain
mutual benefits among them. Previous studies show that through networking firms are able to
concurrently suggest, request, require, perform and adapt their activities (Ford et al., 2003; Świerczek,
2019). Walter, Auer and Ritter (2006), aver that, network abilities allow firms to initiate, maintain and
exploit relationships based on the commitment of various external partners. This suggests that as much
as firms are expected to exhibit network abilities, it is equally imperative that partners in the network
relationship become committed to the agreed course of action for success. Lei and Huang (2014) propose
the degree or the level of network at which firms can avail themselves to opportunities through contact
with other industry stakeholders. Lei and Huang (2014) recognised four ways to assess the degree of
network among cluster. These are density, intensity, reciprocity and multiplexity. In their study, Quaye
and Mensah (2017) introduced the degree of network to mediate the relationship between industrial
cluster and competitive advantage. Their finding revealed that, the relationship between factors of
production and competitive advantage is not mediated by degree of network. However, their study found
that degree of network partially mediated the relationship between horizontal networking and competitive
advantage. We therefore propose the following hypotheses that:
H5: Degree of network mediates the relationship between access to factors of production and
competitive advantage.
H6: degree of network mediates the relationship between horizontal networking and competitive
advantage
H7: Degree of network mediates the relationship between innovation and competitive advantage.
Page 638
SEMIT 2023
Degree of
Network
Factors of
H5
production
H1
H6
Competitive SME
Horizontal H2
advantage H4 Performance
networking
H7
H3
Innovation
Figure 1. Conceptual model and hypotheses
Methodology
This research employed the quantitative research approach to assess the impact of industrial cluster on
SME performance in the wood industry in Ghana. For the current study, the population comprised SMEs
in the wood industry cluster in Ghana. A sample size of 320 SMEs working in clusters was obtained
through the simple random and stratified sampling method. Three main clusters were identified in the
wood manufacturing industry in two main wood industry Regions of Accra and Kumasi in the Ashanti
and Greater Accra Region in Ghana. First and foremost, the researchers obtained permission from the
administrative bodies in charge of these wood clusters. Subsequently, the individual businesses were
identified and grouped into clusters based on their geographical locations. Then a self-administered
questionnaire was sent to these sampled businesses to fill and collected at a later date. It took about two
weeks to circulate the questionnaire to the sampled respondents and another two weeks to collect the
filled questionnaire.
The scale used in the questionnaire design was adapted from earlier research (Amoah et al., 2018; Li,
Han & Shen, 2019; Quaye & Mensah, 2017). As suggested by earlier works, a five-point Likert scale was
utilized to measure the variables for the research constructs. The Likert scale ranged from strongly agree
to strongly disagree. 24 items were used to measure the five multi-item constructs, and they were modified
for our current study
In order to analyze the results of our study, we first cleaned and recoded the data using the Statistical
Package for Social Sciences to eliminate outliers (SPSS V.22). Then, using the direct varimax rotation,
we conducted an exploratory factor analysis (EFA) to identify how many factors to extract for our
investigation. We were able to test the suggested correlations between the study variables as a result. In
Page 639
SEMIT 2023
accordance with Pallant's (2011) theory, the EFA was conducted in three steps: "Assessment of the
suitability of the data for factor analysis, (2) Factor extraction, and Factor rotation and interpretation"
(Yusheng and Ibrahim, 2019). In addition to EFA, we also carried out a confirmatory factor analysis
before transferring the data to SmartPLS 3 for additional investigation. Therefore, the proposed model
was evaluated for model fitness using the structural equation modelling available in SmartPLS 3.
Results of the Study

Confirmatory Factor Analysis
The common method variance was examined using the Harmann's Single-Factor test. Six items were
loaded onto one component during the test's exploratory factor analysis (EFA). At first, 26 items were
entered, however 6 of them were eventually taken out, leaving 20 items. About 95% of the variance in
our sample was explained by the remaining elements or components (see table 1).
Table 1. KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

.952
Bartlett's Test of Sphericity Approx. Chi-Square 9278.068

Df 820
Sig. .000
Reliability and validity of Scales

We first evaluated the convergent and discriminant validity of our scales in order to determine their
dependability. When measuring construct validity, Bartlett's Test of Sphericity was used, and Kaiser-
Meyer-Olkin (KMO) was used to assess the adequacy of individual variable sampling. It is anticipated
that the KMO total score must be 0.6 or higher to do factor analysis (zdamar, 2017). The Bartlett's test of
Sphericity and KMO results, therefore, showed that both are significant and appropriate for the factor
analysis (see Table 1).
The result of the cumulative variance shows 95%, which exceeds the minimum acceptance level of 60%
(Özdamar, 2017). The Bartlett's Test of Sphericity from the table above shows that the correlation
between the variables is 9278.068 which is sufficient, and significant (P> 0.000). Also, the factor loadings
of all the construct from the table 1 exceeded 0.5 (Ringle, Wende et al. 2015, Hair, Ringle & Sarstedt,
2011) which shows the presence of convergent and discriminant validity (See Table 1).
Measurement model reliability and validity
The study’s measurement model reliability and validity were assessed using Factor loadings, Cronbach’s
alpha, Construct reliability and the Average variable extractor. The factor loading is expected load 0.7 or
above whereas the Cronbach’s alpha is expected to be higher than 0.7. from the result, all of these
thresholds were reached meaning that the model used is reliable. The item factor loadings exceeded the
0.70 threshold and shows loading from 0.75 and 0.91. also, the Cronbach’s alpha shows loading from
0.81 and 0.92 which indicates that the items exceeded the threshold of 0.7 (Hair, et al., 2017, Ringle et
al., 2015).
Again, convergent validity was assessed using the average variance extracted (AVEs) and composite
reliability. Since the AVE and CR met the minimum requirements of 0.50 and 0.70, respectively, from
table 2, the convergent validity was also attained (Fornell & Larcker 1981, Ringle et al., 2015).
Page 640
SEMIT 2023
Table 2. Item loading and construct reliability
FL CA CR AVE
CA1 0.838 0.846 0.895 0.681

CA2 0.844
CA3 0.813
CA4 0.806
DoN1 0.839 0.842 0.894 0.678
DoN2 0.826
DoN3 0.827
DoN4 0.801
FP1 0.864 0.927 0.945 0.774
FP2 0.862
FP3 0.890
FP4 0.891
FP5 0.892
FoP1 0.789 0.816 0.878 0.644
FoP2 0.794
FoP3 0.850
FoP4 0.775
HN1 0.784 0.889 0.919 0.693
HN2 0.815
HN3 0.845
HN4 0.841
HN5 0.876
Innov1 0.912 0.884 0.921 0.744
Innov2 0.884
Innov3 0.896
Innov4 0.750
Notes: FL – Item Loadings, CA – Competitive advantage, DoN – Degree of network, FP – Firm performance, FoP
– Factors of production, HN – Horizontal networking, Innov – Innovation, AVE-Average variance extracted, CR-
Composite reliability, CA – Cronbach’s alpha
Table 3. Discriminant Validity

CA DoN FP FoP HN Innov
CA 0.825
DoN 0.749 0.823
FP 0.744 0.838 0.880
FoP 0.668 0.806 0.807 0.802
HN 0.699 0.764 0.787 0.697 0.833
Innov 0.696 0.776 0.865 0.721 0.823 0.863
The discriminant analysis, shown in Table 3, requires a factor to correlate more strongly with one
particular construct than with any other on the scale (Messick, 1989). On their respective scales, each
variable in Table 3 loaded more heavily than any other component. The value of competitive advantage
(0.83) was higher than any other factor on the scale. Degree of Network also had a value of (0.82), FP
(0.88), FoP (0.8), HN (0.83) and Innovation (0.86) respectively in that order.
Page 641
SEMIT 2023
Results of Structural Model

The results of structural model for testing the research hypotheses are presented in Figure 2.
Figure 2. Structural model showing relationship among the variables

Figure 2 shows the structural model’s assessment regarding the relationship between the variables. The
assessment includes the path coefficients that estimate the relationship between the variables. From
Figure 2, factors of production (FoP), Horizontal networking (HN), Innovation (Innov) and Degree of
network (DoN) all related positively with Competitive advantage (CA). From figure 2, Degree of network
had the highest impact on CA (0.399 i.e., 39.9%), followed by Horizontal networking (20.6%), and
Innovation (14.7%). Factors of production on the other hand had little impact on CA (0.098 i.e., 9.8%).
Hypothesis test
In order to test for the hypotheses stated in this study, we carried out bootstrapping with 5000 samples.
Out of the four direct hypotheses stated and tested, three hypotheses showed a positive and significant
relationship among the hypothesised relationships (H2, H3 and H4; p < 0.05). Factors of production had
a positive but insignificant relationship with competitive advantage (β = 0.098; t= 1.304; p >0.05), H1
was therefore not accepted. This means that factors of production (land, labour and equipment) might not
have significant effect on firms’ competitive advantage.
The second hypothesis H2, however, shows a positive and significant relationship between degree of
horizontal networking and competitive advantage (β = 0.206; t= 2.78; p < 0.05) and this led to the
acceptance of H2. This means that, horizontal networking as practiced by SMEs has a positive influence
on SMEs achieving some competitiveness in the industry. The Beta score indicates that, when competitive
advantage increases by about 21% with an adoption of horizontal networking. Hypothesis three also
found a positive and significant relationship between innovation and CA (β = 0.147; t= 1.902; p < 0.10);
this led to the acceptance of H3. The Beta score here means that when innovation level increase by 1%,
Page 642
SEMIT 2023
competitive advantage by SMEs by about 15%. The fourth hypothesis, H4 also shows a positive and
significant relationship between competitive advantage and firm performance (β = 0.744; t= 26.71; p <
0.001); this led to the acceptance of hypothesis H4. The Beta score here means that when competitive
advantage increases by 1%, firm performance increases by about 74%.
Table 4. Path Coefficients

Hypothesis Paths Beta (B) (SD) t-value p-value Action
H1 FoP -> CA 0.098 0.075 1.304 0.192 Not Supported
H2 HN -> CA 0.206 0.074 2.782 ** Supported

H3 Innov -> CA 0.147 0.075 1.950 * Supported
H4 CA -> FP 0.744 0.028 26.713 *** Supported
Note: *** significant at 0.001; ** significant at 0.5; * significant at 0.1
Figure 3. Bootstrapping the path relationship among the hypothesised relationships
Page 643
SEMIT 2023
Table 5. Indirect effects

Paths Beta STDEV t-value p-values
DoN -> CA -> FP 0.296 0.061 4.897 ***
FoP -> CA -> FP 0.069 0.056 1.242 0.214
HN -> CA -> FP 0.151 0.057 2.668 **
Innov -> CA -> FP 0.108 0.057 1.902 *
FoP * DoN -> CA 0.032 0.060 0.543 0.587
HN * DoN -> CA -0.091 0.085 1.069 0.285
Innov* DoN -> CA 0.065 0.092 0.707 0.480
Note: *** significant at 0.001; ** significant at 0.5; * significant at 0.1
Aside the test of the hypothesised direct relationship, we also tested the indirect relationships among the
variables. Degree of network, horizontal networking, factors of production and innovation influenced
firm performance indirectly through competitive advantage. The first path DoN -> CA -> FP shows a
positive and significant indirect effect between DoN and FP (β = 0.296; t= 4.897; p < 0.001). this means
that competitive advantage mediates the relationship between degree of network of SMEs and firm
performance. The second path, FoP -> CA -> FP shows a positive but insignificant indirect effect between
FoP and FP (β = 0.069; t= 1.242; p > 0.05). This result indicates that there is no indirect relationship
between the factors of production and firm performance. It means that firm performance is not dependent
on factors of production. Again, the third path HI -> CA -> FP also shows a positive and significant
indirect effect between horizontal networking and firm performance (β = 0.088; t= 5.33; p < 0.001). here,
competitive advantage plays a key role in ensuring that firms achieve their desired goals which is growth
by mediating the relationship between horizontal networking and firm performance. Finally, the path
Innov -> CA -> FP also shows a positive and significant indirect effect of innovation on firm performance
through competitive advantage. Competitive advantage here also plays an important role in ensuring that
firm’s innovativeness culminates into achieving desired performance which is growth and profitability.
Moderating effect
Figure 4 shows the moderating interventions when the moderating effect of degree of network is imposed
on the interaction between factors of production, horizontal networking, innovation and competitive
advantage. The path coefficient of the direct relationship between FoP and competitive advantage is
1.224, horizontal networking and competitive advantage is 2.644; whereas the influence of the effect of
innovation on competitive advantage is 2.001. The path coefficient value for the moderating influence of
DoN in the relationship between FoP and CA is 0.543 and this is statistically insignificant (FoP * DoN -
> CA; β = 0.032; t = 0.543; ρ > 0.05). This result shows that, degree of network does not moderate the
relationship factors of production and competitive advantage. Again, the path coefficient value for the
moderating influence of DoN in the relationship between HN and CA is 1.069 and this is statistically
insignificant (HN * DoN -> CA; β = -0.091; t = 1.069; ρ > 0.05). This result also shows that, degree of
network does not moderate the relationship between horizontal networking and competitive advantage.
The path coefficient value for the moderating influence of DoN in the relationship between innovation
and CA is 0.707 and this is statistically insignificant (Innov* DoN -> CA; β = 0.065; t = 0.707; ρ > 0.05).
This result also shows that, degree of network does not moderate the relationship innovation and
competitive advantage. Thus, hypotheses, H5, H6 and H7 were not accepted.
Page 644
SEMIT 2023
Figure 4. Moderating effect of DoN in the relationship between FoP, HN, Innov and CA
Discussion
This study assessed the effect of industrial cluster on the performance of SMEs in the wood industry in
Ghana. Findings from this study indicated that factors of production (FoP), horizontal networking (HN),
innovation (Innov) and degree of network (DoN) all related positively with competitive advantage (CA).
Degree of network had the highest impact on competitive advantage (39.9%), this is followed by
horizontal networking (20.6%), and innovation (14.7%). However, factors of production (i.e., land, labour
and equipment) were found to have insignificant effect on competitive advantage. What this means is
that, for firms especially SMEs to achieve competitive advantage in the industry they operate, they should
first and foremost possess or create a certain level of network in their industry that could be leveraged in
terms of using and sharing information, knowledge, skills, experience as well as shared equipment and
tools needed for the day-to-day operations in the wood industry. The wood industry comes with a lot of
cost due to the nature of the machinery needed for the job. SMEs are handicapped in terms of finance and
capital availability needed for the operations and so sharing tools and equipment is one of the effective
means of engaging in this kind of work. Having or been part of group or network thus, gives the individual
firms an opportunity to reduce their cost of doing business and also share experience, knowledge and
skills.
Additionally, SMEs need to be involved in a geographically centred clusters that has the propensity to
contribute to their growth and sustained competitiveness. SMEs wood manufacturers could take
Page 645
SEMIT 2023
advantages provided through the horizontal networking such as finance opportunities and technologies
to exploit new markets, integrate activities or bring their resources and knowledge together to achieve
growth and profitability. SMEs working in clusters are therefore able to garner the necessary knowledge
and skills set to be able to function well in the business. This affirms what McCann and Folta (2011)
acknowledged that SMEs are able to acquire or get access to new ideas due to their involvement with a
geographical cluster with knowledge spill overs propensity.
Also, innovation was found to influence competitiveness of firms. This finding suggests that, SMEs in
the wood manufacturing could obtain their desired growth and sustainability in their business operations
only if they integrate innovation into their day-to-day operations. Innovation here entails the SME firms
to be able to introduce or develop new ideas, concepts and new methods of production that would give
them a sort of competitive edge over the others in the industry. This finding supports earlier findings
(Quaye & Mensah 2017, Lei & Huang, 2014; McCann & Folta 2011) that found a significant effect of
innovation on competitive advantage of SMEs.
The wood manufacturing industry is faced with a number of challenges as mentioned by Ibrahim and
Madichie, (2014). Some of these include the fast depletion of the forest reserve that provides the raw
material in the wood industry as well as the prohibition of governments and their agencies in the export
and purchase of certain scarce and protected wood species. This means that, firms working in the industry
must look for other alternative products to compliment the scarce raw materials needed to sustain their
business.
The findings of this study further revealed a positive and significant relationship between competitive
advantage and firm performance. When firms are able to achieve a sustained competitive advantage in
the industry, this leads to overall performance of the firm in terms of profitability and increased market
share. This finding supports earlier research findings by Zeebaree and Siron (2017) and Rosenbusch et
al. (2011) that, firms that are able to create innovative-resource are more likely to gain competitive
advantage and achieve higher performance. Thus, to achieve higher performance, SMEs in the wood
manufacturing industry should focus on creating or achieving competitive advantage by offering unique
products that are above what the industry provides and also offers value for customers.
Notwithstanding the direct effect of competitive advantage on firm performance, this study also revealed
that, degree of network, horizontal networking, factors of production and innovation influenced firm
performance indirectly through competitive advantage. It is very instructive to note that competitive
advantage does not only influence firm performance, but also acts as a mediator to influence the
relationship between the above-mentioned factors and firm performance. This finding supports earlier
findings by Quaye and Mensah (2017). Finally, the results of this study revealed that degree of network
does not moderate the relationship between factors of production, horizontal networking, innovation and
competitive advantage. However, in previous studies, the degree of networking was found to moderate
the relationship among these variables.
Conclusion
This study looked at how industrial clusters affected the performance of SMEs in Ghana's wood sector.
It assessed the effect of production variables (i.e., land, capital, and labour), horizontal innovation, and
innovation on the competitiveness of SMEs in Ghana's wood manufacturing industry; as well as the effect
of competitive advantage on the performance of small and medium enterprises. Findings from this study
suggests that SMEs in the wood manufacturing industry are able to achieve competitive advantage as a
result of their networking ability as they work in a geographically centred cluster. Also, the findings show
that small and medium enterprises working in the wood industry cluster in Ghana are able to achieve the
desired performance as a result of their competitiveness achieved through their shared knowledge on
industry, market and customer needs in the market.
Theoretical implications
Theoretically, this study contributes to extant literature on SME cluster innovation within the wood
manufacturing sector from a developing economy for broader insights. Firstly, while Horizontal
Page 646
SEMIT 2023
networking (HN), Innovation (Innov) and Degree of network (DoN) are not new in the literature (Haffer,
2021; Lichtenthaler, 2020; Quaye and Mensah, 2017), the contextual nature of this study extends
empirical proof on the need for research into future comparative studies. Secondly, this study has an
important theoretical implication because it highlights not only the individual effects of Horizontal
networking, Innovation and Degree of network on SMEs in the wood manufacturing sector but also
explains that a combination of these factors can contribute to a firm’s sustainability and competitive
advantage. Thirdly, the findings from this study reveal a positive and significant relationship between
competitive advantage and firm performance. This suggests that per the set goals, a firm can decide to
deliberately choose certain cluster options or activities in order to maximize an expected outcome.
Fourthly, an additional theoretical implication resulting from this research lies in the insignificant effects
of factors of production (i.e., land, labour and equipment) on competitive advantage of SMEs within the
wood manufacturing sector. What this means is that, for firms especially SMEs to achieve competitive
advantage in the wood industry, they should first and foremost possess or create a certain level of network
that could be leveraged in terms of using and sharing information, knowledge, skills, experience as well
as shared equipment and tools needed for the day-to-day operations.
Managerial implication
The managerial implications are important in light of the benefits of cluster cooperation to SMEs in the
wood manufacturing sector. The findings show that working in a cluster has a lot of benefits as discussed
above to the participants in the industrial cluster. Consequently, it behoves the individual SMEs
owners/managers to avail themselves to conduct their businesses in groups as working together enhances
their competitiveness through the knowledge sharing with respect to industry, market opportunities as
well as finding and satisfying customer needs. In addition, given that SMEs need to be entrepreneurial in
finding new methods within this era of dynamic business environment (Epede and Wang, 2022) there is
the need to create a common platform for all members regardless of their size, background or education
in the cluster to acquire the necessary skills and knowledge that is constantly updated for the benefit of
members of the cluster. Further, SMEs are handicapped in terms of finance and capital availability
needed for the operations and so sharing of tools and equipment is one of the effective means of enjoying
sustainability and competition. To this end, the findings suggest a need for managers to not only intensify
efforts of cluster partners but also create appropriate space for SMEs in the wood industry to reduce a lot
of cost on the type(s) of machinery needed for the job. Similarly, guiding members within the cluster
network on how to achieve mutually beneficial interaction would not only be a great means of decreasing
the potential negative effect of cluster cooperation but also can offset some wrong perceptions of firms
that do not want to form a cluster.
Directions for future studies
This work represents a novel study into the impact of industrial cluster on SMEs Performance within the
Wood Industry in Sub-Saharan Africa, using Ghana as a case study. Future research should examine the
contextual differences between SMEs in different subject areas for broader insights on the topic. In
addition, considering that SMEs operate in a competitive as well as dynamic environment that is
influenced by a number of forces which affect their operations and threaten their survival, future research
should investigate SMEs differences in motivation towards cluster cooperation and the effect that the
differences in motivation have on the partners level of interaction. Further, SME competitiveness is
determined by their ability to sustain their business. Also, firms that are able to create innovative-resource
are more likely to gain competitive advantage and achieve higher performance (Li et al., 2023). Thus,
future research should focus on the challenges related to SME’s use of only internal resources to provide
depth understanding into how firms can strategize around industrial cluster for success. Furthermore,
this study employed cross-sectional and quantitative methods to unravel Horizontal networking (HN),
Innovation (Innov) and Degree of network (DoN) as critical to cluster cooperation of SMEs within the
wood manufacturing industry. This finding not only sets the agenda for future studies that employ
longitudinal and qualitative methods to investigate the extent to which each or a combination of these
factors can lead to the realisation of expected outcomes but also how the attainment of expected business
outcomes through industrial cluster can influence SME’s future cooperation for broader insights.
Page 647
SEMIT 2023
References
Abbas, J., Raza, S., Nurunnabi, M., Minai, S.M. and Bano, S. (2019), “The Impact of Entrepreneurial
Business Networks on Firms’ Performance Through a Mediating Role of Dynamic Capabilities”,
Sustainability, Vol. 11 No. 3006; doi:10.3390/su11113006
Acolatse, S. (2012), Challenges & Opportunities for SME Development in Ghana, Business Sense.
Adamowicz, M. and Machla, A. (2016), “Small and Medium Enterprises and the Support Policy of Local
Government”, Oeconomia Copernicana, Vol 7 No. 3, pp.405-437.
Agyekum, E. B., Amjad, F., Mohsin, M. and Ansah, M. N. S. (2021), “A bird's eye view of Ghana's
renewable energy sector environment: A Multi-Criteria Decision-Making approach”, Utilities
Policy, Vol. 70 No. 101219.
Ali, B. J. and Anwar, G. (2021), “Strategic leadership effectiveness and its influence on organizational
effectiveness”, International Journal of Electrical, Electronics and Computers, Vol. 6 No. 2.
Amoah-Mensah. A. (2015), “Industrial clustering and competitive advantage of SMEs in the Wood
Industry”, Journal of Business and Enterprise Development, Vol. 5 No. 1, pp.95-124.
Amoah, M., Enoch, G., Dadzie, K. P. and Asomani, J. (2018), “Successes and Challenges in Industrial
Clusters: Some Empirical Evidence from Wood Industrial Cluster in Kumasi, Ghana”, Journal of
Investment and Management, Vol. 7 No. 1, pp.1-12. doi: 10.11648/j.jim.20180701.11
Anshari, M. and Hamdan, M. (2022), “Understanding knowledge management and upskilling in Fourth
Industrial Revolution: Transformational shift and SECI model”, VINE Journal of Information and
Knowledge Management Systems, Vol. 52 No. 3, pp.373-393
Antunes, P. and Tate, M. (2022), “Business process conceptualizations and the flexibility-support trade-
off”, Business Process Management Journal, (ahead-of-print).
Asamoah, O., Kuittinen, S., Abrefa Danquah, J., Quartey, E. T., Bamwesigye, D., Mario Boateng, C. and
Pappinen, A. (2020), “Assessing wood waste by timber industry as a contributing factor to deforestation
in Ghana”, Forests, Vol. 11No. 9, p.939.
Balaz P. and Hamara A. (2016), “Export dependency of Slovakia on German's economy”, Politicka
Ekonomie, Vol. 64 No. 5, pp.573-590. http://doi.org/10.18267/j.polek.1088
Barney, J. (1991), “Firm resources and sustained competitive advantage”, Journal of Management, 17(1),
pp.99-120. doi:http://dx.doi.org/10.1177/014920639101700108
Bateman, S.T. and Snell, S.A. (2007), Management: Leading & Collaborating in a Competitive World.
Singapore: McGraw-Hill/Irwin.
Ben Arfi, W. and Hikkerova, L. (2021), “Corporate entrepreneurship, product innovation, and knowledge
conversion: the role of digital platforms”, Small Business Economics, Vol. 56, pp.1191-1204.
Boschma, R. A. and Ter Wal, A. L. (2007), “Knowledge networks and innovative performance in an
industrial district: the case of a footwear district in the South of Italy”, Industry and Innovation, Vol. 14
No. 2, pp.177-199.
Carpinetti, L.C. R., Galda´mez, E.V.C. and Gerolamo, M.C. (2008), “A measurement system for
managing performance of industrial clusters: A conceptual model and research cases”, International
Journal of Productivity and Performance Management, Vol. 57 No. 5, pp.405-419.
Chin, W. (2010), “How to write up and report PLS analyses”, In Esposito, C. W. W., Vinz, V.,
Henseler, J. and Wang, H. (Eds.), Handbook of partial least squares: concepts, methods and
applications., Springer, Heidelberg, pp.655 – 690.
Cortright, J. and Mayer, H. (2001), “High-Tech Specialization: A comparison of high technology centers.
Center on urban and metropolitan policy”, The Brookings Institution. Survey series, January 2001.
Page 648
SEMIT 2023
Crossan, M. M. and Apaydin, M. (2010), “A multi-dimensional framework of organizational innovation:

A systematic review of the literature”, Journal of Management Studies, Vol. 47 No. 6, pp.1154–1191.
Denga, E. M., Vajjhala, N. R. and Rakshit, S. (2022), “The Role of Digital Marketing in Achieving
Sustainable Competitive Advantage”, Digital Transformation and Internationalization Strategies in
Organizations, pp.44-60.
Drulák, P. and Drulákova, R. (2014), “The richness of the liberal tradition in international relations: Karl
Deutsch on political community and the European integration”, International Relations, Vol. 28 No. 3,
pp.333-349. http://doi.org/10.1177/0047117814545951
Epede, M. B. and Wang, D. (2022), “Global value chain linkages: An integrative review of the
opportunities and challenges for SMEs in developing countries”, International Business Review, 101993.
Foghani, S., Mahadi, B. and Omar, R. (2017), “Promoting clusters and networks for small and medium
enterprises to economic development in the globalization era”, Sage Open, Vol. 7 No. 1,
2158244017697152.
Ford, D., Gadde, L.-E., Håkansson, H. and Snehota, I. (2003), Managing Business Relationships, Whiley,
London. .
Fornell, C. and D. F. Larcker (1981), “Evaluating structural equation models with unobservable
variables and measurement error”, Journal of Marketing Research, Vol. 18 No. 1, pp.39-50.
Frisillo, D. (2007), “An analysis of a potential cluster in an energy sector of Albany, NY”, Journal of
Business & Industrial Marketing, Vol. 22 No. 7, pp.508–516.
Fukugawa, N. (2006), “Determining factors in innovation of small firm networks: a case of cross industry
groups in Japan”, Small Business Economics, Vol. 27 Nos. 2/3, pp.181-193.
Gopalakrishnan, S. and Damanpour, F. (1997), “A review of innovation research in economics, sociology
and technology management”, Omega, Vol. 25 No. 1, pp.15–28.
Haffer, R. (2021), “Supply chain performance measurement system of logistics service providers vs.
supply chain performance: a conceptual framework”, European Research Studies, Vol. 24 No. 2B, pp.
78-97
Hair, J., Ringle C. M. and Sarstedt, M. (2011), “PLS-SEM: Indeed, a Silver Bullet”, Journal of Marketing
Theory and Practice, Vol. 19 No. 2, pp.139–151.
Hair Jr, J. F., Sarstedt, M., Ringle, C. M. and Gudergan, S. P. (2017), Advanced issues in partial least
squares structural equation modeling, Sage Publications.
Hashim, N.A.B., Raza, S. and Minai, M.S. (2018), “Relationship between Entrepreneurial Competencies
and Small Firm Performance: Are Dynamic Capabilities the Missing Link?”, Acad. Strateg. Manag. J.,
Vol. 17, pp.1–10.
Havierniková, K. and Kordoš, M. (2019), “Selected Risks Perceived by SMEs related to Sustainable
Entrepreneurship in case of engagement into Cluster Cooperation”, Entrepreneurship and Sustainability
Issues, Vol. 6 No. 4, pp.1680-1693.
Hurley, R. and Hult, G. (1998). Innovation, market orientation, and organizational learning: an integration
and empirical examination. The Journal of Marketing, 62, 42–54.
Ibrahim, M. and Madichie, N. (2014), “Entrepreneurial orientation among SME Wood Manufacturers in
Ghana”, African Journal of Business and Economic Research, Vol. 9 Nos. 2/3, pp.87-115.
Jiang, X., Liu, H., Fey, C. and Jiang, F. (2018), “Entrepreneurial orientation, network resource
acquisition, and firm performance: A network approach”, Journal of Business Research, Vol. 87, pp.46-
57.
Karlsson, C. (2007), Clusters, Functional Regions and Cluster Policies, CESIS Working Paper, No. 84.
Page 649
SEMIT 2023
Khan, M. A. (2022), “Barriers constraining the growth of and potential solutions for emerging
entrepreneurial SMEs. Asia Pacific”,Journal of Innovation and Entrepreneurship, (ahead-of-print).
Kleynhans, E.P.J. (2006), “The Role of Human Capital in the Competitive Platform of South African
Industries”, SA Journal of Human Resource Management, Vol. 4 No. 3, pp.55-62.
Krugman, P. (1998), “What’s new about the new economic geography?”, Oxford review of Economic
Policy, Vol. 14 No. 2, pp.7-17.
Leckel, A., Veilleux, S. and Dana, L. P. (2020), “Local Open Innovation: A means for public policy to
increase collaboration for innovation in SMEs”, Technological Forecasting and Social Change, Vol. 153,
No. 119891.
Lei, H. S. and Huang, C. H. (2014), “Geographic clustering, network relationships and competitive
advantage: Two industrial clusters in Taiwan”, Management Decision, Vol. 52 No. 5, pp.852-871.
Li, S., Han, S. and Shen, T. (2019), “How Can a Firm Innovate When Embedded in a Cluster?—Evidence
from the Automobile Industrial Cluster in China”. Sustainability, Vol. 11 No. 1837;
doi:10.3390/su11071837
Li, B., Xu, Z., Wu, H., Hong, N. and Skare, M. (2023), “Open innovation: a research framework and case
study of Huawei”, Technological and Economic Development of Economy, Vol. 29 No. 1, pp.278-306.
Lichtenthaler, U. (2020), “Agile innovation: The complementarity of design thinking and lean Startup”,
International Journal of Service Science, Management, Engineering, and Technology, Vol. 11 No. 1,
pp.157–167.
Liu, Y. and Wang, X. (2022), “Promoting competitiveness of green brand of agricultural products based
on agricultural industry cluster”, Wireless Communications and Mobile Computing, pp.1-18.
Lundequist, P. and Power, D. (2002), “Putting Porter into practice? Practices of regional cluster building:
evidence from Sweden”, European planning studies, Vol. 10 No. 6, pp.685-704.
Lundvall, B.-Å. (1985), Product Innovation and User-Producer Interaction, Aalborg, University Press,
Aalborg.
McCann, B. T. and Folta, T. B. (2011), “Performance differentials within geographic clusters”, Journal
of Business Venturing, Vol. 26 No. 1, pp.104-123.
Messick, S. (1989), “Validity”, In Linn R. L. (Ed.), Educational measurement (3rd ed.), pp. 13-104),
Macmillan, New York
Meyer-Stamer, J. (1998), Clustering, Systemic Competitiveness and Commodity Chaines: How Firms,
Business Associations and Government in Santa Catarina (Brazil) Respond to Globalization, prepared
for the International Workshop on Global Production and Local Jobs: New Perspectives on Enterprises,
Networks, Employment and Local Development Policy (Geneva: ILO, 9-10 March).
Mulligan, R., Ramos, J., Martín, P. and Zornoza, A. (2021), “Inspiriting innovation: The effects of leader-
member exchange (LMX) on innovative behavior as mediated by mindfulness and work engagement”,
Sustainability, Vol. 13 No. 10, p.5409
Nadvi, K. (1995), “Industrial Clusters and Networks: Case Studies of SME Growth and Innovation”. Case
studies prepared for the Small and Medium Industries Branch,UNIDO.
Nadvi, K. and Schmitz, H. (1994), Industrial clusters in less developed countries: Review of experiences
and research agenda. IDS Discussion Paper 339 Institute of DevelopmentStudies, Brighton.
Nakalembe, I., Dushime, J., Makuei, Y., Kwitonda, A., Hakizimana, S. and Muathe, S. (2023),
“Financing Start-ups, the Need, Relevance, Facets and Constraints in Kenya Start-ups
Ecosystem”, International Journal of Academic Research in Business and Social Sciences, Vol. 13 No.
1, pp.843-857.
Page 650
SEMIT 2023
Obadi, S. M. and Korcek, M. (2016), “The "Revealed" Comparative Advantage and Competitiveness of
the EU's International Trade visa vis the USA”, Ekonomicky Casopis, Vol. 64 No. 5, pp.397-422.
Özdamar, K. (2017), Ölçek ve test geliştirme yapısal eşitlik modellemesi IBM SPSS, IBM SPSS
AMOS ve MINTAB uygulamalı. [Scale and test development Structural equation modeling IBM SPSS,
IBM SPSS AMOS and MINTAB applied]. Eskişehir, Nisan Kitabevi.
Pan, X., Song, M. L., Zhang, J. and Zhou, G. (2019), “Innovation network, technological learning and
innovation performance of high-tech cluster enterprises”, Journal of Knowledge Management, Vol. 23
No. 9, pp.1729-1746.
Patrick, G. (2017), “Clustering and Product Innovativeness: A Literature Review of Small and Medium-
Sized Enterprises (SMEs) in Kenya”, International Journal of Academic Research in Economics and
Management Sciences, Vol. 6 No. 4, pp.115-145.
Premaratne, S. P. (2001), “Networks, resources, and small business growth: The experience in Sri Lanka”,
Journal of small business management, Vol. 39 No. 4, p.363.
Prifti, R. and Alimehmeti, G. (2017), “Market orientation, innovation, and firm performance— an
analysis of Albanian firms”, Journal of Innovation and Entrepreneurship, Vol. 6 No. 8, pp.1-19. DOI
10.1186/s13731-017-0069-9
Quaye, M.D. and Mensah, I. (2017), “Industrial Cluster and Competitive Advantage of Micro-Firms:
Insight from Wood Industry”, Journal of Creativity and Business Innovation, Vol. 3, pp.170-196.
Ringle, C.M., Wende, S. and Becker, J.M. (2015). SmartPLS 3, Hamburg: SmartPLS.
Retrieved from http://www.smartpls.com
Rosenbusch, N., Brinckmann, J., & Bausch, A. (2011). Is innovation always beneficial? A meta-analysis
of the relationship between innovation and performance in SMEs. Journal of Business Venturing, 26(4),
441–457.
Rosenfield, S. (1995), “Bringing Business Clusters into the Mainstream of Economic
Development”, European Planning Studies, Vol. 5, pp.3–23.
Schmitz, H. and Nadvi. K. (1999), “Clustering and industrialization: introduction”, World development,
Vol. 27 No. 9, pp.1503-1514.
Škerlavaj, M. Song, J. H. and Lee, Y. (2010), “Organizational learning culture, innovative culture and
innovations in South Korean firms”, Expert Systems with Applications, Vol. 37 No. 9, pp.6390–6403.
Świerczek, A. (2019), “Networking and the Competitive Advantage of Supply Chains. Scientific papers
of Silesian University of Technology”, Organization and Management Series, No. 135
Swistak, A. (2015), “Tax penalties in SME tax compliance, Financial Theory and Practice, Vol. 40 No.
1, pp.129–147.
Terziovski, M. (2003), “The relationship between networking practices and business excellence: a study
of small to medium enterprises (SMEs)”, Measuring business excellence, Vol. 7 No. 2, pp.78-92.
Walter, A., Auer, M. and Ritter, T. (2006), “The impact of network capabilities and
entrepreneurial orientation on university spin-off performance”, Journal of Business Venturing,
21, 541-567.
Werekoh, E. A. (2022). The Effects of Taxation on Economic Development: The Moderating Role of
Tax Compliance Among SMEs, Preprints, DOI: https://doi.org/10.21203/rs.3.rs-1238141/v1
Wernerfelt, B., (1984), “A Resource-based view of the firm”, Strategic Management Journal, Vol. 5 No.
2, pp.171-190.
Yusheng, K. and Ibrahim, M. (2020), “Innovation capabilities, Innovation types and Firm Performance:
Evidence from the banking sector of Ghana”, Sage Open, pp.1-12,
https://doi.org/10.1177/215824402092089
Page 651
SEMIT 2023
Yusheng, K. and Ibrahim, M. (2019), “Service innovation, service delivery and customer satisfaction and
loyalty in the banking sector of Ghana”, International Journal of Bank Marketing, Vol. 37 No. 5, pp.1215-
1233. https://doi.org/10.1108/IJBM-06-2018-0142
Zaridis, A.D., (2009), “Competitive Advantage and its Sources in an Evolving Market”, AIP Conference
Proceedings, 1148, 917-921.
Zeebaree, M.R.Y. and Siron, R.B. (2017), “The Impact of Entrepreneurial Orientation on Competitive
Advantage Moderated by Financing Support in SMEs”, International Review of Management and
Marketing, Vol. 7 No. 1, pp.43-52.
Zhu, Q., Cordeiro, J. and Sarkis, J. (2013), “Institutional pressures, dynamic capabilities and
environmental management systems: Investigating the ISO 9000—Environmental management system
implementation linkage”, J. Environ. Manag., Vol. 114, pp.232–242.
Page 652
SEMIT 2023
SP17: Humanities and Social Sciences
Strengthening Women leadership in India & Globe in digital Era

Meenakshi Kaushik1 and Manu Kaushik2
1
Professor, Dean (Research, Trinity Institute of Innovations in Professional Studies, GGSIPU, India
2
Student, BA(Hons) Journalism, Maharaja Agrasen College, University of Delhi, India
Corresponding Author: kaushikmeenakshi36@gmail.com
Abstract
In India and globe the status of women has transitioned over a period of time. Since women are an integral
part of the globe, it is very important to understand significantly the contribution of women in global
economy. Bidding adieu to the traditional women are gradually making their presence felt in the
managerial rungs in a variety of business establishments. National education policy (NEP,2020), several
other governmental schemes & initiatives, continuous technological advancements have set a clear vision
and growing need of digital literacy for Indian women. This research paper is highlighting an urgent need
for strengthening effective leadership in women in India and globe. It is very important to create a
supportive and collaborative work environment by embracing digital transformation, digital literacy and
technological competence in women.
Keywords: Women Leadership, Digital Transformation, Digital literacy, Digital technology, new
education policy (NEP), leadership programs, governmental initiatives.
OBJECTIVES
1.To understand the significance of women leaders and their contribution in business organizations.
2. To study the status of women leaders and their effective leadership performance in present scenario.
3.To study the National Education policy (NEP,2020) and few Indian government initiatives as a way to
strengthen women leadership.
4.analyse the factors and forces contributing to effective leadership in women executives.
5.To study the role of digital transformation and women leaders in business organization
1.Introduction
1.1Conceptual framework of leadership in women executives
According to one survey, when women are empowered to act as leaders, everyone gets benefit. Studies
show that women leaders help increase productivity, enhance collaboration, inspire organizational
dedication, and improve fairness, transparency and productivity. Only 10% of Fortune 500 companies
are led by women. Industrial and organizational psychologists offer a host of evidence-based strategies
to reduce gender gap. Potential women should be imparted coaching, mentoring and sponsorship
programs to exercise leadership in their full capacity. According to Alice Eagly, professor of psychology
emerita at Northwestern University and pioneer in researching women’s leadership, women face
challenges that are greater than those faced by their male counterparts. There is not only an urgent need
to increase the number of women leaders but also to strengthen their capabilities by giving them
opportunities.
Page 653
SEMIT 2023
According to leadership General management 2021 report, society and any institution cannot function
effectively without women’s equal participation in leadership activities. Female leaders are
transformational leaders working as a role model to their subordinates. Women leaders emphasize more
on team work and communication. Females are well known for their caring, protecting and compassionate
nature. A leadership role highly demands to acquire and develop new skills (Ibarra et al.,2013). Few
initiatives should always be taken in to consideration for educating women, developing their managerial
skills through different leadership programs. Women in leadership roles should always be equipped with
latest digital technology like AI, cloud computing and focussed upskilling, reskilling and multi-skilling.
Their aptitude skills, consulting and presentation skills, soft skills, case study, simulations, focussed
group mentorship should be developed. Leadership has been a predominantly masculine role through
centuries that few females have obtained (Eagly & Karau, 2002). Few leadership traits like ambition, self-
reliance and assertiveness, decision making was acquired by women without compromising on their
gender traits. Their emotional intelligence is also found to be higher than men (Eagly & Carli,2007).
Aldrich and Lotito (2020) implied that the leader’s self-help and others to do the right things. Female
leaders are transformational leaders working as a role model to their subordinates. Women leaders
emphasize more on team work and communication. Females are well known for their caring, protecting
and compassionate nature. A leadership role highly demands to acquire and develop new skills (Ibarra et
al.,2013). Few initiatives should always be taken in to consideration for educating women, developing
their managerial skills through different leadership programs. Women in leadership roles should always
be equipped with latest digital technology like AI, cloud computing and focussed upskilling, reskilling
and multi-skilling. Their aptitude skills, consulting and presentation skills, soft skills, case study,
simulations, focussed group mentorship should be developed. Leadership has been a predominantly
masculine role through centuries that few females have obtained (Eagly & Karau, 2002). Few leadership
traits like ambition, self-reliance and assertiveness, decision making was acquired by women without
compromising on their gender traits. Their emotional intelligence is also found to be higher than men
(Eagly & Carli,2007). Leadership is a very important function to maximize efficiency. Aldrich and Lotito
(2020) implied that the leader’s self-help and others to do the right things. Various factors affect the
performance of women such as the gender wage gap, lack of safety policies and flexible work offerings.
It’s important to attract and retain professional women with good leadership qualifications and skills in
senior positions to maintain a good standard and professionalism. India government has launched several
schemes for women empowerment and upliftment. There is an urgent call to develop future ready skills
among women and girls to make them handle their jobs and leadership positions.
According to Putnan & Heinen, an effective leader is supposed to be confident, aggressive, forceful,
competitive and achievement-oriented, but these traits tend to be more often associated with men rather
than women. Women are generally depicted as emotional, passive, dependent and faced with a paradox
that if she displays the culturally defined traits of a women, she is unacceptable as an executive & and if
she acts as a leader, she is called unfeminine. According to Reif, Newston Monezka, said regarding
special training program that this is only needed if women are different from men in capabilities &
qualities but various research studies show that women are in no way different in their capacities &
capabilities. Besides that, women have more abstract visualization than men. This says that there are to
be more women in management at top most positions then men. Women at higher level should help other
women by providing encouragement and explaining them the problems that will be encountered in
attempting to move up the corporate ladder. Thus, women managers could provide insights and share
experience which might help another aspiring female further her career.
Burrow in his studies also observed women managers are most dependable, more concerned & give better
attention to detail than men. Thus, to do away the various myths existing in the society hence to provide
encouragement to improve herself as well as work in a way to change the attitude of the society towards
them. Study by experts revealed that rare endowment of capacity for endurance enables them to explore
the possibilities of compromises and amicable settlements in case of disputes, unlike men who tend to be
little impulsive and even dogmatic on occasions.
According to the experts, the women have an innate desire to build up things, be they business
relationships, collective approaches or cooperative initiatives. In mellowed women, the ego is not as
intrusive as in men.The possession of feminine characteristics is seen as being relatively detrimental to
Page 654
SEMIT 2023
careers (Powell and Butterfield 1979, 1909; Rosen and Jerdee, 1974. Schein 1973 although this
perception is beginning to change among female managers (Brenner, Tomkiewicz and Schein. 1989).
Masculine attributes are viewed as being beneficial (Brenner et. al. 1989; Powell and Butterfield, 1979,
1989; Rosen and Jerdee, 1974; Schein, 1973).
Management guide (2021) evaluated the leadership is the ability to motivate a team to achieve a certain
goal. A leader is a person who works with people by communicating the plans and policies to them.
Leaders help and guide others to do the right things & set the direction and work efficiently and
effectively. As claimed by women’s leadership,2021, a leadership style refers to a leader’s behaviour and
methods while directing, motivating and managing others.
Davidson and Cooper (1986) examined the stresses experienced by women at all levels of management
and compared them to their male counterparts in British organizations. The results suggested that women
managers need to be more flexible in their management style and require both general business skills and
assertion and confidence training skills.Vaz (1988) while studying the women managers in the banking
sector finds out some of the commonly held beliefs regarding women is that they do not display an
objective outlook and lack confidence. It is also held that men value challenges in work more than women
and are willing to undertake any type of difficult assignments. Women in managerial positions are better
able to show objectivity and balanced judgment in decision making and what is even more important is
that they keep an open mind and are less vulnerable to external pressures.Another opinion voiced is that
women tend to be more emotional and temperamental and that men have a more suitable outlook and are
therefore, better suited for executive positions. Furthermore, a view is held that successful women should
adopt a masculine profile and should discard feminine characteristics. But the issue of emotionalism is
itself debatable. Although women are not handicapped intellectually, the stereotype of the ideal woman
discourages expression of abilities.
Morrison 1989 in her book, “Breaking the Glass ceiling: Can women reach the top of America’s largest
corporation?’ has observed that mounting evidence suggests when careers are matched women are
remarkably similar to men in their characteristics, abilities and motives. Yet stereotypical perceptions
lead people to set unrealistic standards for executive women and these expectations are part of the
environment in which women must work and live. In his study Agrawal (1989) found that as individuals,
men and women executives are similar psychologically, intellectually and emotionally. Very few
personality differences are found between the two. In small group training experiences women are found
to be able to lead, motivate and influence other group members in the same way as men do.
1.2 Need for an effective leadership in women executives
According to leadership and general management report 2021, any institution cannot run effectively
without participation of women. Women create a viewpoint that brings healthy competition and a sense
of belongingness among team members and within organizations. John Maxwell expressed that
leadership is influence nothing more and nothing less(kruse,2013).In present scenario, human centred
leadership is to be manifested to empower women to imbibe assertiveness, decisiveness and
persuasiveness by incorporating core competencies and skill-sets focussing on caring and nurturing, soft
skills, interpersonal communication, cooperation, collaboration, based on participative leadership
styles(Syed & Murray,2008).Women are also able to develop an ethic of caring, negotiation and conflict
resolution skills .According to a survey on Human centred leadership, women are considered to be more
transformation oriented than men. Female leaders focus more on trust and team building than men and
so they are more human oriented. Many CEOs and leaders support and accept the promotion of women
to senior management positions & onto the board positions. Yet there a need for cultural shift to guide
women on their career development that leverage their education and skills (Deloitte Global Boardroom
Program,2022).
According to the GEM (global entrepreneurial Monitor ,2022), women leaders have an important impact
on all the levels of activities through creativity, innovation, job creation and competitiveness. Some
authors like Bhatt,2015 consider that there is a higher demand of women leadership in India and globe
due to the changing turbulent times. Men tend to be aggressive while women are tender and promote
teamwork by keeping aside their egos (APA), Mohammadi et al. (2019). As women come across with
Page 655
SEMIT 2023
many hurdles and barriers while aspiring to the higher ladder, women become more stronger and a
positive and inspiring team leader. Women leaders are proven great listeners (Hazarika,2019). Women
are considered as a change agent who not only listen but welcome the ideas of their fellow members and
remain open to changes and challenges in the dynamic market and promote those necessary changes if
required. They are more dedicated towards goal, vision and mission of the organization. Through their
soft skills they create a culture of empathy, compassion promoting each and every member to participate
equally. According to many studies, it has come out that now 1 out of every 4 leaders are women. This
still doesn’t satisfy equality but women came a long way to gain this opportunity (Tandon,2018).
1.3 Strengthening Women in Digital Era
An effective women leadership in organizations in the digital era refers to someone demonstrating
effective leadership attributes leveraging digital technologies by navigating business complexities in
modern digital era to achieve strategic goals. An effective woman leader exhibits strong leadership
qualities such as clear-cut vision, strategic thinking, emotional intelligence, communication skills,
adaptability, and the ability to inspire and motivate others by maintaining work life balance. In digital
era, women need to be digitally skilled and aware of digital technologies by understanding the
implications for business. They have the ability to harness digital tools, data analytics, and emerging
technologies to drive innovation, improve efficiency, and make informed decisions thus empowering
women as future ready employees in India and globe and also navigating the upskilling, multiskilling and
re-skilling. Women leaders create a sense of belongingness and a bond among the employees. An
effective leader should always be regular and consistent with sharing the realities of the business. They
absorb criticism and scepticism fairly and promote consistent quality of conversation permeable across
levels, functions and demographies. An effective leader believes in adequate inclusiveness, empathy and
respect. A leader should build a culture where they really need to be celebrated, not tolerated. Managers
and leaders need to develop a robust reward system and talents management system. It is the
responsibility of women leaders to continuously explore the significant role of technology trends and
using those tools and technology to explore various opportunities to bring necessary changes in
organizations with the integration of artificial intelligence (AI), IOT and immersive tools and thus
enhancing more employee engagement, personalization and efficiency and effectiveness. Technology
evolution fosters a dynamic tech-
Effective women leader formulates & executes the strategic plans aligning with the organization's digital
transformation and digital technology goals. An effective woman leader believes in collaboration and
networking with various stakeholders within and outside the organization to share knowledge, resources,
& explore potential partnerships and opportunities. They learn to be resilient in the face of challenges and
setbacks.Women leaders always follow ethical leadership by adhering to strong ethical principles and
lead with integrity. Women leaders prioritize ethical conduct, fairness, justice and social responsibility,
setting an example for ethical behaviour within the organization. And thus, they coach, mentor and
empower other women within the organization by developing their leadership skills. Effective women
leaders help other women advance and grow their careers. They excel in driving strategic initiatives,
fostering innovation, and adapting to the dynamic nature of the modern business world while upholding
ethical standards and empowering others.
According to Varghese 2021, studied there are fewer women leaders or role model in the limelight to
promote women leadership. People need to praise women leaders. If given opportunities women can
make their experience a way to their success with their deep and broad thinking. As claimed by Moyer et
al. (2018), evaluated one of the of the best ways to give women more opportunities are by seeing them
more than a daughter, wife or mother. Above all they are human beings with separate identities and
qualities.
Effective women leaders always have emotional intelligence and work life balance that refers to the
ability to understand and manage emotions of self and team-members. Leaders must be able to
communicate effectively to convey their vision, provide guidance, and inspire others. They should be
skilled in both verbal and nonverbal communication, active listening, and adapt their communication
style to different audiences and context. An effective leader always believes in making timely decisions
by gathering relevant data and information, consider various perspectives, and use critical thinking to
Page 656
SEMIT 2023
evaluate options before arriving at a decision. They are also willing to take responsibility for their
decisions and learn from mistakes.
2. Strengthening effective leadership in Women executives in business organizations

2.1 National Education Policy (NEP,2020) and Women holistic development
The transformative National Education Policy (NEP), 2020 has set an ambitious vision for Indian youths
and women. As per NEP 2020, new education policy is framed and formulated for the holistic
development of women and to foster critical thinking, visualization development, problem solving skills
and ideation in women aspiring to higher ladder incorporates. The NEP suggests a flexible and
multidisciplinary curriculum that includes coding enabling all to plan and organize thoughts, cognitive
skill improvement, or decision-making skills and to develop 21st century skills among aspiring women
leaders.India’s national curriculum framework (NCF) 4.6.1 pedagogic principles nurture students
including girls and women holistic and integrated understanding. Skill development promotes digital
literacy among women by utilizing ICT tools and resources. New Education Policy believes in A.D.D.I.
E a powerful analytical tool and Four-Fold approach to learning and developing better understanding of
the surroundings, critical thinking, broad understanding of the dynamic market, and developing
conceptual clarity by engaging and interactive games. New Education policy enhances skills and concept
recall, strategic and extended thinking. Thus, NEP promotes problem solving skills, numerical skills,
digital literacy logical and creative thinking, project applications and developing innovative mindset,
ethics and ergonomics, cyber security that are very helpful for women to work in 21st century
organizations and industries. NEP through its various modules promotes media literacy, life skills, design
thinking and future technologies like AR, VR Artificial intelligence, and cloud computing thus promoting
more and more girls and women to get technical and vocational skills to perform well in any area of the
organizations and thus Developing effective leadership skill and embracing digital transformation &
developing digital literacy and technological competence ,empowering and inspiring other women,
balancing work and personal life, emotional intelligence and self-awareness & advocating for diversity
and inclusion and above all creating equal opportunities for women.
2.2 Pradhan Mantri Mudra Yojna (PMMY)
According to Women’s Web (2021) one of the best initiatives taken by Indian government is Pradhan
Mantri Mudra Yojna (PMMY) providing financial assistance to aspiring women leaders and
entrepreneurs for starting their own business and thus bringing confidence as well as ensuring survival,
participation and protection of women leaders. Every sector needs to promote women leadership and
people need to uplift women by providing them opportunities to demonstrate their talent. The pay gap
needs to be reduced to encourage more women to work and show their full potential and effective
leadership style.
2.3 Equal pay Act,1963
Lawnn (2021), equal pay Act was launched in 1963 with the aim to protect against wage discrimination
based on sex. The act protects men and women from the discrimination wages, a great initiative taken by
united states of America government.
2.4 E-Adhigam Scheme one of the most ambitious digital learning programmes in the country. Under this
scheme nearly three lakh tablets distributed among school students including girls and women. The
devices come with pre-loaded content along with personalised and adaptive learning software and 2 GB
free data and thus not only enhancing knowledge and wisdom of girls and women but also enhances their
leadership capability
Factors affecting women leadership in business

Factor that restricts women managers to reach to the top echelon of management are blocked
opportunities, lack of support of employers, limited access to information, restricted access to training,
Page 657
SEMIT 2023
fear of success, marriage and motherhood, prioritization of family over career, stereotyping and
perceptions. Discrimination against women managers exists, relating to remuneration job allocation,
performance appraisal, promotion, training opportunities and reward structure. The exclusion of women
from policy formulation and decision making relegates them to peripheral, supportive activities, women
are better represented in departments like personal, human resource management, marketing, public
relations and training and there is a scarcity of women in engineering and finance.
Skill up-gradation is vital in this digital era. Women managers are being praised by their male
counterparts for their soft skills like caring, understanding, good team work good communication skills,
patience, perseverance etc. Performance appraisals and consequent compensation packages should be so
structured and no room should be left for discrimination. Companies must have objective methods of
performance appraisal and a non-discriminatory compensation and incentive packages ensuring and
enhancing the opportunities for advancement of women and minorities.
Monitoring access to training and development facilities is significant to prevent and break the glass
ceiling. Various researches have proven time and again that the performance of women and minorities go
up a number of times after attending training and development workshops. So, is it not the folly of the
company to allow these minorities and women to develop themselves through training but then hold them
back by the invisible barriers. While recruitment, it is necessary that company should make sure that the
process is not biased against the minorities and women.
Mentoring is also an important aspect of internal development assisting individuals in their advancement.
Counselling help to improvement in effectiveness of women executives. Promotional policies for women
may also prove to be helpful for making them effective. With globalization of the Indian businesses, more
companies are now realizing the value and virtue of gender diversity in their human capital by bringing
in more women managers in the company.
Vidya Chhabaria, the chairperson of Jumbo Group, is a case in point. Among other women mangers,
Indra Nooyi, PepsiCo's worldwide President and Ranjana Kumar, chairperson of Indian Bank is worth
noting. At various departments many women professionals have been promoted in the hierarchy chain.
One such example is the ICFAI Press where the numbers of women employees have seen a healthy
growth in the last couple of years. It has almost doubled from about 20 in 2002 to 37 by November 2003,
and, the list goes on.Kidwai is one of the highest paid women executives in the glamorous highly
competitive, and networked world of investment banking Women like Sucheta Jain, daughter of SP
Oswal, at barely 30 is the executive director of Vardhman Spinning and General Mills. Shobhana
Bhartiya, Vice group chairman of the Hindustan Times. Preetha Reddy, MD, Apollo Hospitals; Mallika
Srinivasan, MD, Tractor and Farm Equipment; Anu Aga, Chairperson, Thermax and Geeta Talwar, MD
Ultimate Automobiles; are hardnosed business women who take tough decisions and run their empires as
well as any other man would have in their position. Sulajja Motwani, the joint managing director of the
Kinetic Group, is self-directed and self-motivated.
Discussion and Conclusion

Women are an integral part of the society and globe thus it is very important to understand the contribution
of women in global economy. National education policy (NEP,2020) and several other governmental
schemes & initiatives, continuous technological advancements have set a clear vision and growing need
of taking more women leaders at higher positions. Organizations and industries are needed to emphasize
more on digital literacy of women as well as reskilling, upskilling and multiskilling to be more competent
and stronger leader. Discrimination against women managers still exists, relating to remuneration, job
allocation, performance appraisal, promotion, training opportunities and reward structure. Women
managers face a dual role responsibility creating a lot of stress, strain, and role conflict. Women leaders
should be given opportunities to attend various training programmes and workshops and take full
privileges of these programs.In conclusion, the researcher wants to convey that various aspects of women
leadership need to be discussed and addressed on time to bring forth new innovative solutions related to
women leadership. Women have displayed an array of abilities like tolerance, patience, forbearance,
concentration, team-spirit, and determination. The leadership effectiveness in women can be developed
Page 658
SEMIT 2023
by providing them multi-dimensional education, training and workshops to women by changing their
vision into reality and encouraging them to develop their full potential. It is not surprising that the
employment scenario is undergoing far-reaching transformation. An effective women leadership in
organizations in the digital era refers to someone demonstrating effective leadership attributes leveraging
digital technologies by navigating business complexities in modern digital era to achieve strategic goals..
In digital era, women need to be digitally skilled and aware of digital technologies by understanding its
implications on business. They have the ability to harness digital tools, data analytics, and emerging
re-skilling. Women leaders create a culture of sense of belongingness and compassion to all. They absorb
criticism and scepticism fairly and promote consistent quality of conversation permeable across levels,
functions and demographies. A leader should build a culture where they really need to be celebrated, not
tolerated. Women leaders being transparent and empathetic develop a robust reward system and talent s
management system and try to understand the core needs of their employees and being sensitive to their
needs. Women in leadership roles should always be equipped with latest digital technology like AI, cloud
computing and focussed upskilling, reskilling and multi-skilling. Their aptitude skills, consulting and
presentation skills, soft skills, case study, simulations, focussed group mentorship should be developed.
Women in leadership roles should continuously attend various training and mentorship programs tailored
to the organizational needs. Policies should be framed in such a manner to encourage more work-life -
balance, emotional balance, resiliency and agility in them. In summary, it is highlighted that continued
progress is required towards diversity and inclusion by supporting more and more women in their
leadership journey to create an equitable society where all can achieve heights of career with right kind
of skillsets and qualifications.
Recommendation:
1.The outcome of the study is to make women aware of the current market trends and developing wisdom
and skill-sets as per the recent trends and dynamic market and accordingly training and developing
women leaders to reinforces workforce diversity, advancement of women in this digital era.
2. Managing a career and a family is clearly a matter of setting priorities and making right choices. So
dual career couples need to manage their life style effectively. The parallel pursuit of career and family
roles necessitates greater understanding and maturity on their parts.
3. Appropriate career planning of women managers where they can rise to the needs of the job and excel
through competence and to reduce the conflict between family and professional life by maintaining a
work-life balance.
4.Organizations should conduct specialized training programs on latest trends and technology for women
aspiring for a higher position to make them more competent and harness their skills as well as coping
strategies. Trainings should also be imparted on stress prevention techniques like yoga, meditation, and
physical exercises to teach women more about work life balance and emotional balance. Various stress
preventing strategies such as prioritizing work, delegating work, laughter exercise, relaxation, diet, proper
sleep, meditation, yoga, holistic therapies can be taken into consideration.
5.The company should provide supportive environment for employees to manage their personal lives by
providing flexi timings, commuting facilities, and sharing of job responsibilities.
6.Re-skilling, Multiskilling and upskilling is required as per the demand of the dynamic market. Women
in organizations need to be digitally literate about digital technology and related aspects.
7. Women at higher level should help other women by providing encouragement and explaining them the
problems to be encountered in attempting to move up the corporate ladder. Thus, women managers could
provide insights and share experience which might help another aspiring female further her career. by
maintaining work life balance.
Page 659
SEMIT 2023
8.In digital era, women need to be digitally skilled and aware of digital technologies by understanding
the implications for business. They have the ability to harness digital tools, data analytics, and emerging
re-skilling.
9.The company should have inbuilt organizational support systems that promote scheduling, flexibility,
flexi time, maternal, and paternal leave policies, resources and referrals for child care reflecting the degree
of support towards minority and women.
10.Various factors affect the performance of women such as the gender wage gap, lack of safety policies
and flexible work offerings. It’s important to attract and retain professional women with good leadership
qualifications and skills in senior positions to maintain a good standard and professionalism. Indian
government has launched several schemes for women empowerment and upliftment. There is an urgent
call to develop future ready skills among women and girls to make them handle their jobs and leadership
positions
11. Organizations need to develop a robust reward system and talents management system. A leader
always tries to understand the needs of the employees and be sensitive to their needs.
References:
Aldrich, A.S. and Lotito, N.J., 2020. Pandemic performance: women leaders in the Covid-19crisis.
Politics & Gender, 16(4), pp.960-967.
APA, 2021. When the Boss is a Woman. [online] https://www.apa.org.
Brooten, B., 2020. Women leaders in the ancient synagogue. Scholars Press.
Carli, L. L., &Eagly, A. H. (2007). Overcoming resistance to women leaders: The importance
ofleadership style. In Women and leadership: The state of play and strategies for change (pp. 127-148).
Jossey-Bass.
Cooper, C & lewis, S. the workplace revolution: Managing Todays dual career families,Kogan Page ,
London,1993.
Cooper ,Cary & Davidson, Marilyn J.high Pressure :Working Lives of Women Managers
Ionlaaana,1982.
Dappa, K., Bhatti, F., &Aljarah, A. (2019). A study on the effect of transformational leadership on job
satisfaction: The role of gender, perceived organizational politics and perceived organizational
commitment. Management Science Letters, 9(6), 823-834.
Dipboye,R.l. and Wily. J.W.(1977), “ reactions of College recruiters a function of Interviewee Sex and
Self Presentation Style” Journal of Vocational Behaviour, 10,1-15
Dipboye,Robert L, Women Managers ,Stereotypes & Realities Survey of Business .May-June 1975
Reprinted In Belle ANN stead. Women in Management
Eagly, A. H., &Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders.
Psychological review, 109(3), 573.
Eagly, A. H., Karau, S. J., &Makhijani, M. G. (1995). Gender and the effectiveness of leaders: a meta-
analysis. Psychological bulletin, 117(1), 125.
https://www.gemconsortium.org/reports/latest-global-report
Hazarika, D., 2019. Women empowerment in India: A brief discussion. International Journal of
Educational Planning & Administration, 1(3), pp.199-202
Heathfield, S., 2021. What Organizations Can Do to Promote Women in Leadership Roles. [Online]
TheBalance Careers.
Page 660
SEMIT 2023
Leadership & General Management, 2021. The 5 Leadership Styles you can use &How to implement
teach. [Online] IMD business school.
Mc Kinsey and Company, 2021. [Online] Available at: <https://www.mckinsey.com/featured-
insights/diversity-and-inclusion/women-in-the-workplace.
Moyer, C.A., Abedini, N.C., Youngblood, J., Talib, Z., Jayaraman, T., Manzoor, M., Larson, H.J.,Garcia,
P.J., Binagwaho, A., Burke, K.S. and Barry, M., 2018. Advancing women leaders in globalhealth: getting
to solutions. Annals of global health, 84(4), p.743.
Naisbitt,j., and Aburdene , P.( 1990),Magatrend 2000,USA , Marrow.
Hackman, M. Z., Furniss, A. H., Hills, M. J., & Paterson, T. J. (1992). Perceptions of gender-role
characteristics and transformational and transactional leadership behaviours. Perceptual and Motor
Skills, 75(1), 311–319. https://doi.org/10.2466/PMS.75.4.311-319
Riger, S., &Galligan, P. (1980). Women in management: An exploration of competing paradigms.
American Psychologist, 35(10), 902.
Page 661
SEMIT 2023
Staff scheduling in the service industry considering qualifications

S. Bouaziz, B. Chaker and D. Dhouib
olidlab.tn@gmail.com
Abstract
Research has demonstrated that effectively managing human resources in the service sector has
significant advantages for managers and employees, particularly when staff is near clients. Due to
the variety of aspects that need to be taken into account, this scheduling challenge has proven to
be a difficult task, for even directors and professional planners. This paper deals with personnel
scheduling problems in the technical control field. To reduce workload and average premiums,
jobs are distributed to inspectors over a time horizon. The solution complies with a variety of
restrictions, such as the maximum tolerated radiation dose exposure limit, staff certifications, and
work hours. A new mathematical framework is developed using the goal programming approach.
Results show a significant performance improvement in the scheduling operation.
Keywords: Personnel scheduling, Qualification, Inspectors, Goal programming
1. Introduction
The Staff scheduling problem, while a subject of extensive research over the years, remains an immensely
relevant and pressing concern in contemporary society. It pervades various sectors, including industries,
production firms, and especially the services sector. This enduring relevance is rooted in the fact that the
efficient scheduling of staff is fundamental to facilitating the functioning of these companies.
An inspection company's strength is directly related to how well it makes decisions and uses critical
thinking then, Flexible and skilled individuals are essential in the dynamic world of consumer
expectations and market changes. In addition to optimizing problems, they also the changing needs of
their customers. Furthermore, these companies can draw in and keep top people, ensuring their continued
success. So, an inspection company's capacity to adapt and endure in a shifting environment depends on
having staff with good decision-making and critical thinking abilities. In such wise, effective employee
scheduling is essential to including inspection firms. On the other hand, efficient staff scheduling lowers
stress and increases job satisfaction. Additionally, badly designed inspections can hurt a company's profits
owing to inefficient resource use, possibly leading to financial instability and team morale problems then,
ineffective scheduling results in lower production and increased absenteeism rates. As a result, organized
scheduling is crucial for the success and sustainability of inspection companies.
The staffing problem insists on the significance of employees in companies to form teams of workers and
assign them a variety of tasks in a way that promotes worker fairness [1]
The overall employee scheduling problem, however, necessitates determining the appropriate staffing
among shifts while taking into account several constraints, such as labor laws in various industries,
employee availability, and skill levels [2].
Finding the best optimal solutions to these complex and heavy problems is a challenge. This study looks
at a staff scheduling issue that occurred at a technical control company and aims to pinpoint managers'
expectations and goals as well as the various types of restrictions placed on the ideal schedule.
2. Staff scheduling: A brief review

Staff scheduling has captivated and engaged scholars for more than 60 years. Despite evolving
technological landscapes and shifting economic paradigms, the scientific community's unwavering
enthusiasm and dedication to this subject. This continual interest may be partially explained by the
perpetual shift in our socioeconomic environment, which necessitates regular revitalization and
adaptation of planning processes. As companies and organizations deal with demands that are getting
Page 662
SEMIT 2023
more complex, from evolving market dynamics to changing labor preferences, the art of staff scheduling
continues to be a critical cornerstone in ensuring operational efficiency. To boot, as is evident from
looking ahead to the future of staff scheduling, the scientific community is ready to keep being very
engaged in this field, modern scheduling methods with efficiency will likely be developed as a result of
the hunt for solutions to this complex problem.
In addition to addressing a company's urgent workforce needs, these solutions are projected to alter
resource allocation and management throughout all operational sectors of a corporation. These
groundbreaking concepts aim to give enterprises a distinct competitive edge in a fiercely competitive
global context by enabling them to seamlessly match their workforce with changing demand, they also
promise to enhance daily and monthly operations. In the current digital era, scheduling is essential across
all industries, particularly those that provide services like healthcare, Manufacturing and construction
industry, call centers, transportation, and telecommunications, these industries understand that personnel
scheduling is crucial for providing high-quality services, maximizing the utilization of human resources,
and raising consumer happiness. Overall, personnel planning and scheduling will continue to be essential
to research and innovation, influencing how businesses in various industries run and serve their
consumers, and highlighting its significance in raising productivity, efficiency, and customer happiness
in any sector in the world.
2.1. Healthcare services:
The healthcare sector has more problems with personnel scheduling, according to scientific studies that
have focused on scheduling issues.
[3]aims to understaffing in nurse scheduling that can cause poor patient care and nurse burnout.
Researchers came up with a novel approach to address this issue that uses a conditional value at risk
constraint to guarantee that the probability of a human resource deficit is below a predetermined level.
Besides, a two-step stochastic program forms the foundation of the suggested approach. The nurse's
schedule is first established. Second, the patient's desire is fulfilled, and the cost of all nurse preparation
is minimized while adhering to all restrictions. Therefore, the researchers presented a deconstruction
approach to handle the issue dividing the main issue into several smaller issues as these problems are
resolved using a variety of methods and were used in an actual case study of a hospital to apply the
suggested methodology.
The findings demonstrate that this approach is capable of a nurse schedule that satisfies every requirement
while lowering the cost of the nursing schedule and the likelihood of understaffing.
[4] focuses on the complex problem of personnel scheduling within a hierarchical management system
for nursing staff. At the edge of this, the authors suggest a model for efficiently allocating workloads that
takes into account practical considerations, and weekend scheduling preferences, to solve the issue, they
combine single heuristic algorithms with hybrid heuristics. In terms of generational distance and Pareto
solution spacing, the experimental results demonstrate how well the hybrid algorithms outperform the
single algorithms in creating high-quality nurse schedules and then optimizing relative fairness. After
that, these results demonstrate the method's potential for handling the complex problems of personnel
scheduling in nurse staff management with hierarchical structures. The authors recommend more research
to improve the algorithm's performance, highlighting the significance of taking hospital managers' and
nurses' preferences into account when designing nurse rostering models and the need for thorough
evaluation metrics that take factors like fairness, workload distribution, and nurse satisfaction into
account to effectively assess schedule quality.
[5] deals with the difficult challenge of nurses scheduling blend data envelope analysis with higher
constraint approaches. Aside, this novel technique is meant to integrate the multiple competing goals of
nursing planning, including the urgent goals of cost reduction, improved patient care quality, and
equitable workload distribution. Besides, the study findings show that this combination technique has
superior effectiveness in improving nurses planning outcomes, and companies in the healthcare industry
are seeking to increase staff motivation and resource allocation.
Page 663
SEMIT 2023
Furthermore, the authors demonstrate the actual use of the proposed model by undertaking an engaging,
realistic case study in a hospital.
Surprisingly, the model produced a nurse's schedule that completely met all predetermined targets.
[6]combines data envelope analysis with increased methods to handle the complicated problem of nurse
scheduling. This innovative approach is intended to integrate the various competing aims of nursing
planning, including the urgent goals of reducing costs, enhancing patient care quality, and distributing
workload equitably. The study findings reveal the superior effectiveness of this combined technique in
considerably enhancing nursing planning outcomes, making it a vital tool for businesses. Healthcare firms
are attempting to improve employee motivation and resource allocation.
Furthermore, the authors demonstrate the suggested model's practical use by conducting an interesting,
realistic case study in a hospital neurology unit.
Surprisingly, in this actual example, the model reached a nursing schedule that perfectly matched all
predetermined goals, while also taking into account individual nursing staff preferences.
[7] delves a multiobjective employee scheduling predicament by a company, this challenge entails
devising employee schedules that align with customer demands while conforming to both company and
employee-imposed constraints. At the side of, the research approach comprises two main components.
Firstly, a mixed-integer linear programming model addresses the multi-objective employee scheduling
problem and considers constraints from customers, employees, and the company that these constraints
are handled simultaneously, encompassing objectives like customer satisfaction, and the reduction of
employee wait and idle times. Secondly, the Prioritize and Improve methodological approach, which
entails sequentially addressing multiple single-objective optimization problems through a hybrid
approach that combines exact and heuristic techniques that these findings used the effectiveness of this
method in generating high-quality schedules while adhering to constraints. Furthermore, the approach is
adaptable and individual optimization methods, suggesting its potential for addressing various multi-
objective optimization challenges in different industrial areas. The authors also call for future research to
enhance the handling of partial prioritization and optional constraints within the even method, thereby
further improving schedule quality.
[8] consider scheduling the nursing team development and nurse planning that nurses must be affected to
specific groups and shifts that respect work regulations and take into consideration nurses' preferences.
This issue is addressed by the signed social structure optimization which explicitly considers the signed
social relationships between nurses, which are known to have a major impact on their performance and
job satisfaction. Overlooking, the mathematical optimization model for shift assignment based on this
social structure. Then, the results show that the framework is effective at reducing unfavorable social
relationships. However, more study is needed to evaluate how well it performs in scenarios involving
larger-scale nurse scheduling problems and to improve the algorithms' effectiveness and scalability. Even
so, it marks a substantial advancement in resolving the difficulties associated with assigning nurse shifts
by recognizing the critical importance of social relationships to nursing performance and well-being.
[9] addresses the problem of nurse scheduling in a hospital, it describes an approach for identifying
scheduling characteristics and restrictions that involves conducting interviews with the head nurse and
regular nurses. A baseline schedule with backup nurses is then developed using a mixed-integer model
to account for absences then, meeting hard requirements is given top priority in this model, and soft
constraint violations are minimized. The authors used a Genetic Algorithm for bigger cases and so there
are two suggested procedures for handling unexpected absences.
What's more, the main contribution is its comprehensive analysis of previously unstudied topics such as
nurses' preferences, task balance, compensation, and absenteeism in scheduling, it also emphasizes how
crucial it is for nurses to combine proactive and defensive methods.
[10] provides a fresh mathematical model to resolve the Nurse Scheduling Problem. Yet, the literature
review examines different exact and approximative approaches to this problem. Along, the suggested
approach includes nurses' preferences while placing a priority on cost-cutting. Results from a genuine
case study show that overtime expenditures can be significantly reduced while also accommodating
Page 664
SEMIT 2023
preferences like taking days off and intermissions between shifts, the advantages of the suggested
methodology for enhancing nurse scheduling in hospitals are emphasized.
2.2. Call centers:
[11]aims to solve the issue of call center personnel scheduling by concentrating on rostering and shift
scheduling the issue entails predicting work loans, figuring out how many workers are required, arranging
shifts, and assigning workers to shifts while taking labor laws into account. Likewise, a two-stage strategy
is suggested in the research, with the artificial bee colony algorithm serving as the primary search
framework. In the first solutions are developed that satisfy hard requirements and in the second stage,
these solutions are improved by satisfying constraints. Finally, the development of multi-objective
algorithms, and the examination of fresh factors for future study.
[12]Merged the scheduling and the staffing personals, this research recommends a two-stage stochastic
integer programming to optimize personnel levels and shift scheduling in light of the various arrival rates.
Similarly, to successfully solve the model, they combine Bender's decomposition strategy with a method
known as mixed-integer rounding inequalities. Their numerical investigation's findings demonstrate that
the integrated model can produce schedules that save costs while maintaining scheduling. For further
research in this field, including the use of more precise taking into consideration factors like client
preferences and server skills.
[13] aimed with a focus on effectively managing break periods, an integrated strategy was designed to
solve shift scheduling and rostering issues in the context of inbound call centers. The challenge lies in
developing ideal work schedules that strike a balance between the call center's operational requirements
and ensuring that agents receive enough downtime to avoid burnout and preserve service quality. Without
mentioning it specifically, the authors applied a novel way to optimize these timetables. The outcomes of
their strategy showed considerable gains in call center productivity, along with decreased personnel
expenses and improved agent well-being. across the future, this integrated strategy shows potential for
addressing comparable labor-management issues across a variety of service-oriented businesses, which
will improve operational effectiveness and employee satisfaction.
2.3. Manufacturing and construction industry:
[14] Aimed for staff scheduling is sometimes complicated by unplanned interruptions like employee
absences and variations in customer demand, as increases a negative impact on productivity and the level
of customer service. Yet, the authors have developed a heuristic method for time staff rescheduling
following minimal disruptions to solve these problems.
Interestingly, their approach is successful when compared to real-world data, the lowered interruption
costs show the ability to quickly provide high-quality solutions in an average of less than one second. In
the future, the suggested method has large potential to improve the effectiveness and robustness of human
scheduling as well as its adaptability.
[15] evaluates potential solutions to the difficult rotating employee scheduling Problem using the ground-
breaking Instance Space Analysis approach. Personnel in test instances are identified, together with their
effect on solution performance, using a visualization-based technique, to build a variety of test instances,
exposing the complexity and gaps in the instance space. They assess three different approaches,
identifying key advantages and disadvantages as well as a change in the phase of case difficulty. The
study emphasizes the importance of the instance space analysis approach in assessing personnel
scheduling options and proposes that it might result in more specialized and effective scheduling
strategies, improving employee management. furthermore, this study makes a substantial addition to the
scheduling of workers and has the potential to revolutionize employee administration in several
industries.
2.4. Transportation:
[16] handles the staff scheduling and planning issues that the civil aviation therefore, the authors present
a customized optimization strategy built on a set covering model to address this issue. this strategy for
the civil aviation authority is utilized to thoroughly assess their methods and create schedules that are
Page 665
SEMIT 2023
both realistic and cost-effective. Then, this would be advantageous for firms like those in the authors'
field that have large auditor pools and a wide range of responsibilities, as well as present prospects for
future research.
[17] takes on the difficult issue of applying linear programming to optimize staff scheduling within an
airplane is a difficult challenge because it necessitates a large workforce, a variety of responsibilities, and
complicated constraints. The authors develop a linear programming model specifically for aviation staff
scheduling to address this issue to lower overall scheduling expenses while ensuring complete job
coverage and adherence to constraints. Their model testing uses actual data from an airline company, and
the results show that it is capable of producing schedules that are both realistic and economical. Moreover,
the authors have considerable potential for building more effective and efficient airline employee
scheduling systems, particularly for airlines with big staffs and a diversity of aircraft operations.
[18] explores the complexities of employee scheduling with inter-departmental transfers, where firms
must effectively manage employee schedules across several days while taking inter-departmental staff
transfers into account. Then, the authors present a heuristic decomposition approach to handle this
challenging issue, and integer linear programming is used in this approach to create anonymous work
schedules, optimize inter-departmental transfers and schedules to the preferences and constraints of
various departments. as well as, the results provided information on the method's potential for increasing
scheduling effectiveness and cutting costs by showing how well it handles complex personnel scheduling
issues in a variety of industries, this study offers contributions to the discipline of employee scheduling
and a potential solution for businesses coping with complex workforce companies.
Telecommunication:
[19]Presents a multifaceted challenge to scheduling for staff rostering, it demands the allocation of staff
members to shifts in a way that properly satisfies the coverage requirements, ensuring that customer
demand, finding this complicated balance while creating personnel rosters is a difficult undertaking that
requires careful preparation and optimization. The literature proposes a hybrid personnel scheduling
model that integrates monthly and daily scheduling processes, it is a mixed integer programming
formulation to solve the problem with monthly and daily scheduling constraints.
Besides, the hybrid scheduling model was assessed using actual data from the company and the results
showed the model was able to produce workable, excellent rosters that complied with all of the
requirements.
[20] aims of scheduling technicians and tasks in a telecommunication company is to assign technicians
then all tasks are assigned to a technician to their qualifications than one didn't task at a time within their
specified time windows and minimized the total travel time and cost of the technicians.
Therefore, the article proposes a new heuristic algorithm for solving the employee and task scheduling
problem that is based on a tabu search framework as tabu search is a metaheuristic algorithm that is often
used to solve combinatorial optimization problems as well as to solve a solution in a fair length of time.
For upcoming research, the algorithm might result in lower expenses, better customer service, and happier
employees.
From reviewing the articles, we found that there are many studies related to personnel scheduling problem
issues in different industrial sectors. Several cases of human resources planning problems with different
characteristics, different limitations, and solution approaches have been presented. but we used the exact
method in this study. This method prioritizes optimality over speed, whereas heuristic methods prioritize
speed. Besides, the exact method produces high-quality results, even though it always guarantees a perfect
solution.
The literature has given rise to many applications in many areas that have been extensively studied, except
for the problem area of personnel planning in engineering control which is discussed in this context. As
far as we are aware, there is a small amount of research focusing on the issue of personnel scheduling in
the technical of control which prompted us to conduct the attention of conducting a thorough investigation
to find out more about this specific field of study and its characteristics
Page 666
SEMIT 2023
This method could be employed, so when an inspection team is required to distribute technical control
missions among inspectors Within a stipulated time duration, these assignments may include tasks that
may involve due dates and prerequisites, and they require trained professionals preferring the most
skilled. We feel that this research contributes to the current literature by providing a solution that handles
the complexity of technical control at a highly practical level. This method could be employed, so when
an inspection team is required to distribute technical control missions among inspectors Within a
stipulated time duration, these assignments may include tasks that may involve due dates and
prerequisites, and they require trained professionals preferring the most skilled. We feel that this research
contributes to the current literature by providing a solution that handles the complexity of technical
control at a highly practical level.
3. Mathematical model
In this section, the suggested mathematical model is introduced, which mainly focuses on the staffing of
technical control firms. The issue frequently surfaces while evaluating or putting into practice personnel
decisions for different units.
3.1. Problem description
For this research project, the primary objective is to optimize the available staff of the organization with
the various tasks at hand, with a concentration on cost-cutting while maintaining employee preferences
and satisfaction. The development of a mathematical model that effectively coordinates the positioning
of inspectors by their numerous inspection tasks. This mathematical model may optimize resource
allocation, increase operational effectiveness, and reduce costs also accounting for the specific
preferences and constraints of the committed personnel.
Doing this, it is required to evaluate the approaches now in use and identify the key objectives as well as
the constraints related to personnel scheduling.
Therefore, field workers' technical control tasks are complex and varied as need a thorough understanding
of numerous significant issues. This vast area includes several non-destructive testing techniques, these
are all designed to thoroughly assess the structural soundness and quality of materials, systems, and
components across many industries. Additionally, it necessitates the strict application of regulatory safety
regulations, which are crucial for safeguarding the integrity of the infrastructure and the workers' safety.
These controls provide a crucial first line of protection against potential dangers, ensuring that every
procedure with the utmost care and precision. Technical control also entails adhering to a variety of
regulations required by standards, legislation, and decrees.
3.2. Goal programming model
Goal programming is a common approach in operations research and has been applied to a wide range of
problems, including personnel scheduling in technical control, it offers adaptability in addressing
personnel scheduling problems by allowing the balance of various objectives according to specific needs.
It provides adaptability in addressing staff scheduling issues by allowing the balance of multiple
objectives based on specific needs. In a personnel scheduling environment, researchers typically use these
methods to develop solutions that match their aims and restrictions.
This method is a multi-criteria decision-making modeling method that enables solving problems when
there are multiple conflicting objectives to be fulfilled at the same time [21]. Then, the personnel
scheduling problem in our case is based on two different objectives the first minimizing the sum of the
weighted deviations from the number of tasks among inspectors, and the second minimizing the overall
bonuses some of these objectives have their own unit, and more specifically, we utilized the weighted
goal programming. This variant of goal programming consists of weighting the different objectives
according to their relative importance. The weightings make it possible to give a higher or lower priority
to each objective and to model priority relationships between them [21]. Also, the stakeholders gave me
the deviations to optimize the problem that this research is to solve the inspection planning problem for
a variable planning horizon. The two objectives are adjusted and weighted into a single objective function
Page 667
SEMIT 2023
to reflect the strategic objectives of the operation. as the weight to reducing costs might be more critical
than minimizing the number of tasks among inspectors, you would assign a higher weight to the more
objective.
3.3. Mathematical formulation
• Sets of the model
Table 24. Sets, subsets, and corresponding domains
Sets Indexes Domain Description
The number of inspectors in the
Inspector (I) i {1,…,I}
company
Day (D) d {1,…,D} The number of days on the schedule
Mission (N) N {1,…,N} The number of missions
Shift (P) P {1,2,3} The 3 shifts to be provided by day
Certificate The number of certificates
C {1,…,C}
(C)
• Mathematical model Parameters

1 if the control mission n requires the certificate c
Q n,c { (𝑛 = 1. . 𝑁, 𝑐 = 1. . 𝐶)
0 if not
1 if the inspector i has the required certificate c

K i,c { (i = 1. . I, c = 1. . C)
0 if not
1 if inspector i is on vacation on day d

RPi,d { (i = 1. . I, 𝑑 = 1. . 𝐷)
0 if not
Pi,n : Bonus recorded for an inspector (i) performing a mission (n), n = 1. . N, i = 1. . I
Pi,e : Bonus recorded of an inspector(i) in the shift (p), i = 1. . I, p = 1. . P
Pm : Inspectors′ average bonus
∑Nn=1 Pi,n
Pm = i = 1. . I
Number of inspectors
Mm ∶ Average number of missions performed by inspectors
∑Ii=1 Pn
Mm = 𝑛 = 1. . 𝑁
Number of mission
REn : estimated radiation dose in a mission (n), n = 1. . N
Page 668
SEMIT 2023
RDi ∶ Dose of radiation already received by the inspector(i), i = 1. . I

R: Maximum allocated radiation dose (20 m sv)
𝑀: A very large variable.
𝛼i = Weight of the objective i
𝑆d,p = starting time of mission d in the shift p(d =1..D , p=1..P)
𝐹d,p = Ending time of mission d in the shift p(d =1..D , p=1..P)
OPTR= Most favorable weekly rest interval for inspector between two consecutive control missions
• Decision variables:
1 if the inspector (i) carries out the mission (n)

X i,n,d,p { on the day (d) at the shift (p) (i = 1. . I, n = 1. . N, d = 1. . D, p = 1. . P)
0 if not
1 if the inspector carries out a mission on day (d) (i

Yi,d { = 1. . I, n = 1. . N, d = 1. . D)
0 if not
DNBi : Negative deviation in relation to the objective function targeting the minimization of the variation
in premiums received by an inspector (i) compared to an average threshold, DNBi ≥ 0 (i = 1. . I)
DPBi : Positive deviation in relation with the objective function targeting the minimization of the variation
in premiums received by an inspector (i) compared to an average threshold DPBi ≥ 0 (i = 1. . I)
DPCi,d,p : Positive deviation in relation with the objective function targeting the minimization of the
variation in the number of rest hours for an inspector (i) between missions performed on day (j) in
destination (d) compared to a most favorable threshold. DPCi,d,p ≥ 0 (i = 1. . I, d = 1. . D, p = 1. . P)
DNCi,d,p : Negative deviation in relation with the objective function targeting the minimization of the
variation in premiums received by an inspector (i) compared to an average threshold, DNBi ≥ 0 (i =
1. . I, d = 1. . D, p = 1. . P)
• Objective function:
min F(x) = ∑Ii=1( α1 (DNBi + DPBi )) + ∑Ii=1 ∑D P
d=1 ∑p=1( α2 (DPCi,d,p + DNCi,d,p )) (1)
• Constraints:
∑N
n=1 X i,n,d,p – Yi,d = 0 avec p = 2, ∀i ∈ {1. . I}, ∀d ∈ {1. . D} (2)
Yi,d ≤ RPi,d 0 ∀i ∈ {1. . I}, ∀d ∈ {1. . D} (3)
Yi,d ≤ RPi,d ∀i ∈ {1. . I}, ∀n ∈ {1. . N}, ∀p ∈ {1. . P} (4)
∑Ii=1 X i,n,d,p = 1 ∀i ∈ {1. . I}, ∀d ∈ {1. . D}, }, ∀n ∈ {1. . N}, ∀p ∈ {1. . P} (5)
∑Ii=1 K i,c × X i,n,d,p ≥ Q n,c ∀c ∈ {1. . C}, ∀d ∈ {1. . D}, }, ∀n ∈ {1. . N}, ∀p ∈ {1. . P} (6)
Page 669
SEMIT 2023
∑Pp=1 X i,n,d,p ≤ 1 ∀i ∈ {1. . I}, ∀d ∈ {1. . D}, }, ∀n ∈ {1. . N} (7)

X ni(p+2)d + X nip(d+1) = 1avec p = 1 ∀i ∈ {1. . I}, ∀d ∈ {1. . D}, }, ∀n ∈ {1. . N} (8)
X nipd + X nip(d+1) + X nip(d+2) = 1 avec p = 3∀i ∈ {1. . I}, ∀d ∈ {1. . D}, }, ∀n ∈ {1. . N} (9)
RDi + ∑D N P
d=1 ∑n=1 ∑p=1( X i,n,d,p × REn ) ≤ R Max ∀i ∈ {1. . I} (10)
∑N D P
n=1 ∑d=1 ∑p=1(X i,n,d,p × Pn) − Pm = 0 ∀i ∈ {1. . I} (11)
∑N D P
n=1 ∑d=1 ∑p=1 X i,n,d,p − Mm = 0 ∀i ∈ {1. . I} (12)
d
∑Dd=1( Sd,p ∗ X i,n,d,p ) − ∑d=1 Fd,p ∗ X i,n,d,p ) + DNCi,d,p − DPCi,d,p = 0 ∀i ∈ {1. . I}, ∀p ∈ {1. . P}, }, ∀n ∈
{1. . N} (13)
∑𝑁 𝑃 𝐷
𝑛=1 ∑𝑝=1 ∑𝑑=1(Pi,p ∗ X i,n,d,p ) − Pm + DNBi − DPBi = 0∀i ∈ {1. . I} (14)
The objective of the model (1) consists in minimizing the sum of the weighted deviations from the number
of tasks among inspectors and overall bonuses.
Constraint (2) ensures that inspectors are available within the planning horizon.
Constraint (3) and (4) guarantees that each inspector is not in his rest period and available for services.
Constraints (5) ensure the execution of all control mission planned.
Constraints (6) verify the qualification of inspectors in the process of scheduling. Constraints (7)
guarantee that all planned control mission are executed.
Constraints (8) and (9) ensure that an inspector cannot be affected for more than one control mission in
one day and that an inspector affected in shift three cannot be affected to the shift one next day.
Constraint (10) guarantees that the maximum authorized radiation dose is respected. Constraints (11)
concerns the premiums and bonuses given to inspectors who completed control missions. It seeks to
reduce the range in these premium amounts relative to an average threshold.
While constraint (12) along aim to minimize the variation in the number of control mission performed by
each inspector according to an average threshold judged as most favorable.
4. Result:
In this section, we first give details about the data used in our experiments. We then compare the solutions
of our integrated model with the solutions obtained by considering staffing and scheduling separately.
The solutions from our integrated model are then compared to the solutions generated by considering
personnel and scheduling individually. Finally, we give performance statistics for the suggested goal
programming.
Besides, we used The program Python to solve the mathematical model then we entered their libraries
numpy and Pulp, successfully employed to bring about these changes on an Intel i7, 2.70 GHz 8 GB
RAM PC.
Following that, we present our numerical experiments, which show the potential benefit of solving the
integrated staffing and scheduling model as well as the computational efficiency of the proposed approach
that the stallholders who gave me the data to put it down in a Python program and automatically the
program give me the results which presented in table 2 and even give me the problem was feasible or
infeasible as in this case the problem was always feasible that’s a result, the mathematical model to
multiobjective was effective, and the challenge was achievable with good answers.
Finally, the model demonstrated its feasibility by providing important solutions and establishing its
effectiveness, which is a positive review of the model. It implies that the model is a useful tool for
improving decision-making and achieving objectives.
Page 670
SEMIT 2023
Table 2: Results of performed tests

Number of test Python
Test Coverage Objective Resolutions
Optimality
N° Days Inspectors Missions rate functions time in (s)
1 2 3 3 100% 15.89 Confirmed 0.03

2 3 3 3 100% 18.60 Confirmed 0.04
3 4 5 5 100% 49.80 Confirmed 0.10
4 6 6 6 100% 66.0 Confirmed 0.60
5 5 7 7 100% 86.60 Confirmed 0.66
6 6 8 8 100% 118.40 Confirmed 0.73
7 7 9 9 100% 144.20 Confirmed 0.80
8 9 11 11 100% 185.48 Confirmed 0.86
9 10 12 12 100% 215.08 Confirmed 0.92
10 13 21 21 100% 356.84 Confirmed 0.98
5. Conclusion and future work

As a result, the final results demonstrated how mathematical modeling and exact procedures may be
coupled to produce good solutions to a people scheduling challenge. This study's technique is unique in
that it improves employee happiness by assuring a balanced workload. Actually, this paper detailed the
formulation and resolution of a multi-skilled staff scheduling problem. The goal was to organize work
shifts and activities for multi-skilled inspectors within the planning horizon while considering all
conceivable restrictions.
In many industries, staff scheduling is one of the biggest problems. Therefore, including employee
preferences in a schedule that takes the company's objectives into account can significantly boost
employee profitability. This research effort's main objective is to develop a framework for a multi-
objective mathematical problem involving the scheduling of inspectors who perform inspections.
Future research will focus on considering that the challenge is thought to be NP-hard, on creating a
metaheuristic to solve the mathematical program created to solve the problems.
References
[1] D. Prot, T. Lapègue, and O. Bellenguez-Morineau, “A two-phase method for the shift design and
personnel task scheduling problem with equity objective,” Int. J. Prod. Res., vol. 53, no. 24, pp.
7286–7298, 2015, doi: 10.1080/00207543.2015.1037023.
[2] L. Kletzander and N. Musliu, “Solving the general employee scheduling problem,” Comput. Oper.
Res., vol. 113, p. 104794, 2020, doi: 10.1016/j.cor.2019.104794.
[3] F. He, T. Chaussalet, and R. Qu, “Controlling understaffing with conditional Value-at-Risk
constraint for an integrated nurse scheduling problem under patient demand uncertainty,” Oper.
Res. Perspect., vol. 6, no. February, p. 100119, 2019, doi: 10.1016/j.orp.2019.100119.
[4] L. Huang, C. Ye, J. Gao, P. C. Shih, F. Mngumi, and X. Mei, “Personnel Scheduling Problem
under Hierarchical Management Based on Intelligent Algorithm,” Complexity, vol. 2021, no. C,
2021, doi: 10.1155/2021/6637207.
[5] A. H. Nobil, S. M. E. Sharifnia, and L. E. Cárdenas-Barrón, “Mixed integer linear programming
problem for personnel multi-day shift scheduling: A case study in an Iran hospital,” Alexandria
Eng. J., vol. 61, no. 1, pp. 419–426, 2022, doi: 10.1016/j.aej.2021.06.030.
[6] M. Hamid, F. Barzinpour, M. Hamid, and S. Mirzamohammadi, “A multi-objective mathematical
model for nurse scheduling problem with hybrid DEA and augmented $ε$-constraint method: A
case study,” J. Ind. Syst. Eng., vol. 11, no. Special issue: 14th International Industrial Engineering
Conference, pp. 98–108, 2018, [Online]. Available:
Page 671
SEMIT 2023
http://www.jise.ir/article_69388_33213b3c7b0ff05ca53540398f38df96.pdf
[7] R. Mansini, M. Zanella, and R. Zanotti, “Optimizing a complex multi-objective personnel
scheduling problem jointly complying with requests from customers and staff,” Omega (United
Kingdom), vol. 114, p. 102722, 2023, doi: 10.1016/j.omega.2022.102722.
[8] A. Farasat and A. G. Nikolaev, “Signed social structure optimization for shift assignment in the
nurse scheduling problem,” Socioecon. Plann. Sci., vol. 56, pp. 3–13, 2016, doi:
10.1016/j.seps.2016.06.003.
[9] R. Otero-Caicedo, C. E. M. Casas, C. B. Jaimes, C. F. G. Garzón, E. A. Y. Vergel, and J. C. Z.
Valdés, “A preventive–reactive approach for nurse scheduling considering absenteeism and
nurses’ preferences,” Oper. Res. Heal. Care, vol. 38, p. 100389, 2023, doi:
10.1016/j.orhc.2023.100389.
[10] A. A. El Adoly, M. Gheith, and M. Nashat Fors, “A new formulation and solution for the nurse
scheduling problem: A case study in Egypt,” Alexandria Eng. J., vol. 57, no. 4, pp. 2289–2298,
2018, doi: 10.1016/j.aej.2017.09.007.
[11] Y. Xu and X. Wang, “A hybrid integer programming and artificial bee colony algorithm for staff
scheduling in call centers,” Comput. Ind. Eng., vol. 171, Sep. 2022, doi:
10.1016/j.cie.2022.108312.
[12] M. Bodur and J. R. Luedtke, “Mixed-integer rounding enhanced benders decomposition for
multiclass service-system staffing and scheduling with arrival rate uncertainty,” Manage. Sci., vol.
63, no. 7, pp. 2073–2091, 2017, doi: 10.1287/mnsc.2016.2455.
[13] T. Türker and A. Demiriz, “An Integrated Approach for Shift Scheduling and Rostering Problems
with Break Times for Inbound Call Centers,” Math. Probl. Eng., vol. 2018, 2018, doi:
10.1155/2018/7870849.
[14] R. Hassani, G. Desaulniers, and I. Elhallaoui, “Real-time personnel re-scheduling after a minor
disruption in the retail industry,” Comput. Oper. Res., vol. 120, p. 104952, 2020, doi:
10.1016/j.cor.2020.104952.
[15] L. Kletzander, N. Musliu, and K. Smith-Miles, “Instance space analysis for a personnel scheduling
problem,” Ann. Math. Artif. Intell., vol. 89, no. 7, pp. 617–637, 2021, doi: 10.1007/s10472-020-
09695-2.
[16] R. W. Gamermann, L. Ferreira, and D. Borenstein, “Long-term audit staff scheduling and
planning: A case study of Brazilian civil aviation authority,” J. Air Transp. Manag., vol. 106, no.
September 2022, p. 102318, 2023, doi: 10.1016/j.jairtraman.2022.102318.
[17] M. Patidar, S. Choudhary, and R. Bhardwaj, “The solution of staff scheduling problem in an
aircraft using linear programming,” Mater. Today Proc., vol. 47, no. xxxx, pp. 6826–6829, 2020,
doi: 10.1016/j.matpr.2021.05.139.
[18] D. Attia, R. Bürgy, G. Desaulniers, and F. Soumis, “A decomposition-based heuristic for large
employee scheduling problems with inter-department transfers,” EURO J. Comput. Optim., vol.
7, no. 4, pp. 325–357, 2019, doi: 10.1007/s13675-019-00119-3.
[19] J. Y. Shiau, M. K. Huang, and C. Y. Huang, “A hybrid personnel scheduling model for staff
rostering problems,” Mathematics, vol. 8, no. 10, pp. 1–20, 2020, doi: 10.3390/math8101702.
[20] J. F. Cordeau, G. Laporte, F. Pasin, and S. Ropke, “Scheduling technicians and tasks in a
telecommunications company,” J. Sched., vol. 13, no. 4, pp. 393–409, 2010, doi: 10.1007/s10951-
010-0188-7.
[21] B. Aouni, “Le modèle de Programmation Mathématique avec Buts dans un Environnement
Imprécis: sa Formulation, sa Résolution et une Application,” p. 157, 1998.
Page 672
SEMIT 2023
Impact of Gender Inequalities on Education in Morocco: An Economic Analysis of Barriers and

Opportunities
Mohammed Alami Chentoufi 1*, Ait Alla Aniss, Naoual Mamdouh

MOHAMMED.ALAMICHENTOUFI@uhp.ac.ma
Laboratory of Mathematical Modeling and Economic Computation
Associated to the Laboratory of Theoretical and Applied Economics Research
Faculty of Economics and Management
aniss.aitalla@uhp.ac.ma
Laboratory of Research in Management and Development
naoualmamdouh@gmail.com
Laboratory of Mathematical Modeling and Economic Computation
Abstract:
Gender inequalities in the field of education pose a major issue that continues to challenge development
efforts in Morocco. Despite significant progress made in the education sector over the past decades,
persistent gender disparities hinder equitable access to quality education. This study aims to examine the
economic impact of these gender inequalities on the Moroccan education system, focusing on the barriers
that impede girls' access to education and the potential opportunities associated with promoting gender
equality.
Drawing on an exhaustive review of recent economic literature, this research explores the socio-economic
determinants of gender inequalities in education in Morocco. It also examines the economic consequences
of these disparities, both in the short and long term, with an emphasis on their impact on productivity,
economic growth, and poverty reduction.
The study employs a rigorous quantitative methodology, analyzing empirical data from various national
and international sources. An advanced econometric model is applied to assess the extent of the impact
of gender inequalities on educational performance and the country's economic outcomes.
Through this analysis, we hope to provide relevant insights for policymakers, researchers, and
development stakeholders concerned with promoting gender equality and improving the education system
in Morocco. The findings will contribute to guiding public policies aimed at strengthening equitable
access to education and creating an enabling environment for the development of all individuals,
regardless of their gender.
Keywords: Gender inequalities, education, economic development, Morocco.
I- Introduction
Over the past decades, Morocco has made significant strides in the education sector, seeking to promote
an inclusive and quality educational system for all its citizens. Despite these advancements, gender
inequalities persist, posing a major challenge to equitable access to education and the country's socio-
economic development. This complex and multidimensional issue has been garnering increasing interest
among economists who seek to understand the economic implications of these disparities and identify
strategies to overcome them.
Page 673
SEMIT 2023
Education is recognized as a fundamental pillar of sustainable development and poverty reduction. Recent
empirical studies have shown that equitable access to quality education, regardless of gender, contributes
to improving a country's economic and social prospects (UNESCO, 2020; World Bank, 2019). However,
despite this evidence, Morocco continues to face persistent challenges concerning gender equality in the
field of education.
In addition to the persistent challenges regarding gender equality in education, there are several key
dimensions and factors that deserve attention. One of these dimensions is the urban-rural divide. While
urban areas in Morocco have seen notable progress in terms of gender parity in education, rural areas lag
behind due to a lack of infrastructure, teacher shortages, and cultural norms that may prioritize boys'
education over girls. Furthermore, the role of teacher training and curriculum development cannot be
underestimated. To address gender inequalities effectively, Morocco must invest in teacher training
programs that sensitize educators to the importance of gender equality in the classroom. This includes
addressing unconscious biases and stereotypes that may affect the way teachers interact with students.
Additionally, revising and modernizing the curriculum to include more diverse perspectives and
narratives can help challenge traditional gender roles and stereotypes. This can empower both boys and
girls to pursue their interests and talents without feeling restricted by societal expectations. Another
critical aspect to consider is the socio-economic status of families. Poverty can disproportionately affect
girls' access to education, as families may prioritize allocating limited resources to boys' education.
Implementing targeted financial assistance programs, scholarships, and incentives can mitigate this
barrier and ensure that all children, regardless of their economic background, have the opportunity to
access quality education. Furthermore, Morocco can benefit from international partnerships and
collaborations to learn from successful strategies implemented in other countries with similar challenges.
Sharing best practices, resources, and experiences can expedite progress in achieving gender equality in
education.
While Morocco has made commendable strides in improving its education system, gender inequalities
persist and hinder the country's socio-economic development. Addressing these challenges requires a
multifaceted approach that encompasses rural-urban disparities, teacher training, curriculum
development, poverty alleviation measures, and international cooperation. By tackling these issues
comprehensively, Morocco can make significant progress towards achieving gender equality in education
and ultimately bolster its economic and social prospects.
1. Context of Gender Inequalities in Education in Morocco
In Morocco, gender inequalities in access to education have historically been rooted in social and cultural
norms (Benradi, 2006; Berker, 2009). Traditional gender roles often assign girls domestic responsibilities,
limiting their access to formal education (Conseil National des Droits de l'Homme, 2015; Rivers et al.,
1988). Gender biases reinforce this situation, leading to underrepresentation of girls in educational
institutions and higher dropout rates compared to boys (Beyene, 2014; Bourqia, 2018).
Geographical disparities also play a significant role in access to education. Rural regions, characterized
by less developed educational infrastructure, face higher dropout rates among girls (Edmonds, 2006; Le,
2011). Economic barriers are another crucial dimension impacting girls' education in Morocco. Low-
income families often struggle financially to support their children's education, leading to prioritization
of boys' education at the expense of girls (Conseil Supérieur de la Formation et de la Recherche
Scientifique, 2015; Hu, 2012).
2. Economic Consequences of Gender Inequalities in Education
The economic consequences of gender inequalities in education are significant for Morocco (Adams et
al., 2010; Amuedo-Dorantes et al., 2010). Firstly, by limiting girls' access to education, the country
underutilizes its human capital, thereby reducing its potential economic contributions through formal
employment and thriving entrepreneurship (Ahmed et al., 2011; Amuedo-Dorantes et al., 2006).
Secondly, these gender inequalities contribute to perpetuating the vicious cycle of poverty (Amuedo-
Dorantes et al., 2010; Belfield, 2000). Girls' education is closely linked to health, nutrition, and family
well-being. When girls are deprived of education, they have fewer opportunities to access economic
Page 674
SEMIT 2023
opportunities and are more likely to remain trapped in a cycle of poverty, perpetuating this situation for
future generations (Benker, 2009; Benradi, 2006).
Gender disparities in education can also lead to broader social costs (Calero et al., 2009; MSFEDS, 2018).
An undereducated and underrepresented workforce results in lower productivity and limits Morocco's
ability to innovate and adapt to global economic challenges. These factors directly impact the country's
economic competitiveness and its capacity to achieve sustainable development (CSEFRS, 2018; Gillot et
al., 2018).
It is essential to take measures to eliminate gender inequalities in education in Morocco. Inclusive and
equitable education policies, as well as programs aimed at raising awareness among families about the
importance of girls' education, can help reverse this trend and foster more balanced and prosperous socio-
economic development for the country (Forum des femmes marocaines, Jossour, 2017; Haut-
Commissariat au Plan, 2017).
3. Research Objectives
Given the significance of gender inequalities in education in Morocco and their substantial economic
impact, this research aims to deepen the understanding of this complex issue. The main objective is to
thoroughly analyze the socio-economic barriers that hinder equitable access to education for girls and
explore the economic opportunities that could arise from promoting gender equality within the Moroccan
education system.
Using a rigorous economic approach and advanced methodologies, we will seek to answer the following
key questions:
What is the economic impact of gender inequalities on education in Morocco?
By providing in-depth insights into these aspects, this research aims to inform public policies and
development initiatives aimed at reducing gender inequalities in education in Morocco and promoting an
equitable and inclusive educational environment for all, regardless of gender.
In conclusion, gender inequalities in education represent a persistent challenge to Morocco's socio-
economic development. This research is part of an effort to better understand the economic factors
underlying these inequalities and identify potential economic opportunities arising from eliminating
gender disparities in the educational system.
II- Literature Review

The literature review examined various research works addressing the impact of gender inequalities on
education in Morocco. Bacchetta and Power (2016) highlighted the influence of social and cultural norms
that limit girls' access to education. Similarly, Afkhami (2017) emphasized the importance of empowering
women through education in Arab societies, including Morocco. The UNICEF report (2018) identified
specific challenges faced by girls in education, especially in rural areas and marginalized communities.
Economic studies such as those by Duflo (2018) and Grown et al. (2020) highlighted the positive
economic implications of women's education for the country's development. Additionally, research
specifically focused on Morocco has been conducted, including the World Bank's study (2019), which
assessed gender equality legislation and its implications for women's access to education. Diop and El
Jouhri (2019) provided an economic analysis of the impact of gender inequalities in education in
Morocco, while Ait Ali and Joumni (2018) used econometric methods to study the determinants of gender
inequalities in education.
Education is a determining factor in a country's socio-economic development, playing an essential role
in improving human capital, productivity, and economic growth (UNESCO, 2020). However, gender
inequalities persist in access to education in many countries, including Morocco, and continue to hinder
society's development potential (World Bank, 2019).
Page 675
SEMIT 2023
1. Social and Cultural Norms Disfavoring Girls' Education

Several previous studies have highlighted the influence of social and cultural norms on gender inequalities
in education in Morocco (Bacchetta & Power, 2016; Afkhami, 2017). Traditional gender roles often
assign girls domestic responsibilities, limiting their access to formal education. As a result, girls'
enrollment rates remain lower than those of boys, especially in rural areas (UNICEF, 2018). Our first
hypothesis is as follows:
H1: Social and cultural norms disfavoring girls have a negative impact on their access to education,
leading to lower enrollment rates compared to boys.
2. Economic Barriers and Girls' Education
Economic constraints pose a significant barrier to girls' education in Morocco, resulting in higher dropout
rates compared to boys (Bacchetta & Power, 2016). This multifaceted issue demands a comprehensive
approach, involving financial incentives, community engagement, and a deep understanding of how
economic factors intersect with cultural norms and societal pressures. Direct costs of education, including
expenses for uniforms, books, transportation, and school supplies, often strain low-income families,
disproportionately affecting girls. Additionally, economic hardships can force girls into child labor,
leading to irregular attendance and eventual dropout. Deep-seated cultural norms that prioritize boys'
(Bacchetta & Power, 2016) education can further exacerbate these challenges. Economic pressures may
also lead to early marriages, interrupting girls' education. Implementing conditional cash transfer
programs, engaging local communities and NGOs, adopting gender-sensitive budgeting, and enhancing
data collection are crucial strategies to alleviate economic barriers, ultimately advancing gender equality
in education and improving overall well-being in Morocco.
H2: Economic barriers are a significant obstacle to girls' education in Morocco, leading to higher dropout
rates compared to boys.
3. Economic Implications of Gender Inequalities in Education
Economic studies have shown that gender inequalities in education have significant economic
consequences for Morocco (Duflo, 2018; Grown et al., 2020). By limiting girls' access to education, the
country underutilizes its human capital, which can hinder economic growth and reduce overall
productivity (World Bank, 2020).
H3: Gender inequalities in education have a negative economic impact on Morocco, resulting in
underutilized economic growth potential and negative social repercussions.
The literature review has highlighted the persistent gender inequalities in education in Morocco and their
potentially negative impact on the country's socio-economic development. Through our hypotheses, we
have formulated specific expectations regarding the factors contributing to these inequalities and their
economic consequences. These hypotheses will serve as the basis for our economic study, which aims to
deepen our understanding of the issue and inform public policies aimed at promoting equitable and
inclusive education for all, regardless of gender.
III- Methodology and Data

1. Sample and Data
The data used for our study were rigorously collected in accordance with scientific principles and
standards. All data sources were secondary data, meaning they were extracted from various existing
sources, including government reports and official publications. These data were made publicly
accessible without any restrictions, ensuring their transparency and availability for in-depth scientific
analysis.
We relied on the websites of government agencies such as the Ministry of National Education and Sports
(MENS), the High Commission for Planning (HCP), and the Ministry of Finance (MF) as reliable and
official sources, thus ensuring the credibility of the collected data. The different summary reports
consulted are published documents covering a wide period ranging from 2012 to 2020.
Page 676
SEMIT 2023
Our methodological approach is based on a rigorous selection of relevant data to address our research
problem. We carefully examined the available sources and chose the most appropriate data based on their
relevance, reliability, and comprehensiveness. This allows us to have a representative and diversified
sample, capturing the various facets of our research problem (see Table 1).
Table 1: "Evolution of key gender indicators in the education sector from 2009 to 2017"
Gender Primary School Life Primary School Delay
Parity Index Expectancy Percentage
Year Boy Girl Boy Girl Boy Girl
2009 0,96 0,86 5,6 5,4 27,70% 26,80%
2010 0,99 0,91 5,8 5,6 26,20% 24,90%
2011 1,02 0,98 6 5,8 24,70% 23,20%
2012 1,02 1,01 6 5,9 25,60% 24,60%
2013 1,04 1,05 6,1 6 23,40% 21,50%
2014 1,04 1,05 6,1 6 23,40% 21,20%
2015 1,11 1,06 5,8 5,7 35,80% 36,40%
2016 1,09 1,05 5,9 5,8 32,30% 31,90%
2017 1,09 1,05 6 5,9 32,30% 31,80%
Sources: Various official reports, MENS, HCP, and MF, between 2012 and 2020.
We used the R language to perform statistical analyses (R Core Team 2021). First, we conducted data
exploration using descriptive statistics to describe the overall trends of different indicators during the
study period. The research also uses econometric analysis, which plays an essential role in studying the
impact of gender inequalities on education in Morocco.
1. Descriptive Statistics
Descriptive statistics are used to present and summarize data on the Gender Parity Index, Primary School
Life Expectancy, and Primary School Delay Percentage for the years 2009 to 2017. These analyses
highlight the persistent gender disparities in access to education, providing valuable insights into temporal
trends and differences between boys and girls.
2. Econometric Analysis
The econometric analysis, on the other hand, delves deeper into the study by examining the relationships
between gender inequalities and education. Using logistic regression, a statistical model suitable for
binary dependent variables, we model the probability of being delayed in primary school based on the
student's gender, Gender Parity Index, and year. This approach allows us to control for the effects of other
variables and specifically quantify the impact of gender on primary school delay.
Our objective was to examine the association between the student's gender (binary variable Girl/Boy) and
the probability of being delayed in primary school, taking into account other explanatory variables such
as the Gender Parity Index and year. We specified the logistic regression model with the following
equation:
Probability (Primary School Delay) = 1 / (1 + exp (-(β0 + β1 * Girl/Boy + β2 * Gender Eq(1)
Parity Index + β3 * Year))).
The coefficients β0, β1, β2, and β3 represent the respective effects of each variable on the probability of
primary school delay. This approach allowed us to quantify the impact of gender on primary school delay
while controlling for the effects of other variables.
Page 677
SEMIT 2023
IV- Results
1. Descriptive Analysis Results
The following table presents indicators related to gender inequalities in the Moroccan education system.
These indicators help to understand how boys and girls may be differently affected by education
disparities. (See Table 2)
Table 2: Descriptive Statistics
Standard
Indicator Mean deviation Median Minimum Maximum Skewness Kurtosis
Parity index (Boys) 1,03 0,045 1,02 0,96 1,11 0,387 -0,446
Parity index (Girls) 1,03 0,045 1,02 0,86 1,06 0,387 -0,446
School life expectancy
5,944 0,206 6 5,6 6,1 -0,091 -0,284
(Boys) Primary
School life expectancy
5,744 0,152 5,9 5,4 6,1 -0,276 -0,593
(Female) Primary
Primary school delay (%)
28,2 4,468 27,7 23,4 35,8 0,914 0,357
(Boy)
Primary school delay (%)
27,47 4,548 26,8 21,2 36,4 1,072 1,218
(Female)
The data presented in this table is about the Gender Parity Index, which measures gender equality in
access to education. The Gender Parity Index values approach 1, suggesting relative equity between boys
and girls over the years. The skewness coefficient is positive (0.387), indicating a slight right-skewness,
with some years having slightly higher Gender Parity Index values than the average. The kurtosis is
negative (-0.446), showing that the distribution of Gender Parity Index values is less leptokurtic than a
normal distribution.
Regarding the Primary School Life Expectancy, it measures the average number of years of education a
student can expect to achieve at the primary level. Boys have slightly higher Primary School Life
Expectancy than girls over the years. The distributions for both groups are approximately symmetrical,
with a skewness coefficient close to zero and less leptokurtic than a normal distribution.
As for the Primary School Delay Percentage, which measures the percentage of students who are older
than expected for their grade level, girls show a slightly lower rate compared to boys over the years, but
both groups have relatively high rates. The skewness coefficient is positive for both groups, indicating a
right-skewness with higher delay rates than the average in certain years. Additionally, the distributions
are more platykurtic than a normal distribution, showing a greater dispersion of values.
In conclusion, the data suggests relative gender equality in access to education, with overall similar
Primary School Life Expectancy but a slight advantage for boys. Primary School Delay remains a
challenge for both groups, although showing a slight downward trend over the years.
2. Logistic Regression Results
To perform the logistic regression, we will use the provided data for the dependent variable (Educational
Indicator - Primary School Delay %) as well as the independent variable (Girl/Boy - binary variable
representing gender). The logistic regression will help us determine if gender (Girl/Boy) has a significant
impact on Primary School Delay % in Morocco. Before moving to the logistic regression, categorization
of the Primary School Delay variable into two groups is necessary (see Table 3).
Page 678
SEMIT 2023
Table 3: Categorization of Primary School Delay into Two Groups

Year Boy Gorl Primary School Delay % Delay Category
2009 0.96 0.86 27.70% 1
2010 0.99 0.91 24.90% 0
2011 1.02 0.98 23.20% 0
2012 1.02 1.01 24.60% 0
2013 1.04 1.05 21.50% 0
2014 1.04 1.05 21.20% 0
2015 1.11 1.06 36.40% 1
2016 1.09 1.05 31.90% 1
2017 1.09 1.05 31.80% 1
Sources: Developed by the authors
In this table, the data for "Primary School Delay %" has been categorized into two groups. If "Primary
School Delay %" is greater than 45%, the value is coded as 1 to indicate "presence of delay." If "Primary
School Delay %" is less than or equal to 45%, the value is coded as 0 to indicate "absence of delay" in
primary school.
Now, we will proceed with the analysis of the logistic regression with the variables "Primary School Life
Expectancy," "Boy" (gender), and the new binary variable "Primary School Delay (0/1)." We will exclude
the years as they are not relevant explanatory variables for this analysis. The results of the multiple logistic
regression are presented in Table 4.
Table 4: Multiple Logistic Regression for Primary School Life Expectancy, Gender, and Primary
School Delay in Morocco
Variable Coefficient Erreur Odds P-value 95% confidence interval
(β) standard ratio (OR) (Upper CI, Lower CI)
Constant 0.020 1.021 - 0.981 (-1.992, 2.032)
Boys -1.065 0.714 0.345 0.140 (-2.478, 0.348)
School life 0.239 0.505 1.270 0.636 (-0.750, 1.229)
expectancy
Primary 1.712 0.963 5.534 0.081 (-0.211, 3.635)
Primary school delay
(0/1)
Sources: Developed by the authors
The multiple logistic regression model is significant (χ² = 7.211, p = 0.066), indicating that the regression
explains some of the variation in Primary School Life Expectancy, gender (Boy/Girl), and Primary School
Delay (0/1) in Morocco. The Hosmer-Lemeshow test suggests a good fit of the model to the data (χ² =
6.124, p = 0.636).
In this multiple logistic regression, we included the binary variable "Boy" to represent gender, as well as
the variable "Primary School Life Expectancy" and the binary variable "Primary School Delay (0/1)"
based on the 45% threshold.
The coefficient for "Boy" is -1.065, showing that boys have a lower probability of experiencing a delay
in primary school compared to girls, but this difference is not statistically significant (p = 0.140).
The coefficient for "Primary School Life Expectancy" is positive (0.239), but the p-value of 0.636
indicates that this coefficient is not statistically significant. Thus, Primary School Life Expectancy is not
a significant predictor of Primary School Delay (0/1) in Morocco.
The coefficient for "Primary School Delay (0/1)" is 1.712. This means that students with a delay in
primary school have a higher probability of being of gender "Boy" compared to girls, but this association
is close to significance (p = 0.081).
Page 679
SEMIT 2023
In summary, the results of the two logistic regression models indicate that gender and Primary School
Life Expectancy do not seem to significantly influence Primary School Delay in Morocco, at least with
the provided data. However, this analysis may be limited by the relatively small sample size and the lack
of other relevant variables that could influence these educational indicators in Morocco. Therefore,
further studies with more variables and data would be necessary to better understand the determinants of
Primary School Life Expectancy and Primary School Delay in Morocco.
V- Discussion
The results of the multiple logistic regression have shed light on the impact of gender inequalities on
education in Morocco, focusing on Primary School Life Expectancy and Primary School Delay (%). Our
hypotheses (H1 and H2) proved relevant in understanding the dynamics between gender and educational
indicators.
Regarding the first hypothesis (H1) on unfavorable social and cultural norms towards girls, our results
did not show a significant relationship between gender (Girl/Boy) and Primary School Delay (%).
Although boys have a slightly lower probability of experiencing a delay in school compared to girls, this
difference did not reach statistical significance (p = 0.140). These findings suggest that social and cultural
norms may not be major determining factors in the rates of primary school delay in Morocco, at least
with the available data.
As for the second hypothesis (H2) on economic barriers, our analysis revealed that the coefficient for the
binary variable "Primary School Delay (0/1)" was positive and significant at a threshold close to
significance (p = 0.081). This indicates that students with a delay in primary school have a higher
probability of being of gender "Boy" compared to girls, suggesting that economic barriers may play a
role in the rates of primary school delay in Morocco.
However, it is important to note that our analysis may be limited by the relatively small sample size and
the lack of other relevant variables that could influence educational indicators in Morocco. Other factors
such as the socioeconomic status of families, access to education, the quality of teaching, educational
policies, and other gender indicators could play an important role in understanding gender inequalities in
education (Diop & El Jouhri, 2019; Grown et al., 2020).
Thus, for a better understanding of the economic impact of gender inequalities on education in Morocco,
future studies should consider including more relevant variables and increasing the sample size.
Additionally, a longitudinal approach could provide more robust insights into the relationship between
gender and educational indicators over time.
By implementing targeted policies and interventions, Morocco can work towards a more equitable and
inclusive education for all, eliminating gender-related barriers. Measures aimed at raising awareness
about the importance of girls' education and combating gender stereotypes can be put in place.
Additionally, ensuring equal access to education for all children, especially in rural and disadvantaged
areas, would require adequate allocation of resources to improve educational infrastructure and ensure
quality teaching (UNICEF, 2018; Bacchetta & Power, 2016).
In conclusion, this study contributes to shedding light on the complex relationship between gender and
education in Morocco, highlighting the importance of considering social, cultural, and economic aspects.
By combining the efforts of government stakeholders, civil society, and the scientific community,
significant progress can be made in promoting equitable and inclusive education, thereby fostering the
socio-economic development of the country as a whole.
VI- Conclusion
This study on the impact of gender inequalities on education in Morocco has deepened our understanding
of the relationships between gender, Primary School Life Expectancy, and Primary School Delay (%).
The results of the two multiple logistic regression models showed that, with the provided data, gender
Page 680
SEMIT 2023
inequalities do not appear to be significant determining factors in the variations of these educational
indicators in Morocco. However, it is important to note that this analysis has some limitations, including
the relatively small sample size and the absence of other relevant variables that could influence these
indicators.
To address gender inequalities in education in Morocco, targeted solutions and interventions are needed.
As recommendations, it is crucial to raise awareness about the importance of girls' education and combat
gender stereotypes. Education programs aimed at promoting gender equality in schools and communities
can be implemented. Access to education must be equitable for all children, regardless of their gender,
with particular attention to rural and disadvantaged areas. Adequate allocation of resources to improve
educational infrastructure and provide quality teaching is essential. Strengthening teacher training to raise
awareness of gender issues in the classroom and promote discrimination-free education is also crucial.
The implementation of policies of inclusion and anti-discrimination in the educational system, ensuring
that girls and boys have equal chances of success, is an important step.
For future in-depth analyses, it is recommended to include more relevant variables. Factors such as the
socioeconomic status of families, access to education, the quality of teaching, educational policies, and
other gender indicators could be explored. Furthermore, a longitudinal study with a larger sample size
could provide more robust insights into these educational indicators and their relationship with gender
inequalities in Morocco. By implementing these recommendations and conducting further in-depth
analyses, Morocco can progress towards a more equitable and inclusive education, eliminating gender-
related barriers and fostering the economic and social development of the country as a whole.
References
Adams, R. H., Cuecuecha, A., & Page, J. (2010). Remittances, household expenditure and investment in
Guatemala. World Development, 38(11), 1626-1641.
Ahmed, J., Mottaleb, K. A., & Khan, H. R. (2011). An empirical analysis of remittances–growth nexus
in Pakistan using bounds testing approach. Journal of Economics and International Finance, 3(8), 494-
505.
Amuedo-Dorantes, C., & Pozo, S. (2006). Migration, remittances, and male and female employment
patterns. American Economic Review, 96(2), 222-226.
Amuedo-Dorantes, C., Georges, A. E., & Pozo, S. (2010). Accounting for remittance and migration
effects on children's schooling. World Development, 38(12), 1747-1759.
Antman, F. M. (2012). Gender, educational attainment, and the impact of parental migration on children
left behind. Journal of Population Economics, 25(4), 1187-1214.
Belfield, C. R. (2000). Economic Principles for Education, Theory of Evidence. Edward Elgar Publishing.
Benadada, A., & El Bouhsini, L. (2015). Le Mouvement des droits humains des femmes au Maroc: étude
réalisée par le centre d’histoire du temps présent, janvier 2013- novembre 2014. Revue de l’administration
de l’éducation, n°10, 2021.
Benker, A. (2009). The impact of internal migration on educational outcomes: evidence from Turkey.
Economics of Education Review, 28(6), 759-765.
Benradi, M. (2006). Prospective Maroc 2030: dynamique social et évolution des statuts des femmes au
Maroc, Rabat, Maroc, Haut-commissariat au plan.
Benradi, M., (Dir.) (2009). Le Féminisme face aux défis du multiculturalisme, actes des travaux du 5ème
congrès des recherches féministes dans la francophonie plurielle. Rabat: Faculté des Sciences Juridiques
Economiques et Sociales-Agdal, Université Mohamed V-Rabat.
Berker, A. (2009). The impact of internal migration on educational outcomes: evidence from Turkey.
Economics of Education Review, 28(6), 759-765.
Page 681
SEMIT 2023
Beyene, B. M. (2014). The effects of international remittances on poverty and inequality in Ethiopia.
Journal of Development Studies, 50(12), 1576-1592.
Bourqia, R. (2018). Penser l'école, penser la société. Casablanca: Editions la Croisée des Chemins.
Bredl, S. (2011). Migration, remittances and educational outcomes: the case of Haiti. International
Journal of Educational Development, 31(3), 267-277.
Calero, C., Bedi, A. S., & Sparrow, R. A. (2009). Remittances, liquidity constraints and human capital
investments in Ecuador. World Development, 37(6), 1143-1154.
Commission Spéciale de l'Education et de la Formation (1999). La Charte nationale de l'éducation et de
la formation. Rabat.
Conseil National des Droits de l’Homme (2015). État de l'égalité et de la parité au Maroc: préserver et
rendre effectifs les objectifs constitutionnels, Rabat.
Conseil Supérieur de la Formation et de la Recherche Scientifique (2015). Vision stratégique de la
réforme 2015-2030: pour une école de l'équité, de la qualité et de la promotion, Rabat, auteur.
CSEFRS (2018). Enseignement supérieur au Maroc: efficacité, efficience et défis du système
universitaire à accès ouvert. Rabat.
Edmonds, E. V. (2006). Child labor and schooling responses to anticipated income in South Africa.
Journal of Development Economics, 81(2), 386-414.
Forum des femmes marocaines, Jossour (2017). Évaluation des mécanismes de promotion de la
représentation politique des femmes au Maroc, Rabat: Fondation Friedrich Ebert, pp. 41-53.
Gillot, G., & Nadifi, R. (Dir.), (2018). La marche vers l'émancipation: travail et éducation des femmes au
Maroc, Casablanca: Afrique Orient.
Haut-Commissariat au Plan (2017). Femmes marocaines en chiffres: tendances d'évolution des
caractéristiques démographiques et socioprofessionnelles. Rabat.
Hu, F. (2012). Migration, remittances, and children's high school attendance: the case of rural China.
International Journal of Educational Development, 32(3), 401-411.
Le, T. (2011). Remittances for economic development: the investment perspective. Economic Modelling,
28(1-2), 477-486.
Ministère de la Solidarité, de la Famille, de l’Egalité et du Développement Social (2015). Le bilan de la
mise en œuvre de la déclaration et de la plateforme d'action de Pékin 1995 après 20 ans, Rabat.
MSFEDS (2018). Plan Gouvernemental pour l'Egalité ICRAM2, 2017-2021, Rabat: AZ Editions.
Rivers, D., Vuong, Q. H., & Vuong, T. T. (1988). Limited information estimators and exogeneity
simultaneous probit models. Journal of Econometrics, 39(3), 347-366.
Page 682
SEMIT 2023
SP18: Advances in Engineering: Machine Learning Techniques
Classification-based Shilling Attack Detection using ODBOT

Halil İbrahim Ayaz1,3[0000-0001-5547-6485] and Zehra Kamişli Öztürk2,3[0000-0003-3156-6464]
1
Necmettin Erbakan University, Department of Industrial Engineering, Konya, Turkey
hiayaz@erbakan.edu.tr
2
Eskisehir Technical University, Department of Industrial Engineering, Eskisehir, Turkey
zkamisli@eskisehir.edu.tr
3
Computational Intelligence and Optimization Laboratory (CIOL), Eskisehir Technical University,
Eskisehir, Turkey
Abstract.
Recommender systems (RS) are vital for predicting future user behavior based on historical actions,
effectively managing information overload. Collaborative Filtering (CF) excels at providing rapid
recommendations for extensive databases. However, RS models are vulnerable to shilling attacks, where
malicious actors introduce fake profiles to manipulate item ratings. This study proactively identifies
future shilling attacks through classification, using Support Vector Machines (SVM) to strengthen RS
models. Despite a smaller dataset with attack instances, an Outlier Detection-Based Oversampling
Technique (ODBOT) addresses class imbalance. Comparative analyses between models with and without
ODBOT reveal that the SVM model with ODBOT outperforms, demonstrating its efficacy in countering
the impact of shilling attacks on recommender systems.
Keywords: Recommender Systems, Shilling Attack, Support Vector Machines, Oversampling.
1 Introduction
Recommendation systems, employing past user behavior for real-time predictions, provide highly
accurate suggestions using various filtering techniques. Collaborative filtering, a prevalent method,
assesses user-product scores to offer recommendations. Widely embraced across diverse domains,
recommendation systems have been the subject of numerous attack detection methods in the literature,
with clustering-based approaches assuming that attack users exhibit distinct scoring distributions as
outliers [1]. However, classification algorithms [2] have proven effective in identifying attack users,
supported by studies involving deep learning[3], machine learning [4], and statistical-based approaches
[5].
This study aims to enhance the proactive detection of future attacks by leveraging insights from prior
research. Building on Bhaumik et al.'s attack profile Error! Reference source not found.[6], this study i
ntegrates knowledge from previous studies [7, 8] into a Support Vector Machines (SVM) model .
Recognizing the challenge of class imbalance in SVM accuracy, particularly with the scarcity of shilling
attacks compared to genuine users, the study employs ODBOT [9], an oversampling algorithm, to
improve the SVM model's precision.
𝐼𝑆 𝐼𝐹 𝐼𝑂 𝐼𝑇
Selected Items Filler Items Unrated Items Target Item
Fig. 19. An attack example
Page 683
SEMIT 2023
2 Computational Evaluation
A comparison for mentioned attacks is given in Table 25 and Table 26 for push and pull attacks,
respectively. In Table 25, mean of performance results are above 99% for all attack types. For push attack
types, the best results are obtained in average attack.
Table 25. Comparison mean of results for attack types on push attacks
Precision Recall F1-Measure Accuracy
With Without With Without With Without With Without
Attack Type
ODBOT ODBOT ODBOT ODBOT ODBOT ODBOT ODBOT ODBOT
Random 0,9954 0,9982 0,9882 0,7321 0,9917 0,7805 0,9976 0,9666
Average 0,9947 0,9976 0,9916 0,7746 0,9931 0,8089 0,9980 0,9736
MAverage 0,9820 0,9955 0,9435 0,5939 0,9614 0,6797 0,9883 0,9371
All Attack 0,9944 0,9985 0,9888 0,7670 0,9915 0,8029 0,9974 0,9723
Similar results are obtained in Table 26 compared to Table 25. It can be said that the proposed method
is validated for investigated attack models for different attack and filler sizes. The relatively best results
are obtained in the modified average attack model for pull attacks.
Table 26. Comparison mean of results for attack types on pull attacks
Precision Recall F1-Measure Accuracy
With Without With Without With Without With Without
Attack Type
ODBOT ODBOT ODBOT ODBOT ODBOT ODBOT ODBOT ODBOT
Random 0,9808 0,9945 0,9179 0,5308 0,9455 0,6165 0,9854 0,9249
Average 0,9810 0,9927 0,9276 0,5869 0,9516 0,6753 0,9863 0,6753
MAverage 0,9820 0,9955 0,9435 0,5939 0,9614 0,6797 0,9883 0,9371
All Attack 0,9810 0,9951 0,9276 0,5710 0,9516 0,6575 0,9863 0,9331
In conclusion, a computational example is given in this section. The proposed model is applied to
investigate different attack types. Four performance metrics and attack and filler sizes are evaluated for
all attack models. Results show that the proposed model is valid for mentioned attack models.
3 Conclusions
In this study, a mixed attack type was introduced to the system. An SVM model was developed to identify
and address these attacks. Two SVM models were created for each replication. The first model comprised
90% genuine and attack users in the learning set. The second model incorporated synthetic users
generated by ODBOT in addition to the first model. Performance metric results indicate that the second
model outperforms the first model. Future research could focus on optimizing the required number of
synthetic users in the ODBOT method, aiming to propose models that deliver quicker results with time
efficiency.
References
1. Zhang, F., Zhang, Z., Zhang, P., Wang, S.: UD-HMM: An unsupervised method for shilling attack
detection based on hidden Markov model and hierarchical clustering. Knowledge-Based Syst. 148,
146–166 (2018).
2. Zhou, W., Wen, J., Xiong, Q., Gao, M., Zeng, J.: SVM-TIA a shilling attack detection method based
on SVM and target item analysis in recommender systems. Neurocomputing. 210, 197–205 (2016).
https://doi.org/10.1016/j.neucom.2015.12.137.
3. Hao, Y., Zhang, F.: An unsupervised detection method for shilling attacks based on deep learning
and community detection. Soft Comput. 25, 477–494 (2021). https://doi.org/10.1007/s00500-020-
Page 684
SEMIT 2023
05162-6.
4. Rani, S., Kaur, M., Kumar, M., Ravi, V., Ghosh, U., Mohanty, J.R.: Detection of shilling attack in
recommender system for YouTube video statistics using machine learning techniques. SOFT
Comput. 27, 377–389 (2023). https://doi.org/10.1007/s00500-021-05586-8.
5. Bhaumik, R., Williams, C., Mobasher, B., Burke, R.: Securing collaborative filtering against
malicious attacks through anomaly detection. AAAI Work. - Tech. Rep. WS-06-10, 50–59 (2006).
6. Adomavicius, G., Bockstedt, J.C., Curley, S.P., Zhang, J.: Do recommender systems manipulate
consumer preferences? A study of anchoring effects. Inf. Syst. Res. 24, 956–975 (2013).
https://doi.org/10.1287/isre.2013.0497.
7. Ayaz, H.İ., Kamışlı Öztürk, Z.: Shilling Attack Detection for Recommender Systems with Quality
Tools. In: International Cumhuriyet Artificial Intelligence Applications Conference. pp. 21–25. ,
Sivas (2022).
8. Ayaz, H.İ., Kamışlı Öztürk, Z.: Online Detection of Shilling Attacks on Recommender Systems wıth
Shiryaev Roberts Procedure. In: Bay, Ö.F. (ed.) 10th International Conference on Advanced
Technologies. pp. 114–117. Taylor and Francis Inc., Van (2022).
9. Ibrahim, M.H.: ODBOT: Outlier detection-based oversampling technique for imbalanced datasets
learning. Neural Comput. Appl. 33, 15781–15806 (2021). https://doi.org/10.1007/s00521-021-
06198-x.
Page 685
SEMIT 2023
Advancing Structural Integrity Assessment through Machine Learning in RC Columns
Aybike Özyüksel Çiftçioğlu*

a
Department of Civil Engineering, Faculty of Engineering, Manisa Celal Bayar University, Manisa,
Türkiye
* Corresponding author: E-mail address: aybike.ozyuksel@cbu.edu.tr.
Abstract
This research presents a comprehensive investigation focused on the predictive modeling of shear
strength in rectangular reinforced concrete (RC) columns using machine learning techniques. Addressing
the intricate challenge posed by shear strength complexity is pivotal for assessing column stability and
structural integrity. The study curates an extensive dataset of 545 experimental observations from diverse
literature sources. Employing four distinct machine learning regression models, the study rigorously
evaluates their performance using established metrics including coefficient of determination, root mean
squared error, mean absolute error, and mean absolute percentage error. The outcomes highlight the
robust predictive capabilities of the selected machine learning models. Notably, high coefficient of
determination values, ranging from 0.936 to 0.939 (93.6% to 93.9% explanatory capacity), underscore
the models' remarkable ability to elucidate shear strength variations. Noteworthy accuracy is observed,
with the Random Forest model demonstrating a slight performance advantage over the Decision Tree
model. Additionally, the efficacy of the XGBoost and LightGBM is highlighted, suggesting their
potential for accurate shear strength estimation. The study validates the precision of the model through
scatter plots and error distribution plots, confirming accurate shear strength predictions in various
scenarios. This work highlights the potential of machine learning models in improving accurate shear
strength estimation, thus significantly contributing to the advancement of structural engineering
methodologies.
Keywords: RC Columns; Shear Strength; Concrete; Machine Learning.
1. Introduction
Reinforced concrete (RC) columns are crucial structural elements in the field of civil engineering,
providing essential support and stability to buildings and infrastructures worldwide [1]. These columns
are designed to withstand a variety of forces, including compression, bending, and shear. Although
significant research has been conducted to accurately predict their behavior under various loading
conditions, one particular aspect that requires the utmost attention is the accurate estimation of shear
strength in RC columns [2,3].
Shear strength plays a pivotal role in determining the overall structural integrity and safety of RC
columns. Unlike compression and bending forces, which are relatively more predictable, shear forces
exert complex effects on these columns, making their accurate estimation a formidable challenge. Shear
failure can occur suddenly and catastrophically, leading to the collapse of the entire structure. Therefore,
a precise prediction of shear strength is of paramount importance to ensure structural reliability and avoid
potential disasters [4].
The implications of underestimating or overestimating the shear strength of RC columns are profound.
If the shear strength is underestimated, it could lead to inadequate reinforcement or a lack of appropriate
design measures. This, in turn, increases the risk of premature shear failure, compromising structural
stability and posing a threat to the safety of occupants. On the other hand, overestimating shear strength
may result in an overly conservative design, leading to unnecessary material and financial costs.
Therefore, achieving an accurate prediction of shear strength is essential not only for safety but also for
optimizing construction practices and resource allocation.
Page 686
SEMIT 2023
The challenges in accurately estimating shear strength arise from the intricate interplay between
various factors influencing the behavior of RC columns. These factors include concrete strength,
reinforcement detailing, column dimensions, loading conditions, and boundary constraints. Researchers
and engineers face the ongoing task of developing reliable models and techniques to capture complex
shear behavior and accurately estimate the shear strength of RC columns.
Machine learning (ML) has emerged as a promising approach to address the challenges associated with
accurately estimating shear strength in RC columns. ML techniques leverage the power of data-driven
analysis and computational algorithms to uncover intricate patterns and relationships within complex
systems [5,6]. In the realm of civil engineering, ML offers a unique opportunity to enhance our
understanding of shear behavior by integrating a multitude of influencing factors and their interactions.
Using historical data on RC column performance and behavior under diverse loading conditions, ML
models can learn from this information and make generalizations accordingly. This allows them to make
predictions about shear strength that go beyond the limitations of traditional analytical methods. ML
algorithms, such as decision trees, support vector machines, neural networks, and ensemble methods, can
be trained on datasets comprising various column geometries, material properties, loading scenarios, and
failure modes. The models can then capture intricate nonlinear relationships that might be challenging to
express through conventional equations. One of the significant advantages of using ML for shear strength
estimation is its adaptability to changing conditions and new data. As new research findings and
experimental data become available, ML models can be updated and refined, continuously improving
their accuracy [7,8].
However, utilizing ML in the context of shear strength prediction comes with its own set of challenges.
High-quality and well-curated datasets are essential for training reliable models. Ensuring that the data is
representative of various real-world scenarios is crucial to prevent bias and overfitting. Furthermore, the
interpretability of ML models is important to gain insight into how different factors contribute to shear
strength predictions, which is crucial to building trust in these models among engineers and stakeholders.
The integration of ML into the realm of civil engineering brings about a significant shift in the approach
to estimating shear strength in RC columns. It allows for a more holistic and data-driven understanding
of this complex phenomenon, enabling more accurate predictions and better-informed design decisions.
As researchers and practitioners continue to bridge the gap between traditional engineering principles and
modern data-driven techniques, the potential to optimize structural safety, construction practices, and
resource allocation remains a compelling prospect.
The main contributions are summarized as follows:
Firstly, this study undertakes a comprehensive investigation to address the intricate challenge of
accurately estimating shear strength in RC columns through the utilization of advanced ML models. By
assembling a substantial dataset comprising 545 experimental observations from diverse sources in the
literature, the research sets the foundation for predictive modeling.
Secondly, the study meticulously assesses the performance of four distinct ML regression models,
namely Random Forest (RF), Decision Tree (DT), XGBoost, and LightGBM. Through a rigorous
evaluation using established metrics, including the coefficient of determination (R2), the mean squared
error (RMSE), the mean absolute error (MAE), and the mean absolute percentage error (MAPE), the
research quantifies the predictive capabilities.
Thirdly, the study provides compelling evidence of the robustness in capturing the nuanced variations
in shear strength. In particular, impressive coefficient of determination values, equivalent to a range of
93.6% to 93.9% explanatory capacity, underscore their exceptional capability in elucidating intricate
shear behaviors.
Furthermore, the comparative analysis highlights the slightly superior performance of the RF model
over the DT model, reaffirming the effectiveness of ensemble methods in this context. In particular, the
study expands its exploration to include XGBoost and LightGBM models, showcasing their potential as
accurate shear strength predictors.
Page 687
SEMIT 2023
2. Database
A comprehensive compilation of 545 experimental data sets on rectangular RC columns has been
meticulously gathered from the literature [9–29]. The investigation requires the provision of ten crucial
input parameters, encompassing geometric dimensions, bar specifications, material properties, and axial
load (P), to accurately estimate the shear strength of RC columns (Vmax). The geometric dimensions
encompass the height of the column (L), the width of the cross-section (B), and the length of the cross-
section (H). Reinforcement details constitute the longitudinal reinforcement ratio (ρl), the transverse
reinforcement bar ratio (ρh), and the spacing of transverse reinforcements (s), all of which significantly
influence the capacity of the column to resist shear forces. Incorporating the properties of the material is
of utmost importance, and we take into account the yield strength of both longitudinal (fyl) and transverse
(fyw) reinforcing bars, as well as the compressive strength of the concrete (fc). These characteristics of
the material directly affect the overall performance and structural integrity of the RC columns under
varying load conditions. The descriptive statistics for the variables are presented in Table 1, covering the
minimum, maximum, mean, and standard deviation values.
Table 1. Descriptive statistics of the dataset

L B H ρl ρh s fc fyl fyw P Vmax
min 225.00 108.50 100.00 0.20 0.01 20.00 20.00 77.06 215.00 0.00 13.34
max 3000.00 610.00 610.00 4.50 4.00 457.20 141.00 745.00 1470.00 5491.75 982.00
mean 1285.87 284.20 301.00 2.15 0.95 101.41 49.12 447.90 496.88 1135.67 228.61
st dv 651.85 109.51 115.65 0.70 0.94 78.54 27.37 79.71 226.32 1083.46 175.22
Figure 1 illustrates a visual representation of the colormap correlation matrix. This matrix effectively
portrays the relationships between variables within the dataset. Notably, the colors red and blue denote
strong positive and negative correlations, respectively. The correlation matrix distinctly reveals that
variables B (cross-section), H (cross-section length), and P (axial load) exhibit a significant correlation
coefficient of +0.6 in relation to the shear strength of the RC columns. Furthermore, it should be noted
that the variable fc (compressive strength of concrete) manifests a negative correlation coefficient of -0.1
with compressive strength.
Figure 1. Correlation matrix of variables in the dataset
Page 688
SEMIT 2023
3. Methodology
3.1. Random Forest
Random Forest (RF) is a powerful and widely used ensemble learning technique in the field of ML
[30]. It is renowned for its robustness, versatility, and excellent predictive performance in both
classification and regression tasks. RF has gained popularity due to its ability to handle complex data sets
and mitigate issues like overfitting. The method is an extension of the DT algorithm, and it combines
multiple individual decision trees to form a more accurate and reliable predictor. The fundamental
principle behind RF lies in aggregating the predictions of multiple decision trees, thereby forming a forest
of trees. Each tree in the forest is trained on a random subset of the data, and at each split, a random subset
of features is considered. This randomness introduces diversity among the trees, reducing overfitting and
enhancing the generalization ability of the model. During the prediction phase, the final output is
determined by averaging (in regression) or voting (in classification) the outputs of individual trees,
resulting in a more robust and accurate prediction. In the classification context, RF constructs multiple
decision trees during the training phase. Each decision tree is built on a random subset of the training
data, and at each split, a random subset of features is considered. During inference, each decision tree in
the forest independently predicts the class label for a given input, and the final class is determined by
majority voting. On the other hand, RF is equally adept at solving regression problems. In regression
tasks, the algorithm assembles multiple decision trees just as in the classification setting, but the
predictions from each individual tree are averaged instead of voting. Consequently, the final prediction
is the mean of the results from all decision trees. This averaging process ensures that the RF regression
model can capture complex nonlinear relationships between features and the target variable while also
mitigating the effects of outliers and noise. RF finds extensive applications in various engineering
domains. In civil engineering, it can help predict structural integrity, soil stability, etc. The proficiency of
the algorithm in handling high-dimensional data and accommodating intricate non-linear relationships
renders it an invaluable instrument for addressing intricate engineering challenges and augmenting
decision-making endeavors within diverse industrial contexts.
3.2. Decision Tree
The Decision Tree (DT) algorithm is a widely employed and interpretable ML technique that serves
as a powerful tool for classification and regression tasks [31]. Its fundamental principle lies in partitioning
the feature space into a hierarchical structure of nodes, where each node represents a decision based on a
particular feature, leading to subsequent splits until the leaf nodes produce the final predictions. Decision
Trees are extensively utilized due to their ability to handle both categorical and numerical data, their ease
of interpretability, and their capability to capture nonlinear relationships between features. The
construction of a DT starts with selecting the most informative feature from the dataset to create the root
node. The subsequent nodes are generated by iteratively choosing the best feature and its optimal split
point, which maximizes information gain or minimizes impurity. Information gain measures the reduction
in uncertainty after a split, whereas impurity refers to the homogeneity of the target values within a node.
The recursive partitioning process continues until predefined termination conditions are met, such as
reaching a specified depth or having a minimum number of data points in a leaf node. To avoid overfitting,
techniques like pruning or setting a minimum number of samples per leaf are commonly used. Decision
trees find wide-ranging applications in engineering disciplines due to their versatility and
comprehensibility. Their capability to handle large datasets and effectively model complex systems has
positioned Decision Trees as a valuable asset in the engineering domain, enabling informed decision-
making and problem-solving across various sectors.
3.3. Extreme Gradient Boosting
Extreme Gradient Boosting (XGBoost) is a state-of-the-art ML algorithm renowned for its exceptional
predictive capabilities [32]. XGBoost has gained widespread popularity in both academia and industry
due to its outstanding performance across a diverse array of tasks, including regression, classification,
and ranking. The efficacy of the algorithm arises from its ability to blend the advantages of boosting and
gradient descent, enabling it to handle complex, high-dimensional data with remarkable accuracy and
efficiency. XGBoost operates by building an ensemble of weak learners, typically decision trees, in a
Page 689
SEMIT 2023
sequential manner. Each subsequent tree aims to correct the errors of the previous ones, progressively
refining the predictions. The algorithm employs a combination of regularization techniques, such as
regularization L1 and L2, to avoid overfitting and improve generalization. Additionally, it employs a
novel "gradient boosting" strategy to minimize a cost function by iteratively fitting weak learners to the
negative gradient of the loss function. This technique facilitates optimizing the performance of the model
by minimizing prediction errors. In the realm of engineering, XGBoost has found extensive applications
across various domains. Its speed and adaptability enable engineers to effectively handle large-scale
datasets and complex relationships, making XGBoost an indispensable tool for solving intricate
engineering challenges and enhancing decision-making processes in diverse engineering applications.
3.4. Light Gradient Boosting Machine
Light Gradient Boosting Machine (LightGBM) is a gradient-boosting framework that has gained
remarkable popularity in the field of ML due to its efficiency and high performance [33]. It is based on
the concept of gradient boosting, which involves sequentially adding weak learners (typically decision
trees) to improve the accuracy of the model. What sets LightGBM apart is its focus on optimizing both
training speed and memory consumption, making it well-suited for handling large-scale datasets and
computationally intensive tasks. The core idea behind LightGBM lies in its novel approach to tree
construction and leaf-wise growth strategy. Unlike traditional depth-first tree growth, LightGBM adopts
a leaf-wise approach, where it selects the leaf node with the maximum decrease in the loss function during
each tree expansion. This technique significantly reduces the number of nodes and the depth of the tree,
thereby reducing computation time and memory usage. Additionally, LightGBM employs a histogram-
based method to quantize feature values into discrete bins, further accelerating the training process. The
algorithm also offers various regularization techniques, such as L1 and L2 regularization, to prevent
overfitting and enhance generalization. LightGBM finds wide applications in various engineering
domains, owing to its ability to handle large-scale datasets and efficiently tackle complex problems. The
superior performance and scalability of the algorithm make it a valuable asset in engineering domains
where data volume and computational resources are paramount concerns, empowering engineers to make
informed decisions and deliver efficient and accurate solutions.

In the study, a comprehensive analysis employing four distinct ML models was conducted to perform
regression analyses. The primary objective of the investigation was to accurately predict the shear
strength (Vmax) of the RC columns. The dataset, comprised of 545 data points, was divided into a 75%
training set and a 25% test set for model evaluation. The regression models used in the study were RF,
DT, XGBoost, and LightGBM. The performance of each model was evaluated based on several key
metrics, including R2, RMSE, MAE, and MAPE. The obtained results are summarized in Table 2 below:
Table 2. Performance Evaluation of Regression Models

Algorithm R2 RMSE MAE MAPE (%)
RF 0.939 48.278 28.972 13.862
DT 0.939 48.327 30.863 16.530
XGBoost 0.937 49.048 27.119 12.215
LightGBM 0.936 49.560 29.155 15.302
The results indicate strong predictive capabilities of the employed ML models, with R2 values ranging
from 0.936 to 0.939. These coefficients of determination suggest that the models can explain
approximately 93.6% to 93.9% of the variability in the shear strength of RC columns. Moreover, the
relatively low values of RMSE, MAE, and MAPE demonstrate the accuracy of the predictions and the
robustness of the models across different algorithms. The RF and DT models yield comparable results in
terms of R2, while the RF model outperforms slightly in terms of RMSE, MAE, and MAPE. This
underscores the effectiveness of ensemble methods in capturing complex relationships within the dataset.
Page 690
SEMIT 2023
Furthermore, both XGBoost and LightGBM exhibit promising results, demonstrating their potential for
accurate shear strength estimation. It is worth noting that the observed MAPE values, ranging from
12.215% to 16.530%, indicate the percentage error in the predictions. This insight is valuable for
assessing the practical utility of the models in engineering applications, as it provides an understanding
of the potential variability in the estimated shear strength.
In Figure 2, scatter plots illustrating the predictions of the four employed ML regression models are
presented. These plots depict the relationship between the predicted shear strength values and the actual
shear strength values for each data point in the test dataset. This visual representation enables a
comprehensive assessment of the predictive performance of the model and provides information on its
accuracy across the range of shear strength values. The scatter plots highlight the alignment of the
predicted values with the ideal diagonal line, which represents perfect predictions. Notably, all four
models exhibit a consistent pattern of predictions closely clustered around this diagonal line, indicating
strong predictive capabilities. Despite slight variations, the scatter plots reveal that the predictions closely
follow the actual values, affirming the models' ability to capture the underlying trends and relationships
within the dataset. This observation is consistent with the quantitative results presented in Table 1, where
metrics such as R2, RMSE, MAE, and MAPE indicated favorable predictive performance. Moreover, the
dispersion of points around the diagonal line suggests that the models perform consistently across the
entire range of shear strength values. This consistency is crucial in engineering applications where
accurate predictions are essential across a broad spectrum of scenarios. The scatter plots further
emphasize the proficiency of the RF and DT models, as their predictions align more closely with the ideal
diagonal line compared to the other models. This alignment signifies the robustness of the models to
handle the various patterns present in the dataset. The XGBoost and LightGBM models also demonstrate
respectable alignment, reinforcing their viability for shear strength estimation.
Figure 2. Scatter Plots for Regression Models
Page 691
SEMIT 2023
Figure 3 displays the error distribution plots corresponding to the predictions generated by the four
distinct ML regression models. These plots illustrate the distribution of errors between the predicted shear
strength values and the actual shear strength values across the test dataset. This visualization provides
insight into the ability of the models to consistently estimate shear strength and the magnitude of errors
associated with their predictions. The error distribution plots depict a symmetric distribution of errors
centered around zero for all four models. This symmetry signifies that the models tend to produce
predictions with balanced overestimations and underestimations, indicating the absence of systemic
biases. While the majority of errors are clustered around zero, there are occasional instances of larger
errors. However, these occurrences are limited and do not deviate significantly from the central trend.
This observation aligns with the quantitative metrics reported earlier, such as the MAE and RMSE, which
demonstrated relatively low magnitudes. Furthermore, the error distribution plots provide evidence of the
consistency of the models in terms of error magnitudes across the entire range of shear strength values.
This uniformity in the error distribution is an essential characteristic in applications where reliable
predictions are crucial in various scenarios. The RF and DT models exhibit narrower error distributions,
indicating their precision in generating predictions close to the actual shear strength values. This precision
is consistent with the visual and numerical findings presented in the scatter plots (Figure 2) and the
performance metrics in Table 1. The XGBoost and LightGBM models also demonstrate commendable
error distribution patterns, indicating their potential utility for shear strength estimation tasks.
Figure 3. Error Distribution Plots for Regression Models
5. Conclusions
This research underscores the formidable potential of ML models in precise shear strength estimation,
thus contributing substantially to advancements in structural engineering methodologies. The research
findings highlight the significant predictive capacity inherent in the utilized ML models utilized,
demonstrating exceptional coefficients of determination ranging from 93.6% to 93.9% in relation to the
variations observed in the shear strength of reinforced concrete columns. The comparative analysis
highlights the effectiveness of ensemble methods, with the RF model exhibiting slight superiority in
predictive performance over the DT model. Moreover, the expansion of the investigation to incorporate
Page 692
SEMIT 2023
XGBoost and LightGBM models accentuates their promising role in precise estimation tasks related to
shear strength. Scatter plots visually validate the models' proficiency in predicting shear strength values
across a range of scenarios, while error distribution plots emphasize the models' balanced overestimations
and underestimations, reinforcing their reliability. In essence, this study underscores the potential of ML
models to significantly enhance shear strength estimation accuracy, thus contributing to the advancement
of structural engineering practices.
References
1. Özyüksel Çiftçioğlu, A. and M.Z. Naser, 2022. Hiding in plain sight: What can interpretable
unsupervised machine learning and clustering analysis tell us about the fire behavior of reinforced
concrete columns? Structures, 40(April), 920–935.
2. Zhou, X. and J. Liu, 2010. Seismic behavior and shear strength of tubed RC short columns. Journal
of Constructional Steel Research, 66(3), 385–397.
3. Park, H.-G., E.-J. Yu and K.-K. Choi, 2012. Shear-strength degradation model for RC columns
subjected to cyclic loading. Engineering Structures, 34, 187–197.
4. Wong, H.F. and J.S. Kuang, 2014. Predicting shear strength of RC interior beam–column joints
by modified rotating-angle softened-truss model. Computers & Structures, 133, 12–17.
5. Moslemi, S., A. Mirzazadeh, G.-W. Weber and M.A. Sobhanallahi, 2021. Integration of neural
network and AP-NDEA model for performance evaluation of sustainable pharmaceutical supply
chain. OPSEARCH.
6. Khalilpourazari, S., A. Mirzazadeh, G.-W. Weber and S.H.R. Pasandideh, 2020. A robust fuzzy
approach for constrained multi-product economic production quantity with imperfect items and
rework process. Optimization, 69(1), 63–90.
7. Naser, M.Z. and A.O. Ciftcioglu, 2022. Causal Discovery and Causal Learning for Fire Resistance
Evaluation: Incorporating Domain Knowledge.
8. Naser, M.Z. and A.Ö. Çiftçioğlu, 2023. Revisiting Forgotten Fire Tests: Causal Inference and
Counterfactuals for Learning Idealized Fire-Induced Response of RC Columns. Fire Technology,
59(4), 1761–1788.
9. Ghannoum, W., B. Sivaramakrishnan, S. Pujol, A.C. Catlin, S. Fernando, N. Yoosuf and Y. Wang,
https://datacenterhub.org/dataviewer/view/neesdatabases:db/aci_369_rectangular_column_datab
ase/. .
10. Belkacem, M.A., H. Bechtoula, N. Bourahla and A.A. Belkacem, 2019. Effect of axial load and
transverse reinforcements on the seismic performance of reinforced concrete columns. Frontiers
of Structural and Civil Engineering, 13(4), 831–851.
11. Marefat, M.S., M. Khanmohammadi, M.K. Bahrani and A. Goli, 2006. Experimental Assessment
of Reinforced Concrete Columns with Deficient Seismic Details under Cyclic Load. Advances in
Structural Engineering, 9(3), 337–347.
12. Li, Y.-A., Y.-T. Huang and S.-J. Hwang, 2014. Seismic Response of Reinforced Concrete Short
Columns Failed in Shear. ACI Structural Journal, 111,.
13. Eom, T.-S., S.-M. Kang, H.-G. Park, T.-W. Choi and J.-M. Jin, 2014. Cyclic loading test for
reinforced concrete columns with continuous rectangular and polygonal hoops. Engineering
Structures, 67, 39–49.
14. Goksu, C., H. Yilmaz, S. Chowdhury, K. Orakcal and A. Ilki, 2014. The Effect of Lap Splice
Length on the Cyclic Lateral Load Behavior of RC Members with Low-Strength Concrete and
Plain Bars. Advances in Structural Engineering, 17, 639–658.
Page 693
SEMIT 2023
15. Zhang, Y., S. Zheng, X. Rong, L. Dong and H. Zheng, 2019. Seismic Performance of Reinforced
Concrete Short Columns Subjected to Freeze–Thaw Cycles. Applied Sciences, 9(13),.
16. Karbasi Arani, K., M. Di Ludovico, M. Marefat, A. Prota and G. Manfredi, 2014. Lateral
Response Evaluation of Old Type Reinforced Concrete Columns with Smooth Bars. Aci
Structural Journal, 111,.
17. Karbasi Arani, K., M. Marefat, A. Amrollahi-Biucky and M. Khanmohammadi, 2013.
Experimental seismic evaluation of old concrete columns reinforced by plain bars. The Structural
Design of Tall and Special Buildings, 22,.
18. Shi, Q., L. Ma, Q. Wang, B. Wang and K. Yang, 2021. Seismic performance of square concrete
columns reinforced with grade 600 MPa longitudinal and transverse reinforcement steel under
high axial load. Structures, 32, 1955–1970.
19. Zhang, J., R. Cai, C. Li and X. Liu, 2020. Seismic behavior of high-strength concrete columns
reinforced with high-strength steel bars. Engineering Structures, 218, 110861.
20. Dinh, N.H., S.-H. Park and K.-K. Choi, 2019. Seismic performance of reinforced concrete
columns retrofitted by textile-reinforced mortar jackets. Structure and Infrastructure Engineering.
21. Hugo, R., F. André and A. António, 2016. Behavior of Rectangular Reinforced-Concrete Columns
under Biaxial Cyclic Loading and Variable Axial Loads. Journal of Structural Engineering,
142(1), 4015085.
22. Li, Y., S. Cao and D.-H. Jing, 2018. Concrete Columns Reinforced with High-Strength Steel
Subjected to Reversed Cycle Loading. ACI Structural Journal, 115,.
23. Melo, J., H. Varum and T. Rossetto, 2015. Experimental cyclic behaviour of RC columns with
plain bars and proposal for Eurocode 8 formula improvement. Engineering Structures, 88, 22–36.
24. Ho, J.C.M., 2012. Experimental Tests on High-Strength Concrete Columns Subjected to
Combined Medium Axial Load and Flexure. Advances in Structural Engineering, 15(8), 1359–
1374.
25. A., W.K. and J.J. O., 1984. Influence of Reinforcement on RC Short Column Lateral Resistance.
Journal of Structural Engineering, 110(1), 90–104.
26. Lam, S., B. Wu, Y. Wong, Z.Y. Wang, Z. Liu and C. Li, 2003. Drift Capacity of Rectangular
Reinforced Concrete Columns with Low Lateral Confinement and High-Axial Load. Journal of
Structural Engineering-Asce - J STRUCT ENG-ASCE, 129,.
27. Hwang, S.-K. and H.-D. Yun, 2004. Effects of transverse reinforcement on flexural behaviour of
high-strength concrete columns. Engineering Structures, 26, 1–12.
28. Ahn, J.-M. and S.-W. Shin, 2007. An evaluation of ductility of high-strength reinforced concrete
columns subjected to reversed cyclic loads under axial compression. Magazine of Concrete
Research - MAG CONCR RES, 59, 29–44.
29. Woods, J., P. Kiousis, M. Ehsani, H. Saadatmanesh and W. Fritz, 2007. Bending ductility of
rectangular high strength concrete columns. Engineering Structures - ENG STRUCT, 29, 1783–
1790.
30. Breiman, L., 2001. Random Forests. Machine Learning, 45(1), 5–32.
31. Quinlan, J.R., 1986. Induction of decision trees. Machine Learning, 1(1), 81–106.
32. Chen, T. and C. Guestrin, 2016. XGBoost: A Scalable Tree Boosting System, In Proc. 22nd ACM
SIGKDD Int. Conf. Knowl. Discov. Data Min., Association for Computing Machinery, pp: 785–
794.
Page 694
SEMIT 2023
33. Ke, G., Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye and T.-Y. Liu, 2017. LightGBM:
A Highly Efficient Gradient Boosting Decision Tree, In Proc. 31st Int. Conf. Neural Inf. Process.
Syst., Curran Associates Inc., pp: 3149–3157.
Page 695
SEMIT 2023
Fuzzy Logic Methods for Determining the Mechanical Behavior of Masonry Walls
Başak Zengin 1 [0000-0003-3719-9423] Pınar Usta 2[0000-0001-9809-3855] and Ozge Onat 3 [0000-0002-4336-0212]
1
Kahramanmaras Istiklal University, Construction Technology, Kahramanmaras/Turkey
2
Department of Civil Engineering, Technology Faculty, Isparta University of Applied Scienes, Isparta,
Turkey
3
Graduate Institute, Isparta University of Applied Sciences, Isparta, Turkey
zenginbasak@gmail.com
Abstract
The mechanical behavior of the wall, in-situ wall tests and numerical analysis are required along with the
material properties. With the support of smart learning techniques, the state of the walls was estimated.
This is possible to obtain healthy data to support the experiment and modeling. The utilization of
mathematical tools like Fuzzy Logic has been demonstrated to be beneficial in resolving intricate
engineering issues, without the need to replicate the studied phenomenon, given that the only available
information consists of the problem's parameters and desired outcomes. To analyses the behavior of the
wall more accurately and in a short time, analyses were made using the fuzzy method, one of the smart
learning techniques, and compared with the data in the studies in which experimental analysis were
applied. The behavior of the wall, the ductility and energy capacity of the wall were tried to be estimated.
In the fuzzy, material parameters and wall load capacities that will affect the properties of the wall are
used as inputs. Thirty-five (35) data sets, experiments and modeling data of different studies were taken.
Estimation results were compared with empirical results.
Keywords: Ductility, Energy Capacity, Fuzzy Logic, and Masonry Wall.
1 Introduction
Unreinforced masonry structures are known for their insufficient seismic performance, which has been
confirmed through numerous investigations of earthquake disasters. This inadequate behavior is typically
attributed to the unsuitable characteristics displayed by the masonry when subjected to tensile stress. In
this regard, unreinforced masonry structures can be enhanced by incorporating vertically and horizontally
reinforced concrete elements (known as confining elements) to create confined masonry structures. These
confined structures offer the advantage of efficiently confining masonry walls and enhancing the
structure's' ability to withstand deformation. [1–4].
There are many factors that affect the compressive strength of clay brick masonry walls connected by
cement mortar. In general, the way masonry responds to compressive loads depends on how the brick
and mortar interact. The properties of the materials themselves (bricks and mortar) also change when
used in a masonry wall compared to when they are used individually. Additionally, it is important to
consider that masonry is a material that has different properties in different directions and is highly
influenced by construction processes. The wide variety of metrics available, some of which are
quantitative (such as brick compression resistance) and others more qualitative (such as construction
processes), make the construction and design of masonry structures more complex.
The deformability of the wall is influenced by the specific properties of the mortar, especially in the joints
of the matrix [5-6]. The mechanical properties of the components that compose the walls and joining
points were determined through laboratory experiments. Wall models were created. The behavior of the
walls was analyzed, and the obtained parameters were utilized in the modeling process for comparisons
[7-10]. The fragility of masonry walls is evident when they are exposed to compressive stresses. However,
under shear stresses, they can exhibit ductile behavior to some extent. Research on the load-displacement
behavior of masonry walls under axial load has shown that the compressive strength of the masonry can
be estimated using the single-axis compressive strength of the masonry unit, while also considering the
compressive strength of the mortar. Further analysis of the district split of the walls revealed that the
power level of the wall is low, leading to reduced ductility. To address this issue, the wall distance
proportion must be carefully designed, and the use of brittle materials should be evaluated. Therefore, it
Page 696
SEMIT 2023
is important to maintain earthquake reliability and ductility through appropriate construction methods
[11]. Research shows that masonry infill walls have an impact on the strength and stiffness of unfilled
frame structures. In previous studies, the design of buildings has often disregarded infill walls due to the
unknown brittle behaviors of these walls. It is recommended to conduct experimental tests and analytical
investigations in order to understand the behavior of the frame and composite infill walls [12, 13 and 14].
Additionally, extensive testing should be conducted to determine the failure at the corner of the walls
during the damage mechanisms. When subjected to lateral loads, an infill wall acts as a diagonal structure
that connects the two corners. The behavior of masonry structures under compression is brittle, and they
also exhibit limited ductile behavior under shear stresses. Experimental examinations of overload
displacement behaviors under axial load in masonry structures indicate that the uniaxial strength of these
structures can only be determined by the uniaxial strength of the masonry elements, and the mortar effect
is restricted. [15-21]. Tomazevic (1999) suggests that the durability effect of the walls can be explained
by considering both inelastic and elastic achievement, using the concept of an excellent plastic curve.
This illustrates a reliance on the equal energy dissipation of actual and theoretical stress-strain curves
[22]. In a study conducted by Essa et al. (2014), the researchers evaluated the impact of the performance
of high strength reinforced concrete beams on the ductility of the infill walls. To achieve this objective,
they prepared different types of materials with varying thicknesses and loaded them with filled and non-
filled beams [23]. Several parameters were obtained during this count. It was found that the type of
scrape-off influences ductility. As a result, the load-bearing walls' behavior could be determined, and it
might be possible to devise techniques for repairing the damage [24].
Modeling techniques are often employed to study the behaviors of walls, as experimental work on the
walls is not always feasible. Composite models are commonly used to represent walls in these
simulations. However, there are instances where quick judgment is necessary. Recent studies have also
incorporated artificial intelligence methods alongside traditional approaches.
Artificial Intelligence (AI) refers to the advanced concept of machines being capable of independently
performing tasks by utilizing algorithms/models that enable computers and machines to function
intelligently. Consequently, AI has been developed at the intersection of various fields, including
computer science, cybernetics, information theory, psychology, linguistics, and neurophysiology [25].
All previous research has demonstrated the effectiveness of AI algorithms using traditional methods
related to assessment, decision-making, prediction, and optimization. These methods are applicable to
the current survey because network issues, especially in the interdisciplinary field of civil engineering,
are influenced by various global factors that are difficult to simulate using math, physics, and mechanics
calculations. Furthermore, the success of AI in solving engineering problems depends on the ability to
learn from data inputs and outputs, ensuring that the intricate relationships between the data are accurately
simulated, even when the interdependencies are uncertain, or the physical phenomena are challenging to
interpret [26]. Over the years, artificial neural networks (ANNs) and customized Neuro-fuzzy inference
systems (ANFIS) have been widely used in civil engineering due to their resilience and high precision.
These models have proven successful in various applications, including the evaluation of compressive
strength and elastic modulus of different concrete types, as well as the analysis of concrete drying
shrinkage and durability [27-36]. Instead of using pragmatic models, the objective of this report is to
determine the compressive strength of a masonry structure made of clay bricks and cement mortar using
two numerical techniques: Artificial Neural Networks and Fuzzy Logic. These techniques solve a
complex problem based on the input and output data of the system, without considering the physical
phenomena that inspired them [36-39].
There have been only a few studies that have utilized artificial intelligence technology in the analysis of
masonry structures. In one study conducted by Plevris and Asteris [40], they employed an artificial neural
network (ANN) to forecast the failure of masonry under biaxial four (4) compression. Their ANN model
not only successfully predicted the failure pattern of masonry at a specific θ (angle between the bed joint
and horizontal compressive stress), but also identified the three-dimensional failure surface formed by
the failure pattern at any θ. Afterwards, Plevris and Asteris [41] proceeded to develop a dimensionless
model for predicting masonry surface failure based on ANN. This model can also be applied to other
masonry materials with similar geometry and mechanical properties. However, there have been very few
studies conducted on the application of artificial intelligence techniques to predict masonry behavior.
Page 697
SEMIT 2023
Zhang et al. [37] established a cellular automata model to determine the cracking patterns of vertically
loaded masonry structures, and Garzón-Roca et al. [38,39] estimated the compressive strength of brick
masonry using both ANN and fuzzy logic methods. Additionally, Plevris et al. [40] developed an artificial
neural network (ANN) model to analyze the failure surface of masonry under biaxial compressive stress.
The main aim of their research was to create ANN and ANFIS models for predicting the compressive
strength of masonry prisms made of hollow concrete blocks. To achieve this goal, they compiled and
evaluated 102 sets of experimental data, which were obtained from their own experiments as well as from
previously published technical literature. Ultimately, the proposed models were compared to several
empirical techniques in order to assess their reliability and accuracy [42-45].
The results of this study, which used Fuzzy Logic methods, were subsequently compared with various
empirical suggestions to demonstrate their accuracy.
2 Material and Method

2.1. Material
Factory bricks, hollow bricks, clay bricks, pumice, cement mortar, and lime mortar were commonly
employed in the research. The mechanical properties of the materials were determined by assessing their
compressive strength and bending strength. (Fig 1).
Fig.1 Masonry unit

The table with the data obtained from the studies is presented (Table1) [46-48].
Table 1. Mechanical strengths of materials
Flexure Strength Flexure Strength Compressive Strength Compressive Strength
Code Brick Mortar Brick Mortar
A2 0.3 2.12 6.38 12.5
A3 0.3 2.12 6.38 12.5
A4 0.3 2.59 6.38 4.59
A5 0.3 2.59 6.38 4.59
A7 0.4 2.12 6.11 12.5
A8 0.4 2.12 6.11 12.5
A9 0.4 2.12 6.11 12.5
A11 0.4 2.59 6.11 4.59
A12 0.4 2.59 6.11 4.59
A13 0.8 2.12 6.05 12.5
A14 0.8 2.12 6.05 12.5
A15 0.8 2.12 6.05 12.5
A17 0.8 2.59 6.05 4.59
A19 0.5 2.12 2.63 12.5
A20 0.5 2.12 2.63 12.5
A22 0.5 2.59 2.63 4.59
A23 0.5 2.59 2.63 4.59
A24 0.5 2.59 2.63 4.59
B1 3 3 10.3 7.2
B2 4 4 8.8 7.1
B3 3.4 3.4 7.5 6
B4 3.3 3.3 7.8 5.7
B5 4 4 8.5 7.3
B6 3.3 3.3 7.9 5.2
C1 1.32 1.32 2.37 2.37
C2 1.32 1.32 2.37 2.37
D2 2 2 10.5 10.5
D3 2 2 11.5 11.5
Page 698
SEMIT 2023
Sample models of the walls were established in the studies, and experiments were conducted
(Figure 2-3). In order to analyze and compare the conditions of these test walls, simulations were
conducted. It was determined that the results of the experiment and the fuzzy logic were compared, and
these results showed similarity.
Fig.2 Experimental of Walls Set

It was discovered that the mechanical properties of the materials used in wall production have an impact
on the ductility and energy capacity of the walls. As a result, the data obtained from these studies were
utilized in this study.
Fig 3. Before-After Experimental System

The data obtained depending on the experiment data of the walls are given in the table. Thirty- five
(35) sample wall works were selected. These walls are coded as follows (Table 2)
Table 2. Walls of Experimental Results
Code Horizontal Δu Δy Ductility Energy
Load Capacity
A1 92.53 34.12 28.15 1.21 2325.41
A2 107.16 29.81 22.85 1.30 1722.03
A3 95.44 32.67 17.7 1.85 1824.57
A4 42.02 35.96 23.61 1.52 786.72
A5 49.07 24.99 12.08 2.07 703.11
A6 42.96 52.27 30.7 1.70 1263.54
A7 43.39 31.08 23.73 1.31 756.31
A8 52.51 22.18 17.85 1.24 565.14
Page 699
SEMIT 2023
A9 44.47 21.87 14.73 1.49 527.72

A10 26.21 18.77 16.37 1.15 306.2
A11 40.57 13.41 11.88 1.29 244.89
A12 34.41 34.43 17.18 2.00 500.43
A13 18.93 22.65 17.8 1.27 295.81
A14 51.27 30.56 9.89 3.09 1169.9
A15 35.77 24.51 15.2 1.61 553.26
A16 12.08 21.93 10.56 2.08 186.83
A17 14.41 25.6 9.43 2.71 270.66
A18 12.42 23.05 8.47 2.72 206.73
A19 46.21 36.21 22.95 1.58 1127.48
A20 96.06 28.15 21.47 1.31 1400.33
A21 53.58 21.79 10.3 2.12 471.375
A22 36.4 21.76 11.83 1.84 441.02
A23 46.69 22.95 14.55 1.57 557.16
A24 43.92 12.65 6.78 1.86 334.93
B1 84.2 8 1.8 4.44 597.82
B2 95.2 7.4 1.6 4.63 628.32
B3 52.4 21.7 2.2 9.86 1079.44
B4 75.3 14.3 1.5 9.53 1020.31
B5 63.8 15.3 1.6 9.56 925.1
B6 94.1 20.9 4.6 4.54 1750.26
C1 35.3 12 2.5 4.80 379.47
C2 63.4 11.9 1.2 9.92 716.42
D1 40 4 0.8 5.00 144
D2 42 3.7 0.8 4.63 138.6
D3 70 3.9 0.6 6.50 252
2.2. Method
Sufficient data has been collected at the start of fuzzy Logic modeling to define the parameters of the
models (Table 3-4). This data is obtained from the experiments being conducted and from previous
studies.
Table 3 Training Data Test

Flexure Flexure Compressive Compressive Ductility Energy
Strength Strength Strength Strength Capacity
Brick Mortar Brick Mortar Wall Wall
A2 0.3 2.12 6.38 12.5 1.30 1722.03
A7 0.4 2.12 6.11 12.5 1.31 756.31
A11 0.4 2.59 6.11 4.59 1.29 244.89
A17 0.8 2.59 6.05 4.59 2.71 270.66
A19 0.5 2.12 2.63 12.5 1.58 1127.48
A22 0.5 2.59 2.63 4.59 1.84 441.02
D2 2 2 10.5 10.5 4.63 138.6
Page 700
SEMIT 2023
Table 4 Testing Data Set

Flexure Flexure Compressive Compressive Ductility Energy
Strength Strength Strength Strength Capacity
Code Brick Mortar Brick Mortar Wall Wall
A1 0.3 2.12 6.38 12.5 1.21 2325.41
A3 0.3 2.12 6.38 12.5 1.85 1824.57
A4 0.3 2.59 6.38 4.59 1.52 786.72
A5 0.3 2.59 6.38 4.59 2.07 703.11
A6 0.3 2.59 6.38 4.59 1.70 1263.54
A8 0.4 2.12 6.11 12.5 1.24 565.14
A9 0.4 2.12 6.11 12.5 1.49 527.72
A10 0.4 2.59 6.11 4.59 1.15 306.20
A12 0.4 2.59 6.11 4.59 2.00 500.43
A13 0.8 2.12 6.05 12.5 1.27 295.81
A14 0.8 2.12 6.05 12.5 3.09 1169.90
A15 0.8 2.12 6.05 12.5 1.61 553.26
A16 0.8 2.59 6.05 4.59 2.08 186.83
A18 0.8 2.59 6.05 4.59 2.72 206.73
A20 0.5 2.12 2.63 12.5 1.31 1400.33
A21 0.5 2.12 2.63 12.5 2.12 471.38
A23 0.5 2.59 2.63 4.59 1.57 557.16
A24 0.5 2.59 2.63 4.59 1.86 334.93
B1 3 3 10.3 7.2 4.44 597.82
B2 4 4 8.8 7.1 4.63 628.32
B3 3.4 3.4 7.5 6 9.86 1079.44
B4 3.3 3.3 7.8 5.7 9.53 1020.32
B5 4 4 8.5 7.3 9.56 925.10
B6 3.3 3.3 7.9 5.2 4.54 1750.26
C1 1.32 1.32 2.37 2.37 4.80 379.48
C2 1.32 1.32 2.37 2.37 9.92 716.42
D1 2 2 10.5 10.5 5.00 144.00
D3 2 2 11.5 11.5 6.50 252.00
2. 2. 1. Fuzzy Logic The fuzzy Logic theory, proposed by Zadeh [50], is a mathematical application that
enables its use in uncertain environments. Unlike classic logic, which categorizes items as either
belonging (1) or not belonging (0) to a set, fuzzy Logic considers a blurred pattern as a set that does not
have a clearly defined limit [51]. This set may only contain basic information with a partial membership
rating (a number between 0 and 1). In 1985, Sugano [52] conducted a study on a fuzzy deduction
technique, which showed that the outcome of the if-then rules can be linear or constant. A fundamental
rule in Sugano's fuzzy model takes the form: "If input 1 = x and input 2 = y, then output is z = ax + by +
c." For each rule, the resulting output zi is weighted by the firing strength wi of the rule [50]. Therefore,
the final output (ofinal) of the technique is the average of all control outputs, which are balanced and
estimated as shown in figure 4.
Fig. 4 Back propagation training fuzzy logic

2.2.2. Fuzzy Logic and Set Data In order to increase the size of the training-testing database, 35 test data
sets were compiled from published information, specifically representing the test results for 35 prisms
Page 701
SEMIT 2023
(as shown in Table 3-4). For the implementation of Fuzzy Logic, the MATLAB Fuzzy Logic Toolbox
[51] was utilized to create a fuzzy deduction system. This system can be employed to determine the
ductility and energy capacity of a masonry structure (f), provided that the compressive strengths of its
bricks and mortar, as well as its flexure strength, are already known. This toolbox includes a membership
function, the if-then rules, and the weight wi. These elements define the fuzzy inference system and
establish the connections between a set of inputs and the output. In this section, a hierarchical evaluation
was first established, which divided the factors that influence the health of the masonry structure into four
layers. Based on the hierarchical model, the judgment matrix of each level was determined using expert
experience and the characteristics of masonry structures. The weight coefficient vectors were then
calculated based on the judgment matrix. The process of fuzzy logic is depicted in figure 5.
Fig. 5 Fuzzy Logic process

The evaluation matrix was determined using the grey method. Consequently, the comprehensive rating
vector for each layer factor was calculated using fuzzy linear transformation, which involves the weight
coefficients vector and the evaluation matrix. Figure 6 depicts the study upload system.
Data Upload Modeling
Testing and Training Data Training Data and FIS out-put
Fig. 6 Fuzzy Logic Study System and Upload Data
3 Finding and Results

After allocating 7 out of 35 data for training, the data underwent training. The preceding section provides
a detailed explanation of how the model operates. A comparison was made between the data obtained
Page 702
SEMIT 2023
from the fuzzy model and the actual data. The graphs generated using the acquired data demonstrate that
an accurate prediction was made using fuzzy logic. The objective of the research was to obtain
information regarding the ductility and energy capacities of the walls. By adding the horizontal loading,
according to the material strengths, the displacement of the wall at the point where it first started to flow
and the final displacement under loading were also estimated to support the study. In Figure 6 and 7 the
study was observed to be demonstrable. Looking at the R2 coefficients, which was obtained as 0.93-0.95
(figure 6-7).
Fig 6. Du- Fuzzy Logic Dat Fig. 7 Dy-Fuzzy Logic Data
The ductility and energy capacity of the wall are important for interpreting its behavior. The main
objective of the study was to analyses the conditions of the walls using intelligent techniques. The figure
8 and 9 values and have obtained in show that the study is supported by fuzzy logic. R2 values were
obtained as 0.99 to 0.97.
Fig. 8 Ductility- Fuzzy Logic Data Fig 9 Energy Capacity- Fuzzy Logic
Experiment and model values were modeled by fuzzy logic. What's more, the results are charted to
be healthy. To compare the Values, here, they are compared with the RSM technique. The results from
the RSM are given in the table of 5. In both estimation methods, results close to the real values were
obtained. When the R coefficients are compared according to the obtained data, it has been determined
that the values are very close. R values were obtained above 0.9
Page 703
SEMIT 2023
Table 5. Experimental, RSM and Fuzzy Logic Compared

Run Independent variables Dependent variables
no.a Compressive Flexure Ductilite Energy Capacity
Strenght Strenght
Brick Mortar Brick Mortar Experimental RSM Fuzzy Experimental RSM Fuzzy
1 0.3 2.12 6.38 12.5 1.21 1.50 1.53 1550 1520.8 1463.0
2 0.3 2.12 6.38 12.5 1.30 1.50 1.71 1722.03 1520.8 1577.0
3 0.3 2.12 6.38 12.5 1.85 1.50 1.76 1824.5 1520.8 1642
4 0.3 2.59 6.38 4.59 1.52 1.76 1.76 786.72 940.59 679
5 0.3 2.59 6.38 4.59 2.07 1.76 1.99 703.11 940.5 562
6 0.3 2.59 6.38 4.59 1.70 1.76 1.76 1263.54 940.5 1220
7 0.4 2.12 6.11 12.5 1.31 1.20 1.71 756.31 849.55 562.0
8 0.4 2.12 6.11 12.5 1.24 1.20 1.57 565.14 849.5 628.0
9 0.4 2.12 6.11 12.5 1.49 1.20 1.76 527.72 849.5 538.0
10 0.4 2.59 6.11 4.59 1.15 1.56 1.53 306.2 331.7 189.0
11 0.4 2.59 6.11 4.59 1.29 1.56 1.57 244.8 331.7 271.0
12 0.4 2.59 6.11 4.59 2.00 1.56 1.99 500.4 331.7 224
13 0.8 2.12 6.05 12.5 1.27 1.62 1.57 295.8 412.7 294
14 0.8 2.12 6.05 12.5 2.03 1.62 1.99 569 412.7 441
15 0.8 2.12 6.05 12.5 1.61 1.62 1.76 553.2 412.76 253
16 0.8 2.59 6.05 4.59 2.08 2.53 1.99 186.8 231.1 172
17 0.8 2.59 6.05 4.59 2.71 2.53 2.4 270.6 231.1 247
18 0.8 2.59 6.05 4.59 2.72 2.53 2.8 206.7 231.1 201
19 0.5 2.12 2.63 12.5 1.58 1.65 1.76 1127.4 1166.3 1012
20 0.5 2.12 2.63 12.5 1.31 1.65 1.71 1400.3 1166.3 1115
21 0.5 2.12 2.63 12.5 2.12 1.65 1.99 997 1166.3 1047
22 0.5 2.59 2.63 4.59 1.84 1.81 1.76 441.02 439.3 487
23 0.5 2.59 2.63 4.59 1.57 1.81 1.76 557.16 439.3 517
24 0.5 2.59 2.63 4.59 1.86 1.81 1.76 334.9 439.38 426
25 3 3 10.3 7.2 4.75 4.78 5 597 597.78 602
26 4 4 8.8 7.1 4.63 4.70 4.63 2112 2207.6 1839
27 3.4 3.4 7.5 6 10.65 10.71 9.91 1079.4 1212.8 1220
28 3.3 3.3 7.8 5.7 7.89 7.92 7.94 1020.3 999.53 833
29 4 4 8.5 7.3 9.56 9.70 9.56 264 156.30 162
30 3.3 3.3 7.9 5.2 5.28 5.39 5.55 1347 1273.0 1042
31 1.32 1.32 2.37 2.37 7.36 7.37 7.34 379.4 551.09 448
32 1.32 1.32 2.37 2.37 7.36 7.37 7.34 716.4 551.09 512
33 2 2 10.5 10.5 5.00 4.79 5.55 144 124.02 140
34 2 2 10.5 10.5 4.63 4.79 4.81 138.6 124.02 125
35 2 2 11.5 11.5 6.50 6.44 6.5 1038 1084.2 980
4. Conclusions
In this study, the researchers utilized the numerical procedure of fuzzy logic to determine the compressive
strength and tension of masonry prisms. They gathered a reliable database of published experimental
results, from which they randomly selected 35 samples for instructional purposes.
Page 704
SEMIT 2023
The remaining 28 samples were then used to test the suggested modeling. The research resulted in the
subsequent outcome, which shows that the Fuzzy Logic modeling trained by the algorithm in the hidden
layer performs well in predicting commodities. The projected significance is important in relation to
experimental consequences for both the training and testing sets in the formulated models.
The Fuzzy Logic model, which utilized bell-shaped participation functions, displayed a high level of
accuracy. Additionally, the performance indices indicated that the Fuzzy Logic model slightly
outperformed the RSM model. The comparison revealed that, on average, the empirical methods
underestimated ductility and energy capacity by around 10.
On the other hand, the predicted results obtained from the models developed in this study closely align
with experimental values. Generally, the proposed Fuzzy Logic models exhibit great applicability and
reliability in predicting the ductility and energy capacity of various masonry prisms.
The estimation of ductility's energy has been found to be more accurate compared to its capacity.
According to the study, selecting material properties allows for estimating the mechanical behaviors of
the wall.
In subsequent research, various mechanical strengths can be examined, or empirical formulas can be
created to describe the behaviors of the wall. Experimental data is open to further increases and
experimentation with various methods in smart learning techniques.
5. References
1. Furtado, A., Ramos, T., Rodrigues, H., Arêde, A., Varum, H., & Tavares, P. In-plane
Response of masonry infill walls: experimental study using digital image correlation. Procedia
Engineering, 114, 870–876. (2015).
2. Gautam, D., & Chaulagain, H. Structural performance and associated lessons to be learned from
world earthquakes in Nepal after 25 April 2015 (MW 7.8) Gorkha earthquake. Engineering
Failure Analysis, 68, 222–243. (2016).
3. Pinar, U., Nihat, M., Evci, A., & Ergün, S. Assessment of seismic damage on the exist buildings
using Fuzzy Logic. IOP Conference Series: Materials Science and Engineering, 300(1), 012062.
(2018).
4. Losanno, D., Ravichandran, N., Parisi, F., Calabrese, A., & Serino, G. Seismic performance of a
Low-Cost base isolation system for unreinforced brick Masonry buildings in developing
countries. Soil Dynamics and Earthquake Engineering, 141, 106501. (2021).
5. Zengin B., Toydemir B., Ulukaya S., Oktay D., Yüzer N. & Kocak A., Determination of
mechanical properties of blend bricks used in historical structures. 3rd International Conference
on Protection of Historical Constructions, Portugal. (2017).
6. Vasconcelos G & Lourenço PB. Experimental characterization of stone masonry in shear and
compression. Construct Build Material; 23(11):3337–45. (2009).
7. Sarhosis V. & Sheng Y. Identification of material parameters for low-bond-strength masonry.
Engineering Structures, 60,100–11. (2014).
8. Eslami A., Ronagh H.R., Mahini S.S. & Morshed R. Experimental investigation and nonlinear
analysis of historical masonry buildings – A case study. Construction and Building Materials, 35,
251–260. (2012).
9. Aktan S. & Doran B., Constitutive modeling of masonry walls under in-plane loadings. Sigma
J-Eng & Nat Sci 7 (2), 165-171. (2016).
10. Doran, B., Köksal, H. O., Aktan, S., Ulukaya, S., Oktay D. & Nabi Yüzer, In-Plane shear
behavior of traditional masonry walls. International Journal of Architectural Heritage
Conservation, Analysis, and Restoration ISSN: 1558-1558-3066. (2016)
11. Tomazevic, M., Lutman, M. & Bosıljkov, V. Robustness of hollow clay masonry units and
seismic behavior of masonry walls. Construction and Building Materials, 20 (10): 1028-1039.
(2009).
12. Onat O., Lourenco P.B. & Koçak A. Nonlinear analysis of RC structure with massive infill wall
exposed to shake table. Earthquakes and Structures, 10, 811-828. (2016).
Page 705
SEMIT 2023
13. Koçak A. & Yildirim M.K. Effects of infill wall ratio on the period of reinforced concrete framed
buildings. Advances in Structural Engineering, 14, 731-743. (2011).
14. Crisafulli, F.J., Carr, A.J. & Park, R. Analytical modeling of infilled frames structures, A general
review. Bulletin of the New Zealand Society for Earthquake Engineering, 33, (1), 30-47. (2000).
15. Koçak A. and Yildirim M.K. Effects of infill wall ratio on the period of reinforced concrete
framed buildings. Advances in Structural Engineering, 14, 731-743. (2011).
16. Crisafulli, F.J., Carr, A.J. & Park, R. Analytical modeling of infilled frames structures, A general
review. Bulletin of the New Zealand Society for Earthquake Engineering, 33, (1). 30-47. (2000).
17. Asteris, P.G., Giannopoulos, I.P. & Chrysostomou, C.Z. Modeling of infilled frames with
openings. Open Construction and Building Technology Journal, 6, 81-91. (2012).
18. Asteris P.G. & Cotsovos, D.M. Numerical investigation of the effect of infill walls on the
structural response of RC frames. Open Construction and Building Technology Journal, Bentham
Science Publishers, 6, 164-181. (2012).
19. Asteris, P.G., Antoniou, S.T., Sophianopoulos, D.S. & Chrysostomou, C.Z. Mathematical macro-
modeling of infilled frames: State of the art. Journal of Structural Engineering, 137 (12), 1508-
1517. (2011).
20. Chrysostomou, C.Z. & Asteris, P.G. On-the in-plane properties and capacities of infilled frames.
Engineering Structures, 41, 385-402. (2012).
21. Asteris, P.G., Chrysostomou, C.Z., Giannopoulos, I.P. & Smyrou, E. Document masonry infilled
reinforced concrete frames with openings. In: III ECCOMAS thematic conference on
computational methods in structural dynamics and earthquake engineering, Corfu, Greece.
(2011).
22. Tomaževi, M. Earthquake -Resistant design of masonry buildings, series on innovation in
structures and construction. Vol 1, Imperial College Press, London. (1999).
23. Sayed A., Essa A., T., Badr M. R. K & El-Zanaty A., H. Effect of infill wall on the ductility and
behavior of high-strength reinforced concrete frames. cHBRC Journal 10, 258–264. (2014).
24. Bennett, R.M., Boyd, K.A., & Flanagan, R.D., Compressive properties of structural clay tile
prisms. Journal of Structural Engineering, 123(7): 920-926. (1997).
25. Lu P, Chen S, & Zheng Y. (2012). Artificial intelligence in civil engineering. Math Probl Eng
2012.
26. Shahin MA. State-of-the-art review of some artificial intelligence applications in pile
foundations. Geosci Front;7(1):33–44. (2016).
27. Tayfur, G. Erdem T.K. & Kirca, Ö. Strength prediction of high-strength concrete by fuzzy
logic and artificial neural networks, J. Mater. Civil. Eng. 26 (11) 04014079, (2014).
28. Madandoust, R. Bungey, J.H. Ghavidel, R. Prediction of the concrete compressive strength by
means of core testing using GMDH-type neural network and ANFIS models. Comput. Mater.
Sci. 51 (1) 261–272, (2012
29. Aydin, A.C. Tortum, A. Yavuz, M. Prediction of concrete elastic modulus using adaptive neuro-
fuzzy inference system. Civ. Eng. Environ. Syst. 23 (4) 295–309, (2006).
30. Duan, Z.H. Kou, S.C. Poon, C.S. Using artificial neural networks for predicting the elastic
modulus of recycled aggregate concrete. Constr. Build. Mater. 44 524–532, (2013).
31. Ahmadi-Nedushan, B. Prediction of elastic modulus of normal and high-strength concrete using
ANFIS and optimal nonlinear regression models. Constr. Build. Mater. 36 665–673, (2012
32. Yuan, Z. Wang, L.N. & Ji, X. Prediction of concrete compressive strength: Research on hybrid
models genetic based algorithms and ANFIS, Adv. Eng. Softw. 67 156–163, (2014).
33. Duan, Z.H. Kou, S.C. Poon, C.S. Prediction of compressive strength of recycled aggregate
concrete using artificial neural networks. Constr. Build. Mater. 40 1200–1206, (2013).
34. Öztas, A. Pala, M. Özbay, E. Kanca, E. Çag˘lar, N. & Bhatti, M.A. Predicting the compressive
strength and slump of high strength concrete using neural network. Constr. Build. Mater. 20 (9)
769–775, (2006).
35. Bal, L. Buyle-Bodin, F. Artificial neural network for predicting drying shrinkage of concrete.
Constr. Build. Mater. 38 248–254, (2013).
36. Parichatprecha, R. Nimityongskul, P. Analysis of durability of high-performance concrete using
artificial neural networks. Constr. Build. Mater. 23 (2) 910–917, (2009).
Page 706
SEMIT 2023
37. Zhang, Y. Zhou, G. Xiong, Y. & Rafiq, M. Techniques for predicting cracking pattern of
masonry wallet using artificial neural networks and cellular automata. J. Comput. Civ. Eng. 24
(2) 161–172, (2010).
38. Garzón-Roca, J. Marco, C.O. Adam, J.M. Compressive strength of masonry made of clay bricks
and cement mortar: Estimation based on neural networks and fuzzy logic. Eng. Struct. 48 21–27,
(2013).
39. Garzón-Roca, J. Adam, J.M. Sandoval, C. & Roca, P. Estimation of the axial behaviour of
masonry walls based on Artificial Neural Networks, Comput. Struct. 125 145–152, (2013).
40. Plevris, V. & Asteris, P.G. Modeling of masonry failure surface under biaxial compressive stress
using neural networks, Constr. Build. Mater. 55 447–461, (2014).
41. Asteris P G & Plevris V. Anisotropic masonry failure criterion using artificial neural networks.
Neural Comput Appl; 28:2207–2229. (2016).
42. Masonry Standards Joint Committee (MSJC) of the Masonry Society. Building code
requirements and specification for masonry structures: containing building code requirements for
masonry structures (TMS 402–11/ACI 530–11/ASCE 5–11), specification for masonry structures
(TMS 602–11/ACI 530.1- 11/ASCE 6–11), and companion commentaries. Reston (VA):
American Society of Civil Engineers, p. 236. (2011).
43. Canadian Standards Association (CSA), CSA S304.1-04 (R2010) – Design of Masonry
Structures, Canadian Standards Association, Toronto (Ontario), p. 64, 2004.
44. British Standards Institution (BSI), BS EN 1996 (Eurocode 6): Design of Masonry Structures,
British Standards Institution, p. 128, (2005).
45. Sarhat, S.R. Sherwood, E.G. The prediction of compressive strength of ungrouped hollow
concrete block masonry. Constr. Build. Mater. 58 111–121, (2014).
46. Singhal V. &Rai D. Behaviour of Confined Masonry Walls with Openings under In-Plane and
Out-of-Plane Loads. Earthquake Spectra, Volume 34, No. 2, pages 817–841 (2018).
47. Bui T-T., Limam A.& Sarhosis V. Failure analysis of masonry wall panels subjected to in-plane
and out-of-plane loading using the discrete element method. European Journal of Environmental
and Civil Engineering (2019)
48. Abdulla K. F. Cunningham L S, Gillie M., Simulating masonry wall behaviour using a simplified
micro-model approach. Engineering Structures 151 349–365, (2017).
49. Zengin B. Investigating elastic-plastic behaviour of masonry walls by experimental and
numerical methods, Department of Civil Engineering, Ph.D. Thesis, Turkey. (2018)
50. Zadeh LA. Fuzzy sets. Inform Control 1965; 8:338–53. (1965).
51. Matlab 7.0. User’s Guide. The MathWorks, Inc., Natick, MA, (2004).
52. Takagi T., Sugeno M. Fuzzy identification of systems and its application to modeling and control.
IEEE Transactions on Systems, Man and Cybernetics. Pp.116-132. (1985).
Page 707
SEMIT 2023
SP19: Artificial Intelligence Innovations
Deep Learning and Probe Data for Urban Traffic Flow Prediction
Serap Ergün1[0000-0003-2504-5101]
1
Isparta University of Applied Sciences, Department of Computer Engineering, 32260, Turkey
serapbakioglu@isparta.edu.tr
Abstract
This paper explores an innovative approach to predicting traffic flow in urban areas using deep learning
and probe information. With advancements in Intelligent Transportation Systems, the integration of probe
data provides a richer source of real-time information. The study focuses on developing a deep learning
model for accurate and real-time traffic flow predictions, comparing it with traditional methods. The goal
is to enhance prediction accuracy and contribute to more efficient intelligent transportation systems.
Keywords: Traffic flow prediction, Intelligent Transportation Systems, Probe information, Deep learning
Introduction
Urban areas face escalating traffic congestion challenges, necessitating effective traffic management for
commuter convenience, safety, and environmental sustainability. The synergy of Deep Learning and
Probe Information holds promise in addressing these challenges by improving traffic flow prediction.
Intelligent Transport Systems advancements enable seamless data collection from various sources,
including probes on vehicles, offering real-time insights into traffic conditions. (Tao et al., 2020; Sun et
al., 2020).
Urban Traffic Challenges:
Managing urban traffic complexities, influenced by factors like time, weather, events, and infrastructure,
is a critical issue. Traditional systems relying on fixed sensors struggle to adapt to dynamic conditions,
resulting in suboptimal traffic management and increased pollution (Duan et al., 2018).
Deep Learning's Potential:
Deep Learning, a subset of machine learning, autonomously learns intricate patterns from large datasets.
In urban traffic, it can predict flow dynamics based on diverse factors, including historical data and real-
time probe information (Tran et al., 2020; Perez-Murueta et al., 2019).
Harnessing Probe Information:
Probe data from GPS-equipped vehicles provide real-time insights into traffic conditions. Integrating this
data into Deep Learning models enhances the accuracy of traffic flow predictions, enabling informed
decisions for commuters and dynamic traffic management for authorities (Essien et al., 2021; Lv et al.,
2014).
Benefits and Outcomes:
Enhancing traffic flow prediction benefits commuters with accurate information, aids urban planners in
dynamic traffic management, and contributes to environmental sustainability by reducing congestion and
emissions (Koesdwiady et al., 2016).
Page 708
SEMIT 2023
Key Research Areas and Contributions

1. Harnessing Probe Data: The research capitalizes on the potential of probe data, a valuable resource for
gaining insights into traffic patterns and dynamics. By integrating probe data into the prediction model,
it is aimed to improve the accuracy of traffic flow forecasts.
2. Deep Learning Advancements: Deep learning has emerged as a powerful tool in the field of machine
learning. In this study, its application in the context of traffic flow prediction is explored. Deep learning's
ability to autonomously learn feature representations offers promise in achieving more precise predictions
without the need for prior site-specific knowledge.
3. Comparative Analysis: A comparative analysis with conventional machine learning techniques,
shedding light on the effectiveness of the deep learning-based approach are conducted. This comparative
assessment provides valuable insights into the advantages and limitations of various prediction methods.
Conclusion and Future Prospects

The research contributes insights into traffic flow prediction, advocating for refining models by exploring
temporal granularity, optimizing input data configurations, and addressing data diversity challenges.
Despite progress, ongoing research is crucial for advancing urban mobility solutions and creating smarter,
more efficient transportation systems.
References
Duan, Z., Yang, Y., Zhang, K., Ni, Y., & Bajgain, S. (2018). Improved deep hybrid networks for urban
traffic flow prediction using trajectory data. Ieee Access, 6, 31820-31827.
Essien, A., Petrounias, I., Sampaio, P., & Sampaio, S. (2021). A deep-learning model for urban traffic
flow prediction with traffic events mined from twitter. World Wide Web, 24(4), 1345-1368.
Koesdwiady, A., Soua, R., & Karray, F. (2016). Improving traffic flow prediction with weather
information in connected cars: A deep learning approach. IEEE Transactions on Vehicular
Technology, 65(12), 9508-9517.
Lv, Y., Duan, Y., Kang, W., Li, Z., & Wang, F. Y. (2014). Traffic flow prediction with big data: A deep
learning approach. IEEE Transactions on Intelligent Transportation Systems, 16(2), 865-873.
Perez-Murueta, P., Gómez-Espinosa, A., Cardenas, C., & Gonzalez-Mendoza Jr, M. (2019). Deep
learning system for vehicular re-routing and congestion avoidance. Applied Sciences, 9(13),
2717.
Saleem, M., Abbas, S., Ghazal, T. M., Khan, M. A., Sahawneh, N., & Ahmad, M. (2022). Smart cities:
Fusion-based intelligent traffic congestion control system for vehicular networks using machine
learning techniques. Egyptian Informatics Journal, 23(3), 417-426.
Sun, S., Wu, H., & Xiang, L. (2020). City-wide traffic flow forecasting using a deep convolutional neural
network. Sensors, 20(2), 421.
Tao, Y., Sun, P., & Boukerche, A. (2020, June). A delay-based deep learning approach for urban traffic
volume prediction. In ICC 2020-2020 IEEE International Conference on Communications
(ICC) (pp. 1-6). IEEE.
Tran, L., Mun, M. Y., Lim, M., Yamato, J., Huh, N., & Shahabi, C. (2020). DeepTRANS: a deep learning
system for public bus travel time estimation using traffic forecasting. Proceedings of the VLDB
Endowment, 13(12), 2957-2960.
Page 709
SEMIT 2023
Multi-Agent Deep Q-Network for Optimal Signal Control in Urban Intersections

Serap Ergün1[0000-0003-2504-5101]
1
Isparta University of Applied Sciences, Department of Computer Engineering, 32260, Turkey
serapbakioglu@isparta.edu.tr
Abstract
This research addresses escalating urban traffic congestion by proposing a multi-agent system integrated
with Deep Q-Network (DQN) for optimal signal control in multi-intersection road environments.
Traditional signal control methods rely on manual estimations, leading to suboptimal outcomes. The
study leverages DQN, a reinforcement learning technique, to adapt signal timings based on real-time
traffic data. The integration of a multi-agent system facilitates coordination between intersections,
enhancing overall traffic flow. The findings provide valuable insights for effectively tackling traffic
congestion.
Keywords: Intelligent transport systems, Traffic congestion and control, Signal switching, Multi-agent
system.
Introduction
Urban areas face increasing time and economic losses due to traffic congestion, necessitating innovative
solutions (Hu et al., 2023). This study focuses on optimizing traffic flow through signal control,
particularly at intersections, crucial points in urban road networks. Improper signal switching contributes
to congestion, making the optimization of traffic signal control paramount. Manual parameter
adjustments based on human experience have limitations in adapting swiftly to changing traffic patterns.
To overcome these challenges, the study proposes a cutting-edge control system using a multi-agent
system with DQN.
Intersections are crucial points in urban road networks, serving as crossroads where multiple streams of
traffic converge. To ensure safety and regulate vehicle flow, traffic lights are commonly employed at
these intersections. However, improper signal switching can exacerbate congestion issues. For instance,
overly long red-light durations or the absence of adaptive signal timing can lead to extended wait times
for drivers and result in traffic gridlock. Conversely, the slow pace of traffic signal changes can lead to
traffic jams, especially during right turns, where minor delays can accumulate to create significant
bottlenecks. Therefore, the optimization of traffic signal control is of paramount importance in managing
urban traffic congestion (Ge et al., 2021; Haddad et al., 2022; Kim and Jeang, 2019).
Presently, signal control parameter adjustments are often based on manual estimations rooted in human
experience. Traffic engineers typically rely on historical data, visual observations, and their expertise to
fine-tune signal timings. However, this manual approach has limitations, as it cannot adapt swiftly to
changing traffic patterns, unexpected events, or variations in demand throughout the day. Moreover, it
can be challenging to optimize signal timings for multiple intersections that are interconnected and
influence each other's traffic flow.
To address these challenges and alleviate traffic congestion effectively, this study proposes a cutting-
edge control system. This system combines the Deep Q-Network (DQN) method introduced by Joo and
Lim (2021) with a multi-agent system to facilitate appropriate parameter adjustments in a multi-
intersection road environment. The integration of DQN, a powerful reinforcement learning technique,
enables the traffic control system to learn and adapt its decisions over time based on real-time traffic data.
DQNs are a type of artificial neural network that have demonstrated remarkable success in solving
complex decision-making tasks. In the context of traffic signal control, a DQN-based system can
continuously analyze incoming traffic data, such as vehicle counts, speeds, and congestion levels, to make
intelligent decisions about signal timings (Zhang et al., 2023). By learning from historical and real-time
data, the system can dynamically adjust traffic light parameters, such as green-light durations, cycle
lengths, and phase sequences, to optimize traffic flow (Ge et al., 2021).
Page 710
SEMIT 2023
Furthermore, the integration of a multi-agent system allows for coordination and communication between
traffic signal controllers at different intersections. This coordination is essential to avoid conflicts and
ensure smooth traffic flow throughout the entire network. For instance, when one intersection detects a
surge in traffic demand, it can communicate with neighboring intersections to adapt their signal timings
accordingly, preventing congestion from propagating through the road network.
In conclusion, the study's proposed control system represents a significant step toward addressing the
escalating issue of urban traffic congestion. By leveraging advanced techniques such as Deep Q-
Networks and multi-agent systems, this approach promises to revolutionize traffic signal control, making
it more adaptive, efficient, and responsive to the complex dynamics of urban traffic. As cities continue to
grow and traffic volumes increase, innovative solutions like this are vital to enhancing urban mobility,
reducing travel times, and improving the overall quality of life for urban residents.
Conclusion
The proposed control system represents a significant step in addressing urban traffic congestion.
Leveraging advanced techniques like DQN and multi-agent systems makes signal control more adaptive,
efficient, and responsive to urban traffic dynamics. Future research directions include exploring complex
road scenarios, sensor-enabled signal control, and the integration of emerging technologies. Collaboration
with urban planners and policymakers is emphasized for effective implementation, considering
environmental impacts and sustainability goals.
References
Ge, H., Gao, D., Sun, L., Hou, Y., Yu, C., Wang, Y., & Tan, G. (2021). Multi-agent transfer reinforcement
learning with multi-view encoder for adaptive traf-fic signal control. IEEE Transactions on
Intelligent Transportation Sys-tems, 23(8), 12572-12587.
Haddad, T. A., Hedjazi, D., & Aouag, S. (2022). A deep reinforcement learning-based cooperative
approach for multi-intersection traffic signal con-trol. Engineering Applications of Artificial
Intelligence, 114, 105019.
Hu, T., Hu, Z., Lu, Z., & Wen, X. (2023). Dynamic traffic signal control using mean field multi‐agent
reinforcement learning in large scale road‐networks. IET Intelligent Transport Systems.
Joo, H., & Lim, Y. (2021, October). Intelligent Traffic Signal Control System using Deep Q-network. In
2021 IEEE 3rd Eurasia conference on IOT, Communi-cation and Engineering (ECICE) (pp. 285-
287). IEEE.
Kim, D., & Jeong, O. (2019). Cooperative traffic signal control with traffic flow prediction in multi-
intersection. Sensors, 20(1), 137.
Zhang, Y., Wang, S., & Jiang, R. (2023). Large-Scale Traffic Signal Control by a Nash Deep Q-network
Approach. arXiv preprint arXiv:2301.00637.
Page 711
SEMIT 2023
SP20: Artificial Intelligence: The Future of Everything
Energy Consumption Forecasting using Time Series Analysis Methods

Fatma Yildirim1 and Hilal Arslan2
1
2
Abstract
Energy is an indispensable element of societies. Technology has been developing based on
population growth and people's expectations, and as a result, the need for energy is increasing day
by day. With the increasing need, the dependence on energy reaches high levels and it is predicted
that it will reach even higher levels. With the developments in the field of industrialization and the
increase in population, energy consumption can be reduced by measures that can be taken within
the framework of habits. In this study, electricity consumption estimation was achieved with time
series methods.
Keywords: time series, ARIMA.
1. Introduction
The usage of renewable energy sources is extremely important socially and economically. They are
constantly found in nature and are sensitive to the environment. Therefore, developing countries take
important steps to produce renewable energy sources. Turkey aims to reduce energy consumption via
increasing energy efficiency without reducing the production quality as well as reducing the living
standard. In this study, small measures that can be taken by using the energy data consumed by electrical
household appliances used in a residence in district of Ankara and the effect of energy consumption in
the next periods are estimated with time series algorithms.
Osman et al. [1] carried out energy efficiency improvement studies using 42 percent of randomly selected
houses for the intervention. Energy protection endeavor has been made to improve energy efficiency in
UK households. As a result of the estimations, if the efforts to save a one fourth of the energy currently
used in the UK households (reducing technology costs, making energy efficiency studies in a wider area)
are continued, 50 percent energy efficiency will be achieved by 2035 [2]. In the study [3], Internet of
Things (IoT) focused smart Home Energy Management System (HEMS) studies were carried out to
identify electrical house-hold appliances based on the crossbred Unsupervised Automatic Clustering-
Integrated Neuro-Fuzzy Classification (UACNFC) model. In the study [4], it is aimed to decrease the
effect of energy consumption on the environment as far as feasible. Meeting energy efficiency targets
encompasses the work of investments that require almost all residences to improve their performance [5].
In the study [6], studies were carried out on energy efficiency measures on detached houses with higher
energy density. To forecast the energy and spreading decrease potential in Turkey, the dataset was
forecasted based on the single-family housing stock from single-family residences in Turkey’s three
largest cities, Ankara, Istanbul and Izmir. Studies have been carried out on energy use and emission
effects by switching from natural gas to electrical appliances in Australian homes. They found that
electricity changes led to large decreases in residential natural gas consumption and smaller increases in
electricity consumption from the grid in terms of energy content [7]. Jo and Yoon [8] suggested three
smart models as IoT platform application labor implementations for a smart home to manage IoT
Page 712
SEMIT 2023
efficiently. In the study [9] also broadens the application areas of bootstrap aggregation and forecasting
methods to the electric power industry in order to get more accurate requisition estimates.
3. Method
Time series estimating is a method for predicting incidents from start to finish a time series. They predict
forthcoming events by analyzing past propensities with the hypothesis that future propensities will be
parallel to historical propensities. They are efficiently used in weather estimating, energy forecasting and
engineering areas. They are also used to eliminate trends and seasonality is to take the difference of the
surveillance at a certain moment with the previous moment. A correlation function is obtained between
two time steps in a series in such a way that it is related to other time steps. Today's data depends on
yesterday's and yesterday's data depends on the next day. In this study, we performed a time series
method, ARIMA that achieves 0.315 RMSE value.
4. Conclusion
This study details the estimates of total energy consumption in a home. The ARIMA model is one of the
models that can be used at this point. Predictions were made after the necessary work on the ARIMA
model was carried out. Time series forecasting studies for the use planning of the future role of energy
are at the beginning of the studies to be carried out at this point.
References
1. OSMAN, Liesl M., et al. A randomised trial of home energy efficiency improvement in the homes
of elderly COPD patients. European Respiratory Journal, 2010, 35.2: 303-309.
2. ROSENOW, Jan, et al. The remaining potential for energy savings in UK households. Energy Policy,
2018, 121: 542-552.
3. LIN, Yu-Hsiu. Design and implementation of an IoT-oriented energy management system based on
non-intrusive and self organizing neurofuzzy classification as an electrical energy audit in smart
homes. Applied sciences, 2018, 8.12: 2337.
4. BENAVENTE-PECES, César. On the energy efficiency in the next generation of smart buildings—
Supporting technologies and techniques. Energies, 2019, 12.22: 4399.
5. BONE, Angie, et al. Will drivers for home energy efficiency harm occupant health? Perspectives in
public health, 2010, 130.5: 233-238.
6. GUGUL, Gul Nihal; KOKSAL, Merih Aydinalp; UGURSAL, V. Ismet. Techno-economical analysis
of building envelope and renewable energy technology retrofits to single family homes. Energy for
Sustainable Development, 2018, 45: 159-170.
7. HAMMERLE, Mara; BURKE, Paul J. From natural gas to electric appliances: Energy use and
emissions implications in Australian homes. Energy Economics, 2022, 110: 106050.
8. JO, Hana; YOON, Yong Ik. Intelligent smart home energy efficiency model using artificial
TensorFlow engine. Human centric Computing and Information Sciences, 2018, 8.1: 1-18.
9. DE OLIVEIRA, Erick Meira; OLIVEIRA, Fernando Luiz Cyrino. Forecasting mid-long term electric
energy consumption through bagging ARIMA and exponential smoothing methods. Energy, 2018,
144: 776-788.
Page 713
SEMIT 2023
Recognition
Allam Fatima Zohra
Avenue des Frères Oudek, El Harrach, 16200 Algiers, Algeria- ORCID: 0000-0002-2273-8658
fatima_zohra.allam@g.enp.edu.dz
Hamami Mitiche Latifa
latifa.hamami@g.enp.edu.dz
Bousbia-Salah Hicham
hicham.bousbia_salah@g.enp.edu.dz
Abstract
Facial recognition is an image processing technology that identifies or verifies the identity of an
individual based on their unique facial features. It uses algorithms to analyze features such as eye
distance, face shape, eye, nose, and mouth contours to create a digital signature specific to each person.
Our work focuses on the implementation and comparison of two descriptors, widely developed in the
literature, in this case, the Local Binary Patterns (LBP) and the Hough Transform. The objective is to
evaluate their respective performance in facial recognition in terms of accuracy, robustness to lighting
variations and processing speed.
The two descriptors are implemented and applied on a specific database to extract relevant facial
features. The results obtained made it possible to determine the advantages and limits of each descriptor
in the field of facial recognition. We compared the performance of these two descriptors in terms of
precision, calculation speed and robustness.
Keywords: Facial Recognition, Local Binary Pattern, Hough Transform.
Introduction
Biometric facial identification and facial recognition are rapidly expanding areas of research and
application, offering exciting possibilities in many industries. Facial biometrics exploits the unique
characteristics of the face for the identification and verification of individuals. Facial recognition is a
specific application of facial biometrics, which aims to identify and authenticate a person by comparing
their face with data stored in a database. We propose to apply two methods, commonly used in various
fields, and widely developed in the literature, in this case, the Hough Transform and the Local Binary
Patterns (LBP).
Hough's Transform is a computer vision technique that allows the detection of specific geometric shapes
in an image. It is often used to extract geometric features from the face, such as straight lines or circles.
Hough's method is appreciated for its accuracy in detecting these shapes, even in complex or noisy
images.
On the other hand, Local Binary Patterns are used to describe local intensity variations in an image. They
are commonly used in facial recognition to capture facial textural information. LBP compares the
intensity levels of neighboring pixels with respect to a central pixel and generates binary patterns to
represent these local variations. This method is appreciated for its simplicity of calculation and its
robustness to lighting variations.
Page 714
SEMIT 2023
The comparison of these two methods is essential to evaluate their respective performances. Criteria, such
as recognition accuracy, processing speed, robustness to lighting variations, ability to manage large data
sets or resource consumption, are often considered for the comparison.
In this article, we will implement and compare the Hough Transform and Local Binary Patterns for face
recognition. We will evaluate their respective performance using a diverse dataset and appropriate metrics
such as recognition accuracy, robustness to lighting variations, and processing speed. This comparison
will allow us to understand the advantages and limitations of each method and to determine which is more
suitable for practical face recognition applications.
In this article, we present, in section 2, some notions on biometrics, in which we approach the main steps
of a facial recognition system and the most commonly developed techniques. Thereafter, we explain, in
section 3, the two descriptors that we applied for the implementation of our recognition system, in this
case, the Hough Transform and the LBP. Section 4 includes the implementation phases of each algorithm,
the results obtained as well as the comparison between these two descriptors in terms of robustness,
computation time and in terms of reliability and precision.
Facial Biometric
Principle
Facial biometrics is a technology that uses the unique physical characteristics of a person's face to reliably
identify them. It relies on the use of image acquisition techniques and facial recognition algorithms to
extract distinctive features, such as face shape, eye, nose and mouth proportions, facial contours, folds,
and wrinkles. This information is processed to create a biometric template which can be compared to a
database of registered facial templates, in order to verify or identify a specific person [1][2]. Figure 1
illustrates the general principle of a facial recognition system.
Facial recognition finds applications in various fields such as the security of protected environments,
marketing, monitoring and human resource management. It is also used in airports to verify the identity
of passengers, in security systems to control access to buildings, in criminal investigations to identify
suspects, in video surveillance, unlocking devices for smartphones, etc.
Figure 1. Block diagram of a Facial Recognition System
However, facial recognition raises concern about privacy and security. Indeed, it is important to consider
privacy and data protection issues, as biometric information, once collected, can be sensitive and requires
adequate security measures to preserve it. It should be noted that some people are concerned about the
collection and use of their biometric data, as well as the risk of misuse of this technology for surveillance
purposes.
Steps of facial recognition
The main steps of facial recognition are summarized as follows [1][3][4]:
1. Acquisition: The image of the face is acquired using a camera or other imaging device.
2. Preprocessing: The image is processed and normalized to remove distortions and unwanted noise.
Page 715
SEMIT 2023
3. Face detection: The system uses face detection algorithms to detect the position and contours of
the face in the image.
4. Feature extraction: Facial features such as distance between eyes, size and shape of eyes, nose and
mouth are extracted from the image.
5. Comparison: The extracted face features are compared with those stored in a database of known
faces to find a match.
6. Decision making: Based on the match found, the system decides on who the person is.
In some applications, there may be additional steps, such as machine learning to improve recognition
performance and verifying the identity of the person using biometric methods [1][3].
Main facial recognition techniques
Face recognition techniques can be divided into three approaches [4][5][6]:
• Global approaches.
• Local approaches.
• Hybrid approaches.
The main methods of these three approaches are described in the following. The diagram summarizes the
classification of the main algorithms used in facial recognition.
Figure 2 represents facial recognition techniques.
Figure 2. Facial Recognition Techniques
1. Global approaches
Facial recognition methods use commonly employed statistical analysis techniques, thus avoiding the
need to identify characteristic points of the face such as the centers of the eyes, the nostrils, and the center
of the mouth. These images are processed as a whole and are often converted into vectors [3][5][6].
These methods use a face subspace analysis technique, which is based on the principle that a class of
shapes is in a subspace of the input image space. The representation of the original image is very
redundant, and its dimensionality can be reduced by considering only relevant attributes.
The main algorithms applied in the global approach are:
• Principal Component Analysis (PCA)
• Linear Discriminant Analysis (LDA)
• Independent Component Analysis (ICA)
2. Local approach
Local techniques, based on models, exploit prior information on facial morphology and are based on
characteristic points of the face. Local methods include [4][7][8]:
• Active Appearance Models (AAM).
• Support Vector Machine (SVM).
Page 716
SEMIT 2023
• Local Binary Patterns (LBP).

3. Hybrid approaches
Hybrid facial recognition approaches combine both global and local features to improve performance.
These characteristics have distinct properties, and their complementarity can be exploited to obtain a
better classification. The goal is to take advantage of the advantages of each approach in order to
compensate for the limitations of the other.
However, there remains a challenge to effectively combine these two types of features in order to improve
facial recognition, and there are few works specifically dedicated to this application.
Application of the LBP algorithm and the Hough Transform to the Facial Recognition
Our approach consists in using images of faces of different people for our facial recognition system, then
assessing the performance and evaluating the effectiveness of our method by verifying the identity of
each individual. We will also discuss the limitations and advantages of this approach, as well as future
directions for research in this area.
Face Recognition System Architect
Figure 3 illustrates the architecture of the facial recognition system. We will explain the main blocks of
the system.
Figure 3. Face Recognition System Architect
1. Perceptual Channels Decomposition

To decompose the image into channels, we use the Wavelet Transform, an image processing technique
that divides the image into different frequency domains.
Wavelet Decomposition consists of decomposing the image into several frequency bands, each providing
specific information about the image. To do this, we use the “dwt2” function of the Matlab software
which performs the Wavelet Decomposition of the “img” image using the “db6” wavelet family
(Daubechies 6) with four domains.
The four domains correspond to the Wavelet Decomposition coefficients of the image, each having a
different angular selectivity. The decomposition coefficients are divided into four matrices:
• LL1: Matrix of approximation coefficients (Low-Low) which contains the low frequency
information of the image. This matrix represents the image at a lower scale with reduced resolution.
• HL1: Matrix of horizontal detail coefficients (High-Low) which contains the high frequency
information in the horizontal direction of the image.
• LH1: Matrix of vertical detail coefficients (Low-High) which contains the high frequency
information in the vertical direction of the image.
Page 717
SEMIT 2023
• HH1: Matrix of diagonal detail coefficients (High-High) which contains the high frequency
information in the diagonal direction of the image.
2. Feature extraction
After dividing the image into four domains with distinct levels of selectivity, we used a technique to
extract features using varied feature extraction cell sizes (Figure 4).
Figure 4. Feature extraction using different cell sizes.
3. Feature vector
After extracting the features, we will apply the method of Principal Component Analysis (PCA) to
select the relevant features and reduce the dimension of the feature vectors.
4. Classifier
The classifier plays a crucial role in learning, being an algorithm or model that learns to classify or predict
data based on features or input variables.
In our application, we used the MSVM (Multi Support Vector Machine) classifier to perform data
classification. The MSVM is a classifier that aims to find the optimal hyperplane to separate the different
classes in the feature space.
Database
We used the Yale B Extensive Face Database. This is a collection of face images widely used in facial
recognition research. It includes images from approximately 133 subjects, with 20 images per subject.
Images are captured under different pose angles, various facial expressions, and varying lighting
conditions, allowing the effectiveness and performance of facial recognition algorithms to be tested in
varied, real-world scenarios.
In our case, to test our application, we considered a database that contains 10 classes. We have divided
this database into two distinct parts:
1. The training base is used to train the system. It leverages part of the database to allow the system
to learn patterns and relationships between features and classes.
2. The test base is used to evaluate system performance. It is composed of the remaining classes in
the database and is used to test the system against new data not trained by the system.
Implementation, Comparison and Result

In this comparative study, we analyze the performance of two techniques widely used in the field of object
detection:
Page 718
SEMIT 2023
• The Hough Transform [11].

• Local binary patterns (LBP).
The Hough Transform is known for its high accuracy in detecting geometric shapes, while LBP is
appreciated for its computational simplicity and robustness to lighting variations.
Implementation with Hough Transform
The Hough Transform is a powerful technique used in the field of computer vision and image processing
to detect geometric shapes, such as lines or circles, in an image [11].
The basic idea of the Hough Transform is to represent geometric shapes in a separate parameter space
called "Hough Space". Instead of directly detecting the pixels corresponding to a specific shape in the
image, the Hough Transform makes it possible to represent these shapes as points in this parameter space.
Let us take the example of line detection. In “Hough Space,” each point represents a potential line in the
image. For each pixel in the original image, the Hough Transform calculates all possible line parameters
that pass through that pixel and increments the corresponding coordinates in “Hough Space”. Thus, the
more a line is present in the image, the closer the corresponding points in Hough space will be to each
other.
Once all potential rows have been accumulated in “Hough Space,” a thresholding step is often performed
to detect the most significant rows. Local maxima in “Hough Space” are identified, which correspond to
the most dominant lines in the image.
Once the parameters of the detected lines have been obtained from the “Hough Space,” it is possible to
reconstruct them in the original image to visualize the results of the detection.
The Hough Transform can be extended to detect other geometric shapes such as circles or ellipses by
adjusting the parameter space and vote accumulation.
Implementation with Local Binary Pattern
It is a technique commonly used in the field of computer vision for the description and recognition of
local patterns in an image. It is a simple and efficient method to extract local features from an image. It
works by comparing the intensity of neighboring pixels to a central pixel in a given window. This
comparison generates a unique binary pattern that represents the local intensity variations in that region
of the image.
The LBP process consists of three main steps:
• Selection of a central pixel: For each pixel of the image, a central pixel is chosen.
• Comparison of neighboring pixels: The neighboring pixels of the central pixel are compared to the
latter. If the neighbor has an intensity greater than or equal to that of the central pixel, it is coded
as "1". Otherwise, it is encoded as "0".
• Generation of the LBP pattern: The results of comparisons of neighboring pixels are concatenated
to form a Local Binary Pattern.
Once the LBP patterns are extracted from the image, they can be used for pattern recognition, image
classification, object detection, etc.
Results and evaluation
To assess the effectiveness and reliability of our facial recognition system, we must first create a
validation base. Next, we will check if our system is able to identify face images presented to it as input
that have already been used to train the system through an SVM classifier.
We have structured the database as follows:
• 10% for validation.
• 20% For testing.
• 70% For learning.
Page 719
SEMIT 2023
To properly visualize the confusion matrix, we use 10 classes of 20 images each.

The figures (from 5 to 10) illustrate the results obtained for the two descriptors, respectively, Hough
Transform and LBP. The results show the classifier predictions, confusion matrices, and accuracy rates
of each algorithm.
Figure 5. Classifier predictions for 10 classes (case of Hough Transform)
Figure 6. Confusion Matrix for 10 classes (case of Hough Transform)
Figure 7. Result of the implementation: Accuracy in % (case of Hough Transform)
Page 720
SEMIT 2023
Figure 8. Classifier predictions for 10 classes (case of LBP)
Figure 9. Confusion Matrix for 10 classes (case of LBP)
Figure 10. Result of the implementation: Accuracy in % (case of LBP)
We find that, for both approaches, our system correctly identifies all the people presented as inputs, with
an accuracy of 100% in the case of the Hough Transform and 95% in the case of LBP. These results
demonstrate the efficiency and reliability of the system produced. The latter is able to recognize each
individual that exists in the database on which it has been trained with great precision.
Comparison
In this comparative study, we analyze the performance of two techniques widely used in the field of object
detection: The Hough Transform and Local Binary Patterns (LBP). The Hough Transform is known for
its high accuracy in detecting geometric shapes, while LBP is appreciated for its computational simplicity
and robustness to lighting variations.
In terms of accuracy, the Hough Transform stands out by accurately capturing the geometric shapes
present in an image. It is particularly effective for detecting lines, circles, or ellipses. In contrast, LBP
focuses more on describing local variations in intensity, which makes it less accurate in detecting specific
geometric shapes, but more suitable for detecting textural patterns or local structures.
Page 721
SEMIT 2023
When it comes to computational speed, LBP is advantageous due to its simplicity. Local bit pattern
calculations can be performed quickly, making it an efficient method for real-time applications. The
Hough Transform, on the other hand, can be more computationally expensive due to its vote accumulation
process in the “Hough Space.”
In terms of robustness to lighting variations, LBP has a significant advantage. By using local thresholds,
it is able to capture texture information independent of global intensity variations. The Hough Transform,
on the other hand, can be more sensitive to lighting variations, which can affect its detection accuracy.
Conclusion
Facial recognition is a powerful and promising technology that uses the unique characteristics of the
human face for the identification and verification of individuals.
In this article, we are interested in developing a robust and efficient biometric system using the face
modality. We proposed to apply two algorithms, the first one is based on the Hough Transform, and the
second one on LBP Local Binary Patterns.
In our approach, we conducted, beforehand, a Perceptual Channels Decomposition. We applied a wavelet
transform for different frequency scales. Then we extracted the features using the Histogram of Oriented
Gradients (HOG) method and the Hough Transform. This step allowed us to represent each image by a
vector of significant features. To reduce the dimensionality of the data and improve the efficiency of the
model, we applied Principal Component Analysis (PCA). This technique allowed us to select the most
discriminating components and to use them to represent the faces in a space of reduced dimension.
Finally, we used a classifier to perform face prediction. Different classification methods exist in the
literature. We used Multi Support Vector Machines (MSVM).
In the last step, we compared the two developed methods. We found that the Hough Transform is an
efficient technique for the detection of specific geometric shapes, while the Local Binary Patterns (LBP)
are useful for capturing the textural information of the face.
Their implementation in facial recognition systems can offer significant performance in terms of
accuracy: The Hough Transform gives an accuracy rate equal to 100% and that of LBP around 95%.
At the end of this work, we can conclude that the choice between the two approaches will depend on the
specific requirements of the application. If geometric accuracy is paramount, the Hough Transform will
be preferred, while LBP will be more appropriate in the case of textural pattern detection or in the case
of real-time applications where computational speed and robustness to variations in lighting are essential.
References
[1] A. Abaza, A. Ross, CH. Hebert, M.A.F. Harrison, and M.S. Nixon, “A survey on ear biometrics”,
ACM Embedded Computing Systems, vol. 45, no. 2, article 22, pp. 1-35, February 2013.
[2] S. Guerfi, “Authentication of individuals by recognition of biometric characteristics linked to 2D/3D
faces”, PhD Thesis at the University of Evry-Val d'Essonne, France, 2008.
[3] N. Morizet, “Biometric recognition by multimodal fusion of Face and Iris”, PhD Thesis, Specialty:
Signal and Images. National School of Telecommunications. Paris, France, March 2009.
[4] C. Fiche, “Pushing the limits of facial identification in the context of video surveillance”, PhD Thesis
from the University of Grenoble, Speciality: Signal - Images - Speech - Telecoms (SIPT), January
2012.
Page 722
SEMIT 2023
[5] B. Chu, “Neutralization of facial expressions to improve face recognition”, Thesis of the University
of Lyon, Ecole Centrale de Lyon, specialty: Computer Science, March 2015
[6] A. Sharma and K. K. Paliwal, « Linear Discriminant Analysis for the small sample size problem: an
overview », International Journal of Machine Learning and Cybernetics, vol. 6, no. 3, June 2014.
[7] M. Tafar and S. Miguet, “Local Appearance Modeling for Object Class Recognition”, Pattern
Analysis and Applications, vol. 22, pp. 439–455, 2019.
[8] J.-Y. Zhang, H.-P. Zhao and S. Chen, “Face recognition based on weighted Local Binary Pattern
with adaptive threshold”, Journal of Electronics and Information Technology, vol. 36, no. 6, pp. 1327-
1333, June 2014.
[9] E. Owusu, Y. Zhan and Q. R. Mao, “A Facial Expression Recognition System SVM-AdaBoost”,
Applied Intelligence,, vol. 40 , pp. 536–545, 2014.
[10] J. Soldera, C.A. Ramirez Behaine, J. Scharcanski, "Customized Orthogonal Locality Preserving
Projections With Soft-Margin Maximization for Face Recognition", Instrumentation and
Measurement, IEEE Transactions, vol. 64, no. 9, pp.2417-2426, September 2015.
[11] C. Tu, “Improvements of the Hough Transform in image processing”, PhD Thesis in Signal,
Image, Automation, Images, Signals and Intelligent Systems Laboratory (Créteil) LISSI, Paris-Est,
September 2014.
Page 723
SEMIT 2023
Multi-Objective Optimization based Efficient Healthcare Management System

1
Harinandan Tunga, 1Surjendu Pal, 2*Samarjit Kar, 3Debasis Giri
1
harinandan.tunga@rcciit.org.in, surja.pal104@gmail.com
2
Department of Mathematics, National Institute of Technology Durgapur
samarjit.kar@maths.nitdgp.ac.in
3
Department of Information Technology, Maulana Abul Kalamazad University of Technology
debasis_giri@hotmail.com
Abstract
A multi-objective based efficient healthcare system model is formulated in a comprehensive manner
which has two objectives namely Maximization of Quality of Service (QoS) and Minimization of Total
Cost of Operation. This model is solved using an elitist Non-Dominated Sorting Genetic Algorithm
(NSGA II) and Multi-Objective Particle Swarm Optimization. The proposed model uses two meta-
heuristic models to do a performance analysis of the algorithms used and generate a pareto-optimal front
and analysis is performed on the given data. The first objective function focuses on minimizing the overall
cost associated with various resources in healthcare, including doctors, nurses, technicians, beds, rooms,
and medical equipment. The second objective function aims to maximize the quality of service provided,
considering the expertise of doctors, nurses, and the availability of beds. Therefore, improvement in the
Quality of Service (QoS) is obtained along with optimization of the operational cost. The proposed
method is implemented on the Python working platform and its performance is verified. The proposed
method’s performance is confirmed by contrasting it with MOPSO and NSGA II algorithms in terms of
approximate Pareto optimal results, hypervolume (HV), generational distance (GD), inverted
generational distance (IGD), space and spread. As a result, the obtained results showed that the proposed
method is efficient for improving quality of service (QoS) and reducing the total cost of healthcare
management system performance.
Keywords: Quality of Service; Multi-objective optimization; Healthcare management system; MOPSO,
NSGA-II; Cost minimization; Quality maximization
1. Introduction:
Healthcare is crucial for individuals and society, encompassing prevention, disease management, and
medical services. Hospitals play a vital role in providing these services, but operational efficiency faces
challenges, especially during COVID-19, leading to reduced efficiency [1]. Resource management is a
significant challenge, requiring effective utilization of personnel, equipment, and facilities to provide
seamless services. Overcrowding hampers service quality and efficiency [3], and managing patient inflow
is crucial [4].
The World Health Organization predicts trillions in global healthcare spending by the decade's end,
posing challenges for lower-income individuals and countries like South Africa and India [6][7][8].
Despite investments, resource constraints, allocation inefficiencies, and coordination issues persist.
Balancing cost minimization and service quality maximization is challenging [9].
This study introduces a cost-optimization algorithm for healthcare establishments, classifying and
organizing critical resources to mitigate overcrowding and overutilization. Two primary objectives are
considered:
o Total Cost of Operation: This includes the operational cost of the healthcare establishment,
encompassing all functionaries and resources.
o Maximization of Quality of Service (QoS): This objective focuses on delivering the highest
quality of service to patients.
Page 724
SEMIT 2023
Both objectives are analyzed using two separate genetic algorithms, NSGA-II and MOPSO. NSGA-II
identifies non-dominated solutions across all objectives, while MOPSO extends Particle Swarm
Optimization for multi-objective challenges.
The study aims to balance cost minimization and service quality maximization, comparing the
performance of NSGA-II and MOPSO in healthcare management. Extensive experimentation
demonstrates NSGA-II's potential to enhance efficiency.
The research benefits healthcare organizations, policymakers, and decision-makers. The multi-objective
optimization-based system enables informed resource allocation decisions, leading to cost savings,
improved service quality, and higher patient satisfaction. Ultimately, it advances healthcare management,
addressing cost-effective, high-quality delivery challenges.
1.1. Literature Survey
The healthcare sector plays a pivotal role in society, delivering critical emergency services to patients in
need. Unfortunately, it currently grapples with severe challenges, including overcrowding and a shortage
of essential healthcare professionals, such as doctors and nurses. These issues not only compromise the
Quality of Service (QoS) provided to patients but also limit the number of individuals who can receive
care. The COVID-19 pandemic underscored the unpreparedness of the healthcare sector to effectively
serve the growing population. This deficiency poses a significant threat, given that the healthcare sector
is an emergency service provider, and any shortfall in services can have life-threatening consequences.
Despite significant advancements in medical sciences in recent decades, meeting the increasing demands
of the population remains a formidable challenge. In many middle- or low-income countries, insufficient
healthcare services result in a large segment of the population being unable to access emergency care
when needed. Therefore, the present situation necessitates better utilization of healthcare resources and
a substantial enhancement in QoS to ensure that no patient is left unattended, and healthcare resources
are optimally deployed.
In our research, we have examined various studies addressing critical issues across the healthcare sector.
Rotenstein explores the relationship between work overload, burnout, and job attrition during COVID-
19 within the healthcare workforce [1]. De Souza highlights the challenge of hospital bed management
[2], while Fu presents a Stochastic Programming Model for Service Scheduling with Uncertain Demand
[3]. Gualandi proposes a model to enhance hospital patient flow [4], and Dixit focuses on healthcare
supply issues [5]. The World Health Statistics conducts research on monitoring health for Sustainable
Development Goals [6], and Batun contributes to healthcare delivery optimization [7]. Demeester and
Souffriau introduce a hybrid tabu search algorithm using AI for patient bed assignments [8]. Bhargava
addresses the global health worker shortage and international migration [9], and Buchan tackles nursing
shortages [10].
The Gupta investigates health workforce surge capacity during pandemics [11], while Pulido-Fuentes
explores primary care strategies during COVID-19 [12]. Liu and Bellamy study healthcare systems in
low- and middle-income countries [13]. Yan and McClure use Genetic Algorithms for Health Service
planning [14], and Deb works on the NSGA-II multi-objective genetic algorithm [15]. Chiang focuses on
optimizing operating room and nursing unit schedules [16], and Fathollahi-Fard addresses home
healthcare optimization [17]. Gorunescu analyzes bed-occupancy management in hospitals [18], and
Gong presents a multi-objective optimization approach for bed allocation [19]. Chang emphasizes
inpatient bed allocation optimization [20], while AbdelAziz tackles hospital admission optimization [21].
Finally, de Souza addresses hospital bed management [22]. These studies have provided valuable insights
and identified limitations that we address in this paper.
1.2. Contribution
The outcomes of this research hold profound implications for healthcare organizations, policymakers,
and decision-makers. Our proposed multi-objective optimization-based healthcare management system
promises to empower healthcare administrators with informed decision-making capabilities, ultimately
leading to cost savings, elevated service quality, and heightened patient satisfaction. As a result, this
Page 725
SEMIT 2023
research contributes significantly to the advancement of healthcare management practices, addressing the
critical challenges associated with achieving cost-effective and high-quality healthcare delivery.
Key Contributions of the Proposed Work:
• Introduction of an optimization algorithm aimed at reducing the overall operational costs of
healthcare establishments.
• Provision of the highest possible Quality of Service (QoS) to patients.
• Pioneering a multi-objective optimization-based approach for healthcare management.
• Offering guidelines to healthcare establishments to prevent resource overutilization challenges.
• Comparative analysis of two genetic algorithms (GAs), namely MOPSO and NSGA-II, to assess
their performance in achieving cost minimization and service quality maximization objectives.
• Enhancing the work-life balance for clinicians and healthcare workers, enabling them to serve
patients with the utmost care.
• Paving the way for an efficient and effective healthcare management system capable of meeting
the evolving needs and demands of the population.
• Identification of optimal solutions that strike a balance between cost minimization and service
quality maximization.
• The potential to significantly enhance the efficiency and effectiveness of healthcare systems on
a global scale.
• In order to find the Pareto-optimal solution for the present situation, two MOEAs - NSGAII, and
MOPSO, are used.
• To compare the convergence and variety of two MOEAs, a data analysis technique including
performance measures such as GD, IGD, space, spread, HV, and MPFE is employed.
There is considerable convergence and diversity among the MOEAs, as illustrated by the significant
comparative results for these performance indicators. Fig. 1 shows the framework's evaluation.
Assumption Mathematical Performanc Statistical

Methodology
Model e Measures Studies
Efficient Bi-Objective - GD&IGD

Healthcare functions based Evolutionary Validation of
- Space
Management Efficient Algorithms – result and
- Spread
System Healthcare NSGA II & comparison
- HV
Management MOPSO
- MPEF
System
Fig.1. The framework's evaluation
2. Problem Formulation
In this section we aim to discuss the problem statement in detail along with the concerned objectives. we
formulate the problem of healthcare management as a multi-objective optimization task. Our objective is
to efficiently allocate healthcare resources while simultaneously minimizing costs and maximizing
service quality. We consider multiple decision variables, notations, and equations to represent the
problem accurately.
The decision problem, for our problem, is to maximally coordinate the utilization of multifold resources
within the hospital. In our case the multifold resources are doctors, nurses, beds and rooms. In this case
the most appropriate model to be created is a linear integer programming problem.
Firstly, we define the notations to be used in the model Sets: -
I- Set of departments
J - Set of doctors type
Page 726
SEMIT 2023
K - Set of nurses type

L - Set of available beds
R - Set of rooms
E - Set of medical equipment
H - Set of non-medical staff
Decision Variables: -
XDij - Number of type j ∈ J doctors in department i ∈ I
XNik - Number of type k ∈ K nurses in department i ∈ I
XDAij - Number of type j ∈ J doctors added in department i ∈ I
XNAik - Number of type k ∈ K nurses added in department i ∈ I
XTih - Number of type h ∈ H non-medical staff in department i ∈ I
XBil - Number of beds l ∈ L in department i ∈ I
XBAil - Number of beds added l ∈ L in department i ∈ I
XMie - Number of medical equipment e ∈ E in department i ∈ I
Xrav – Number of rooms of each type under use r ∈ 𝑅
Binary Variables: -
Zilr - Equals to 1, if bed l ∈ L are allocated for room r ∈ R in department i ∈ I, equals to 0 otherwise
Yir - Equals to 1, if room r ∈ R is used in department i ∈ I equals to 0 otherwise
Parameters : -
bdij - Cost of j ∈ j type doctors in department i ∈ I
bsik - Cost of k ∈ K type nurses in department i ∈ I
bthi - Cost of h ∈ H type technicians in department i ∈ I
btil - Cost of operating beds l ∈ L in department i ∈ I
bwil - Waiting cost for beds l ∈ L in department i ∈ I
brri - Cost to operate room r ∈ R if used in department i ∈ I
bmier - Cost to operate medical equipment e ∈ E for room r ∈ R if used in department i ∈ I
mdij - Maximum number of doctors for each type can be allocated to each department
mnik - Maximum number of nurses each type can be allocated in department
mbil - Maximum number of beds l ∈ L can be allocated to each department
ρi - Maximum fund available for department i ∈ I
Qj – Quality of Service index of doctors of type j ∈ J
Qk - Quality of Service index of nurses of type k ∈ K
Qh - Quality of Service index of non-medical staff of type h ∈ H
Qb - Quality of Service index of the type of bed for l ∈ L
Qr - Quality of Service index of room of type r ∈ R
Qop - Quality of Service index of the type of operation bed for l ∈ L
Page 727
SEMIT 2023
Qe - Quality of Service index of equipment of type e ∈ E

mQj – Minimum Quality of Service value required for doctors
mQk – Minimum Quality of Service value required for nurses
mQh – Minimum Quality of Service value required for technicians
mQe – Minimum Quality of Service value required for equipment
mbl - Minimum Quality of Service value required for beds used
mwbl - Minimum Quality of Service value required for waiting beds
mr - Minimum Quality of Service value required for rooms being used
Objective Functions: -
The objectives of the healthcare management system is to find all the decision variables such that to
maximize the Total Quality of Service (TQoS) and minimize total operating cost (ToC) for doctors, nurse,
non-medical staffs, bed, waiting bed, rooms and medical equipment. The Quality of Services for doctors,
nurse, non-medical staffs, bed, rooms and medical equipment is denoted as QoSu and operational cost for
doctors, nurse, nurse-aids, non-medical staff, beds, rooms and medical equipment is denoted as OCu
where u is the type of service and u > 0 and u is u ∈ N. Total number of services is denoted by ‘no’ where
no > 0 and no ∈ 𝑁.
Mathematically, the objective function can be written as follows: -
Objective 1: Maximize the Total Quality of Services (TQoS)
TQoS = Maximum ∑𝑛𝑜
𝑢=1 𝑄𝑜𝑆𝑢 (1)
Where,
QoS1=∑𝑖∈𝐼 ∑𝑗∈𝐽 𝑄𝑗(𝑋𝐷𝑖𝑗 + 𝑋𝐷𝐴𝑖𝑗) (2)
QoS2=𝑄𝑘(𝑋𝑁𝑖𝑘 + 𝑋𝑁𝐴𝑖𝑘) (3)
QoS3=𝑄ℎ ∗ 𝑋𝑇𝑖ℎ (4)
QoS4=∑𝑖∈𝐼 ∑𝑙∈𝐿 𝑄𝑏 ∗ (𝑋𝐵𝑖𝑙) (5)
QoS5=∑𝑖∈𝐼 ∑𝑙∈𝐿 𝑄𝑤𝑏 ∗ 𝑋𝐵𝐴𝑖𝑙 (6)
QoS6=∑𝑒∈𝐸 ∑𝑟∈𝑅 ∑𝑖∈𝐼 (𝑄𝑒 ∗ 𝑋𝑀𝑖𝑒) (7)
QoS7=∑𝑖∈𝐼 ∑𝑟∈𝑅 Qr ∗ (Xrav) (8)
Objective 2: Minimize the Total Operational Cost (ToC)
ToC=Minimize∑𝒏𝒐
𝒖=𝟏 𝑶𝑪𝒖 (9)
Where,
OC1=∑𝑖∈𝐼 ∑𝑗∈𝐽 𝑏𝑑𝑖𝑗 (𝑋𝐷𝑖𝑗 + 𝑋𝐷𝐴𝑖𝑗) (10)
OC2=∑𝑖∈𝐼 ∑𝑘∈𝐾 𝑏𝑠𝑖𝑘(𝑋𝑁𝑖𝑘 + 𝑋𝑁𝐴𝑖𝑘) (11)

OC3=∑𝑖∈𝐼 ∑ℎ∈𝐻 𝑏𝑡ℎ𝑖 ∗ 𝑇ℎ𝑖 (12)
OC4=∑𝑖∈𝐼 ∑𝑙∈𝐿 𝑏𝑡𝑖𝑙(Zilr ∗ 𝑋𝐵𝑖𝑙) (13)
OC5=∑𝑖∈𝐼 ∑𝑙∈𝐿 (𝑏𝑤𝑖𝑙 ∗ 𝑋𝐵𝐴𝑖𝑙) (14)
OC6=∑𝑒∈𝐸 ∑𝑟∈𝑅 ∑𝑖∈𝐼 (𝑏𝑚𝑖𝑒𝑟 ∗ 𝑋𝑀𝑖𝑒) (15)
OC7=∑𝑖∈𝐼 ∑𝑟∈𝑅 𝑏𝑟𝑖𝑟 ∗ 𝑌𝑖𝑟 (16)
Page 728
SEMIT 2023
Constrains : -
XDij+XDAij ≤ mdij (17)
State that the number of doctor should not be greater the maximum number of doctors can allocated in
each department.
XNik+XNAik≤mnik (18)
XDij+XDAij≤XTih (19)
XBil+XBAil≤mbil (20)
∑ Zilr ≤ Yir [are to make sure that beds are allocated to the available room] (21)
ToC ≤ ∑𝑖=𝐼 𝜌(𝑖) [State that the budget spent should not be greater than the fund provided] (22)
XDij, XDAij, XNik, XNAik, XTih, XBil, XBAil, XMier ≥ 0 and integer (23)
Zilr , Yir ∈ {0,1} (24)
∑𝑖∈𝐼 ∑𝑗∈𝐽 𝑄𝑗(𝑋𝐷𝑖𝑗 + 𝑋𝐷𝐴𝑖𝑗)≥ mQj (25)
∑𝑖∈𝐼 ∑𝑘∈𝐾 𝑏𝑠𝑖𝑘(𝑋𝑁𝑖𝑘 + 𝑋𝑁𝐴𝑖𝑘) ≥ mQk (26)
∑𝑖∈𝐼 ∑𝑙∈𝐿(𝑄𝑏 ∗ 𝑋𝐵𝑖𝑙 + 𝑄𝑤𝑏 ∗ 𝑋𝐵𝐴𝑖𝑙) ≥ mQh (27)
𝑄𝑒(𝑏𝑚𝑖𝑒𝑟 ∗ 𝑋𝑀𝑖𝑒)≥ mQe (28)
Qr ∗ (Xrav) ≥ mr (29)
𝑄ℎ ∗ 𝑋𝑇𝑖ℎ ≥ mQh (30)
The above mathematical model has two objective functions and some constraints. The first objective
function (Eq .1) to Maximize the Total QoS for healthcare managememt system takes into consideration
seven services (QoSu=1..8). We Calculate Maximum Quality of Services of each of the service in the Eq.
2 through Eq. 9. The second objective function (eq. 9) is to minimize the total operational cost (ToC)
of the healthcare management system that takes into consideration eight services (OCu=1..8). We Calculate
the Minimum Cost of Operation of each of the services in the Eq. 11 through Eq. 18.
First part of this objective 1, consists of Equation (eq.2) is a measure of the QoS provided by the Doctors
of each department. It is the sum of the QoS provided by doctors of different types and ranks. The equation
(3) is a measure of the QoS provided by the nurses. It is the sum of the QoS provided by nurses of different
categories. The fourth equation (eq. 4) provides the QoS provided by the non-medical staff/technicians
in the hospital in the hospital. Equation (5) measures the QoS provided by the beds present in the hospital.
The sixth Equation (eq. 6) gives a measure of the Quality of Service for waiting beds or unoccupied beds.
In the eq. 7 we measure the Quality of Service provided by the machineries or equipment used. In eq. 8
we calculate the QoS of the rooms of the hospital.
The eq. 9 gives the second objective function and it aim is to minimize the total operational cost. It has 7
parts OC1 for the Cost of Operation of Doctors (Eq .11), OC2, OC3, OC4, OC5, OC6, OC7. Each of the 7
Equation are a measure of the operational cost of the functionaries of the hospital. The Equation (10)
gives a measure of the total operational cost of the Doctors of the hospital. Equation (11) provides the
operational cost of nurses in the hospital. Equation (12) is a measure of the cost of operation of the
technicians/non-medical staff. The Equation (13) gives the total operational cost associated with the
various categories of beds in the hospital. Equation (14) provides the operational cost of the beds in the
hospital that are under the waiting state. Equation (15) provides the operational cost of the equipment
used in the hospital. The Equation (16) and provides the cost for allocating rooms in the hospital.
The model is also subject to various constraints. In equation (17) we see that the number of doctors should
not be greater the maximum number of doctors can allocated in each department.
Similarly, equation (18) is subject to the constraint that the number of doctors should not be greater the
maximum number of doctors that can be allocated in each department. Equation (19) and (20) we see that
Page 729
SEMIT 2023
the maximum number of technicians and the maximum number of beds that can be allocated are subject
to the constraint of the maximum number of technicians and maximum number of beds respectively.
The equation (21) ensures that the beds are allocated in the available rooms. The equation (22) ensures
that the totals cost of operation does not exceed the maximum available funds.
Equation (23) ensures that the allocations of doctors, nurses, beds, equipment etc made should always
have integral values. The equation (24) ensures that the binary variables Zilr and Yir always takes the values
0 for allocation not possible and 1 for allocation possible respectively.
Equations (25),(26),(27),(28),(29),(30) ensures that the minimum quality required or bare-minimum QoS
required to be maintained in terms of the doctors, nurses, beds, equipment, rooms and non-medical staff
respectively for making the system run smoothly without making any compromise on the quality of
service.
3. Proposed Methodology
In this research paper, we propose a bi-objective approach to optimize service quality while minimizing
operational costs. We employ Multi-objective Evolutionary Algorithms (MOEAs) to identify the optimal
Pareto front, representing our candidate solutions. Our model considers various parameters, including
department size, medical staff types, bed availability, equipment, and non-medical staff. We start by
randomly generating parameter values and decision variables and then calculate fitness values based on
both objective functions. The next generation's population selection prioritizes individuals with superior
service quality and lower operational costs. Evolutionary Algorithm (EA) operators like selection,
crossover, and mutation create a child population iteratively until a termination condition is met.
Ultimately, the final population comprises Pareto-optimal solutions, balancing service quality and
operational cost.
3.1. MOEAs Used
3.1.1. NSGA II
NSGAII is a sorting-based MOEA that maintains solution elitism and diversity by sorting solutions in
each iteration and preserving their crowding distance. It starts with a random parent population (𝑃𝑔 ) and
creates a child population (𝑄𝑔 ) using EA operators. A front formation procedure combines (𝑃𝑔 )and
(𝑄𝑔 )to select the next generation's parent population (𝑃𝑔+1 ) based on nondominated sorting. This
procedure arranges solutions into different fronts, with 𝐹1 containing the best solutions.
Crowding distance is calculated to preserve diversity, favouring solutions with larger crowding distances.
𝐹1 represents Pareto optimal solutions, and these steps are repeated until a termination condition is met.
The complete NSGAII approach is described in the provided pseudocode, and input variables
[𝐼, 𝐽, 𝐾, 𝐿, 𝑅, 𝐸, 𝐻] are specified in section 4.2.2, with the initial population generated by randomly
assigning values to these input parameters.
Table 1 NSGAII pseudocode
Algorithm: NSGAII(P)
Start NSGAII
Input: Population(𝑃𝑔 )of size N at generation g = 0
Output: First nondominated front (𝐹1 ).
Steps:
Step 1: Apply selection, crossover and mutation to generate child population 𝑄𝑡 for g = 0
Step 2: While 𝑔 ≤ 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛 Do
𝑅𝒈 = 𝑃𝒈 ∪ 𝑄𝒈
Calculate fitness value for every individual of 𝑅𝑔
Apply nondominated sorting to generate nondominated fronts 𝐹𝑖 for 𝑖 = 1,2,3, … , 𝑛.
While size of (𝑃𝑔+1 ∪ 𝐹𝑖 ) ≤ 𝑁
Page 730
SEMIT 2023
𝑃𝑔+1 = 𝑃𝑔+1 ∪ 𝐹𝑖
𝑖 =𝑖+1
End while
Calculate crowding distance of 𝐹𝑖 .
𝑧 ← 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑃𝑔+1
𝑃𝑔+1 = 𝑃𝑔+1 ∪ 𝐹𝑖 [1: (𝑁 − 𝑧)]
Perform selection, crossover and mutation on 𝑃𝑔+1 to generate 𝑄𝑔+1
Step 3: End while
Step 4: End NSGA II
3.1.2. MOPSO
The Multi-objective Particle Swarm Optimization (MOPSO) algorithm is widely used in various fields
for tackling multi-objective problems. It begins with swarm initialization, where initial solutions are
generated randomly using input parameters. Fitness is determined based on Total Quality of Service
(TQoS) and Total Operational Cost (TOC). Local (pbest) and global best (gbest) are then identified for
each individual, guiding swarm selection for the next generation. The archive, containing global and local
best points, is updated before proceeding to the next iteration. See the algorithm pseudocode below.
Table 2 MOPSO Pseudocode
Algorithm: MOPSO(P)
Input: I - Set of departments, J - Set of doctors type, K - Set of nurses type, L - Set of available beds,
R - Set of rooms, E - Set of medical equipment, H - Set of non-medical staff
Output: The non-dominated particles in the SWARM and save them in the EX-ARCHIVE external
archive. The first non-dominant front .
Steps
Step 1: Generate initial particles solution (x0i).
Step 2: Pbest0i is memorized for each solution (fitness value) and Gbest0, which is the global best, is
the solution which provides best Fitness Value (in our case it is TQos and minimum TOC).
In each iteration, Pbest and Gbest values get updated by comparing the current value with the
previous values.
Select the gbest for each particle from EX-ARCHIVE
Step 3: The velocity array (v0i) is obtained by choosing continuous numbers between 0 and the
number of departments(I)
Step 4: The following equation is used to update the velocity to get v1i, Where, w0= 0.6, C1= C2= 0.2,
D1=0.3 and D2=0.5
vi1 = w(0) vi(0) + C1 D1 (pbest(0)i – xi(0)) + C2 D2 (gbest(0)i – xi(0))
Step 5: Update the particles configuration x1i., x1i = x0i + θ*v1i ; Now,If x1i> xMAX then x1i = xMAXand
If x1i< xMIN then x1i = xMIN, , Where θ is known as the intensification (or diversification)
parameter. Smaller value of θ leads to diversification of the Pareto front and large value of θ
causes intensification of the Pareto front.
Step 6: Next, the solutions are compared by higher TQos. If two swarm have same TQos the lower
TOC is considered first.
Step 7: Go to step 1 until all the number of iterations gets completed otherwise go to step 8.
Step 8: In each generation, update the content of EX-ARCHIVE. Here we need to determine whether
the new solution should be added to the archive or not. This is determined by performing a
non-domination test on the external archive.
Step 9: Finish
Page 731
SEMIT 2023
4. Results Analysis
4.1. Performance Measure
There exist several metrics for evaluating the performance of Multi-Objective Evolutionary Algorithms
(MOEAs). Assessing MOEAs' convergence and diversity is essential in this evaluation process.
Convergence refers to how closely the solutions on the algorithm's nondominated front resemble the
actual Pareto front, while diversity relates to the distribution of solutions along the Pareto front.
It's important to note that relying on just a few metrics is insufficient to determine the superiority of one
algorithm over another. To comprehensively assess the performance of the three MOEAs mentioned
earlier, we consider the following metrics:
Generational Distance (GD, Inverted Generational Distance (IGD), Spacing, Spread , Maximum Pareto
Front Error (MPFE) and Hyper-volume. These metrics collectively provide a comprehensive assessment
of MOEAs' performance, taking into account convergence, diversity, and dispersion of solutions along
the Pareto front. Using this array of metrics ensures a robust evaluation of the algorithms.
4.2. Simulation Studies
4.2.1. Simulation Setup
The proposed method is mainly depends upon seven matrices. They are set of departments (I) , Set of
doctors type (J), Set of nurses type (K), Set of available beds (L), Set of rooms (R), Set of medical
equipment (E) and Set of non-medical staff (H). For all of the algorithm the initial values of these matrices
are kept same. The parameters for a single instance are retained as constant in the case of the three
MOEAs employed in the proposed study. Nondominated sorting is utilized for selection in NSGAII.
Other important MOEA factors include the number of generations, population size, crossover rate,
mutation rate, and so on. Table 1 shows the parameters used in the simulation.
Table 1: Parameters used in simulation
Parameters Value
Number of decision variables 9
Maximum number of generation in MOEA 1000
Population Size 100
Swarm size in MOPSO 100
Crossover Rate 0.7
Mutation Rate 0.02
4.2.2. Dataset used in MOEAs

The proposed work mainly dependent on 7 parameters discussed earlier. As we are unable to find suitable
dataset for the model we generated the values of the paramerts randomly. Table 2 shows the parameter
utilised in the simulation.
Table 2: Dataset used in simulation
Variables Description Values
I Set of departments 1-20
J Set of doctors type 1-50
K Set of nurses type 1-20
L Set of available beds 1-30
R Set of rooms 1-30
E Set of medical equipment 1-50
H Set of non-medical staff 1-10
Values of all decision variable are generated randomly from the respective range of parameters. All the
cost values are generated in the range 100-10000 randomly. The values of are also generated in random
Page 732
SEMIT 2023
in range [1-20], [1-50], [10-1000] respectively. All the values of quality of service (QoS) index is
generated in range [1-10].
4.2.3. Instance setting and its Evaluation
Proposed work is evaluated in 10 instances where each instance has different parameter value. Table 3
represents the result of the evaluation. Here the value of objective-1 which is quality of service is shown
where the value of objective-2 which is total operational cost is nearly same. Another value Δ is calculated
which is the difference of Qos of NSGAII and MOPSO for a instance. If greater than zero then NSGAII
perform better in that instance otherwise MOPSO. The evaluation result for instance “I3” is shown in the
Fig. 1.
Table 3: Evaluated sample result of 10 instances
Instances n(I) n(J) n(K) n(L) n(R) n(E) n(H) Total Op. Cost
Gap Total Quality of Gap
(Δ1) Service (Δ2)
NSGAII MOPSO NSGAII MOPSO
I1 5 5 10 5 5 12 2 12000 12002 -2 526 525 1

I2 5 8 15 20 14 15 12 24256 24256 0 1015 1020 5
I3 8 10 15 25 20 20 20 59028 59027 1 3029 3030 -1
I4 10 12 20 50 25 25 30 92255 922556 -1 8265 8263 2
I5 14 15 26 60 30 30 42 112503 112500 3 10245 10265 -20
I6 15 20 30 70 40 40 59 206025 206022 3 19696 19690 6
I7 20 28 32 100 50 42 63 325496 325496 0 30549 30530 19
I8 22 30 35 115 70 50 80 536984 536988 4 56894 56860 34
I9 25 39 40 130 100 56 94 759840 759836 4 800002 80000 2
I10 30 40 50 200 150 70 100 2068259 2068258 1 100012 100000 12
Fig 1 (a) Evaluation of proposed model I6 Fig 1 (b) Evaluation of proposed model I7
In the above table, the columns n(I), n(J), n(K), n(L), n(R), n(E) and n(H) denotes the number of
departments, doctor type, nurse type, available beds, rooms, medical equipment and non-medical
staff respectively. For the instance “I5”, for nearly same operational cost which is 112503 the total
quality of service for NSGAII and MOPSO are 10245 and 10265. Here the Δ is negative which
indicates for this instance MOPSO performs better than NSGAII. Similarly, for instance “I6” the Δ
value is positive which denotes that NSGAII performs better than MOPSO.
Page 733
SEMIT 2023
4.2.4. Statistical Analysis
The proposed MOEA-based models' outputs are evaluated using previously performance indicators.
We assesses performance metrics for central tendency (mean and median) and variability (standard
deviation and interquartile range) on ten instances after 20 runs each. ANOVA tests gauge algorithm
convergence and diversity, with a significance level of 0.05. Table 4 summarizes the results, and Fig.
2 presents corresponding box plots. Fig. 2(a) shows the box plot of GD value of two MOEAs which
indicates GD NSGAII's superiority over MOPSO. IGD determines diversity. From 2(b) we observe
that the solution for MOPSO is more diverse than NSGAII. Fig 2(d), 2(e) also revels the same. From
Fig 2(c) we see that the dolution space for MOPSO is lower. So solution of MOPSO is well
distributed. For both the MOEAs MPFE preforms similar to GD. From Fig 2(f), the performance of
NSGAII is better than MOPSO.
Table 4 ANOVA Test Summaries

Critical value (at 𝛼 = 0.05) = 3.2
Degree of freedom = 242
Performance ((𝑴𝒆𝒂𝒏 − 𝒙𝒊 )𝟐 ) (𝑺𝑺𝑸𝒃 ) F-ratio Conclusion
metrics NSGAII MOPSO
GD 5.625E-6 5.135E-6 6.789E-06 7.005 The results are significant at the 5%
significance level since 7.005 > 3.220.
IGD 4.296E-5 5.826E-5 0.03256 122.598 In light of the fact that 122.598 > 3.220,
the results are statistically significant at
the 5% level of significance.
Space 5.34489E-5 7.796E-5 8.7845E-06 1.995 At 5% significance level, the results are
not significantly different for three
groups as 1.995<3.220.
Spread 0.09218 0.34647258 5.964879 263.597 263.597 > 3.220, so it's significant at 5%.
HV 0.7887549642 0.581267 0.6971675 13.478 The results are significant at 5% because
13.478> 3.220.
MPFE 0.002569764 0.00998678 0.000951 4.404 A 5% significance level indicates that
4.404 > 3.220 indicates significance.
Page 734
SEMIT 2023
(a) (b)
(c) (d)
(e) (f)
Fig. 8 Box plot of the performance metrics with respect to NSGAII and MOPSO[8(a)-8(f)]
From the above discussion, it can be shown that the overall performance of proposed approach is
consistent across all instances. Each of the 10 instances is subjected to an ANOVA test, which examines
all the performance indicators with three groups: NSGAII and MOPSO. This implies that there are 20
Page 735
SEMIT 2023
samples in the test. Table 6 displays the square between groups (𝑆𝑆𝑄𝑏 ) as well as the sum of squares for
the three various groups ((𝑀𝑒𝑎𝑛 − 𝑥𝑖 )2 ) . In this case, the F-ratio defines the variety of MOEA outcomes
in terms of several performance indicators.
5. Conclusion
In our proposed research, we have introduced an innovative model for efficient healthcare management
and have addressed it using two well-established Multi-Objective Evolutionary Algorithms (MOEAs),
NSGA-II and MOPSO. To validate the effectiveness of our work, we conducted a rigorous statistical
analysis, including the ANOVA test, which consistently produced compelling results. Consequently, we
conclude on the originality and uniqueness of our proposed model, as outlined below:
• We have developed an efficient healthcare management system and applied both NSGA-II and
MOPSO algorithms to solve it. Moreover, we conducted a statistical analysis to assess the
performance of these algorithms in terms of convergence and diversity. Key performance metrics,
such as GD, IGD, Space, Spread, HV, and MPFE, were analysed to make this assessment.
• Our work introduces a data analysis-driven approach to address a problem that goes beyond existing
methodologies. This novel approach advances the field by leveraging data analytics in solving
complex healthcare management challenges.
• After closely examining the solutions across ten instances, our findings indicate that NSGA-II
demonstrates superior convergence, while MOPSO generates well-distributed solutions.
By adopting this comprehensive analytical approach, our research contributes to the body of knowledge
in multi-objective studies from a theoretical standpoint. Consequently, this unique technique holds
promise for the development of effective healthcare management systems.
Reference
[1] Lisa S. Rotenstein, Roger Brown, Christine Sinsky, and Mark Linzer (2023). The Association of
Work Overload with Burnout and Intent to Leave the Job Across the Healthcare Workforce During
COVID-19, J Gen Intern Med 38(8):1920-7, doi: 10.1007/s11606-023-08153-z.
[2] M.C. de Souza, Thiago Souza, and Guilherme Luís Roehe Vaccaro (2016). Hospital bed
management: An analysis from the perspective of the theory of constraints, Revista ESPACIOS,
Vol. 37, Published by Editorial Espacios (ISSN: 0798-1015).
[3] Yu Fu, and Amarnath Banerjee (2021). A Stochastic Programming Model for Service Scheduling
with Uncertain Demand: an Application in Open-Access Clinic Scheduling, , SN Operation
Research Forum(2021) 2(3): 43. Published online 2021 Aug 27, doi: 10.1007/s43069-021-00089-6.
[4] Raffaella Gualandi, Cristina Masella, and Daniela Tartaglini (2019), Improving hospital patient
flow: a systematic review, Business Process Management Journal ahead-of-print(ahead-of-print)
doi:10.1108/BPMJ-10- 2017-0265.
[5] Anuj Dixit, Srikanta Routroy, and Sunil Kumar Dubey, (2019), A systematic literature review of
healthcare supply chain and implications of future research, International Journal of Pharmaceutical
and Healthcare Marketing Vol. 13(4):405-435. Published by Emerald (ISSN: 1750-6123),
doi:10.1108/IJPHM-05-2018-0028.
[6] World Health Organization, (2019), World Health Statistics 2019: Monitoring Health for the SDGs,
Sustainable Development Goals, World Health Organization, World health statistics 2019:
monitoring health for the SDGs, sustainable development goals, ISBN: 9789241565707.
[7] Sakine Batun, and Mehmet A. Begen, (2013), Optimization in Healthcare Delivery Modeling:
Methods and Applications, Handbook of Healthcare Operations Management (pp.75-119) 26,
doi:10.1007/978-1-4614-5885-2_4.
[8] Peter Demeester, Wouter Souffriau, Patrick De Causmaecker, and Greet Vanden Berghe (2010), A
hybrid tabu search algorithm for automatically assigning patients to beds, Elsevier: Artificial
Intelligence in Medicine, Volume 48, Issue 1, January 2010, Pages 61-70,
doi.org/10.1016/j.artmed.2009.09.001.
Page 736
SEMIT 2023
[9] J. Bhargava, J. Docquier, and F. Moullan (2014), Global shortage of health workers and international
migration: Evidence from the Philippines, Published in: Human Resources for Health, Glob Health
Action. 2014; 7: 10.3402/gha.v7.23611. Published online 2014 Feb 13. doi: 10.3402/gha.v7.23611.
[10] L. Buchan and J. Aiken (2008), Solving nursing shortages: a common priority, Journal of Clinical
Nursing, Volume 17, Issue 24 p. 3262-3268, doi.org/10.1111/j.1365-2702.2008.02636.x.
[11] Neeru Gupta, Sarah A. Balcom, Adrienne Gulliver, and Richelle L. Witherspoon (2021), Health
workforce surge capacity during the COVID‐19 pandemic and other global respiratory disease
outbreaks: A systematic review of health system requirements and responses, Int J Health Plann
Manage. 2021 May; 36(Suppl 1): 26–41. Published online 2021 Mar 1. doi: 10.1002/hpm.3137.
[12] Montserrat Pulido‑Fuentes, Luisa Abad González, Isaac Aranda Reneo, Carmen Cipriano‑Crespo1,
Juan Antonio Flores‑Martos and Ana Palmar Santos (),Towards a liquid healthcare: primary care
organisational and management strategies during the COVID-19 pandemic - a qualitative study,
Pulido‑Fuentes et al. BMC Health Services Research (2022) 22:665, doi.org/10.1186/s12913-022-
07855-0.
[13] J. J. Liu, E. A. Bellamy, and A. B. Barnet, (2016), Healthcare systems in low- and middle-income
countries: An economic perspective, Published in: International Journal of Health Economics and
Management Year: 2016, J Med Eng Technol. 2016 Oct-Nov; 40(7-8): 336–341. doi:
10.1080/03091902.2016.1256081.
[14] Charles Yan, Nathan McClure, Sean P. Dukelow, Balraj Mann, and Jeff Round (2022), Optimal
Planning of Health Services through Genetic Algorithm and Discrete Event Simulation: A Proposed
Model and Its Application to Stroke Rehabilitation Care, MDM Policy Pract. 2022 Jul-Dec; 7(2):
23814683221134098. Published online 2022 Oct 22. doi: 10.1177/23814683221134098.
[15] K.Deb, A. Pratap , S. Agarwal, and T. Meyarivan (2002), A fast and elitist multiobjective genetic
algorithm: NSGA-II, Journals & Magazines, IEEE Transactions on Evolutionary Computation,
Volume: 6 Issue: 2, 1089-778X/02$17.00 © 2002 IEEE.
[16] An Jen Chiang, Angus Jeang, Po Cheng Chiang, and Chien-Ping Chung, (2019), Multi-objective
optimization for simultaneous operating room and nursing unit scheduling, International Journal of
Engineering Business Management Volume 11: 1–20, Sage Journals,
doi.org/10.1177/1847979019891022.
[17] Amir M. Fathollahi-Fard, Abbas Ahmadi, and Behrooz Karimi, (2021), Multi-Objective
Optimization of Home Healthcare with Working-Time Balancing and Care Continuity, MDPI -
Publisher of Open Access Journals, Sustainability 2021, 13(22), 12431,
doi.org/10.3390/su132212431.
[18] F. Gorunescu, S. I. McClean and P. H. Millard (2002), A queueing model for bed-occupancy
management and planning of hospitals., The Journal of the Operational Research Society, Vol. 53,
No. 1 (Jan., 2002), pp. 19-24 (6 pages), Published By: Palgrave Macmillan Journals.
[19] Yue-jiao Gong, Jun Zhang and Zhun Fan, (2010), A multi-objective comprehensive learning particle
swarm optimization with a binary search-based representation scheme for bed allocation problem in
General Hospital, 978-1-4244-6588-0/10/$25.00 ©201 0 IEEE,
[20] Jian Chang, and Lingjuan Zhang, (2019), Case Mix Index weighted multi-objective optimization of
inpatient bed allocation in general hospital, Journal of Combinatorial Optimization, Springer, vol.
37(1), pages 1-19, doi: 10.1007/s10878-017-0204-3.
[21] Amr Mohamed AbdelAziz, Louai Alarabi, Saleh Basalamah and Abdeltawab Hendawi (2021),
Multi-Objective Optimization Method for Hospital Admission Problem—A Case Study on Covid-
19 Patients., Algorithms 2021, 14(2), MDPI - Publisher of Open Access Journals,
doi.org/10.3390/a14020038.
[22] M.C. de Souza, Thiago Souza, Guilherme Luís and Roehe Vaccaro (2016), Hospital bed
management: An analysis from the perspective of the theory of constraints, ESPACIOS Vol. 37(30),
ISSN 07981015.
Page 737
SEMIT 2023
Exploring the Elastic, Electronic, Dynamical, and Optical properties of cubic ZrCoAs half –
Heusler Semiconductor: An ab-initio Study
Lynet Allan,1, * R.E Mapasha2 Winfred M. Mulwa,3 Julius M. Mwabora,1
and Robinson J. Musembi1
1
Department of Physics, Faculty of Science and Technology, University of Nairobi, P.O.Box
30197-00100 Nairobi Kenya.
2
Department of Physics, University of Pretoria, Private Bag x 20, Hatfield, Republic of South
Africa.
3
Department of Physics, Faculty of Science, Egerton University, P.O Box 536-20115 Egerton
Kenya
*Corresponding author: E-Mail: allanlynet3@students.uonbi.ac.ke
Abstract
The electronic, mechanical, elastic, dynamical, and optical properties of ZrCoAs half-Heusler compound
have been systematically studied using the plane wave self-consistent field approach. The exchange
correlation functional used is Perdew-Burke-Erzerhof of the generalized gradient approximation (GGA-
PBE). Structural and electronic properties have been studied with and without Spin Orbit Coupling
(SOC) effects. The incorporation of SOC effects leads to a decrease in the Kohn-Sham band-gap. The
formation of the electronic bandgap was attributed to Co 3d, Zr 4s, and As 4p, for the conduction band,
and Co 3d and As 4p for the valence band without SOC effects. Co 5d, Zr 8d, and As 4p, became dominant
in the conduction band while Co 3d and As 4p dominated the valence band when SOC was taken into
account. The lattice constant also decreased by 0.063% when SOC was factored into the computations,
showing a much-improved agreement with experimental observations. The material exhibited ductility,
mechanical stability, and dynamical stability, making it promising for a variety of applications. Also, it
was discovered that the material's optical properties are excellent for photovoltaic applications, indicating
a potential application in solar energy conversion technology.
Keywords: DFT, Spin Orbit Coupling (SOC), electronic, dynamical, mechanical, elastic, and optical
properties, half Heuslers ZrCoAs.
1. Introduction
The global demand for sustainable and renewable energy sources has spurred extensive research in the
field of photovoltaic materials aimed at harnessing solar energy for power generation [1]. Among the
various materials investigated, Heusler compounds have garnered significant attention due to their
captivating electronic, magnetic, and thermoelectric properties [2]. These intermetallic compounds,
named after Friedrich Heusler, exhibit a cubic structure and comprise three elements: an alkali or alkaline
earth metal (A), a transition metal (B), and a metalloid (X), leading to the formations of ABX for half
Heuslers and AB2X for full Heuslers [3].
Heusler compounds have exhibited potential for use in photovoltaics, providing a unique avenue for
enhancing solar energy conversion efficiency. Zirconium-based alloys, with applications ranging from
automobiles to energy sectors, have gained attention [4]. Specifically, ZrCoY (Y = Sb, Bi) half Heusler
compounds have emerged as captivating candidates for photovoltaic and thermoelectric applications due
to their tunable properties. [2], [5]–[8],
Among the Zr-based half Heusler semiconductors, ZrCoAs compound underscores its significance within
the realm of intermetallic compounds, spotlighting its distinct electronic, magnetic, and structural
attributes [4]. The compound's potential for superconductivity has been investigated, alongside its
response to extreme conditions. These findings hold promise for applications in diverse fields, facilitated
by advancements in synthesis and characterization techniques, paving the way for potential use in
quantum materials, superconducting electronics, and spintronics[4]. Despite notable strides in
Page 738
SEMIT 2023
understanding half Heusler compounds[2], [5]–[8], to the best of our knowledge, limited research has
focused on ZrCoAs compound specifically for photovoltaic applications.
In this paper, we present an ab-initio study aimed at analysing the electronic, optical, elastic, dynamical,
and structural properties of the ZrCoAs half Heusler compound, aiming to bridge this research gap. We
employ the first-principles calculations to accurately determine ZrCoAs's fundamental properties. This
approach unveils key electronic characteristics like the band structure and optical absorption properties,
enriching our understanding of the material's potential as a light-harvesting medium. Additionally, the
study of elastic properties and lattice dynamics offers insights into the mechanical and dynamic stability
of ZrCoAs, critical for the reliability and longevity of energy-harvesting devices. The investigation of
ZrCoAs's properties further contributes to the understanding of its crystal symmetry and lattice constants,
crucial for device optimization and design.
Through this comprehensive ab initio analysis of ZrCoAs half Heusler compounds, aimed at renewable
energy applications, we endeavour to contribute to the advancement of sustainable energy technologies
and the development of high-efficiency solar cells through novel materials. The paper is organized as
follows: Section 2 outlines the calculations made for future reference. In Section 3, we delve into the
details of our findings. Structural characteristics are discussed in Section 3.1, while Section 3.2 covers
the projected density of states (PDOS) and electronic band structures. The discussion on elastic constants
and mechanical stability is presented in Section 3.3. Section 3.4 explores the dynamical properties of the
compounds, including phonon dispersions, followed by the analysis of the optical response of ZrCoAs in
Section 3.5. The summary is outlined in Section 4.
2. Calculations
To conduct our calculations, we obtain the crystal structure information files from the Materials Project
[9]. Subsequently, input files are generated using the Quantum Espresso tools available on the Materials
Cloud platform [10]. The GGA-PBE exchange-correlation functional is chosen, along with ultra-soft
nonlinear core corrections and scalar relativistic treatment for Zr, Co, and As elements. A 50Ry cutoff
energy was employed, and the integration of the Brillouin zone was executed utilizing a Monkhorst-Pack
grid with dimensions of 5x5x5. The structural properties of the ground state were determined by
conducting a geometry optimization procedure, which involved evaluating the total energy per unit cell
across various lattice constant values. The lattice parameters and other structural properties, providing
essential insights for device design and optimization, were determined. The optimization process involved
minimizing the total energy with respect to lattice parameters and fitting the data to the Birch–Murnaghan
equation of state[11]. The optimized parameters are further used to calculate the electronic, optical,
elastic, and dynamical properties of ZrCoAs following the procedure outline in references[12] and [13].
By comprehensively analyzing these properties, this study aims to shed light on the potential of as highly
efficient materials for photovoltaic applications. The results obtained from this research will contribute
to the advancement of sustainable energy technologies, addressing the global demand for clean and
renewable power sources.
3. Results And Discussions

3.1. Structure of ZrCoAs
The geometry optimization of the ZrCoAs half Heusler (HH) compound was carried out by calculating
the lattice parameter as a function of total energy. The optimized structure is shown in Figure 20a. The
cubic half-Heusler compounds ZrCoAs, has equal lattice parameters a = b = c and angles α = β = γ = 90°,
with its atoms located at Zr (1/2, 1/2, 1/2), Co (1/4, 1/4, 1/4), and As (0, 0, 0) Wyckoff positions. The
calculated Zr-Co bond length was 2.64 Å, this was found to be equal to Co-As bonds but slightly less
than Zr-As bonds which was obtained as 3.05Å. The calculated bond lengths were found to be in good
agreement with those for similar Zr based HH compounds studied by Allan et al(2023) [14], an indication
that the structure was accurately defined.
Page 739
SEMIT 2023
Figure 20: (a) Optimized structure (b) Energy vs lattice constant curve without SOC, and (c) Energy vs
lattice constant curve with SOC for cubic ZrCoAs
Figure 20b and c, respectively, shows the lattice constant (a0) and energy relationship. Subsequently, the
obtained data from energy lattice parameter relationship were fitted to the Birch–Murnaghan equation of
state [15]. The ground state properties of ZrCoAs after the fitting process yielded a lattice parameter of a
= 6.0238 Å and 5.954 Å respectively, without and with SOC effects. Additional ground state properties
obtained from the fitting are listed in Table 27.
Table 27: Lattice constants, its derivatives and electronic bandgap of ZrCoAs compared to similar Zr
based half Heusler compounds
Comp Calculated bulk modulus a0 Volume Emin Band Ref
(B in GPa) (Å) (Å)3 (eV) gaps
Birch -Murnaghan Thermo- (eV)
equation of state pw
ZrCoAs 143.7 107.9 6.023 49.75 -5660.23 1.0957 This work
Without
SOC
ZrCoAs 145.7 107.9 5.954 49.66 -5593.84 0.85 This work
With SOC
ZrCoSb 127.6 133.3 6.109 45.01 -7928.71 1.0765 [14]
ZrCoBi 114.8 207.6 6.199 46.25 -7932.11 1.0416 [14]
As illustrated in Table 27, we observe the calculated bulk modulus of 143.7 GPa, and the lattice constant
is 6.023 Å for ZrCoAs without SOC, which are inconformity with the reported values for ZrCoSb and
ZrCoBi [14]. Since the bulk modulus is a measure of a material's resistance to compression or volume
change under pressure. A higher bulk modulus indicates greater stiffness and a more compact structure.
In this case, ZrCoAs has the highest bulk modulus, followed by ZrCoSb and ZrCoBi.
The volume for ZrCoAs is approximately 49.75 Å3 without SOC and 49.66 Å3 with SOC, this conforms
with what was realised for ZrCoSb and ZrCoBi in ref [14]. The volume represents the size of the unit
cell and provides information about the material's density. In this case, ZrCoAs has the largest volume
among the three materials, while ZrCoSb and ZrCoBi have slightly smaller volumes. For the minimum
Page 740
SEMIT 2023
energy(Emin) and electronic band gaps, ZrCoAs has Emin of -5660.23 eV without SOC and -5593.84
eV with SOC. The corresponding band gap for ZrCoAs is 1.09 eV without SOC and 0.85 eV with SOC.
ZrCoSb and ZrCoBi, respectively, have been reported to have Emin of -7928.71 eV, with the band gap
being 1.0765 eV and Emin of -7932.11 eV, and the band gap is 1.04 eV. The minimum energy (Emin)
represents the equilibrium energy of the material, and the band gap indicates the energy range between
the highest occupied and lowest unoccupied electronic states. In this case, ZrCoSb and ZrCoBi have
lower minimum energies (an indication that they are more stable) and slightly smaller band gaps
compared to ZrCoAs. The significance and nature of the band gaps will be discussed in section 3.2.
3.2. Electronic properties
To gain insight into the electronic bandgap size of ZrCoAs, electronic structure calculations were
performed along the Г-X-K-Г-L-W-X high symmetry points, with and without SOC. From figure 2c, an
indirect W symmetry point (conduction band minima) and the L symmetry point (valence band maxima)
without SOC effects is observed. Figure 2a confirms the indirect gap, located between the Г symmetry
point (conduction band minima) and the L symmetry point (valence band maxima) with SOC effects.
When SOC is not included, ZrCoAs exhibits a band gap of 1.09 eV. However, when SOC is incorporated,
the band gap reduces to 0.85 eV. This reduction in the band gap is attributed to the spin-orbit splitting
caused by SOC. Spin-Orbit Coupling introduces energy corrections based on the interaction between the
electron's spin and its orbital motion, leading to a splitting of energy levels. In the presence of SOC, the
energy bands near the band edges shift differently for spin-up and spin-down electrons, breaking the spin
degeneracy and modifying the electronic structure as seen in Figure 21a. The SOC-induced band splitting
results in the change in the magnitude of band gap, thereby decreasing it. Even though the material retains
its indirect bandgap nature, the position of Valence Band Maxima and Conduction Band Minimum
changes, an indication that the SOC effect enhances the overlap between the valence and conduction
bands, making the material more favourable for certain electronic and optoelectronic applications[16],
[17].
It is therefore important to consider the SOC effect when studying and designing devices based on
ZrCoAs, especially those that rely on its electronic properties, such as solar cells, transistors, or other
energy-harvesting and electronic devices. SOC can significantly influence the material's behavior and
functionality, making its accurate inclusion in simulations and calculations crucial for a comprehensive
understanding of ZrCoAs' electronic properties.
Figure 21:Plots for (a) Electronic band structure with SOC effects (b) density of states with SOC
effects, (c) Electronic band structure without SOC, and (d) density of states without SOC as function of
energy for ZrCoAs.
Further analysis of the density of states with SOC effects demonstrated how different states from the
constituent elements contribute to the formation of the compound's conduction and valence bands.
In Figure 21d, it is evident that the valence band formation is predominantly influenced by Co 3d and Zr
3d states, with As 2p states making minimal contributions around the fermi level (EF) when SOC was not
incorporated in the computation. Conversely, the conduction band formation is primarily attributed to Zr
Page 741
SEMIT 2023
8d states, along with additional states from Co5d and As 3p when SOC is incorporated. The change in
dominant states contributing to the valence and conduction bands of ZrCoAs when SOC is incorporated
in the electronic structure calculations can be attributed to the relativistic effects introduced by SOC.
Without SOC, the electronic structure calculations consider only the static interactions between the
electrons and the atomic nuclei, neglecting the relativistic effects caused by the electron's motion relative
to the atomic nucleus and its own spin. In this case, the dominant states influencing the valence and
conduction bands are mainly determined by the 3d orbitals of cobalt (Co) and zirconium (Zr) atoms for
the valence band and the 4d orbitals of zirconium (Zr) atoms for the conduction band.
However, when SOC is incorporated into the calculations, the relativistic effects become significant. SOC
considers the coupling between the electron's orbital motion and its intrinsic spin angular momentum due
to the electron's charge and mass. This relativistic effect leads to spin-orbit splitting, which means that
the energy levels of the electronic states are affected differently for electrons with different spin
orientations.
As a result of SOC, the contribution of additional atomic orbitals becomes more pronounced. In the
conduction band, in addition to the dominant contribution from Zr 4d states, the Co 5d and As 3p orbitals
also play a more substantial role. The SOC-induced splitting modifies the electronic states near the band
edges, causing these additional orbitals to contribute to the conduction band formation.
The relativistic effects of SOC also impact the valence band, where the contributions of Co 3d and Zr 3d
states remain dominant, but the As 2p states contribute minimally around the Fermi level (EF).
Overall, the inclusion of SOC in the electronic structure calculations alters the electronic band structure
by introducing spin-orbit splitting, which changes the dominant atomic orbitals contributing to the
valence and conduction bands of ZrCoAs. SOC provides a more accurate description of the electronic
properties, especially in materials with heavy elements and strong spin-orbit coupling interactions, like
ZrCoAs. Given this, SOC effects were incorporated in the rest of the computations in this study.
Additionally, Table 27 compares the band gap with those of selected ZrCoSb/Bi compounds. To determine
which material is the most favourable for solar energy harvesting, we need to consider the material's band
gap and its compatibility with solar energy conversion processes. In solar energy harvesting, materials
with appropriate band gaps that can efficiently absorb solar photons and generate charge carriers
(electrons and holes) are preferred[18]. A suitable band gap allows the material to absorb a significant
portion of sunlight and enables efficient energy conversion. Based on the reported ans calculated band
gaps reported in Table 27, we can compare the band gaps of the materials. ZrCoAs (both with and without
SOC) has the widest band gap, which means it requires photons with higher energy (shorter wavelengths)
to be efficiently absorbed. On the other hand, ZrCoSb and ZrCoBi have slightly narrower band gaps,
making them capable of absorbing a broader range of solar photons. In solar energy harvesting, it is
generally more favourable to have a material with a band gap close to the optimal value for solar
absorption, which is around 1.1 to 1.4 eV. This range corresponds to the energy of visible light, where a
significant portion of solar radiation lies[19]. Among the three materials, ZrCoSb and ZrCoBi have band
gaps that are closer to the optimal range for solar absorption. Therefore, ZrCoSb and ZrCoBi are likely
to be more favourable for solar energy harvesting compared to ZrCoAs. However, it's important to note
that the material's band gap is just one factor and other material properties and engineering considerations
may also play a role in determining the overall performance and efficiency of a solar energy harvesting
device. Also, the Kohn-Sham DFT is well known to underestimate the electronic band gaps[20]. Among
the calculated optical properties of ZrCoAs in section 3.5, the optical gap will be discussed.
3.3. Mechanical Stability
The crystal structure of ZrCoAs half Heusler compound is characterized by a face-centered cubic
arrangement. Within the cubic crystal system, there are three independent elastic constants, namely C 11,
C12, and C44 [21]. The calculated values for these elastic constants were found to be C11 = 180.07 GPa,
C12 = 26.8 GPa, and C44 = 71.13 GPa, indicating that the ZrCoAs semiconductor is mechanically stable
according to the Born criteria[22]. The derived values using the Voigt–Reuss–Hill approximation for bulk
modulus (B), Young's modulus (E), shear modulus (G), and Poisson's ratio (ν) as presented in Table 28.
Page 742
SEMIT 2023
The Pugh's ratio (B/G) of 1.47 indicates that the material is ductile. Additionally, the Poisson's ratio (ν),
a mechanical property of materials that represents the ratio of lateral strain to axial strain when a material
is subjected to uniaxial stress, is a dimensionless quantity that characterizes a material's response to
deformation in different directions. From our calculations, the Poisson's ratio suggests that the compound
exhibits ionic properties. There results are in good agreement with reported Zr based half Heusler alloys
tabulated in Table 28.
Table 28: The derived values using the Voigt–Reuss–Hill approximation for bulk modulus (B), Young's
modulus (E), shear modulus (G), and Poisson's ratio (ν)
Comp C11 C12 C44 B(GPa) G(GPa) B/G E(GPa) v θD(K) Ref
ZrCoAs 180.1 26.8 71.1 107.9 73.2 1.47 167.3 0.14 403.39 This
work
ZrCoSb 262.3 68.1 67.7 113.3 79.9 1.66 199.8 0.25 393.12 [14]
ZrCoBi 322.2 150.3 52.7 207.6 64.3 3.23 174.7 0.35 312.75 [14]
3.4. Dynamical Stability

Lattice dynamics is the study of how atoms in a crystal lattice vibrate and interact with each other[23],
[24]. The phonon dispersion and Density of States (DOS) are key components of lattice dynamics analysis
and provide valuable insights into the thermal and vibrational behaviour of materials like ZrCoAs. The
phonon dispersion curve obtained for ZrCoAs represents the relationship between phonon frequencies
and wavevectors in the Brillouin zone as shown in Figure 22(a).
Different branches of the dispersion curve correspond to various phonon modes, and the dispersion along
each branch provides information about the material's vibrational properties. In Figure 22a at Г point of
the Brillouin zone there are three acoustic phonon modes, one for each crystallographic direction. These
modes have low frequencies and correspond to the collective vibrations of atoms in the crystal lattice.
They play a crucial role in determining the material's thermal conductivity, as they carry heat energy
through lattice vibrations[25]. On the other hand, the optical phonon modes are higher in frequency and
correspond to vibrations that involve the relative motion of atoms within the unit cell[25].
Figure 22b shows the Density of States (DOS) plot, which has the distribution of phonon states with
respect to their energies. It provides information about the number of phonon modes that are available at
different energy levels. The peaks in the DOS curve represent the allowed phonon modes, and the area
under the curve corresponds to the total number of phonon states. The phonon frequencies and modes
dictate the vibrational behaviour of the material. The presence of specific phonon modes can give rise to
characteristic vibrational patterns, influencing various properties.
Figure 22:(a) Phonon dispersion curves and (b) density of states for ZrCoAs
In this study, the phonon branches near the Brillouin zone boundaries are of particular interest. These
modes are associated with specific lattice distortions and can give insight into the stability and mechanical
properties of ZrCoAs. For example, soft phonon modes near the Brillouin zone boundary indicates a
phase transition or structural instability. Additionally, the absence of negative frequencies is a clear
Page 743
SEMIT 2023
indication that the ZrCoAs is dynamically stable. The lattice dynamics analysis based on the phonon
dispersion and DOS for ZrCoAs provides an understanding of its vibrational and thermal properties which
are crucial in designing and optimizing the material for various technological applications, such as
thermoelectric and optoelectronic devices.
3.5. Optical Properties
The optical properties of ZrCoAs, which are essential for understanding the material's suitability for
optoelectronic and photovoltaic applications, have also been investigated. The complex dielectric
function was employed to determine properties such as the refractive index (n), extinction coefficient (k),
absorption coefficient (α) and energy loss (L). The results are presented in
Figure 24 and Figure 25 and summarized in Table 29.
Figure 23 Frequency-dependent complex dielectric functions ε1(ω) and ε2(ω)
Page 744
SEMIT 2023
Figure 24: Frequency-dependent refractive index n(ω) and extinction coefficient k(ω)
Figure 25: Frequency-dependent absorption coefficient α (ω) and energy loss L(ω)
In
Figure 23, major peaks in the complex dielectric functions are observed between 1.9 eV and 4.5 eV, with
additional peaks at 6.1 eV and 7.3 eV. Regions where ε1(ω) < 0 indicated that incident photons are fully
attenuated. The refractive index and extinction coefficient, illustrated in
Page 745
SEMIT 2023
Figure 24, are obtained from the dielectric functions. They provide valuable information about the
material's optical behavior. From Figure 25, ZrCoAs is a good absorber of photon energy with very
minimal energy loss.
Table 29: Optical properties of ZrCoAs compared to selected Zr based half Heusler compounds
Comp ε1(0) ΔEOG n(0) R(0) L(0) Ref
ZrCoAs 22.67 1.78 2.560 0.19205 0.000352 This
work
ZrCoSb 19.7 1.36 4.435 0.399 0.000347 [14]
ZrCoBi 21.1 1.55 4.593 0.413 0.000344 [14]
Based on the reported results in Table 29, ZrCoSb has the lowest optical bandgap indicating that it is
more efficient in absorbing a wider range of solar photons compared to ZrCoAs with a bandgap of 1.78
eV. However, both ZrCoSb and ZrCoBi are reported to have refractive indices compared to the calculated
refractive index of ZrCoAs (2.560), suggesting better light bending and trapping abilities suitable for light
management in solar energy harvesting applications. ZrCoSb has the lowest reflectivity (0.399) among
the three compounds, indicating that it is less prone to reflecting incident sunlight. Lower reflectivity
enhances light absorption, making ZrCoSb more favorable for solar energy harvesting.
The differences in optical conductivity (L(0)) among the three compounds are relatively small. All three
materials have low optical conductivities, which is advantageous for solar energy harvesting as it reduces
energy loss due to absorption of light with high-frequency photons.
Based on the optical properties, ZrCoSb appears to be the most favorable material for solar energy
harvesting among the three compounds. Its relatively lower bandgap, high refractive index, low
reflectivity, and low optical conductivity suggest its potential for efficient light absorption and utilization
in solar cells or other solar energy conversion devices. None the less, the optical properties of ZrCoAs
are comparable to those of the other two reported Zr based HH compounds.
We are studying ZrCoAs and comparing it with ZrCoSb and ZrCoBi to evaluate their potential for solar
energy harvesting. The choice of these materials stems from their unique properties, specifically their
bandgaps, refractive indices, reflectivity, and optical conductivity, which directly impact their
performance in solar energy conversion. Such comparative analysis is vital for optimizing material
selection and design, contributing to the advancement of sustainable energy technologies and paving the
way for more efficient and environmentally friendly solar energy solutions.
4. Conclusion
In conclusion, the comprehensive study of ZrCoAs has provided valuable insights into its structural,
electronic, optical, lattice dynamical, and mechanical properties. The compound exhibits a stable crystal
structure, making it promising for practical applications. Electronic band structure analysis indicates
contributions primarily from Co 3d and Zr 3d states, influencing the material's electronic behavior with
an indirect bandgap of 1.0957 eV without SOC and 0.85 eV with SOC, suggesting its potential for solar
energy harvesting and optoelectronic applications. The lattice dynamical analysis reveals phonon
frequencies and modes, contributing to thermal and vibrational behavior, with particular interest in
phonon branches near the Brillouin zone boundaries.
In terms of mechanical properties, ZrCoAs exhibits a Young's modulus of 167.3 GPa, a shear modulus of
73.2 GPa, and a bulk modulus of 107.9 GPa, indicating its mechanical strength and resistance to
deformation. The material's B/G ratio of 1.47 suggests its ability to withstand volume changes compared
to shear deformation. Additionally, ZrCoAs has a relatively low Poisson's ratio of 0.14, signifying its
limited lateral contraction under tensile stress. ZrCoAs 's optical properties were excellent for
photovoltaic applications, indicating a potential for used in solar energy harvesting devices. Its optical
properties are comparable to those of the other two reported Zr based HH compounds, ZrCoSb and
ZrCoBi.
Page 746
SEMIT 2023
Overall, the collective investigation of these properties provides a comprehensive understanding of

ZrCoAs, offering opportunities for optimizing its performance in various technological applications.
These insights contribute to the potential use of ZrCoAs in solar cells, optoelectronic devices,
thermoelectric applications, and structural components, promoting advancements in renewable energy
and materials science fields. As researchers continue to explore and harness the potential of ZrCoAs, it
holds the promise of contributing to advancements in renewable energy technologies and materials
science because of its versatility and suitability for a wide range of practical applications.
Acknowledgment
The authors gratefully acknowledge the support of the Partnership for Skills in Applied Sciences,
Engineering and Technology (PASET)-Regional Scholarship Innovation Fund (RSIF), which provided
funding. The authors would also like to acknowledge the Centre for High-Performance Computing
(CHPC) in RSA for providing the necessary computing resources that were essential for conducting the
computational simulations and analyses in this research. The university of Pretoria is also acknowledged
for support.
Data Availability
The data that support the findings in this work are available upon reasonable request.
Author Contribution
Lynet Allan, R.E Mapasha, Winfred M. Mulwa, Julius M. Mwabora, and Robinson J. Musembi:
Conceived and designed the analysis, performed the calculations, contributed analysis tools, performed
the analysis, compiled the paper.
Deceleration of Interest
The authors have no conflict of interest to declare.
References
[1] O. Tahvonen and S. Salo, “Economic growth and transitions between renewable and
nonrenewable energy resources,” European Economic Review, vol. 45, no. 8, pp. 1379–1398, 2001.
[2] T. Graf, S. S. Parkin, and C. Felser, “Heusler compounds—A material class with exceptional
properties,” IEEE Transactions on Magnetics, vol. 47, no. 2, pp. 367–373, 2010, [Online]. Available:
https://ieeexplore.ieee.org/abstract/document/5654590/
[3] T. Graf, C. Felser, and S. S. Parkin, “Simple rules for the understanding of Heusler compounds,”
Progress in solid state chemistry, vol. 39, no. 1, pp. 1–50, 2011, [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S0079678611000021
[4] D. Chattaraj, S. Dash, and C. Majumder, “Structural, electronic, elastic, vibrational and
thermodynamic properties of ZrNi and ZrNiH3: A comprehensive study through first principles
approach,” International Journal of Hydrogen Energy, vol. 41, no. 44, pp. 20250–20260, Nov. 2016, doi:
10.1016/j.ijhydene.2016.09.046.
[5] S. R. Culp, J. W. Simonson, S. J. Poon, V. Ponnambalam, J. Edwards, and T. M. Tritt, “(Zr, Hf)
Co (Sb, Sn) half-Heusler phases as high-temperature (> 700° C) p-type thermoelectric materials,” Applied
Physics Letters, vol. 93, no. 2, p. 022105, 2008.
[6] H. Xiao et al., “Superconductivity in the half-Heusler compound TbPdBi,” Physical Review B,
vol. 97, no. 22, p. 224511, 2018, [Online]. Available:
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.97.224511
Page 747
SEMIT 2023
[7] D. Zhao, L. Wang, L. Bo, and D. Wu, “Synthesis and thermoelectric properties of Ni-doped
ZrCoSb half-Heusler compounds,” Metals, vol. 8, no. 1, p. 61, 2018, [Online]. Available:
https://www.mdpi.com/254598
[8] L. Jodin, J. Tobola, P. Pecheur, H. Scherrer, and S. Kaprzyk, “Effect of substitutions and defects
in half-Heusler FeVSb studied by electron transport measurements and KKR-CPA electronic structure
calculations,” Physical Review B, vol. 70, no. 18, p. 184207, 2004, [Online]. Available:
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.70.184207
[9] A. Jain et al., “Commentary: The Materials Project: A materials genome approach to accelerating
materials innovation,” APL materials, vol. 1, no. 1, p. 011002, 2013, [Online]. Available:
https://aip.scitation.org/doi/10.1063
[10] L. Talirz et al., “Materials Cloud, a platform for open computational science,” Scientific data,
vol. 7, no. 1, pp. 1–12, 2020, [Online]. Available:
https://scholar.googleusercontent.com/scholar.bib?q=info:7a2oXMCiqkkJ:scholar.google.com/&output
=citation&scisdr=CgVAeMUUEODwnOY1R3k:AAGBfm0AAAAAY34zX3mZg-xDb_isclxjTmw-
MxZsPutG&scisig=AAGBfm0AAAAAY34zX2ROAjMsOuesV-lG-
7ET9nxz17n9&scisf=4&ct=citation&cd=-1&hl=en
[11] V. Tyuterev and N. Vast, “Murnaghan’s equation of state for the electronic ground state energy,”
Computational materials science, vol. 38, no. 2, pp. 350–353, 2006, [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S0927025606000735
[12] L. Allan, W. M. Mulwa, J. M. Mwabora, R. J. Musembi, and E. Mapasha, “An Ab-Initio Study
of P-Type Zrcoy (Y= Sb and Bi) Half-Heusler Semiconductors,” Available at SSRN 4443887.
[13] L. Allan, W. M. Mulwa, R. J. Musembi, and B. O. Aduda, “First Principles Study of the
Structural, Mechanical, Electronic, and Lattice Dynamical Properties of the Half-Heusler Alloys ZrCoY
(Y= Sb, Bi),” arXiv preprint arXiv:2204.03759, 2022, [Online]. Available:
https://arxiv.org/abs/2204.03759
[14] L. Allan, W. M. Mulwa, J. M. Mwabora, R. J. Musembi, and E. Mapasha, “An Ab-Initio Study
of P-Type Zrcoy (Y = Sb and Bi) Half - Heusler Semiconductors.” Rochester, NY, May 10, 2023. doi:
10.2139/ssrn.4443887.
[15] L. Allan, R. Musembi, and B. Aduda, First Principles Study of the Structural, Mechanical,
Electronic, and Lattice Dynamical Properties of the Half-Heusler Alloys ZrCoY (Y=Sb, Bi ). 2022.
[16] A. M. Boldin, “Analysis of Spin Polarization in Half-Metallic Heusler Alloys,” Macalester
Journal of Physics and Astronomy, vol. 3, no. 1, p. 3, 2015, [Online]. Available:
https://digitalcommons.macalester.edu/mjpa/vol3/iss1/3/
[17] M. Levy, “Universal variational functionals of electron densities, first-order density matrices,
and natural spin-orbitals and solution of the v-representability problem,” Proceedings of the National
Academy of Sciences, vol. 76, no. 12, pp. 6062–6065, 1979.
[18] A. Zahedi, “Maximizing solar PV energy penetration using energy storage technology,”
Renewable and sustainable energy reviews, vol. 15, no. 1, pp. 866–870, 2011.
[19] X. Jiang, Z. Zang, Y. Zhou, H. Li, Q. Wei, and Z. Ning, “Tin halide perovskite solar cells: an
emerging thin-film photovoltaic technology,” Accounts of Materials Research, vol. 2, no. 4, pp. 210–219,
2021.
[20] E. S. Goh, J. W. Mah, and T. L. Yoon, “Effects of Hubbard term correction on the structural
parameters and electronic properties of wurtzite ZnO,” Computational Materials Science, vol. 138, pp.
111–116, 2017.
[21] J. Pokluda, M. Černỳ, M. Šob, and Y. Umeno, “Ab initio calculations of mechanical properties:
Methods and applications,” Progress in Materials Science, vol. 73, pp. 127–158, 2015.
Page 748
SEMIT 2023
[22] M. Born, “Thermodynamics of crystals and melting,” The Journal of Chemical Physics, vol. 7,
no. 8, pp. 591–603, 1939, [Online]. Available: https://aip.scitation.org/doi/abs/10.1063/1.1750497
[23] C. Çoban, K. Çolakoğlu, and Y. Ö. Çiftçi, “First principles study of the structural, mechanical,
phonon, optical, and thermodynamic properties of half-Heusler (HH) compound NbFeSb,” Physica
Scripta, vol. 90, no. 9, p. 095701, 2015.
[24] S. Califano, V. Schettino, N. Neto, S. Califano, V. Schettino, and N. Neto, Lattice Dynamics.
Springer, 1981.
[25] M. T. Dove, Introduction to lattice dynamics. Cambridge university press, 1993.
Page 749
SEMIT 2023
Performance Enhance and Management of Renewable Based Microgrid with backup Fuel Cell
Saqib Waris 1, Linta Khalil 2 Syed Mustafa Hassan Naqvi3, Kamran Liaquat Bhatti 4,
Muhammad Zahid Shafique 5, Malik Hamza Attiq 6
1
Polytechnic Institute Australia
2
NFC Institute of Engineering & Technology Multan, Pakistan
3
NFC Institute of Engineering & Technology Multan, Pakistan
4
University of Wollongong, Australia
5
Deakin University Waurn ponds.
6
Pakistan Engineering Council, Islamabad Pakistan
Saqibwaris17@gmail.com, khalillinta@gmail.com,
Mustafanaqvi8765@gmail.com,dr_mklbhatti@nfciet.edu.pk,
, zahidshafique3@gmail.com, malikhamzaattiq@gmail.com
Abstract
Solar and wind energy have proved to be the major sources, with fast growth rates in the last two to three
decades, of renewable energy for electricity generation. This study focuses on several studies of the new
integration of wind energy from a permanent magnetic synchronous generation system (PMSG) [1] with
a grid connected system. The aim is focused on to extract maximum power from wind and solar energy
system is a novel technique called incremental conductance method to get maximum power point
tracking (MPPT) [1], it gives faster responses to the changes in PV voltage and current. In addition, wind
and solar power supplies are integrated in an appropriate manner to always guarantee the continuity of
demand supplies. The overall operation of the wind generator is measured by the speed calculation,
basically a rotor speed estimator that is sensor free and prevents all mechanical sensors. Rotor speed has
been estimated for boosters for the control of the turbine speed by keeping voltage and current (input dc).
Furthermore, under critical conditions of demand and PVA variation on wind farm power, the test system
is supported by fuel cell resources.
Key words: Photo voltaic array (PVA), maximum power point tracking (MPPT), Permanent magnet
synchronous generator (PMSG), Synchronous reference frame (SRF), Photovoltaic (PV), Direct current
(DC).
1. Introduction
Countries all over the world are striving to come up with green energy solutions to secure the fossil fuel
reserves for future generations. The technology with the following behind wind and water is widely
accepted to be the best among renewables for developing future energy sources because it offers long-
term potential, especially in regards to electricity and thermal power, but also as regards alternative fuels.
Wind energy, in and on its own, is able to provide a lot of power, but is subject to extreme weather
conditions and is rare in areas of low vegetation. The light from the sunlight varies because of cloud
cover, buildings, trees, and such like, which prevents as much of it from being received by solar panels,
is not allowing some of the solar radiation to pass through to the sky. Also, they suffer from the common
issue of needing to be refueled, thus they are unable to provide constant power. wind- or photovoltaic
power (also known as solar power) is typically employs two or more different resources, typically wind
power and PV. Such a hybrid system possesses a considerable advantage when these two sources are
connected: using two sources of power increases the overall system's dependability at the load. Although
there is a significant amount of wind, the air is typically having a more prominent light ray of the sun.
When combining both the wind and solar systems, the effectiveness of their transfer of energy and their
operational stability can be improved. If one or more of the loads cannot be met or doesn't supply
sufficient energy, the other energy source can supplement the deficit. Several hybrid wind and solar
power systems that use conventional controllers for lower ratings are addressed in [2] in [the authors
address hybrid wind and PV systems that use several hybrid controllers] as an additional or
supplementary power system features to existing or previous power generation, consisting of wind
Page 750
SEMIT 2023
turbines and solar panels Wind power, connected to the grid through a Buck-Boost converter, is then
used to step-down to AC voltage for small-scale energy production. It is necessary to use maximum
MPPT in order to obtain maximum power output. in this approach, MPPT fuzzy logic controller are used
for wind turbine energy systems Fully intend Purb and Observe MPPT (Maximum power point tracking)
is used in addition to the method of measuring expansion to provide information on solar panels.
Figure 1 represents the architecture, as presented in the proposal's proposed block diagram. Buck-coupled
PMSG connected to the wind turbine is also follows a Diode Bridge, and then an Uncontrolled three-
three phase transformer, which is connected to a Grid-Power system diode-bridge, where a regulated-
three phase is applied. using MPPT, to increase voltage while limiting the number of duty cycles and
reduces cost by controlling DC/DC voltage using MPPT technique and limits pulse duration and
decreases cost with the ability to multiply voltage via (most effective) PWM (Permissible PWM (HCS
& FL). [3]
The photo-voltaic cell is connected to a boost converter to produce a higher voltage Since it was proven
that maximum power by using MPPT Pert and Observe method provides better results, it is now widely
used. Three-phase voltage is applied to this output voltage, which in turn goes to the AC grid voltage.
Fig. 1. Block diagram of Hybrid system (consisting of WECS and Solar PV System) [1]
Using photovoltaic cells and solar panels, this system enables light to be used to produce direct current
(DC). A solar panel is identical to a photovoltaic array, so when a panel or a set of solar cells is connected,
it can generate the same amount of power as one of connected panels. Connecting in a series tends to
yield a higher voltage, connecting in parallel is most often yields equal voltage, but with a decreased
current, and connecting in parallel is often yields lower current. Series is commonly selected because it
offers a greater voltage while having the same current as a parallel connection yields the same voltage
but a lower current. This is a specific instance of series-parallel circuits, for example where two panels
that are wired in series would double their voltage while the current remains the same.
The number of solar panels used in an array can be as few as few panels connected in an urban
environment or as well as many as in a grouping in the countryside. A flexible photovoltaic array (PV)
allows designers to develop solar systems to serve any size need.
It is important to note that photovoltaic panels from different manufacturers should not be mixed in the
same array. Even if their voltage and current outputs are within a couple of one another, the two modules
should not be interconnected. Because differences in the VI characteristic curves and their spectral
response may result will often contribute to mismatch losses in the array's efficiency, some guidelines
specify a different rule for linear arrays. The photo-voltaic array has various electrical characteristics,
mechanical, and optical features, but they all produce excess light. In the case of a photovoltaic array,
the output current is given as a relationship between the voltage and the current output. the amount and
intensity of solar irradiance (sunlight) controls the current output of the PV array, and the temperature of
Page 751
SEMIT 2023
the PV arrays cells does the same. linear voltage-exponential current relationships can be given by the
manufacturers which may also be termed current intensity vs. current curves, or current intensity vs.
current or voltage and current response functions adjustable area, in expansible field. This is open circuit
voltage: the array provides when the terminals are not connected to a load. This parameter has a fixed
load operating limit setpoint voltage (Vmax) that is greater than the maximum which describes the array
voltage. The capacity of this solar power system is affected by the number of PV panels in a series. The
ISC, when the output connectors are shorted together. This circuit capacity is a lot greater than the design
current value, which is referred to as Imax.
This relates to the point where the array is connected to the load where power is supplied by the battery
is at its maximum, where the power point of a photovoltaic array is given in Watts (W), which measures
the maximum power output (Wp). the fill factor is the maximum power the array can provide under
normal conditions and short-circuit current times open circuit ( Voc x Isc ). This fill factor provides an
idea of the quality of the array and how close the fill factor is to unity, the more power the array can
generate. on the RMS-expanded values are between 0.7 and 0.8 RMS values. Solar arrays have
efficiencies ranging from 10-12 percent at monocrystalline, polycrystalline, amorph, and thin-film.
photovoltaic I-V characteristics give designers an opportunity to control system designs that provide the
maximum power handling capability within tolerable operational parameters. the power point is equal to
the module's maximum power output when exposed to 1000 W/m2 of radiation.
According to the PVA maximum power point tracking method, a new pulse is produced by the
incremental conductance. The Voltage and Current source methods are the same, except that with respect
to how much information they are able to provide. In order to reach the current value of power, a
comparison is made between the previous level of power and the new power required (maximum power
point). The equations involved in MPPT is provided for use in the output for when they can be found.
(1)
In PMSG wind farm ensuring that carbon emissions are kept to a minimum could be reduced by utilizing
renewable sources such as wind energy that harnessing a sustainable wind power, as with new,
environmentally friendlier techniques such as the wind turbines [1], or by using environmentally friend
wind technologies like turbines [1] In contrast to that regard to conventional energy sources, wind power
is not contributing to our greenhouse emissions as it is to our environment because it does not produce
those gases. Even with recent progress in conservation and the cost of other low-carbon energy sources,
such as nuclear, there are still problems with wind energy, and its cost has not yet fallen."
Various types of wind turbine generators exist, for the various speeds of wind. Some sources note that
while DFIG is more commonly used nowadays than PMSG, it offers advantages recognized advantages
for the double-feed process (DFIG). PMSG has a few advantages. PMSG in particular is simple, without
a rotor current, is rigidly rotates at low speed, direct, and without a gearbox. Fewer costs are anticipated
because high-efficiency and low-maintenance assets are expected to be obtained. While there are
advantages to PMSG, there are also some disadvantages. To expand, it must have the highly flexible
structured electromagnetic field, necessitating the excellent standard of manufacturing and operation. It
is important to know about the generator's power inverter, as well, because it changes the generator's
speed.
In this experiment, the entire wind system will be modelled, including the physical components of a
synchronous generator and its control strategy for PMSG effects and strategy for grid synchronous
generators and permanent magnet based PMSGs, as well as the pitch controllers, will be incorporated
into the simulation. The model system and the implementation strategies include a PMSG wind turbine,
a pitch model, an inverter control generator side, and a controller; also, a control method to adjust the
grid voltage called Grid-scale Inverter active. Generator side PWM back-BV back vector control and
Page 752
SEMIT 2023
decoupling is employed, alongside the addition of an independent axis-axis of power inverter/ power
inverter with side decoupling controls MPPT in order to better MPPT performance and decoupling of
the reactant side with it The Additionally, the traditional PI controller is used to further the controller
strategy by broadening its range of functionality. The generator has three integrated inverter controllers,
which are available as an option. The experiment is performed using Simulink 2016a from which these
simulation results are derived is completed with Simulink 2012b for development and performance
analysis. The findings show the influence and effectiveness of the centralized PMSG on the state of wind
turbines with these innovations.
It is shown in Figure 2 that the basic structure of a wind turbine diagram is represented. Because it turns
the rotor, the wind turbine produces a rotational torque. Direct the output of the generator's engine via
the drive shaft to the generator rotor. the generator adds electrical torque to or subtracts it from the
mechanical torque generated by the wind turbines, determining the effect on the vehicle's speed.
Fig. 2. General wind turbine PMSG system with control schemes [1]
In grid side inverter controller, the objective of the grid-side inverter is to preserve stability and to control
the active and reactive power [4] of the DC line voltage [5,6]. The model is given as follows for
grid/transformer inductance:
(2)
Here, is the grid's -axis output voltage, is the grid's angular frequency expressed in electrical degrees, is
the grid's resistance, is the grid's inductance, and and are the grid's -axis and -axis currents, respectively.
The current of -axis and -axis can be regulated to control active and reactive power. The loop voltage
and current are depicted in Figure15. Similar to the generator-side inverter controller, the inner current
loop is controlled via a PI controller. The output voltage loop generates a PI controller that is used to
calculate the difference between and to generate. As a result, the -axis current is set to zero in order to
decouple active and reactive power control by moderating the -axis current and the -axis current.[7]
Page 753
SEMIT 2023
Fig. 3. Schematic diagram of grid-side inverter controller.[7]
A fuel cell is an electrochemical cell that converts fuel to electricity via an electrochemical reaction. To
sustain the reactions that generate electricity, these cells require a constant supply of fuel and an oxidizing
agent (generally oxygen). As a result, these cells can continue to generate electricity indefinitely until
the supply of fuel and oxygen is interrupted. Despite being invented in 1838, fuel cells did not enter
commercial use until the early twentieth century, when NASA began using them to power space capsules
and satellites. Today, these devices serve as the primary or backup power source for a variety of facilities,
including industries, commercial buildings, and residential structures. Similar to electrochemical cells, a
fuel cell consists of a cathode, an anode, and an electrolyte. The electrolyte in these cells enables protons
to move.
1.1. Different Types of Fuel Cells:
Despite their similarity in operation, fuel cells come in a wide variety of configurations. This subsection
discusses several of these different types of fuel cells.
1.1.1. The Fuel Cell with a Polymer Electrolyte Membrane (PEM)
Additionally, these cells are referred to as proton exchange membrane fuel cells (or PEMFCs). These
cells operate between 50˚C and 100˚C. The electrolyte in PEMFCs is a polymer capable of conducting
protons. A PEM fuel cell is typically composed of bipolar plates, a catalyst, electrodes, and a polymer
membrane. Despite their environmentally friendly transportation applications, PEMFCs can also be used
for stationary and portable power generation.
1.1.2. Fuel Cells Based on Phosphoric Acid
These fuel cells utilize phosphoric acid as an electrolyte to conduct the H+. These cells operate at
temperatures between 150˚C and 200˚C. Due to the non-conductive nature of phosphoric acid, electrons
are forced to travel to the cathode via an external circuit. Because the electrolyte is acidic, the components
of these cells corrode or oxidize over time.
1.1.3. Sodium Hydroxide Fuel Cell
These fuel cells use a solid acid material as the electrolyte. At low temperatures, the molecular structures
of these solid acids are ordered. At elevated temperatures, a phase transition occurs, resulting in a
dramatic increase in conductivity. CSHSO4 and CH2PO4 are examples of solid acids (cesium hydrogen
sulphate and cesium dihydrogen phosphate respectively)
1.1.4. Fuel Cells Made of Alkaline
This was the primary source of electricity for the Apollo space program's fuel cell. In these cells, an
aqueous alkaline solution is used to saturate a porous matrix that serves as a barrier between the
electrodes. These cells operate at relatively low temperatures (approximately 90˚C). These are extremely
efficient cells. Along with electricity, they generate heat and water.
Page 754
SEMIT 2023
1.1.5. Fuel Cells Based on Solid Oxide

These cells utilize a solid oxide or a ceramic electrolyte as their electrolyte (such as yttria-stabilized
zirconia). These fuel cells are extremely efficient and relatively inexpensive (theoretical efficiency can
even approach 85 percent).
These cells operate at extremely high temperatures (lower limit of 600˚C, standard operating
temperatures lie between 800˚C and 1000˚C). Due to their high operating temperatures, solid oxide fuel
cells are limited to stationary applications.
1.1.6. Fuel Cells Based on Molten Carbonate
Lithium potassium carbonate salt is used as the electrolyte in these cells. At elevated temperatures, this
salt becomes liquid, allowing the carbonate ions to move freely. Like SOFCs, these fuel cells operate at
a relatively high temperature of 650˚. Corrosion is a possibility on both the anode and cathode of this
cell due to the high operating temperature and the presence of the carbonate electrolyte. Carbon-based
fuels such as natural gas and biogas can be used to power these cells.
1.1.7. Fuel cell applications
Fuel cell technology is extremely versatile. Currently, extensive research is being conducted in order to
develop a cost-effective fuel cell-powered automobile. A few examples of this technology's applications
are listed below. Fuel cell electric vehicles, or FCEVs, run on clean fuel and are thus more
environmentally friendly than vehicles powered by internal combustion engines. They have been used to
power numerous space missions, most notably the Apollo space programmer. Generally, these cells
produce heat and water as byproducts. The portability of certain fuel cells is critical in certain military
applications. Additionally, these electrochemical cells can be used to power a variety of electronic
devices. In many remote areas, fuel cells are also used as primary or backup sources of electricity.
2. Proposed Methodology
The proposed method is consisting of combination PVA, wind turbine and Fuel cell, PVA is connected
to the boost converter and then connected to MPPT to extract maximum power generated from the.
Windfarm is connected at the DC bus, inside the wind turbine there is PMSG (permanent Magnet
Synchronous Generator), which connected to the diode bridge rectifier and then it connected buck boost
converter. Diode bridge rectifier (DBR) will convert variable AC to variable DC, but variable DC cannot
be directly connected to the system .Hence by using buck boost converter stabilizing the Dc voltage by
500 V with help of fuzzy logic controller MPPT technique, Fuzzy logic MPPT controller take feedback
of the power generated by the windfarm and the speed of the windfarm and its generated the pulses to
IGBT .Now the windfarm will generate the power at 500V and PVA also generate the power at 500V.
Temperature and irradiance are two of two inputs for the PV model. If the input voltage and the output
wattage remain constant, the PV power does as well. The variables in real life don't stay constant; in fact,
they are constantly changing from time to time. thus, so the PV output is unable to maintain constant
output in shading conditions the primary operating modes of the PV model are divided into two
categories. If temperature and irradiance is kept constant, the x-expansion will take place in the first
instance. If the second case were to hold true, the temperature is held constant. The Simulink model is
created by using MATLAB 2016a, The Simulink model of the combination of PVA and wind turbine as
shown in the figure 4, and the combination of PVA, WIND and FUEL cell is connected as shown in the
figure 5. When the critical condition if the PVA and wind farm power generation failed there is a chance
of breakdown in order to avoid the damage to the system and providing uninterrupted supply at load side
there is connected fuel cell across dc bus, whenever the critical condition if PVA and wind farm failed
then the fuel cell will be activated and inject power to the grid system.11kW load is connected at
consumption end.
Page 755
SEMIT 2023
Parameters of solar PV array

The design parameters of PVA and wind farm is shown in below tables, the Simulink models developed
according the parameters given in the tables 1 and 2 below:
Table 1. Solar PV module Parameters [1]
Solar PV array Parameters Value
Rated Power (W) 600

Open circuit voltage (V) 63.6
Short circuit current (A) 12.5
Optimum voltage (V) 42.4
Optimum current (A) 6.75
The parameters for PMSG and Wind are shown in table 2.
Table 2. Wind Turbine PMSG Parameters [1]

Wind Turbine Parameters Value
Rated Power (MW) 2.3
Blade diameter (m) 71
Rated wind speed (m/s) 14
Number of blades 3
Turbine Inertia Jr (kg.m2) 670
PMSG Parameters Value
Rated Power (MW) 2
Rated Voltage (V) 690
Rated Stator Frequency (Hz) 11.25
Number of poles P 30
Stator Winding resistance Rs(mȍ) daxis 0.73051
Synchronous Inductance Ld (mH) q-axis 1.21
Synchronous Inductance 2.31
Lq (mH) 4.696 1.2
Flux Leakage (V.s) 2.356
Stator Leakage Inductance Ls (mH) 8000
Rated Rotor Speed (rad/sec) 14.15
Generator Inertia JG (kg.m2) Peak line to
neutral back emf constant
Km(V/rpm)
Page 756
SEMIT 2023
Fig. 4. Simulation diagram of combined PVA and wind turbine
Fig. 5. Simulation Diagram of combined PVA, Wind turbine and fuel cell
3. Simulation Results
The proposed system the capacity of the PVA is 100 kW, the capacity of the wind farm is 50kW and the
capacity of the fuel cell is 30 kW. The windfarm and PVA generates at 500V (DC), The maximum power
is injected to the grid is 180kW when the solar irradiation is very high, and the wind farm is having
optimal speed of 12-14 m/s and the fuel cell is operating at full scale. Total power of combination of
PVA and wind is generated during the optimal conditions is about 140 kW as shown in the figure 6. The
total power generated during optimal conditions of PVA, windfarm and the fuel cell id 180kW as shown
in the figure 7. Power generated from the fuel cell is 30kW as shown in the figure 8. figure 9 shows the
total power generated by fuel cell operating at full scale. Figure 10 represents the total power
consumption of the is 11kW, figure 11 is the total power injected to grid system. Figure 12 represents
Page 757
SEMIT 2023
the solar irradiation from morning to evening it starts slowly from 6am in the morning and deceases in
the evening around 6pm, the temperature given is 26 Degree Celsius. DC voltage maintained at 500V
and the VI characteristics of the total system as shown in figure 13.
Whenever during the critical conditions if PVA source and the Windfarm fails the fuel cell will be
activated and start injecting power to the grid system otherwise the system leads to failure. If system
failed during critical condition the power injected to the Grid system will be zero.
Fig. 6. Total power generated by PVA and Wind turbine
Fig. 7. Total power generated by PVA, Wind turbine and Fuel cell
Page 758
SEMIT 2023
Fig. 8. Power generated by fuel cell
Fig. 9. Total power consumed by load
Page 759
SEMIT 2023
Fig. 10. Total power injected to grid system
Fig. 11. Solar irradiation and temperature
Page 760
SEMIT 2023
Fig. 12. Dc volage maintained at 500v
Fig. 13. Voltage and current
In this proposed system there is no system failure. In critical condition if the PVA and Wind turbine fails
Fuel cell will be activated and power will be injected to the grid, hence no grid failure even in critical
conditions. Extraction of maximum power from PVA by using incremental conductance MPPT. Simple
and less complexity in implementation.
4. Conclusion
This paper illustrates how different methods can be used to combine a ground-turbine solar PV (or
ground-based PV) system with a grid-connected PMSG wind system. Simulation results vary depending
on different cases are given for different cases connected to each other. The simulation results are verified
using the MATLAB Simulink platform. The rate of expansion of the energy delivered by the modules is
graphed using time on the y-axis. To keep track of the rate of working and atmospheric exposure over
time, different variations are established in the test system, including solar irradiation and wind speed.
The fuel cell provides a second power source to assist when wind and solar power have failed or not
enough electricity has been generated.
Page 761
SEMIT 2023
5. Future Recommendation
Combining renewable sources like PVA, Wind turbine along fuel cell as a backup source. In simulation
where the critical condition if PVA and windfarm fails the whole system will failure which leads to
system inefficiency. In this research approach if these two-failure fuel cell will get activated and start
injecting power to grid, which can control the system failure. And also I have used to incremental
conductance technique to extract more power than regular MPPT technique, this can be still optimized
and extract the more power by using optimal control techniques. In order to make system more efficient
and better we have to focus future research on energy management control techniques. Enhancement of
energy management control with more optimization techniques like particle form optimization, fire play
algorithms, adaptive fuzzy, adaptive fuzzy pi, adaptive PI can help to system to become more stable and
efficient.
These techniques will play a vital role when the critical power failures and avoid failure of grid system.
References
[1] Lahari N V and H. Vasantha Kumar Shetty “Integration ‫خ‬f Grid Connected Pmg Wind Energy ‫ش‬nd
Solar Energy Systems Using Different Control Stratagies,” 2015 International Conference on Power
and Advanced Control Engineering (ICPACE), 978-1-4799-8371-1/15/$31.00 c 2015 IEEE.
[2] H. H. El-Tamaly, Adel A. Elbaset Mohammed “Modeling and Simulation of Photovoltaic/Wind
Hybrid Electric Power System Interconnected with Electrical Utility” 978-1-4244-1933-3/08/$25.00
©2008 IEEE.
[3] Divya Teja Reddy Challa1, I. Raghavendar 2 “Implementation of Incremental Conductance MPPT
with Direct Control Method Using Cuk Converter”, Vol.2, Issue.6, Nov-Dec. 2012 pp-4491-4496,
SSN: 2249-6645.
[4] Ting-Chung Yu Yu-Cheng Lin “A Study on Maximum Power Point Tracking Algorithms for
Photovoltaic Systems”,2010
[5] "Modeling of a Fuel Cell with PI Controller in Simulink - Matlab", The 2016 3rd International
Conference on Systems and Informatics, pp. 160-165, 2016.
[6] Dr. Abu Tariq1, Mohammed Asim2 and Mohd.Tariq3 “Simulink based modeling, simulation and
Performance Evaluation of an MPPT for maximum power generation on resistive load”, IPCBEE
vol.6 (2011) © (2011) IACSIT
Press, Singapore
[7] Chia-Nan Wang, Wen-Chang Lin, and Xuan-Khoa Le,” Modelling of a PMSG Wind Turbine with
Autonomous Control”, Published 27 May 2014.
Page 762
SEMIT 2023
SP21: Inventory Control, Production Planning and Scheduling
A New Integrated Criteria Weighting Approach for Determining Production Priority in Demand
Management
Elif Kılıç Delice1[0000-0002-3051-0496] and Elif Akviran2[0000-0003-2776-2374]
1
Atatürk University, Erzurum, 25240, Turkey
2
A101 Yeni Mağazacılık A.Ş., İzmir, Turkey
elif.kdelice@atauni.edu.tr
Abstract
In the demand management process; it is tried to balance the customers' needs with the company's
supply possibilities. Effective demand management starts with effective stock control and tracking
of customer orders and ends with customer satisfaction. Correctly determining production priorities
is one of the ways to increase customer satisfaction. In this context, a new approach has been
developed to determine the production priority needed in the demand management process. A new
integrated criteria weighting approach based on Multi-Criteria Decision Making (MCDM) methods
is proposed to determine production priorities. In the first stage of this three-stage approach,
customer alternatives are determined. In the second stage, production priority criteria are
determined depending on customer characteristics. In the third stage; with LBWA (Level Based
Weight Assessment) and CRITIC (The Criteria Importance Through Intercriteria Correlation)
methods, objective and subjective criteria weights are calculated by considering expert opinions
and real data. Then, integrated criteria weights are determined with the SOWIA (Subjective and
Objective Weight Integrated Approach) method using objective and subjective criteria weights. In
the literature; few studies determine production priority using the MCDM method. Additionally,
no study has been found in the literature in which subjective and objective weighting methods are
used integrated in the ceramic industry. Therefore, it is thought that this study will contribute to
the literature.
Keywords: Production Priority, LBWA, CRITIC, SOWIA.
1. Introduction
In today's world, there are many businesses that have decided to downsize due to economic difficulties.
Contrary to the decision of businesses to downsize, the increasing population and the resulting increased
demand caused the returns to customers to slow down. In order for the business not to decide to downsize
and to survive in the face of economic difficulties, it must use its limited production resources in a way
that will achieve maximum efficiency and increase its profitability, considering the interests of the
business. Managers must increase customer satisfaction and manage demand effectively, considering
limited production resources and business profitability. However, determining production priorities in
demand management is extremely important. When making this decision, it has been observed that most
businesses avoid using too many mathematical methods in real life and act based on the experience of the
decision maker. However, depending on the knowledge and understanding of decision makers, the
decisions taken may not always be correct. Customers' real data should also be considered in these
decisions. Therefore, within the scope of the study, a new solution approach based on Multi-Criteria
Decision Making (MCDM) is proposed, which allows businesses to easily determine production priorities
by benefiting from both the experiences of decision makers and real data about customers [1].
MCDM can be defined as the decision maker's selection process using at least two criteria from a set
consisting of a finite countable or an infinite number of uncountable alternatives, or in other words,
making a choice by evaluating based on two or more criteria [2]. In decision-making problems where
Page 763
SEMIT 2023
there are many alternatives and criteria, MCDM methods are used in order to keep the decision
mechanism under control and to reach the decision as quickly and easily as possible [3]. When the studies
in the literature are examined, it is seen that MCDM methods are grouped under two headings
"Weighting" and "Ranking". In this study, a new integrated criteria weighting approach is proposed by
using LBWA (Level Based Weight Assessment), CRITIC (The Criteria Importance Through Intercriteria
Correlation) and SOWIA (Subjective and Objective Weight Integrated Approach) methods, which are
MCDM weighting methods.
In the proposed integrated criteria weighting method; the LBWA method was chosen to weight the
criteria to be used in determining demand priority based on expert opinions. The LBWA method was
developed by Žižović and Pamučar [4] and was used to determine weight coefficients for evaluating
supplier performances. LBWA, a subjective weighting method based on the principle of pairwise
comparison, achieves importance weights by grouping decision criteria and evaluating the criteria in each
group with a simple algorithm. The CRITIC method proposed by Diakoulaki et al. [5] was developed due
to the need for an objective criteria weighting method and was used in the selection of Greek
pharmaceutical companies. CRITIC eliminates the influence of decision makers on the decision and
produces results by acting directly on the data. The SOWIA method was first used by Das et al [6] along
with the MOORA (Multi-Objective Optimization by Ratio Analysis) method for performance evaluation
and ranking of seven Indian Institutes of Technology. SOWIA, on the other hand, is a combined
weighting method that allows a single weight to be obtained by considering both objective and subjective
criteria weights. With the integrated approach created using these three methods, the importance weights
of the criteria to be considered in determining the production priority were determined and their order of
importance was revealed. Customer demand priority can be determined by using the order of importance
of the criteria.
In a company operating in the ceramics industry, an integrated criteria weighting approach consisting
of three stages was applied to determine production priority according to customer characteristics. The
ceramics company exports floor and wall ceramics, glazed porcelain and technical granite. Production
demands brought about by the rapidly increasing customer portfolio in recent years have remained above
the operating capacity. This situation has created the need to prioritize customer demands.
MCDM literature is quite comprehensive, but the number of studies in which production priorities are
determined using the MCDM method is few. In addition, no other study has been found in the literature
in which subjective and objective methods are considered together in-demand prioritization and criteria
weights are determined. Therefore, it is thought that this study will contribute to the literature.
2. Methods
2.1. LBWA method
The LBWA method, which is based on weighting the relationships between criteria in line with expert
opinions, is presented as a subjective weighting method. The LBWA method was developed because
there was a need for a method based on the principle of a small number of pairwise comparisons and
containing rational expressions when other subjective criteria weighting methods were examined. The
method was developed for four purposes [4]:
1. Comparing a small number of criteria (n-1).
2. Creating a model independent of the number of criteria with a simple algorithm.
3. Defining a model in which criteria weights give reliable results for rational evaluation.
4. Developing a model that can be easily explained to decision makers so that problems can be
easily solved.
When determining criteria weights with the LBWA method, the most important criterion is determined
first. For example; among the 𝑆 = {𝐶1 , 𝐶2 , … , 𝐶𝑛 } criteria, the 𝐶1 criterion was chosen by experts as the
most important criterion in the decision-making process. Decision-makers create subsets of criteria:
Page 764
SEMIT 2023
Level S1: The most important criterion 𝐶1 must be either equally essential or at most twice as important
as the criteria in this group.
Level S2: The most important criterion 𝐶1 must be at least 2 times and at most 3 times more important
than the criteria in this group.
Level S3: The most important criterion 𝐶1 must be at least 3 times and at most 4 times more important
than the criteria in this group.
…
Level Sk: The most important criterion 𝐶1 must be at least k times more important and at most k+1 times
more important than the criteria in this group (Demir, 2020).
By applying the rules mentioned above, decision-makers determine the rough classification of the
criteria and group the criteria according to their level of importance. If the significance of the criterion 𝐶𝑗
is denoted by 𝑠(𝐶𝑗 ), where 𝑗 ∈ {1,2, … , 𝑛}, then we have 𝑆 = 𝑆1 ∪ 𝑆2 ∪ … ∪ 𝑆𝑘 , where for every level
𝑖 ∈ {1,2, … , 𝑘}, the Equation (1) applies.
(1)
Within the formed subsets (levels) of the influence of the criteria it is performed the comparison of
criteria by their significance. Each criterion 𝐶𝑖𝑝 𝑆𝑖 in the subset is assigned with an
integer 𝐼𝑖𝑝 ∈ {0,1,2, … , 𝑟} such that the most important criterion 𝐶1 is assigned with 𝐼1 = 0, and if 𝐶𝑖𝑝 is
more significant than 𝐶𝑖𝑞 then 𝐼𝑝 < 𝐼𝑞 , and if 𝐶𝑖𝑝 is equivalent to 𝐶𝑖𝑞 then 𝐼𝑝 = 𝐼𝑞 . Where the maximum
valueon the scale for the comparison of criteria is defined by applying the expression (2)
𝑟 = 𝑚𝑎𝑥{|𝑆1 |, |𝑆2 |, … , |𝑆𝑘 | } (2)
The elasticity coefficient is obtained according to the r value determined according to Equation (2). 𝑟0 ,
defined as the elasticity coefficient, is determined by the 𝑟0 > 𝑟 (𝑟0 ∈ 𝑅). The influence function is
defined as in Equation (3).
𝑟0
𝑓 (𝐶𝑖𝑝 ) = 𝑖∙𝑟 +𝐼 (3)
0 𝑖𝑝
Where i is the level number in which the criterion is classified; 𝑟0 is the elasticity coefficient; 𝐼𝑖𝑝
indicates the integer value assigned to criterion 𝐶𝑖𝑝 within the relevant level. Using Equation (4), first,
the weight of the most important criterion is found. Then, the weight of the remaining criteria is obtained
using Equation (5).
1
𝑤1 = 1+𝑓(𝐶 (4)
2 )+⋯+𝑓(𝐶𝑛 )
𝑤𝑗 = 𝑓(𝐶𝑗 ) ∙ 𝑤1 (5)
Where j=2,3,…,n, and n present total number of criteria.
2.2. CRITIC method
The CRITIC method, which is used to find criterion weights, is among the objective weighting methods
because it reaches the result by directly processing the data. When calculating the criterion weight, both
the standard deviation of a criterion and its correlation with other criteria are considered. When
calculating the weights, decision-makers have no influence on the calculation; importance weights are
determined with simple statistical calculations [7]. The steps to calculate criterion weight with the
CRITIC method are as follows:
The initial decision matrix is created as in Equation (6) using numerical data regarding the criteria to be
used in decision making.
Page 765
SEMIT 2023
𝑥11 𝑥21 ⋯ 𝑥𝑖1

𝑥12 𝑥22 ⋯ 𝑥𝑖2
⋯ ⋯ ⋯ ⋯
𝑋= 𝑥 (i=1, 2, …, m and j=1,2,…,n) (6)
1𝑛 𝑥2𝑛 ⋯ 𝑥𝑖𝑛
⋯ ⋯ ⋯ ⋯
[ 𝑥𝑚1 𝑥𝑚2 ⋯ 𝑥𝑚𝑛 ]
Where xij represents the altenative performance value of ith alternative for jth criterion. Benefit and cost
criteria are standardized with the help of linear normalization. Equation (7) and Equation (8) are used to
normalize the benefit and cost criteria, respectively, and a normalized decision matrix is created.
𝑥𝑖𝑗 −𝑥𝑗𝑚𝑖𝑛
𝑟𝑖𝑗 = 𝑥 −𝑥𝑗𝑚𝑖𝑛
(7)
𝑗𝑚𝑎𝑘
𝑥𝑗𝑚𝑎𝑘 −𝑥𝑖𝑗
𝑟𝑖𝑗 = 𝑥 −𝑥𝑗𝑚𝑖𝑛
(8)
𝑗𝑚𝑎𝑘
Where 𝑟𝑖𝑗 is the normalized value of the ith alternative on jth criterion. The standard deviation of the
criteria is obtained with Equation (9).
2
∑𝑛 ̅̅̅̅)
𝑖=1(𝑟𝑖𝑗 −𝑟
𝜎𝑗 = √ 𝑛−1
𝑖𝑗
(9)
Where ̅̅̅
𝑟𝑖𝑗 shows the average value of 𝑟𝑖𝑗 and 𝜎𝑗 shows the standard deviation of the jth criterion.
The correlation coefficient is calculated using the values obtained as a result of normalization in
Equation (10). 𝑃𝑗𝑘 indicates the correlation value between criterion j and criterion k.
∑𝑚
𝑖=1(𝑟𝑖𝑗 −𝑟̅𝑗 )∙(𝑟𝑖𝑘 −𝑟
̅̅̅)
𝑘
𝑃𝑗𝑘 = (10)
2
√∑𝑚 𝑚
𝑖=1(𝑟𝑖𝑗 −𝑟̅𝑗 ) ∙∑𝑖=1(𝑟𝑖𝑘 −𝑟
̅̅̅)
𝑘
2
The criterion weight is calculated according to Equation (11). When specified as 𝜎𝑗 ∙ ∑𝑛𝑘=1(1 − 𝑃𝑗𝑘 )=𝐶𝑗 ,
it can be simplified as stated in Equation (12).
𝜎𝑗 ∙∑𝑛
𝑘=1(1−𝑃𝑗𝑘 )
𝑤𝑗 = ∑𝑛 𝑛 (11)
𝑘=1(𝜎𝑗 ∙∑𝑘=1(1−𝑃𝑗𝑘 ))
𝐶𝑗
𝑤𝑗 = ∑𝑛 (12)
𝑘=1 𝐶𝑘
2.3. SOWIA method

Criteria weights obtained by subjective and objective weighting methods are integrated with the SOWIA
method in 3 steps. The initial decision matrix created in the first step is normalized. In the second step;
criteria weights are determined using the normalized matrix according to the established MCDM method.
Since normalization and criteria weight determination calculations will be different for each method,
special formulations are not given in the first and second steps. For example, Zaher [8] used the Entropy
method as the objective method and the RR (Rank Reciprocal) weighting method, which is based on the
decision maker interpreting and ranking the criteria in order of importance, as the subjective method. If
the objective and subjective criteria weights are available to the decision maker, steps 1 and 2 should be
skipped and the method should be continued directly from the third step.
In the third step of the SOWIA method; The integrated criterion weight 𝑊𝑗𝑖𝑛𝑡 is calculated using
Equation (13).
𝑊𝑗𝑖𝑛𝑡 = 𝛼 ∙ 𝑂𝑗 + (1 − 𝛼) ∙ 𝑆𝑗 (13)
Where 𝑂𝑗 and 𝑆𝑗 are the criteria weights obtained by subjective and objective weighting methods. α is
the objective decision weight and takes a value in the range (0,1).
Page 766
SEMIT 2023
2.4. Integrated method

In this study, the integrated approach proposed to determine production priority consists of 3 stages as
shown in Figure 1.
Fig. 26. Flowchart of the proposed integrated approach
As seen in Figure 1, customer alternatives are determined in the first stage. In the second stage, the criteria
that will affect the production priority depending on customer alternatives are determined by the decision-
makers working in the enterprise. In the third stage; criteria weights obtained both subjectively and
objectively by LBWA and CRITIC methods are combined using the SOWIA method to obtain integrated
criteria weights. Finally; the weights obtained from subjective, objective, and integrated criteria
weighting methods are compared. Thus, the most important criteria to be used in determining the
customer to whom production priority will be given are determined.
3. Application
In this study, the criteria used to determine customer priority in a company that exports ceramics are
prioritized with the proposed integrated criteria weighting approach. In the company; The Export
Planning and Tracking department should manage demand and stock in a way that increases customer
satisfaction and business profitability, considering the technical conditions of the factories. Production
plan prepared by the factory; It is created by considering various parameters such as customer priority,
product cost, production capacity, as well as technical constraints in the business. For demand
management that depends on more than one criterion, customer priority must be determined, which is of
great importance at the planning stage. The proposed approach was implemented in three stages as shown
in Figure 1.
Stage 1. Determination of customer alternatives
The company in question has approximately 50 customers to whom it actively sells. Among these
customers, 22 customers with the most sales and production demands were taken into consideration.
For customer confidentiality, customer names are not shared and are named as 𝑀1, 𝑀2...𝑀22.
Stage 2. Determination of criteria and their values
7 criteria that will affect business profitability and be used in determining customer rankings were
determined by 6 experts. these criteria:
Average selling price (C1): It is the sales price given to the customer.
Stock (C2): It refers to the stock produced for the customer and waiting in the factory.
Page 767
SEMIT 2023
Demand amount(C3): Demand quantity refers to the total amount of products that the customer states
she will purchase in 2022.
Sales amount (C4): The total sales amount made to the customer in 2021.
Payment performance (C5): It refers to the customer's payment performance after receiving the
product.
Delivery time (C6): The period during which the product must be delivered on the deadline determined
by the customer. The business is fined for each day that this period passes. Customers whose penalty
period is stated as 600 in Table 3 do not have any criminal sanctions, but the period is entered into the
Canias ERP system as 600 days, which is 1.5 years.
Stock conversion rate (C7): It varies depending on the time the customer purchases the product. It
indicates how many days the product is kept in stock.
When evaluating customer priority; the average selling price should be taken as a benefit criterion as it
increases business profitability and customers with the maximum selling price are given priority.
Likewise, demand amount, sales amount, and payment performance criteria are also benefit criteria.
Stock, delivery time, and stock conversion rate criteria are cost criteria, and minimizing these values will
positively affect customer priority. The criterion values shown in Table 1 were obtained using the data in
the Canias ERP program of the 22 customers to whom the company sells the most.
Table 1. Criteria and their values
Stage 3. Determination of Integrated Weights of Criteria

Step 1. Determination of subjective weights of criteria with the LBWA method
First, the order of importance of the criteria is determined. Criteria are scored as 1-highest and 9-lowest.
The most important criterion by decision makers was determined as sales amount. The second criterion,
Page 768
SEMIT 2023
which is at least as important as the sales amount, is determined as the average selling price. Scoring was
made by experts as shown in Table 2.
Table 2. Order of importance of criteria
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
1 2 5 1 4 3 9
After the importance of the criteria is determined, the criteria are compared pairwise and ranked from
most important to least important. This ranking made by decision-makers is shown in Table 3.
Table 3. Sorting criteria according to their importance
𝐶1 𝐶4 𝐶2 𝐶6 𝐶5 𝐶3 𝐶7
1 1 2 3 4 5 9
As seen in Table 3, the most important criterion was determined to be C4. Decision-makers created subsets
of criteria:
Level S1: The most important criterion 𝐶4 is of equal importance to the 𝐶1 , 𝐶4 , 𝐶5 , 𝐶6 . If it is shown as
given in Equation (1);
𝑆1 = {𝐶1 , 𝐶4 , 𝐶5 , 𝐶6 }
Level S2: The most important criterion 𝐶4 is at least twice and at most three times more important than
the 𝐶2 , 𝐶3 , 𝐶7 . If it is shown as given in Equation (1);
𝑆2 = {𝐶2 , 𝐶3 , 𝐶7 }
Then, the criteria are compared with each other and integer assignment is made according to their
importance levels. The most important criterion 𝐶4 was assigned a value of 0. Integer values of other
criteria are as follows. After integer value assignment, the maximum integer value assigned is found.
Equation (2) is used for this process.
𝐶1 3
𝐶2 1
𝐶4 0
𝑆1 = { }, 𝑆2 = {𝐶3 2 }
𝐶5 1
𝐶7 3
𝐶6 2
𝑟 = 𝑚𝑎𝑥{|𝑆1 |, |𝑆2 |, … , |𝑆𝑘 | } = 𝑚𝑎𝑥{4,3} = 4
According to the r value, the elasticity coefficient is determined from the 𝑟0 > 𝑟 equation as 𝑟0 > 4.
i.e. 𝑟0 = 5. Impact functions for each level were calculated as follows using Equation (3) and shown in
Table 4.
5
𝑓(𝐶15 ) = = 0,833
1∙5+1
Table 4. Influence functions of criteria
Page 769
SEMIT 2023
The criteria weights are calculated using Equation (4) and Equation (5). First, the weight of the most
important criterion is determined. Then other criteria weights are calculated. Table 5 shows the weights
determined for all criteria. The 𝑤4 and 𝑤5 values of criteria 𝐶4 and 𝐶5 , which are determined as the most
important criteria, are calculated as follows:
1
𝑤4 = = 0,226
1 + 0,625+. . . +0,385
𝑤5 = 0,833 ∙ 0,226 = 0,188
Table 5. Criteria weights determined according to LBWA method
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝒘𝒋 0,141 0,103 0,094 0,226 0,188 0,161 0,087
Rank 4 5 6 1 2 3 7
As seen in Table 5, according to the LBWA method, the sales amount criterion with the highest weight
score was determined as the most important criterion, and the least important criterion was determined as
the stock conversion ratio. It was stated that the resulting ranking was meaningful when evaluated
together with experts. It has been interpreted that the importance of other criteria is also low for a customer
with a low sales volume. In addition, since the profit a customer with a low payment performance will
provide to the business will be low, it is deemed appropriate to have a higher degree of importance
compared to other criteria. The business should give priority to the customer with a low delivery time
among customers with good sales amount and payment performance. Therefore, the fact that this criterion
is ranked 3rd has been approved by experts. According to experts, another important criterion that will
increase business profits among other criteria is the average selling price. A customer who does not have
a product stock that is produced and ready for shipment is more valuable than a customer whose demand
is high but whose stock is not yet converted to shipment, because it does not create inventory costs for
the business. For this reason, the ranking of this criterion was deemed appropriate.
Step 2. Determination of objective weights of criteria with the CRITIC method
In this step, the initial decision matrix must first be created as shown in Equation (6). In this content,
the values given in Table 1 were used to create the initial decision matrix. The initial decision matrix was
normalized using Equation (7) and Equation (8), and the normalized decision matrix was obtained as
shown in Table 6.
Table 6. Normalized decision matrix
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝑀1 0,529 0,145 1,000 1,000 0,767 0 0,340
𝑀2 0,790 0,290 0,925 0,873 0,698 1,000 0,600
𝑀3 0,648 0,918 0,816 0,830 0,930 0 0,840
𝑀4 0,932 0,562 0,705 0,682 0,233 0 0,500
𝑀5 0,506 0,725 0,892 0,573 0,349 0 0,200
𝑀6 1,000 0,580 0,545 0,448 0 0 0,300
𝑀7 0,932 0,957 0,326 0,427 0,814 0 0,900
𝑀8 0,447 0 0,992 0,342 0,233 0 0
𝑀9 0,628 0,971 0,326 0,342 0,721 1,000 0,300
𝑀10 0,275 0,482 0,358 0,342 0,930 0 0,700
𝑀11 0,347 0,928 0,323 0,327 0,814 1,000 0,900
𝑀12 0,435 0,722 0,298 0,299 1,000 1,000 1,000
𝑀13 0,975 0,870 0,270 0,274 0,953 0,889 1,000
𝑀14 0,932 0,188 0,214 0,236 1,000 1,000 0,800
𝑀15 0 0,290 0,261 0,214 1,000 0 0,500
Page 770
SEMIT 2023
𝑀16 0,209 0,529 0,208 0,193 0,465 1,000 0,500

𝑀17 0,398 1,000 0,165 0,151 0,698 1,000 0,600
𝑀18 0,790 0,725 0,193 0,130 0,698 1,000 0,800
𝑀19 0,478 0,844 0,061 0,062 0,930 1,000 0,900
𝑀20 0,932 0,942 0,036 0,045 0,349 0 0,800
𝑀21 0,963 0,653 0,071 0,003 0,535 0,926 0,200
𝑀22 0,315 0,597 0 0 0,698 0,926 0,700
Standard deviation (𝜎𝑗 )values of the criteria were calculated with Equation (9) and shown in Table 7.
Table 7. Standard deviation values

𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝜎𝑗 0,295 0,297 0,328 0,282 0,290 0,499 0,288
Then, the correlation coefficients of the criteria were calculated using Equation (10). Correlation
values (𝑃𝑗𝑘 ) are given in Table 8.
Table 8. Correlation values of criteria
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝐶1 1,000 0,178 0,045 0,131 -0,295 -0,041 0,073
𝐶2 0,178 1,000 -0,463 -0,282 0,122 0,224 0,471
𝐶3 0,045 -0,463 1,000 0,884 -0,296 -0,516 -0,469
𝐶4 0,131 -0,282 0,884 1,000 -0,065 -0,439 -0,185
𝐶5 -0,295 0,122 -0,296 -0,065 1,000 0,364 0,652
𝐶6 -0,041 0,224 -0,516 -0,439 0,364 1,000 0,325
𝐶7 0,073 0,471 -0,469 -0,185 0,652 0,325 1,000
(1 − 𝑃𝑗𝑘 ) calculations to be used to find the importance weights of the criteria are given in Table 9.
Table 9. (1 − 𝑃𝑗𝑘 ) values
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝐶1 - 0,822 0,955 0,869 1,295 1,041 0,927
𝐶2 0,822 - 1,463 1,282 0,878 0,776 0,529
𝐶3 0,955 1,463 - 0,116 1,296 1,516 1,469
𝐶4 0,869 1,282 0,116 - 1,065 1,439 1,185
𝐶5 1,295 0,878 1,296 1,065 - 0,636 0,348
𝐶6 1,041 0,776 1,516 1,439 0,636 - 0,675
𝐶7 0,927 0,529 1,469 1,185 0,348 0,675 -
Total 5,908 5,749 6,814 5,956 5,518 6,083 5,132
The final importance weights were calculated with the help of Equation (11) and are shown in Table 10.
Table 10. Criteria weights determined according to CRITIC method
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝑤𝑗 0,129 0,127 0,166 0,125 0,119 0,225 0,110
Rank 3 4 2 5 6 1 7
Page 771
SEMIT 2023
As seen in Table 10, the criterion with the highest importance was determined as delivery time. Experts
stated that the delivery time is the most important criterion because the customer who is not given
importance even though the deadline is short, causes harm to the business and negatively affects the
profit. The least important criterion is the stock conversion ratio.
Step 3. Determination of integrated criteria weights with the SOWIA method
In steps 1 and 2; The importance rankings of the criteria obtained using the LBWA and CRITIC methods
were different from each other. While the order of importance of the sales amaount criterion is in the first
place with the LBWA method; It was ranked fifth by the CRITIC method. Considering this difference in
criterion importance weights and rankings, a single weight and importance order was determined for each
criterion with the SOWIA method in the final stage.
In this study, the criteria weights obtained by the CRITIC method are considered as 𝑂𝑗 obtained by the
objective method. criteria weights obtained by the LBWA method are considered as 𝑆𝑗 obtained by the
subjective method. In this context, integrated weights were obtained using Equation (13). As seen in
Table 11, the criteria weights determined according to different α values were compared with each other.
Thus, the most appropriate α value was tried to be determined.
Table 11. Determination of optimal α value according to criteria weights

𝛼 0,2 0,4 0,6 0,8
𝑊𝑆𝑂𝑊𝐼𝐴 𝑊𝑆𝑂𝑊𝐼𝐴 𝑊𝑆𝑂𝑊𝐼𝐴 𝑊𝑆𝑂𝑊𝐼𝐴
𝐶1 0,139 0,136 0,134 0,132
𝐶2 0,107 0,112 0,117 0,122
𝐶3 0,101 0,123 0,137 0,152
𝐶4 0,206 0,185 0,165 0,145
𝐶5 0,181 0,160 0,146 0,132
𝐶6 0,174 0,187 0,200 0,212
𝐶7 0,091 0,096 0,101 0,105
For example, when the order of importance of the criteria is examined, the delivery time criterion
comes first and the sales amount criterion comes second. When criteria with the same sales amount
are compared, priority should be given to the customer with good payment performance. In this case,
𝛼 = 0,2 and 𝛼 = 0,8 are excluded from evaluation. Then, when two customers with good payment
performance are compared, priority should be given to the customer with the higher sales price. Thus,
the 𝛼 = 0,4 value was determined as the most appropriate α value for this problem according to the
decision makers. Accordingly, the integrated criteria weights obtained by the SOWIA method are given
in Table 12. For example, 𝑊1𝑖𝑛𝑡 is calculated as follows.
Table 12. Criteria weights and rankings calculated with the SOWIA method for α=0.4
𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7
𝑊𝑗𝑖𝑛𝑡 0,136 0,112 0,123 0,185 0,160 0,187 0,096
Rank 4 6 5 2 3 1 7
As seen in Table 12, the criterion with the highest importance according to the SOWIA method was the
delivery time criterion, while the criterion with the lowest weight was the Stock Conversion Ratio.
Step 4. Comparison of criteria weights and alternative rankings
In this step, the priority orders of the criteria obtained according to the criteria weights determined by
the LBWA, CRITIC and SOWIA methods were compared. Thus, the criteria that should be taken into
consideration first when determining the priority order of customer demands were determined.
Comparison results are shown in Table 13.
Page 772
SEMIT 2023
Table 13. Comparison of criteria weights and rankings
CRITIC
𝑊𝐶𝑅𝐼𝑇𝐼𝐶 𝑊𝐿𝐵𝑊𝐴 𝑊𝑆𝑂𝑊𝐼𝐴 LBWA SOWIA
Ranking
𝐶1 0,129 0,141 0,136 3 4 4
𝐶2 0,127 0,103 0,112 4 5 6
𝐶3 0,166 0,094 0,123 2 6 5
𝐶4 0,125 0,226 0,185 5 1 2
𝐶5 0,119 0,188 0,160 6 2 3
𝐶6 0,225 0,161 0,187 1 3 1
𝐶7 0,110 0,087 0,096 7 7 7
As seen in Table 13, according to the CRITIC method, the most important criterion is the delivery time,
and according to the LBWA method, the most important criterion is the sales amount. According to the
SOWIA method, the most important criterion is determined as the delivery time, as in the CRITIC
method. Regardless of whether the customer's order quantity is large or small, the business will be
penalized if the product is not delivered within the specified deadline. As long as the payment
performance and average selling price of customers with high sales volume are also good, business
profitability will increase. By paying attention to the priority order of the criteria, especially the delivery
time and sales amount criteria, it can be determined which customer's demand should be met first.
4. Conclusion and Recommendations

Since business profitability increases as customer satisfaction increases, ensuring customer satisfaction
is an important element in demand management. Nowadays, with the increasing competition conditions,
maintaining customer satisfaction and loyalty has become very difficult. Businesses that want to increase
customer satisfaction and therefore profitability need to determine customers' production priorities
correctly in order to meet customer orders on time and effectively. Multiple conflicting criteria are
effective in determining production priority. For this reason, within the scope of the study, the importance
weights and therefore the order of importance of the criteria to be used in determining the production
priority were determined with the integrated criteria weighting approach developed using MCDM
methods. According to this order, it can be determined which customer's request will be given priority.
The approach proposed in the study was used to determine the production priorities of a company
operating in the ceramics industry. However, the proposed integrated weighting approach can also be
applied to different sectors. In future studies, the approach proposed in this study can be improved by
using different MCDM methods or various fuzzy-based MCDM methods in cases where information is
incomplete and uncertain.
References
1. Bayram, E.; A Multi Criteria Decision Making Based Integrated Approach for Determining
Production Priority in Demand Management. Master's Thesis, Ataturk University, Graduate School
of Natural and Applied Sciences, Erzurum (2022).
2. Çelikyay, S.: Application of Multi Criteria Decision Making Methods in Multi-Purpose Combat
Aircraft Selection. Master's Thesis, Istanbul Technical University, Graduate School of Natural and
Applied Sciences, Istanbul (2002).
3. Orhan, M.: The Place of AHP in Decision Making Problems and Comparison of European Union
Member Countries with Development Criteria. Master's Thesis, Marmara University, Graduate
School of Social Sciences, Istanbul (2013).
Page 773
SEMIT 2023
4. Žižović, M., Pamučar, D.: New model for determining criteria weights: level based weight assessment
(LBWA) model. Decision Making: Applications in Management and Engineering, 2(2), 1-20
(2019).
5. Diakoulaki, D., Mavrotas, G., Papayannakis, L.: Determining objective weights in multiple criteria
problems: The critic method. Computers & Operations Research, 22(7), 763-770 (1995).
6. Das, M. C., Sarkar, B., Ray, S.: On the performance of Indian technical institutions: a combined
SOWIA-MOORA approach. Opsearch, 50(3), 319-333 (2013).
7. Ecer, F.: Çok Kriterli Karar Verme Geçmişten Günümüze Kapsamlı Bir Yaklaşım. Seçkin
Yayıncılık, Ankara (2020).
8. Zaher, H., Khalifa, H. A., Mohamed, S.: On Rough Interval Multi Criteria Decision Making.
International Journal of Scientific & Technology Research, 9(1), 3694-3698.
Page 774
SEMIT 2023
Multi-Depot Vehicle Routing Problem with Drone Collaboration in Humanitarian Logistic

Zeynep Yuksel1, Dursun Emre Epcim2, Suleyman Mete3
1
Gaziantep University, Faculty of Engineering, Department of Industrial Engineering, Gaziantep 27310,
Turkey, zeynpyukselll@gmail.com
2
Turkey, dursunemreepcim@gmail.com
3
Turkey, suleyman489@gmail.com
Abstract
Routing problems are used in many areas to obtain the most appropriate results in terms of time and cost.
An attempt is made to address the issue by formulating mathematical models that incorporate multiple
variables such as capacity, time, cost, and demand, tailored to the specific area of application. Natural
disasters are one of these applications. In natural disasters, especially time management is a critical issue.
For this reason, routing models play a crucial role in delivering aid to disaster victims and transporting
disaster victims to hospitals. In this study, a mathematical model is proposed to be applied in post-disaster
humanitarian aid logistics. The model, which aims to minimize the total distribution time, also considers
the distribution of perishable commodities. Drones are integrated into this operational framework to
facilitate the dissemination of perishable commodities. Thus, a new mathematical model for the multi-
depot vehicle routing problem, which includes both truck-drone collaboration and perishable
commodities, has been introduced to the literature. The proposed model is solved with an illustrative
example using GAMS/CPLEX software and the results are tested.
Keywords: Drone-Truck Collaboration, Humanitarian Logistic, Multi-Depot Vehicle Routing Problem,
Perishable Commodities.
1. Introduction
Disasters are tragic events that can cause great losses. Many people around the world lost their lives due
to natural disasters in their region. When their effects on people and buildings are examined, earthquakes
are among the disasters with the most potential impact [1]. According to EM-DAT (The International
data Disaster Database-Center for Research on the Epidemiology of Disaster), 242,000 people lost their
lives in the 1976 China earthquake, 222,570 in the 210 Haiti earthquake and 165,708 in the 2004
Indonesia earthquake. As a result of the tsunami disaster in the Indian Ocean, the number of people who
died in the Indonesian earthquake increased to 226,408. Recently, 50,783 people lost their lives as a result
of the earthquake that occurred in Turkey on February 6, 2023 [2].
One of the vital points that increases the magnitude of the effects of disasters such as earthquakes is the
distribution of humanitarian aid after the earthquake. Lack of supplies that do not arrive on time may
increase the effects of the disaster by causing people to survive the disaster but die afterwards. Therefore,
the post-disaster supply chain is very important to reduce post-disaster deaths and prevent the suffering
of disaster victims from increasing [3]. There are many difficulties in carrying out post-disaster
humanitarian aid logistics activities. One of these difficulties is that blood and some food materials spoil
in a short time. Considering the importance and perishability of these foods, the distribution of such
perishable commodities becomes a vital issue [4].
Routing is one of the most important parts of post-disaster humanitarian logistics. Routing is vital to
manage the process and reduce distribution time. For this reason, many academic studies have been
conducted on the delivery of supplies to disaster victims as soon as possible, including routing the
distance between need points and warehouses. However, when the studies are examined, it is understood
that there is a gap in the literature for a situation that includes multi-warehouse truck-drone cooperation
in post-disaster logistics and also takes perishable commodities into consideration. The aim of this study
is to deliver the products to the disaster victims in the most effective way with drone-truck collaboration
in vital post-disaster logistics. With a different perspective from the literature, a mathematical model is
Page 775
SEMIT 2023
developed for the multi-depot vehicle routing problem with drone , which includes the safe delivery of
perishable products to disaster victims.
The remaining sections of this paper, Section 2 presents the literature. Section 3 presents the problem
definition and the proposed mathematical model. Section 4 contains the testing of the model with the
illustrative case. Section 5 contains the conclusion and finally the references.
Vehicle routing problem (VRP), one of the most important problems in the field of disaster management,
is delivery points starting from one or more warehouses [5]. The purpose of VRP in humanitarian logistics
is to determine the most appropriate routing considering the constraints such as vehicle capacity, number
of vehicles, demand and time. It is anticipated that the integration of drones into traditional VRP in
humanitarian logistics will maximize immediate distribution. Hence, Multi-Depot Vehicle Routing
Problem with Drone are presented in this section.
Stodola and Kutej [6] tackle the MDVRP-D in their study. Drones are paired with vehicles to make
deliveries from warehouses to customers. They developed a mathematical model for this problem whose
objective is to minimize the duration of the entire logistics operation. In addition, Adaptive Node
Clustering Ant Colony Optimization with Node Clustering algorithm, a metaheuristic algorithm based on
Ant Colony, is proposed as a solution. The multi-depot unmanned aerial vehicle (UAV) routing issue was
addressed by Li et al. [7]. They suggested a mathematical model that does not impose restrictions on the
depot where UAVs launch and land. The proposed model includes multiple objective functions to
minimize the time and number of UAVs on the route. Hybrid large neighborhood search was suggested
as a potential remedy. Rathinam and Sengupta's [8] paperwork deals with the issue of path selection for
a number of UAVs traveling from various depots to particular terminals and locations. Each UAV starts
from a depot and travels to at least one location as part of their proposed solution to the issue. For the
multi-depot UAV routing problem, the authors additionally provide a 2-approximation approach and a
lower bound algorithm. Manyam et al. [9] focused on the persistent intelligence, surveillance, and
reconnaissance routing problem. Their model aims to optimize the data collection and delivery tasks of
multiple UAVs. As a solution methodology, they presented complex computational methods as well as
heuristic approaches. Habib et al. [10] studied a real-time optimization problem for UAV path planning
in dynamic situations. They considered this problem as a variant of a multi-depot vehicle routing problem
(MDVRP) and proposed a mixed integer linear programming (MILP) model for solving this problem.
Kim et al. [11] studied the distribution of drugs and test kits by drones for patients with chronic diseases
who need to go to clinics for routine health examinations in rural areas. They proposed two models for
this situation. In the first of these models, they used the closure approach to find the optimal number of
drone centers. In the second model, they proposed a MDVRP model that minimizes the operating costs
of drones. Haller [12] worked on two models to optimize the US Marine Corps' use of UAVs. In the first
model, author aimed to find the optimal depot locations for the charging and supply of UAVs. In the
second model, author proposed a model for the MDVRP for UAVs by improving a single depot model
in the literature. Hamid et al. [13] worked on optimizing the homemade food delivery process. They used
drones and crowdsourcing as two novel approaches for this process. To solve this problem, they
developed a multi-warehouse vehicle routing model using transport costs, freshness of the delivered food
and delivery date satisfaction as objective functions. Due to the complexity of the problem, they proposed
a self-regulating hyper-heuristic method to obtain a solution. This method is based on genetic algorithm
and modified particle swarm optimization and includes new selection and mutation mechanisms.
As a result of the research, there are very few studies of the multi-depot vehicle routing problem involving
drone collaboration in disaster logistics. In this context, the consideration of the distribution of perishable
commodities in drone collaborative studies of the multi-depot vehicle routing problem differs from other
studies. Thus, the problem definition and mathematical model proposal of this study contribute to the
literature.
Page 776
SEMIT 2023
3. Problem Definition and Formulation

The problem addressed in this study is distribution in post-disaster humanitarian aid logistics. The focus
is on the fastest and most efficient delivery of relief supplies to the affected areas. The materials to be
distributed are divided into 2 groups as other needs (Type 1) and perishable commodities (Type 2). Type
1 supplies include needs such as clothes and blankets, while type 2 includes needs that are at risk of
spoilage. The fleet of logistics vehicles used in distribution is homogeneous.
Figure 27 Problem Definition

In this problem, relief supplies are safely distributed to the demand points by logistics vehicles (trucks
and drones) travelling from the depots. As shown in Figure 1, drones and trucks move separately and
have different routes. The problem is considered as an open form of a multi-depot vehicle routing problem
(MDVRP). Therefore, a vehicle leaving a depot can return to any depot. The assumptions of the problem,
which deals with post-disaster humanitarian aid in MDVRP open form for two different product types,
are as follows.
• There are sufficient numbers of logistics vehicles and relief supplies in warehouses.
• Distribution points and demands are known.
• The maximum time required to distribute the Type 2 product group is known.
• Only one product type is in demand at each distribution point.
3.1 Mathematical Model

In this study, MDVRP-D model is proposed to be implemented in post-disaster humanitarian aid logistics.
The notations of the developed model are as indicated in table 1.
Table1. Notations of MDVRP-D Model
NOTATIONS
Indices
i,j Tasks and depots
v Fleet of trucks
u Fleet of drones
d Depots
p Product type
𝑄𝑉 Capacity of truck
𝑄𝑈 Capacity of drone
Take-off Time for the takeoff of the drone
Landing Time for the landing of the drone
MF Maximum flight time
L Maximum delivery time for perishable product type
Page 777
SEMIT 2023
Parameter
𝑅𝑗𝑝 Demand for product p at point j (𝑗 ∈ 𝑗 − 𝑑)
𝑇𝑖𝑗𝑢 The time of reaching from point i to point j by drone u
𝑇𝑖𝑗𝑣 The time of reaching from point i to point j by truck v
𝐷𝑖𝑗 Distance between point i and point j
Variables
𝐶𝑣 Total cost of the truck
𝐶𝑢 Total cost of the drone
𝐸𝑢 The initial total payload weight carried by the drone u
𝑆𝑖𝑣 Arbitrary numbers
𝑆𝑖𝑢 Arbitrary numbers
𝑊𝑖𝑗𝑢 Payload carried by the drone u traveling from point i to point j
Binary Variables
𝑋𝑖𝑗𝑣 1, if truck v arrives at point j after leaving point i (I≠j);
otherwise, 0
𝑌𝑖𝑗𝑢 1, if drone u arrives at point j after leaving point i (I≠j);
otherwise, 0
𝑍𝑗𝑝 1, if there is any demand of product type p at point j
(𝑗 ∈ 𝑗 − 𝑑) ; otherwise, 0
Objective function (1) minimizes the time it takes logistics vehicles to deliver relief supplies.
𝑀𝑖𝑛 ∑ ∑ ∑(𝑇𝑖𝑗𝑣 ∗ 𝑋𝑖𝑗𝑣 ) + ∑ ∑ ∑(𝑇𝑖𝑗𝑢 ∗ 𝑋𝑖𝑗𝑢 ) (1)

𝑖 𝑗 𝑣 𝑖 𝑗 𝑢
Constraint (2-33) ensures the same number of trucks and drones leaving and returning to the
depot.
∑ ∑ 𝑋𝑗𝑖𝑣 = ∑ ∑ 𝑋𝑖𝑗𝑣 ∀𝑣 (2)

𝑖∈𝑖−𝑑 𝑗∈𝑗∩𝑑 𝑖∈𝑖−𝑑 𝑗∈𝑗∩𝑑
∑ ∑ 𝑌𝑗𝑖𝑢 = ∑ ∑ 𝑌𝑖𝑗𝑢 ∀𝑢 (3)

𝑖∈𝑖−𝑑 𝑗∈𝑗∩𝑑 𝑖∈𝑖−𝑑 𝑗∈𝑗∩𝑑
Constraints (4-5) allow the same truck or drone to leave the depot only once.
∑ ∑ 𝑋𝑖𝑗𝑣 = 1 ∀𝑣 (4)
𝑗∈𝑗−𝑑 𝑖∈𝑖∩𝑑
∑ ∑ 𝑌𝑖𝑗𝑢 = 1 ∀𝑢 (5)
Constraint (6-7) ensures that there is only one entry into a task by a drone or a truck.
∑ ∑ 𝑋𝑖𝑗𝑣 + ∑ ∑ 𝑌𝑖𝑗𝑢 = 1 ∀𝑗 ∈ 𝑗 − 𝑑, 𝑖 ≠ 𝑗 (6)

𝑖 𝑣 𝑖 𝑢
∑ ∑ 𝑋𝑖𝑗𝑣 + ∑ ∑ 𝑌𝑖𝑗𝑢 = 1 ∀𝑗 ∈ 𝑗 − 𝑑, 𝑖 ≠ 𝑗 (7)

𝑗 𝑣 𝑗 𝑢
Page 778
SEMIT 2023
Constraint (8-9) ensures that the truck or drone entering the same point leaves that node.
∑ 𝑋𝑖𝑗𝑣 = ∑ 𝑋𝑗𝑖𝑣 ∀𝑣, 𝑖 , 𝑖 ≠ 𝑗 (8)

𝑗 𝑗
∑ 𝑌𝑖𝑗𝑢 = ∑ 𝑌𝑗𝑖𝑢 ∀𝑢, 𝑖 , 𝑖 ≠ 𝑗 (9)

𝑗 𝑗
Constraint (10-11) provides that the payload carried by each truck and each drone along its route does
not exceed its capacity.
∑ ∑ ∑ 𝑅𝑗𝑝 ∗ 𝑋𝑖𝑗𝑣 ≤ 𝑄𝑉 ∀𝑣 (10)

𝑝 𝑗∈𝑗−𝑑 𝑖
∑ ∑ ∑ 𝑅𝑗𝑝 ∗ 𝑌𝑖𝑗𝑢 ≤ 𝑄𝑈 ∀𝑢 (11)

𝑝 𝑗∈𝑗−𝑑 𝑖
Constraint (12) allows perishable commodities to be delivered to distribution points on a

drone or a truck without exceeding the spoilage time.
∑ ∑ ∑ 𝑍𝑗𝑝 ∗ (𝑇𝑖𝑗𝑣 ∗ 𝑋𝑖𝑗𝑣 + 𝑇𝑖𝑗𝑢 ∗ 𝑌𝑖𝑗𝑢 ) ≤ 𝐿 ∀𝑗 ∈ 𝑗 − 𝑑, 𝑖 ≠ 𝑗, 𝑝 = 2 (12)

𝑖 𝑣 𝑢
Constraint (13) ensures that the time required by each drone along its route, including landing and
take-off time, does not exceed the maximum flight time.
∑ ∑ 𝑌𝑖𝑗𝑢 ∗ (𝑇𝑖𝑗𝑢 + 𝑡𝑎𝑘𝑒 − 𝑜𝑓𝑓) (13)

+ ∑ ∑ 𝑌𝑖𝑗𝑢 ∗ (𝑇𝑖𝑗𝑢 + 𝑡𝑎𝑘𝑒 − 𝑜𝑓𝑓 + 𝑙𝑎𝑛𝑑𝑖𝑛𝑔)

𝑗∈𝑗−𝑑 𝑖∈𝑖−𝑑
+ ∑ ∑ 𝑌𝑖𝑗𝑢 ∗ (𝑇𝑖𝑗𝑢 + 𝑙𝑎𝑛𝑑𝑖𝑛𝑔) = 𝑀𝐹 ∀𝑢

𝑖∈𝑖−𝑑 𝑗∈𝑗∩𝑑
Constraint (14-15) prevents routes that do not start and end at the depot.
𝑆𝑖𝑣 − 𝑆𝑗𝑣 + 𝑄𝑉 ∗ 𝑋𝑖𝑗𝑣 ≤ 𝑄𝑉 − ∑ 𝑅𝑗𝑝 ∀𝑖 ∈ 𝑖 − 𝑑, ∈ 𝑗 − 𝑑, 𝑣, 𝑖 ≠ 𝑗 (14)

𝑝
𝑆𝑖𝑢 − 𝑆𝑗𝑢 + 𝑄𝑈 ∗ 𝑌𝑖𝑗𝑢 ≤ 𝑄𝑈 − ∑𝑝 𝑅𝑗𝑝 ∀𝑖 ∈ 𝑖 − 𝑑, 𝑗 ∈ 𝑗 − 𝑑 , 𝑢, 𝑖 ≠

𝑗, 𝑅𝑗1 ≤ 𝑄𝑈 , 𝑅𝑗2 ≤ 𝑄𝑈, 𝑅𝑖1 ≤ (15)
𝑄, 𝑈𝑅𝑖2 ≤ 𝑄𝑈
Constraints (16-17) provide upper and lower bounds for each truck and drone.
∑ 𝑅𝑖𝑝 ≤ 𝑆𝑖𝑣 ≤ 𝑄𝑉 ∀𝑖, 𝑣 (16)

𝑝
∑ 𝑅𝑖𝑝 ≤ 𝑆𝑖𝑢 ≤ 𝑄𝑈 ∀𝑖, 𝑢, 𝑅𝑖1 ≤ 𝑄, 𝑈𝑅𝑖2 ≤ 𝑄𝑈 (17)

𝑝
3.2 Illustrative Example

The proposed model was tested by applying the mathematical model to the case in the article by Song et
al. [14]. The case was solved by taking 5% of the demands in the case and solving it with the first 10
tasks. First of all, by using the p-center method, 2 temporary depot locations are established from the 10
Page 779
SEMIT 2023
tasks and the case is made suitable for the multi-depot vehicle routing problem. Then, routes were created
by applying the mathematical model developed with the data of the logistics vehicles given below. The
obtained results show that the model works smoothly in a small case consisting of 10 points. However,
in problems with 38 and 50 duty points, the model does not give a solution within 3600 seconds. Due to
the complexity of the model, a heuristic algorithm is required to solve the model in cases with more points
as mentioned. In the case with 10 points, the paths and values obtained are as shown in table 2.
Figure 2. Routes of the Illustrative Example
As shown in Figure 2, two temporary depot locations are selected among the 10 nodes in the case. The
selected depot locations are nodes 1 and 6. For the 10-node case, 1 drone and 1 vehicle from the depot at
node 1, 1 drone from the depot at node 6 and a total of 3 logistics vehicles from the depots were used. In
the case study, all logistics vehicles returned to the depot from which they left the case. The objective
function of the model solved in collaboration with the drone and truck, whose information is given in the
results obtained, is 140.257.
Table 2. Results of Illustrative Example
Capacit Max. delivery Max. Takeoff Landing Routes
y time for Flight
product type 2 time
Truck 100 30 - - - 1-7-2-8-10-3-1
Drone 35 30 30 2 1 6-9-6;1-5-4-5
4. Conclusion
The model developed for use in humanitarian aid logistics is nonlinear. The novelty of the study is that
the open-form multi-depot vehicle routing problem includes the transportation of perishable commodities
to the demand points without spoilage. In addition, truck-dron collaboration in the model helps the
problem to give effective results. While vehicles can provide advantages in terms of capacity, drones are
more advantageous in terms of time, which is especially important for perishable products, and access to
difficult areas.
While the model proposed in this study can provide results for small cases, it is not possible to obtain
results for large cases because the model is complex. Therefore, implementing heuristic methods to the
model is one of the future works. In addition, improving the model with the use of heterogeneous logistics
vehicle fleets can also be among the future studies.
References
1. Mavrouli, M., Mavroulis, S., Lekkas, E., & Tsakris, A. (2023). The impact of earthquakes on public
health: A narrative review of infectious diseases in the post-disaster period aiming to disaster risk
reduction. Microorganisms, 11(2), 419.
Page 780
SEMIT 2023
2. EM-DATA CRED (2023) Earthquakes in Türkiye A Review from 1900 to Today, EM-DATA CRED
Crunch n°72 / UCLouvain. Available at: https://www.emdat.be/publications/
3. Diabat, A., Jabbarzadeh, A., & Khosrojerdi, A. (2019). A perishable product supply chain network
design problem with reliability and disruption considerations. International journal of production
economics, 212, 125-138.
4. Rashidzadeh, E., Hadji Molana, S. M., Soltani, R., & Hafezalkotob, A. (2021). Assessing the
sustainability of using drone technology for last-mile delivery in a blood supply chain. Journal of
Modelling in Management, 16(4), 1376-1402.
5. Laporte, G. (1992). The vehicle routing problem: An overview of exact and approximate
algorithms. European journal of operational research, 59(3), 345-358.
6. Stodola, P., Kutěj, L., “Multi-Depot Vehicle Routing Problem with Drones: Mathematical
Formulation, Solution Algorithm and Experiments”, Solution Algorithm and Experiments.
7. Li, X., Li, P., Zhao, Y., Zhang, L., Dong, Y., 2021, “A Hybrid Large Neighborhood Search Algorithm
for Solving the Multi Depot UAV Swarm Routing Problem”, IEEE Access, 9, 104115-104126.
8. Rathinam, S., Sengupta, R., 2006, December. “Lower and upper bounds for a multiple depot UAV
routing problem”, In Proceedings of the 45th IEEE Conference on Decision and Control, pp. 5287-5292,
IEEE.
9. Manyam, S. G., Rasmussen, S., Casbeer, D. W., Kalyanam, K., Manickam, S., 2017, June, “Multi-
UAV routing for persistent intelligence surveillance & reconnaissance missions”, In 2017 international
conference on unmanned aircraft systems (ICUAS), pp. 573-580, IEEE.
10. Habib, D., Jamal, H., Khan, S. A., 2013, “Employing multiple unmanned aerial vehicles for co-
operative path planning”, International Journal of Advanced Robotic Systems, 10(5), 235.
11. Kim, S. J., Lim, G. J., Cho, J., Côté, M. J., 2017, “Drone-aided healthcare services for patients with
chronic diseases in rural areas”, Journal of Intelligent & Robotic Systems, 88, 163-180.
12. Haller, J. R., 2021, “Expanding Optimization of Energy Efficient UAV Routing in Support of Marine
Corps Expeditionary Advanced Base Operations with Multiple Supply Depots” , Doctoral dissertation,
Monterey, CA; Naval Postgraduate School.
13. Hamid, M., Nasiri, M. M., Rabbani, M., 2023, “A mixed closed-open multi-depot routing and
scheduling problem for homemade meal delivery incorporating drone and crowd-sourced fleet: A self-
adaptive hyper-heuristic approach”, Engineering Applications of Artificial Intelligence, 120, 105876.
14. Song, B. D., & Ko, Y. D. (2016). A vehicle routing problem of both refrigerated-and general-type
vehicles for perishable food products delivery. Journal of food engineering, 169, 61-71.
Page 781
SEMIT 2023
SP22: Multi-Criteria Decision-Making
A Bibliometric Analysis of Multi-Criteria Decision-Making Techniques in Disaster Management

and Transportation in Disaster Conditions
Mehmet Kabak1[0000-0002-8576-5349] and Ezgi Aktaş Potur2[0000-0003-0192-8655]
1 Gazi University, 06570, Ankara, Turkey.
mkabak@gazi.edu.tr
2 Gazi University, 06570, Ankara, Turkey
ezgiaktas@gazi.edu.tr
Abstract
Disaster management is a crucial discipline aimed at minimizing the potential damage stemming from
natural calamities and protecting communities against their possible impacts. To achieve effective and
beneficial solutions, a key fundamental tool that can be employed is the utilization of multi-criteria
decision-making techniques within this process. Disaster management is a multifaceted process that
involves significant uncertainties, risks, and intricacies. This complexity arises from the necessity to make
decisions not only based on cost-effectiveness but also taking into consideration the social,
environmental, and public health impacts. The utilization of multi-criteria decision-making techniques
in pre-disaster planning, intervention strategy development, and recovery initiatives empowers disaster
managers with a broader perspective, enabling comprehensive scenario analyses and contributing to risk
reduction efforts. The application of multi-criteria decision-making techniques contributes to the
formulation of enhanced and sustainable strategies in the field of disaster management. These techniques
should be acknowledged as crucial tools, aiming to bolster community resilience against disasters and
mitigate the detrimental impacts they entail. In addition, transportation vehicles play a crucial role in the
prompt rescue of the injured and the speedy distribution of emergency aid in disaster situations. These
transportation assets support the coordinated delivery of rescue personnel and essential supplies to
affected regions, expediting the mobilization of communities. The primary objective of this paper is to
perform a bibliometric analysis of the extant academic literature concerning the application of multi-
criteria decision-making (MCDM) techniques in the field of disaster management and transportation in
disaster conditions. This study is envisaged to provide a guideline for researchers in their future studies
in the field of disaster management and transportation in uncertain environment under disaster conditions.
Keywords: Disaster Management, Decision-Making, Bibliometric Analysis, Transportation in Disaster
Conditions.
1. Introduction
Disasters are defined as circumstances resulting in scenarios where there can be loss of life, loss of
property, environmental detriment, or vital disruption of societal equilibrium. Disaster management is a
cross-disciplinary area concentrated on minimizing the consequences resulting from events of either
natural or human-caused origins [1]. These events cover crucial incidents for instance earthquakes,
floods, tsunamis, fires, and hurricanes, which have the potential to negatively impact human lives, the
environment, and economic systems.
In the event of a natural disaster, the necessity of utilizing transportation resources to provide
immediate and efficient assistance to the population, thereby reducing the number of affected individuals
and alleviating economic consequences cannot be overstated. Disaster management is defined as a
process or practice involving mitigation, preparedness, response, and recovery, after all the specific nature
of the unfolding disaster or emergency [2].
Page 782
SEMIT 2023
Disasters, which can occur as either seasonal and predictable or suddenly and unpredictably, arise in
various forms. Regions affected by unpredictable natural disasters require more comprehensive
preparation plans that encompass actions to be taken before, during, and after the disaster, compared to
areas impacted by seasonal disasters [3].
Unpredictable natural disasters have the potential to cause substantial harm to live and properties in the
areas they strike. Therefore, it is highly crucial for all nations to emphasize disaster preparedness and
dedicate attention to the field of disaster management. While some countries are well-prepared for
emergencies and disaster response, comparatively less affluent ones are susceptible to encountering more
significant losses due to their deficiency in planning and preparatory endeavors [4].
On February 6, 2023, two consecutive earthquakes occurred in Turkey, with magnitudes of 7.8 Mw and
7.5 Mw, respectively, within a 9-hour interval. According to official reports, a minimum of 50,783
individuals lost their lives, and over 122,000 people sustained injuries. Turkey is a seismic-prone country,
with 98% of its territory located on active fault lines and various seismic zones. The presence of
earthquakes is inevitable in nations like Turkey, which are situated along active tectonic plate boundaries,
thus necessitating the development of disaster management strategies.
Given the prevalence of numerous natural and anthropogenic disasters like earthquakes, pandemics, and
fires that countries have encountered over the past few decades, it is clear that the utilization of scientific
approaches at all phases of disaster management is essential in order to effectively address these
challenges [5]. Emerging as a significant domain within the field of operations research, multi-criteria
decision-making (MCDM) techniques can be readily applicable in the fundamental and ancillary
activities of disaster management to assist decision-makers and policymakers in facilitating the decision-
making process [5]. MCDM is an approach that deals with the complexity of decision-making among
various alternative options, allowing for the assessment of conflicting criteria during the selection
process. In the complex and multidimensional area of disaster management, which involves crucial
aspects such as efficient resource utilization, transportation and logistics, post-disaster management
process, and ensuring safety, it can be stated that the utilization of MCDM techniques will play a pivotal
role. There is a limited body of bibliometric research concerning the utilization of MCDM techniques in
the field of disaster management. This study is anticipated to provide researchers and policymakers with
a comprehensive viewpoint, encompassing both statistical and quantitative aspects, regarding the
implementation of these techniques in the realm of disaster management and transportation in disaster
conditions.
The subsequent section of the study presents information pertaining to the bibliometric analysis. Section
3 includes the data sources and statistical analysis methods utilized for disaster management. Section 4
provides a similar analysis for transportation in disaster conditions. In section 5, the study concludes by
presenting the results of the research and discussing potential enhancements for future studies.
2. Bibliometric Analysis
Bibliometric analysis, which is progressively gaining magnitude in scientific research, is commonly
favored due to its capacity to unveil trends in research areas of interest, process scientific data, create
considerable research impact, and summarize a substantial amount of bibliometric data [6]. Furthermore,
the emergence of databases such as Web of Science, Science Direct, and Scopus has facilitated the
utilization of bibliometric data. Therefore, recent advancements in these databases have contributed to an
increased interest in bibliometric analyses.
Bibliometric analysis methods are structured into two principal classes: performance analysis and science
mapping. The first one clarifies the contributions of research elements, while the second one focuses more
on the interactions between research elements [6]. In this research, the methodology of science mapping
has been utilized. Science mapping covers the following five types of analysis:
Citation analysis: Citation analysis describes a technique of uncovering connections that arise when one
study references another study, commonly referred to as citation-based mapping. This method facilitates
the identification of influential works within a given field objectively. Through the utilization of this
Page 783
SEMIT 2023
analysis, which allows for the detection of studies that wield influence in a particular domain, noteworthy
findings can be revealed by analyzing the frequencies of citations in scholarly works. This approach
enables the unveiling of studies that have achieved notable outcomes [7].
Co-authorship analysis: Co-authorship analysis, which involves the examination of collaborative
relationships among authors, contributes to the understanding of connections and social dynamics within
academia by revealing networks of collaborating scholars [8].
Co-word analysis: Co-word analysis involves the investigation of frequently recurring words across a
collection of publications, with the intention of revealing thematic patterns and relationships [9].
Co-citation analysis: Co-citation analysis, signifies a methodological approach based on the premise that
two distinct publications, both of which are concurrently present in the reference lists of diverse studies,
are related with each other [10].
Bibliographic coupling: If two studies are citing the same sources, this circumstance is assessed as
bibliographic coupling [6].
In this study, the VOS viewer software was utilized to conduct a bibliometric analysis of publications on
disaster management applying MCDM techniques.
3. Data Sources and Statistical Analysis for Disaster Management

The exploration was conducted on August 13, 2023, involving all fields using the keywords "multi-
criteria decision-making" and "disaster management" resulting in the identification of 69 findings [11-
79]. Publications indexed within the Web of Science database were included in the evaluation process.
Number of Publications
15 13 14
10 7 7 7
5 4
5 3 2 2 2
1 1 1
0
Fig. 1. The number of relevant publications in the Web of Science database between 2006-2023
Figure 1 illustrates the distribution of publications spanning the years 2006 to 2023. It is evident that the
number of studies conducted in the last 2 years exceeds one-third of the total publication count. It is
anticipated that the year 2023 will witness the highest volume of research activities by the end of the
same year. This projection underscores the increasing significance of the research topic. To provide a
more comprehensive exploration of the annual results, the study establishes an analysis of citation
patterns by employing various citation thresholds. Table 1 presents the results.
Page 784
SEMIT 2023
Table 1. Annual WoS citation counts for disaster management

Year Total Publication WoS Total Citation >50 >25 >10
2006 1 2
2009 1 21 1
2010 1 8
2011 0 0
2012 0 0
2013 3 19
2014 2 46 1
2015 5 23
2016 2 24 1
2017 4 111 1 1
2018 7 69 1 2
2019 7 155 1 4
2020 7 64 1 1
2021 2 75 1 1
2022 13 85 3
2023 14 14
According to Table 1, it is observed that a total of 3 studies have received more than 50 citations. These
studies were conducted in 2017 [28], 2019 [21], and 2021 [41]. The study with the highest number of
citations per total publication occurred in the year 2021, with a score of 37.5. This was followed by 2017
and 2019 with rates of 27.75 and 22.14, respectively.
Document Categories
Engineering Multidisciplinary
Computer Science Artificial Intelligence
Operations Research Management…
Computer Science Interdisciplinary…
Remote Sensing
Geography Physical
Meteorology Atmospheric Sciences
Environmental Sciences
Geosciences Multidisciplinary
Water Resources
0 2 4 6 8 10 12 14 16
Fig. 2. Distribution of publications based on the Web of Science categories.

In Figure 2, the foremost four categories encompassing the highest number of publications according to
the Web of Science classifications are delineated as water resources, geosciences multidisciplinary,
environmental sciences, and meteorology atmospheric sciences. The percentages of these categories are
20.290%, 15.942%, 14.493%, and 14.493%, respectively.
30 24 22
20
7 4 4 4 4
10 3 3 2
0
Fig. 3. Top 10 countries with the most publications
Page 785
SEMIT 2023
In Figure 3, the ranking of the top 10 countries based on publication counts is presented. It can be observed
that Italy and Turkey, exhibit a markedly greater number of publications compared to the rest of the
countries.
Fig. 4. Geographic positions of all participating institutions
Figure 4 illustrates the geographical distribution of organizations associated with the literature that
contributes to multi-criteria decision-making applications in disaster management. The graphic was
generated using the coordinates of affiliations data obtained from Web of Science and processed on the
website www.gpsvisualizer.com.
The dimensions of the red circles expand commensurate with the contribution ratio of each organization.
It is noticeable that Turkey and Iran exhibit the highest density of circles among the countries.
The data extracted from the Web of Science Core Collection database was subjected to analysis
employing the VOS viewer bibliometric mapping program. In this study, a bibliometric analysis is
performed utilizing citation analysis, co-authorship analysis, and co-word analysis.
Authors with a minimum of two publications and at least two citations were filtered for author citation
analysis. Out of the initial 152 authors, a subset of 14 authors met the established threshold. Authors
situated outside the cloud were removed from the map. A density map was generated for the remaining 6
authors is given in Figure 5.
Fig. 5. Density visualization of author citation analysis

The density map in Figure 5 highlights the most contributory author in terms of citation count, Mahmoud
Reza Delavar, with 11 publications and 59 citations in this regard. Following closely, Moradi, M. and
Moshiri, B. share the second position with 14 citations each in terms of citation count.
Page 786
SEMIT 2023
Fig. 6. Density visualization of source citation analysis

According to the density map in Figure 6, the source with the highest density in terms of citations is the
International Journal of Disaster Risk Reduction, with 150 citations, followed by the Journal of
Hydrology with 90 citations.
To comprehensively reveal collaborations and associations among countries, organizations, and authors,
an analysis of co-authorship has been conducted. In addition to the number of publications attributed to
each country, the citation count also offers valuable insights to researchers. In order to explore citation
counts and identify countries engaged in collaborative efforts, a filter has been applied to include only
those 'countries' with a minimum of 2 citations and 2 documents. Consequently, from the initial group of
12 countries, two countries lying outside this cloud have been excluded, resulting in a network structure
for the remaining 10 countries, as depicted in Figure 6.
Fig. 7. Network visualization of country analysis

The sizes of nodes in Figure 7 are correlated with the citation counts of the respective countries, with
each color denoting a distinct cluster. Iran stands out with the highest publication count and citation
frequency, and it is notable that Iran's most frequent collaborative country is the Netherlands. Turkey,
positioned in the blue-colored cluster and ranking second in citation count, engages primarily in
collaborative efforts with Italy. Within the red-colored cluster, Spain, France, and India are situated;
similarly, USA, Bangladesh, and Australia grouped in the green-colored cluster due to their enhanced
collaborative interactions with one another.
Figure 8 presents a density map concerning organizations that possess a minimum of 2 publications and
at least 2 citations.
Page 787
SEMIT 2023
Fig. 8. Density visualization of organizations
Analysis of Figure 8 reveals that the University of Tehran stands out with the highest density, marked by
13 publications and 175 citations. Following this, Istanbul Technical University ranks second,
demonstrating a density with 5 publications and 91 citations. Yıldız Technical University, positioned in
the third place in terms of density, is associated with 2 publications and 54 citations.
Figure 9 showcases the network configuration of six authors from a pool of 152, fulfilling the criterion
of at least 2 publications and 2 citations, and situated within the cloud.
Fig. 9. Network visualization of co-author analysis
The sizes of the circular node shapes in Figure 9 are directly proportional to the respective publication
counts. Each cluster is composed of authors who collaborate most frequently with each other. The author
with the highest number of publications, Delavar, M.R., is situated within the green cluster. Delavar,
M.R., closely collaborates with authors Sheikhian, H., and Stein, A., all clustered together. In the red
cluster, authors Moshiri, B., Khamespanah, F., and Moradi, M., form a collaboration network.
Table 2 provides an overview of the research areas to which authors in the green and red clusters have
directed their focus. In Table 3, the contributions of each author to the keywords associated with the areas
outlined in Table 2 are outlined.
Page 788
SEMIT 2023
Table 2. Research focus areas of clusters for disaster management

Number of
Cluster Papers Areas of Research Focus
Earthquake
Decision Making
Vulnerability Assessment
Cluster 1 and Earthquake Vulnerability Assessment
16
Cluster 2 Seismic vulnerability assessment
Disaster Management
Decision Making
Uncertainty management
Table 3. Author-Keyword Analysis for Disaster Management

No of Paper Contributed to each Keyword
Cluster 1 Cluster 2
Khamespana Moradi, Moshiri, Delavar, M. Sheikhian, Stein,
h, F M. B. R. H. A.
MCDM ( (Multi Criteria Decision
1 1 6 2 1
Making)
Earthquake 1 1 1
Vulnerability Assessment 1 1 1
Granular Computing Model 2 2
Granular Computing 1 1 1
Seismic Vulnerability Assessment 2 5 1 1
Uncertainty management 1 1 1
Decision Making 1 1
Group Multi Criteria Decision
Making 1 2 2 2
Earthquake Vulnerability
Assessment 2 2 2
Disaster Management 1 1 1
GIS 1 1 1 4 1 1
When Table 3 is examined, it becomes clear that the most commonly used keywords in the works of the
authors presented in Figure 9 are “MCDM (Multi Criteria Decision Making)”, “Seismic Vulnerability
Assessment”, and “GIS”. These keywords are notably prevalent in the research conducted by Delavar,
M.R.
As a final analysis, the VOS viewer program was employed to explore the associations among keywords,
whereby publications with each keyword occurrence of 3 times or more were filtered. Out of the 190
keywords, 11 keywords satisfying this threshold criterion are presented in the network map provided in
Figure 10.
Page 789
SEMIT 2023
Fig. 10. Network visualization of co-word analysis

In Figure 10, where each color represents a distinct cluster, the sizes of nodes are proportionate to the
occurrence frequencies of keywords. Among 11 keywords, “disaster management” and “multi-criteria
decision making” are the most commonly employed terms, as confirmed by their highest total link
strength. In addition, the keyword “earthquake” exhibits an association with these commonly used terms.
Following, the keyword “seismic vulnerability assessment” is situated within the yellow-colored cluster,
while the “TOPSIS” keyword is combined with “Fuzzy AHP” in the blue-colored cluster. The red-colored
cluster encompasses “risk”, “MCDM”, “AHP (Analytic Hierarchy Process)”, and “GIS (Geographic
Information System)” keywords.
4. Data Sources and Statistical Analysis for Transportation in disaster conditions

In this section, bibliometric analysis has been conducted with a concentrate on transportation under
disaster conditions, which constitutes a critical component in the rescue of lives and the delivery of
emergency aid during disaster situations, leading to a more refined focus for this study. On October 9,
2023, an exploration was carried out, centered around the keywords "multi-criteria decision-making,"
"disaster," and "transportation," which resulted in the identification of 24 key findings [18, 22, 31, 37,
39, 43, 46, 48, 55, 64, 68, 72, 75, 77, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89]. The assessment encompassed
publications that were indexed within the Web of Science database.
6 5
4 4
4 3
2 2 2
2 1 1
0
2012 2016 2017 2018 2019 2020 2021 2022 2023
Fig. 11. The number of transportation related publications in disaster conditions in the Web of Science
database between 2012-2023
More than half of the 24 studies conducted between 2012 and 2023 were carried out in the year 2020 and
beyond. This trend is indicative of the growing importance of multi-criteria decision-making methods in
recent years in the context of making decisions concerning transportation under disaster conditions.
Table 4 provides a citation analysis based on citation thresholds of “>50,” “>25,” and “>10” over years.
Page 790
SEMIT 2023
Table 4. Annual WoS citation counts for transportation related publications

Year Total Publication WOS Total Citation >50 >25 >10
2012 2 7
2016 1 16 1
2017 1 1
2018 2 30 1
2019 3 22 1
2020 5 87 2 2
2021 2 26 1
2022 4 18
2023 4 0
According to Table 4, it can be observed that two studies have received more than 25 citations. These
particular studies, [43] and [80], attained the highest scores per publication in the year 2020, with a
maximum of 17.4 points. Following these, the years 2016 and 2018 are ranked next with 16 and 15 points
per publication, respectively.
Document Categories
Engineering Civil
Computer Science Information Systems
Management
Green Sustainable Science Technology
Computer Science Interdisciplinary Applications
Environmental Sciences
Engineering Multidisciplinary
Water Resources
Meteorology Atmospheric Sciences
Geosciences Multidisciplinary
0 1 2 3 4 5 6 7 8
Fig. 12. Distribution of transportation related publications based on the Web of Science categories.
In Figure 12, it is observed that three prominent Web of Science categories, namely Geosciences
Multidisciplinary, Meteorology Atmospheric Sciences, and Water Resources, each make an equal
contribution with seven studies (Each of them comprises 29.27% of the total 24 studies).
10
6 6
5
2 2
0
Iran Turkey China Thailand
Fig. 13. Top 4 countries with the most transportation related publications
Figure 13 presents the four countries with the highest number of publications among the 24 studies.
Among these countries, Turkey and Iran take the lead with six studies each, surpassing the other countries
to secure the top position.
Page 791
SEMIT 2023
Fig. 14. Geographic positions of all participating institutions for transportation related publications
Figure 14 provides the geographic locations of the institutions associated with studies contributing to
multi-criteria decision-making applications in disaster-related transportation. It is evident that Turkey and
Iran remain the countries with the highest density of circles.
This part of the research involved the performance of citation analysis, co-authorship analysis, and co-
word analysis on 24 studies. These studies employed multi-criteria decision-making techniques related
to transportation in disaster situations, and the data were sourced from the Web of Science Core
Collection database.
Sources with a minimum of two publications and two citations have been filtered for source citation
analysis, and the journals meeting the threshold value are presented in Figure 15. Among the 19 sources,
three met the threshold value.
Fig. 15. Density visualization of transportation related publications for source citation analysis
Among the 3 sources, the International Journal of Disaster Risk Reduction is ranked first in terms of both
the number of documents (4 documents) and citations (66 citations). Following that, the journal
Sustainability secures the second place with 2 documents and 11 citations, whereas the Disaster Advances
journal occupies the third spot with 2 documents and 7 citations.
To investigate the relationships among the countries, organizations, and authors associated with the 24
studies, a Co-authorship analysis was utilized. Among the 21 countries that were filtered for having at
least 2 citations and 2 publications, Turkey, Iran, and Thailand exceeded the prescribed threshold. Figure
16 presents a density map specific to these countries.
Page 792
SEMIT 2023
Fig. 16. Density visualization of transportation related publications for country analysis
Among the countries represented in Figure 16, Turkey leads in terms of density, backed by 6 publications
and 108 citations. Iran comes in second place with 6 publications and 28 citations, followed by Thailand
in third place with 2 publications and 18 citations.
Figure 17 illustrates the connections between the 9 authors, among a total of 52 authors, who have at least
1 publication and one citation and are situated within the cloud.
Fig. 17. Network visualization of transportation related publications for co-author analysis
When examining the nodes in proportion to the number of relevant publications, it is evident that within
the blue cluster, Gül, M., with the highest number of publications (3 publications, 55 citations),
collaborates most frequently with authors Yücesan, M., and Saner, H. S.
Similarly, authors in the green and red clusters have also regularly engaged in collaborative work with
fellow cluster members within the same clusters, thereby enhancing their network strength and,
consequently, increasing the visibility of their work.
In Table 5, the research domains that the authors in the green, blue, and red clusters presented in Figure
17 have focused on are provided. In Table 6, the contributions of each author to the keywords associated
with the fields in Table 5 are delineated.
Table 5. Research focus areas of clusters for transportation related publications
Cluster 1 Number of Papers Research Focus
1 3 Hospital disaster preparedness, Fuzzy decision making
2 3 Hospital disaster preparedness
3 3 Transportation planning, Supply chain network
Page 793
SEMIT 2023
Table 6. Author-Keyword Analysis for Disaster Management

No of Paper Contributed to Each Keyword
Cluster 1 Cluster 2 Cluster3
Ortiz- López- Navarro- Gul, Yucesa Saner, Lo, Liaw, Lin,
Keyword Barrios, M Meza, P Jiménez, E M n, M HS HW CF KY
Disaster
1 1 2 2 1
preparedness 1
Fuzzy decision
1 1 1 1 1
making
TOPSIS 1 1 1 2 2 1
Transportation
1
planning 1 1 1
Supply chain
1
network 1 1 1
MCDM 1 1 1 1
As indicated in Table 6, it is apparent that the authors in the blue and green clusters showcased in Figure
17 have placed their emphasis on the subject of hospital disaster preparedness in their studies.
Additionally, it can be observed that the keywords “Disaster preparedness” and “TOPSIS” received
greater emphasis from Gul, M, and Yucesan, M.
For keyword analysis, keywords that occurred a minimum of two times within the set of 104 keywords
were filtered, and the connections between the 8 keywords that met the threshold value are presented in
Figure 18.
Fig. 18. Network visualization of transportation related publications for co-word analysis
The keyword 'multi-criteria decision making' holds the highest total link strength of 5 and shares the same
cluster as the keyword 'transportation network performance.' Following closely, the keyword 'TOPSIS'
has a total link strength of 4 and is clustered together with 'AHP.' The keywords 'disaster management'
and 'multi-criteria decision-making' are grouped within the yellow cluster, while the subsequent green
cluster features the coexistence of 'disaster preparedness' and 'humanitarian logistics,' interconnected by
their link strength.
Page 794
SEMIT 2023
5. Conclusion
In recent years, the escalating natural disasters worldwide, coupled with unprecedented challenges
imposed upon societies, such as pandemics and earthquakes, have accentuated the greater importance of
disaster management and rapid transportation. This bibliometric analysis has been conducted with the
aim of guiding researchers and decision-makers in comprehending the potential of Multiple Criteria
Decision Making (MCDM) techniques in the realm of disaster management and transportation in disaster
conditions. Within the first part of the study, a science mapping of publications containing the keywords
“multi-criteria decision-making” and “disaster management”, retrieved from the Web of Science
database, has been carried out using the VOS Viewer software. In the second part, particular attention is
given to the topic of transportation, which plays a vital role in ensuring the swift movement of relief
supplies, emergency response teams, and evacuees in disaster situations, with the aim of minimizing the
consequences of the disaster. This portion of the study includes a bibliometric analysis of studies
containing the keywords 'multi-criteria decision-making,' 'disaster,' and 'transportation. The findings of
the study reveal a holistic view of the trends, patterns, and dynamics within the academic discourse of
disaster management and transportation that playing one of the most effective role in this process
concerning research on Multiple Criteria Decision Making (MCDM).
While this study illuminates forthcoming research opportunities regarding the application of MCDM
techniques in disaster conditions, with their significant potential for decision-making and effective
strategy development, it is essential to note that the study has certain limitations. Only academic studies
retrieved from the Web of Science database were examined within the scope of this study. However, in
future bibliometric analysis studies, the inclusion of other databases for scrutiny could potentially broaden
the study's scope and contribute to making disaster management processes more effective and sustainable.
Another limitation pertains to the lack of detailed examinations incorporating geographic and cultural
variations that might impact the role of MCDM techniques in disaster management and transportation in
disaster conditions. In spite of these limitations, it is deemed that this study constitutes a substantial step
towards achieving its aims and encouraging further research in this field.
References
1. Seedorf, J., Tagami, A., Arumaithurai, M., Koizumi, Y., Melazzi, N. B., Kutscher, D., ... & Psaras, I.
(2015). The benefit of information centric networking for enabling communications in disaster
scenarios. In 2015 IEEE Globecom Workshops (GC Wkshps) (pp. 1-7). IEEE.
2. Daud, S. M. S. M., Yusof, M. Y. P. M., Heo, C. C., Khoo, L. S., Singh, M. K. C., Mahmood, M. S.,
& Nawawi, H. (2022). Applications of drone in disaster management: A scoping review. Science &
Justice, 62(1), 30-42.
3. Burling, W. K., & Hyle, A. E. (1997). Disaster preparedness planning: policy and leadership issues.
Disaster Prevention and Management: An International Journal, 6(4), 234-244.
4. Manyaga, F., Nilufer, N., Hajaoui, Z. (2020). A systematic literature review on multi-criteria decision
making in disaster management. International Journal of Business Ecosystem & Strategy (2687-
2293), 2(2), 1-7.
5. Gul, M., Yucesan, M., & Erdogan, M. (2022). (Eds.). Multi-Criteria Decision Analysis: Case Studies
in Disaster Management. CRC Press.
6. Donthu, N., Kumar, S., Mukherjee, D., Pandey, N., & Lim, W. M. (2021). How to conduct a
bibliometric analysis: An overview and guidelines. Journal of business research, 133, 285-296
7. Hota, P. K., Subramanian, B., & Narayanamurthy, G. (2020). Mapping the intellectual structure of
social entrepreneurship research: A citation/co-citation analysis. Journal of business ethics, 166(1),
89-114.
Page 795
SEMIT 2023
8. Cisneros, L., Ibanescu, M., Keen, C., Lobato-Calleros, O., & Niebla-Zatarain, J. (2018). Bibliometric
study of family business succession between 1939 and 2017: mapping and analyzing authors’
networks. Scientometrics, 117, 919-951.
9. Leung, X. Y., Sun, J., & Bai, B. (2017). Bibliometrics of social media research: A co-citation and co-
word analysis. International Journal of Hospitality Management, 66, 35-45.
10. Hjørland, B. (2013). Facet analysis: The logical approach to knowledge organization. Information
processing & management, 49(2), 545-557.
11. Delavar, M. R., Moradi, M., & Moshiri, B. (2015). Earthquake vulnerability assessment for hospital
buildings using a GIS-based group multi criteria decision making approach: a case study of Tehran,
Iran. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information
Sciences, 40, 153-157.
12. Asghar, S. (2009, October). A survey on multi-criteria decision making approaches. In 2009
International conference on emerging technologies (pp. 321-325). IEEE.
13. Mohammadnazari, Z., Mousapour Mamoudan, M., Alipour-Vaezi, M., Aghsami, A., Jolai, F., &
Yazdani, M. (2022). Prioritizing post-disaster reconstruction projects using an integrated multi-
criteria decision-making approach: A case study. Buildings, 12(2), 136.
14. Atmaca, E., Aktaş, E., & Öztürk, H. N. (2023). Evaluated Post-Disaster and Emergency Assembly
Areas Using Multi-Criteria Decision-Making Techniques: A Case Study of Turkey. Sustainability,
15(10), 8350.
15. Moheimani, A., Ishizaka, A., Hosseini, S. M. H., & Mangla, S. K. (2023). Sorting radiology
departments in a disaster management assessment with G-ARASsort. International Journal of
Production Research, 1-24.
16. Nain, A., Jain, D., & Trivedi, A. (2023). Multi-criteria decision-making methods: application in
humanitarian operations. Benchmarking: An International Journal.
17. Caroleo, B., Palumbo, E., Osella, M., Lotito, A., Rizzo, G., Ferro, E., ... & De Gregorio, D. (2018).
A knowledge-based multi-criteria decision support system encompassing cascading effects for
disaster management. International Journal of Information Technology & Decision Making, 17(05),
1469-1498.
18. Fariza, A., Basofi, A., Prasetyaningrum, I., & Pratiwi, V. I. (2020). Urban flood risk assessment in
sidoarjo, indonesia, using fuzzy multi-criteria decision making. In Journal of Physics: Conference
Series (Vol. 1444, No. 1, p. 012027). IOP Publishing.
19. Sayyed Ali , A., Abolfazl , M. ., Mohammad , E. (2020). Operational site selection for disaster
management bases in Tehran, Iran. MAUSAM, 71(3), 431–442.
20. Palomares, I., Galway, L., Haran, M., Neef, M., Woods, C., & Wang, H. (2015, November). A
knowledge management and need-capacity matching approach for community-based disaster
management and recovery. In 2015 10th International Conference on Intelligent Systems and
Knowledge Engineering (ISKE) (pp. 389-396). IEEE.
21. Kanani-Sadat, Y., Arabsheibani, R., Karimipour, F., & Nasseri, M. (2019). A new approach to flood
susceptibility assessment in data-scarce and ungauged regions based on GIS-based hybrid multi
criteria decision-making method. Journal of hydrology, 572, 17-31.
22. Ghavami, S. M., Maleki, J., & Arentze, T. (2019). A multi-agent assisted approach for spatial Group
Decision Support Systems: A case study of disaster management practice. International Journal of
Disaster Risk Reduction, 38, 101223.
23. Tezcan, B., Alakaş, H. M., Özcan, E., Eren, T. (2021). Afet sonrası geçici depo yeri seçimi ve çok
araçlı araç rotalama uygulaması: Kırıkkale ilinde bir uygulama. Politeknik Dergisi, 26(1), 13-27.
Page 796
SEMIT 2023
24. Nyimbili, P. H., & Erden, T. (2018). Spatial decision support systems (SDSS) and software
applications for earthquake disaster management with special reference to Turkey. Natural Hazards,
90, 1485-1507.
25. Choukolaei, H. A., Ghasemi, P., & Goodarzian, F. (2023). Evaluating the efficiency of relief centers
in disaster and epidemic conditions using multi-criteria decision-making methods and GIS: A case
study. International Journal of Disaster Risk Reduction, 85, 103512.
26. Barutcu, I., & Ic, Y. T. (2022). Selecting the field hospital location for earthquakes: an application
for Ankara Province in Turkey. International Journal of Emergency Services, 11(1), 168-187.
27. Sotoudeh-Anvari, A., Sadjadi, S. J., Hadji Molana, S. M., & Sadi-Nezhad, S. (2019). A stochastic
multi-objective model based on the classical optimal search model for searching for the people who
are lost in response stage of earthquake. Scientia Iranica, 26(3), 1842-1864.
28. Hoque, M. A. A., Phinn, S., Roelfsema, C., & Childs, I. (2017). Tropical cyclone disaster
management using remote sensing and spatial analysis: A review. International journal of disaster
risk reduction, 22, 345-354.
29. Peker, I., Ar, I. M., Erol, I., & Searcy, C. (2022). Leveraging blockchain in response to a pandemic
through disaster risk management: an IF-MCDM framework. Operations Management Research, 1-
26.
30. Sheikhian, H., Delavar, M. R., & Stein, A. (2015). Uncertainty handling in disaster management using
hierarchical rough set granulation. ISPRS Annals of the Photogrammetry, Remote Sensing and
Spatial Information Sciences, 2, 271-276.
31. Sabzevari, A. R., & Delavar, M. R. (2017). Gis-based site selection for underground natural resources
using fuzzy ahp-owa. The International Archives of the Photogrammetry, Remote Sensing and Spatial
Information Sciences, 42, 463-468.
32. Üstün, A. K., Anagün, A. S. (2016). Determination of importance weights of Istanbul's distrıcts using
analytic hıerarchy process.
33. Seo, J. P., Cho, W., & Cheong, T. S. (2015). Development of priority setting process for the small
stream restoration projects using multi criteria decision analysis. Journal of Hydroinformatics, 17(2),
211-225.
34. Sushil. (2019). Efficient interpretive ranking process incorporating implicit and transitive dominance
relationships. Annals of Operations Research, 283(1-2), 1489-1516.
35. Khamespanah, F., Delavar, M. R., & Zare, M. (2013). Uncertainty management in seismic
vulnerability assessment using granular computing based on covering of universe. The International
Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 40, 121-126.
36. Khamespanah, F., Delavar, M. R., Alinia, H. S., & Zare, M. (2013). Granular computing and
Dempster–Shafer integration in seismic vulnerability assessment. Intelligent Systems for Crisis
Management: Geo-information for Disaster Management (Gi4DM) 2012, 147-158.
37. Saksrisathaporn, K., Bouras, A., Reeveerakul, N., & Charles, A. (2016). Application of a decision
model by using an integration of AHP and TOPSIS approaches within humanitarian operation life
cycle. International Journal of Information Technology & Decision Making, 15(04), 887-918.
38. Pahlavani, P., Sheikhian, H., & Bigdeli, B. (2020). Evaluation of residential land use compatibilities
using a density-based IOWA operator and an ANFIS-based model: A case study of Tehran, Iran.
Land Use Policy, 90, 104364.
39. He, Y., & Jung, H. (2018). A voting TOPSIS approach for determining the priorities of areas damaged
in disasters. Sustainability, 10(5), 1607.
Page 797
SEMIT 2023
40. Sadrykia, M., Delavar, M. R., & Zare, M. (2017). A GIS-based decision making model using fuzzy
sets and theory of evidence for seismic vulnerability assessment under uncertainty (case study:
Tabriz). Journal of Intelligent & Fuzzy Systems, 33(3), 1969-1981.
41. Ekmekcioğlu, Ö., Koc, K., Özger, M. (2021). Stakeholder perceptions in flood risk assessment: A
hybrid fuzzy AHP-TOPSIS approach for Istanbul, Turkey. International Journal of Disaster Risk
Reduction, 60, 102327.
42. Chauhan, A., Debnath, R. M., & Singh, S. P. (2018). Modelling the drivers for sustainable agri-food
waste management. Benchmarking: An International Journal, 25(3), 981-993.
43. Yılmaz, H., & Kabak, Ö. (2020). Prioritizing distribution centers in humanitarian logistics using type-
2 fuzzy MCDM approach. Journal of Enterprise Information Management, 33(5), 1199-1232.
44. Alam, M. S., & Haque, S. M. (2022). Multi-dimensional earthquake vulnerability assessment of
residential neighborhoods of Mymensingh City, Bangladesh: A spatial multi-criteria analysis based
approach. Journal of Urban Management, 11(1), 37-58.
45. Otay, I., & Jaller, M. (2020). Multi-expert disaster risk management & response capabilities
assessment using interval-valued intuitionistic fuzzy sets. Journal of Intelligent & Fuzzy Systems,
38(1), 835-852.
46. Erden, C., Ateş, Ç., & Esen, S. (2023). Distribution Center Location Selection in Humanitarian
Logistics Using Hybrid BWM–ARAS: A Case Study in Türkiye. Journal of Homeland Security and
Emergency Management, (0).
47. Nekooie, M. A., & Tofighi, M. (2020). Resilient and sustainable modular system for temporary
sheltering in emergency condition. VITRUVIO-International Journal of Architectural Technology
and Sustainability, 5(2), 1-15.
48. Parizi, S. M., Taleai, M., & Sharifi, A. (2022). A GIS-Based Multi-Criteria Analysis Framework to
Evaluate Urban Physical Resilience against Earthquakes. Sustainability, 14(9), 5034.
49. Sadrykia, M., Delavar, M. R., & Zare, M. (2017). A GIS-based fuzzy decision making model for
seismic vulnerability assessment in areas with incomplete data. ISPRS International Journal of Geo-
Information, 6(4), 119.
50. Mouhoumed, R. M., Ekmekcioğlu, Ö., & Özger, M. (2023). A hybrid MCDA approach for
delineating sites suitable for artificial groundwater recharge using drywells. Journal of Hydrology,
620, 129387.
51. Gharakhlou, M. A. H. D. I., Faraji Sabokbar, H. A., & Givehchi, S. (2010). Access Enhancement by
Making Changes in the Route Network to Facilitate Rescue perations in Urban Disasters.
International Journal of Environmental Research, 4(2), 183-192.
52. Koc, K., Ekmekcioğlu, Ö., & Işık, Z. (2023). Developing a probabilistic decision-making model for
reinforced sustainable supplier selection. International Journal of Production Economics, 259,
108820.
53. Sheikhian, H., Delavar, M. R., & Stein, A. (2015). Integrated estimation of seismic physical
vulnerability of Tehran using rule based granular computing. The International Archives of the
Photogrammetry, Remote Sensing and Spatial Information Sciences, 40, 187-193.
54. Moradi, M., Delavar, M. R., Moshiri, B., & Khamespanah, F. (2014). A novel approach to support
majority voting in spatial group MCDM using density induced OWA operator for seismic
vulnerability assessment. The International Archives of the Photogrammetry, Remote Sensing and
Spatial Information Sciences, 40, 209-214.
55. Timperio, G., Tiwari, S., Lee, C. K., Samvedi, A., & de Souza, R. (2020). Integrated decision support
framework for enhancing disaster preparedness: A pilot application in Indonesia. International
Journal of Disaster Risk Reduction, 51, 101773.
Page 798
SEMIT 2023
56. Yang, C. L., Shieh, M. C., Huang, C. Y., & Tung, C. P. (2018). A derivation of factors influencing
the successful integration of corporate volunteers into public flood disaster inquiry and notification
systems. Sustainability, 10(6), 1973.
57. López, C., Ishizaka, A., Gul, M., Yücesan, M., & Valencia, D. (2022). A calibrated Fuzzy Best-
Worst-method to reinforce supply chain resilience during the COVID 19 pandemic. Journal of the
Operational Research Society, 1-24.
58. Ortiz-Barrios, M., Silvera-Natera, E., Petrillo, A., Gul, M., & Yucesan, M. (2022). A multicriteria
approach to integrating occupational safety & health performance and industry systems productivity
in the context of aging workforce: A case study. Safety science, 152, 105764.
59. Taherizadeh, M., Niknam, A., Nguyen-Huy, T., Mezősi, G., & Sarli, R. (2023). Flash flood-risk areas
zoning using integration of decision-making trial and evaluation laboratory, GIS-based analytic
network process and satellite-derived information. Natural Hazards, 1-27.
60. Li, J., Li, Q., Liu, C., Khan, S. U., & Ghani, N. (2014). Community-based collaborative information
system for emergency management. Computers & operations research, 42, 116-124.
61. Munpa, P., Kittipongvises, S., Phetrak, A., Sirichokchatchawan, W., Taneepanichskul, N.,
Lohwacharin, J., & Polprasert, C. (2022). Climatic and Hydrological Factors Affecting the
Assessment of Flood Hazards and Resilience Using Modified UNDRR Indicators: Ayutthaya,
Thailand. Water, 14(10), 1603.
62. Silavi, T., Malek, M. R., & Delavar, M. R. (2006). Multicriteria map overlay in geospatial information
system via intuitionistic fuzzy AHP method. In Applied Artificial Intelligence (pp. 401-408).
63. Yucesan, M., Başhan, V., Demirel, H., & Gul, M. (2022). An interval type-2 fuzzy enhanced best–
worst method for the evaluation of ship diesel generator failures. Engineering Failure Analysis, 138,
106428.
64. Gergin, R. E., & Peker, İ. (2019). Literature review on success factors and methods used in warehouse
location selection. Pamukkale üniversitesi mühendislik bilimleri dergisi, 25(9), 1062-1070.
65. Ekmekcioğlu, Ö. (2023). On the identification of most appropriate green roof types for urbanized
cities using multi-tier decision analysis: A case study of Istanbul, Turkey. Sustainable Cities and
Society, 104707.
66. Akbarian, H., Gheibi, M., Hajiaghaei-Keshteli, M., & Rahmani, M. (2022). A hybrid novel
framework for flood disaster risk control in developing countries based on smart prediction systems
and prioritized scenarios. Journal of environmental management, 312, 114939.
67. Mondal, T., Bhattacharya, I., Pramanik, P., Boral, N., Roy, J., Saha, S., & Saha, S. (2018). A multi-
criteria evaluation approach in navigation technique for micro-jet for damage & need assessment in
disaster response scenarios. Knowledge-Based Systems, 162, 220-237.
68. Ghavami, S. M. (2019). Multi-criteria spatial decision support system for identifying strategic roads
in disaster situations. International Journal of Critical Infrastructure Protection, 24, 23-36.
69. Yadav, D. K., & Barve, A. (2019). Prioritization of cyclone preparedness activities in humanitarian
supply chains using fuzzy analytical network process. Natural Hazards, 97, 683-726.
70. Kaya, C. M., & Derin, L. (2023). Parameters and methods used in flood susceptibility mapping: a
review. Journal of Water and Climate Change.
71. Zhu, J., Orooji, F., & Chelson, D. (2018). Optimal material selection algorithm for post-disaster
shelters applying analytic hierarchy process: a case study of refugee shelters in Germany. In
Construction Research Congress 2018 (pp. 563-572).
72. Littidej, P., Uttha, T., & Pumhirunroj, B. (2022). Spatial Predictive Modeling of the Burning of
Sugarcane Plots in Northeast Thailand with Selection of Factor Sets Using a GWR Model and
Machine Learning Based on an ANN-CA. Symmetry, 14(10), 1989.
Page 799
SEMIT 2023
73. Celik, E., Yucesan, M., & Gul, M. (2021). Green supplier selection for textile industry: a case study
using BWM-TODIM integration under interval type-2 fuzzy sets. Environmental Science and
Pollution Research, 28, 64793-64817.
74. Debnath, B., Shakur, M. S., Bari, A. M., Saha, J., Porna, W. A., Mishu, M. J., ... & Rahman, M. A.
(2023). Assessing the critical success factors for implementing industry 4.0 in the pharmaceutical
industry: Implications for supply chain sustainability in emerging economies. Plos one, 18(6),
e0287149.
75. Lo, H. W., Liaw, C. F., Gul, M., & Lin, K. Y. (2021). Sustainable supplier evaluation and
transportation planning in multi-level supply chain networks using multi-attribute-and multi-
objective decision making. Computers & Industrial Engineering, 162, 107756.
76. Moradi, M., Delavar, M. R., & Moshiri, B. (2013). Sensitivity analysis of ordered weighted averaging
operator in earthquake vulnerability assessment. The International Archives of the Photogrammetry,
Remote Sensing and Spatial Information Sciences, 40, 277-282.
77. Saner, H. S., Yucesan, M., & Gul, M. (2021). A Bayesian BWM and VIKOR-based model for
assessing hospital preparedness in the face of disasters. Natural hazards, 1-33.
78. Polat, E. G., Yücesan, M., & Gül, M. (2023). A comparative framework for criticality assessment of
strategic raw materials in Turkey. Resources Policy, 82, 103511.
79. Talukdar, S., Naikoo, M. W., Mallick, J., Praveen, B., Sharma, P., Islam, A. R. M. T., ... & Rahman,
A. (2022). Coupling geographic information system integrated fuzzy logic-analytical hierarchy
process with global and machine learning based sensitivity analysis for agricultural suitability
mapping. Agricultural Systems, 196, 103343.
80. Ortiz-Barrios, M., Gul, M., López-Meza, P., Yucesan, M., & Navarro-Jiménez, E. (2020). Evaluation
of hospital disaster preparedness by a multi-criteria decision making approach: The case of Turkish
hospitals. International journal of disaster risk reduction, 49, 101748.
81. Khalid, M. A., & Ali, Y. (2020). Economic impact assessment of natural disaster with multi-criteria
decision making for interdependent infrastructures. Environment, Development and Sustainability,
22, 7287-7311.
82. Munyaka, J. C. B., & Yadavalli, V. S. S. (2021). Decision support framework for facility location
and demand planning for humanitarian logistics. International Journal of System Assurance
Engineering and Management, 12, 9-28.
83. Pouryari, M., Mahboobi Ardakani, A. R., & Hassani, N. (2022). A Multi-Criteria Vulnerability of
Urban Transportation Systems Analysis Against Earthquake Considering Topological and
Geographical Method: A Case Study. Iranian Journal of Science and Technology, Transactions of
Civil Engineering, 1-14.
84. Xu, W., Lu, Y., & Proverbs, D. (2023). An evaluation of factors influencing the vulnerability of
emergency logistics supply chains. International Journal of Logistics Research and Applications, 1-
34.
85. Anvari, M., Anvari, A., & Boyer, O. (2023). A prepositioning model for prioritized demand points
considering lateral transshipment. Journal of Humanitarian Logistics and Supply Chain Management.
86. Rahat, R., Pradhananga, P., Calle Muller, C., & ElZomor, M. (2023). Multi-criteria evaluation of
Safe-to-Fail flood solution alternatives for developing resilient coastal cities. Sustainable and
Resilient Infrastructure, 1-16.
87. Fengwei, L., Xiuli, D., & Mingju, Z. (2012). Transitive Comparison Method and its application in
Collapse Risk Identification. Disaster Advances, 5(4), 609-613.
88. Vaani, N., & Sekar, S. K. (2012). Regional landslide hazard zonation and vulnerability analysis using
AHP and GIS-a case study of Nilgiris district, Tamil Nadu, India. Disaster advances, 5(4), 171-176.
Page 800
SEMIT 2023
89. Drakaki, M., Gören, H. G., & Tzionas, P. (2018). An intelligent multi-agent based decision support
system for refugee settlement siting. International Journal of Disaster Risk Reduction, 31, 576-588.
Page 801
SEMIT 2023
Food Distribution in Times of Famine

Mete Gündoğan1[0000-0002-6753-9954] and Murat Ata2[0000-0002-3770-9983]
1
Ankara Yıldırım Beyazıt University, Department of Industrial Engineering, Ankara, Turkey
2
TUBİTAK, Ankara, Turkey
Abstract
Regardless of the ongoing debate surrounding the existence of 'global warming,' the heightened attention
to this issue in recent years underscores the heightened likelihood of 'famine' or 'drought' occurrences.
These events may arise from a variety of factors, including human actions, environmental challenges, or
natural phenomena that manifest intermittently. Consequently, it becomes imperative to proactively take
measures to mitigate the risks associated with famine and drought.
In addition to the imperative of averting resource wastage, there is a pressing need to ensure the judicious
and equitable utilization of the limited available resources, at least for a certain duration. Addressing this
challenge necessitates effective solutions for identifying, procuring, and distributing essential resources.
Key elements in this context include water resources, long-lasting dry food supplies, seasonally adaptable
fresh food production, and access to healthcare resources and medications. Moreover, the duration and
predictability of shortages emerge as crucial considerations.
This study proposes a model for the procurement and distribution of such essential resources in the
medium term, under the assumption of open and well-developed communication channels. The model is
based on the Structured Systems Analysis and Design Methodology (SSADM), offering adaptability to
diverse scenarios.
Keywords: Famine, Resource Management, Climate Change, Food Security, Humanitarian Aid,
Emergency Response, Disaster Preparedness, Supply Chain Management, Crisis Management, Aid
Organizations, Humanitarian Relief
1. Introduction
Famine stands as one of the most formidable challenges in human history. Disruptions in food production
due to natural disasters, economic hardships, and other factors can threaten access to essential nutrition.
During periods of famine, the development of an effective food distribution and resource management
strategy becomes paramount for survival and societal stability.
In the Ottoman Empire, 19 periods of famine and starvation were identified between 1564 and 1612.
According to archival documents, in 1676, 1679, 1696, 1715, 1725, 1746, 1746, 1757, 1797, 1815, 1887,
seven times in the Mediterranean Region and five times in the Black Sea Region, droughts lasting at least
two years were experienced. Drought waves in 1844-1846, 1879-1881 and in 1887, 1892-1893, 1905-
1906 and 1907-1910 in the Tripolitania Province, and droughts in 1922 Sinop, 1923 Haymana, 1928
Central Anatolia, 1929 Malatya, 1933 Central Anatolia and Eastern Anatolia, 1935, 1939, 1942, 1945,
1947-1950, 1955-1962, 1970-1973, 1977, 1982-1984, 1994 in the Republican period.[1]
In the years between the First and Second World Wars, a number of measures were taken and legal
arrangements were made due to the poverty and shortages. While trying to heal the wounds of the First
World War and the period of National Struggle, a law was passed in 1920 and it was decided to impose
some sanctions on those who did not comply with the rules of mobilization.[2]
One of the first things the state did in times of disaster was to determine the extent of the disaster. This
practice was not a one-off, but was based on the reports of local administrators and sometimes on the
results of the investigations carried out in the region by high-level officials sent by Istanbul in order to
update the information at certain periods of the year. The aim of the state in obtaining information about
the disaster area was to determine the extent of the disaster and to provide more effective assistance to
the disaster victims. As a result of these investigations, the needy people in the disaster area were
Page 802
SEMIT 2023
identified, these people were recorded in the registers called “inan defteri” and the aid was provided
according to these records.[3]
In addition, it is expected that the production part of the countries will develop in this process and reduce
dependence on exports. In other words, the adoption of the concept of “field to fork”, that is, from
production to consumption, will make this period more comfortable. Localization of the supply chain is
also envisaged, along with the conversion of production to internal resources. It is expected that countries
with products in their hands will beat the pandemic faster and become stronger.[4]
In this study, we will try to develop an alternative food distribution management plan will can be
implemented by the society. We have developed a model for managing large crises in the future.
1.1. The Problem
What is famine?
The United Nations employs a five-step scale known as the Integrated Food Security Phase
Classification (IPC), which was developed in collaboration with non-governmental organizations like
Oxfam, to evaluate a country's food security. In Stage 5, referred to as "famine/humanitarian catastrophe,"
several critical criteria must be met, including:
1. Mortality Rate: More than two individuals per 10,000 die daily.
2. Acute Malnutrition: Acute malnutrition rates exceed 30 percent.
3. Livestock: All livestock in the affected area are deceased.
4. Food Availability: There is less than 2,100 kilocalories of food and 4 liters of water available per
person daily.
Stage 5 represents the most dire and catastrophic level of food insecurity and requires immediate and
substantial humanitarian intervention.
Fig. 28. Acute Food Insecurity Classification

People facing food insecurity raised Jan 2020 135.000.000 to 265.000.000 on Dec 2020. In spite of a
2.3% rise in global cereal production in 2019 compared to 2018, the count of individuals confronting
acute food insecurity is projected to potentially double, escalating from 135 million in January 2020 to
Page 803
SEMIT 2023
265 million by the conclusion of 2020. This highlights a concerning trend in the midst of apparent food
production increases.
The production of wheat, corn and rice in Turkey has been static or even declining over the years.
According to the UN fifth national report on Biological Diversity (2014), the growing area has been
restricted and allowed to be grown in a narrower area due to policies applied to certain types of grain in
Turkey over the years. Static or decreasing production amounts in these products were tried to be
compensated by an increase in imports. [5]
According to the Turkish Grain Board (TMO) 2019 grain sector report, 2.6 million tons of wheat
(including durum wheat), 0.05 million tons of barley, 0.4 million tons of corn, 0.2 million tons of rice
were imported in 2010. In 2019 the same products, wheat 11.3, barley 0.86, corn 3.9, rice 0.2 million tons
were imported (Figure 2). Although the numbers have varied over the years, we can say that there have
been dramatic increases in imports of products other than rice.
Fig. 29. 2010-2020 Grain Imports

The production of wheat, corn and rice in Turkey has been static or even declining over the years.
Fig. 30. Turkish Grain Board, Grain Sector Report 2020
When we say “disaster”, it is meant natural events such as earthquakes, floods, fires, which usually
happened in a not very long time and the effects of which are also relatively short. But, for example, there
is no action plan in Turkey for situations of prolonged famine or difficult access to vital needs for a period
Page 804
SEMIT 2023
of time. Also we cannot found anything about neighborhood or agriculture and sustainability of fresh
food.
Home gardening, urban and peri-urban agriculture (UPA) can be very important. The concept of UPA
pertains to an industry situated within or at the periphery of an urban area. It revolves around the
utilization of locally available products, services, human resources, and natural resources for the
cultivation, processing, and distribution of agricultural products. [5], [6]
The need for such long distances and long and complex supply chains to reach basic foodstuffs will
especially worsen the effects of crises. It will make conditions that are already difficult enough to become
even more unbearable. All segments of society will be adversely affected by such situations. Social
solidarity and healthy communication will be of vital importance. Therefore, it is necessary to plan these
and project them into the future. We think our model will especially respond to this need.[7]
In such cases, the actions taken must be within a certain plan and the results must be calculable. In order
for the steps taken to be accurate, possible results should have been calculated within reasonable margins
of error beforehand. If it is worked within a plan, the probability of encountering such problems is
reduced.[7]
2. The Method
2.1. SSADM
Structured Systems Analysis and Design Method, commonly abbreviated as SSADM, is a widely used
systems development methodology. It provides a framework for designing and implementing information
systems.
The Structured Systems Analysis and Design Method (SSADM) was the prevalent structured approach
employed for information systems in the United Kingdom's government departments during the early
1980s. SSADM's distinctive feature is its segregation of logical design from physical design. It focuses
on creating a hardware/software-agnostic logical design that can subsequently be translated into an initial
physical design.
This approach offers several advantages. It enables developers to tackle one problem at a time, avoiding
the premature introduction of unnecessary constraints during the development process. It also facilitates
communication with users who might not possess advanced computer literacy but can still validate a
logical specification or design for their system. This approach ensures that system development proceeds
efficiently and aligns with the needs and capabilities of both developers and users.[8]
The primary goal of SSADM was to improve the analysis and design of information systems within the
UK government. The main objective of SSADM is to produce high-quality information systems that meet
the needs of the end-users while staying within time and budget constraints. It aims to provide a clear and
comprehensive framework for the analysis and design of information systems.
In the construction of SSADM, several key components are utilized, as exemplified in Figure 4. The
stages of the process are delineated in Figure 5. The context boundary plays a pivotal role in defining the
system's boundaries. It allows for the segregation of the system's internal and external environments,
enabling them to be treated separately. Within the system, activity boxes serve as sub-elements, further
partitioning and detailing its components and functions.
Page 805
SEMIT 2023
Fig. 31. Components of SSADM
Fig. 32. Stages of SSADM [8]
Provides a systematic and structured approach to system development. Ensures that user requirements
are thoroughly understood and incorporated into the system design. Helps in managing and controlling
the development process, ensuring that projects stay on track. Facilitates documentation, making it easier
for developers to understand and maintain the system.
Can be time-consuming, especially for small projects where a less formal methodology might be more
appropriate. May require extensive training for team members to become proficient in the methodology.
Some critics argue that it can be too rigid and may not adapt well to rapidly changing requirements.
Originally developed for use in government IT projects, SSADM has found applications in various
industries, including finance, healthcare and telecommunications. It is particularly suited for large-scale,
complex information systems where a structured approach is essential.
Page 806
SEMIT 2023
Overall, SSADM remains a prominent methodology in the field of systems analysis and design, providing
a robust framework for developing high-quality information systems.
2.2. System
A few people realize it, but we live in an enormous world and universe whose every detail is delicately
calculated, measured, and based on an extremely delicate balance. When we look at the construction and
structure of many events or objects, we can see that they actually have much more detail and finesse than
they appear from the outside. We can often evaluate them like “the natural course of life, how else could
it have happened anyway?” As much as we are familiar with the characteristics of this environment that
we have been in since we were born, even so, we are somewhat blind to them.
In particular, we link almost all kinds of corruption in nature to the evil deeds that people do and try to
learn some lessons. So, when we turn our heads from these wonders that have not been done by human
hands and look at what human beings have done, we can identify many mistakes and negativity made by
those responsible, whether you call it smugness or jealousy. Especially the subject in one of the areas
where we are experts “here should be like this, why did they do this here” we can make many criticisms.
They even say “the master doesn't like the other's work.”
On one side are formations created by Allah (CC) and which everyone believes to be perfect. Although
these have been studied and investigated for years, they still have not been reached to the end and no
errors have been found in the discovered parts. On the other hand, there are structures that people make
and almost everyone needs correction or change, including who make them.[7]
When you mention the term "System," it evokes diverse interpretations for different individuals. The
connotation of this concept undeniably hinges on the knowledge we gain, the occupation we pursue, the
milieu we inhabit, the experiences we engage with, in essence, our accumulated insights. "System" can
signify an operating framework or a web of connections for those fascinated by computing, a
comprehensive assembly of components constituting a device or a manufacturing site for engineers, and
a learning process or a logistical network for servicemen. Nevertheless, notwithstanding its immense and
intricate nature, encompassing realms like the "healthcare system, educational infrastructure, and
economic framework," to which we are all accustomed, there also exist establishments that serve as a
unifying force for nearly everyone.
The concept of "the system" can typically be distilled as an entirety that engages with its surroundings,
the constituents of that entirety, and the intricate interplay among them. In line with Aristotle's wisdom,
"The whole is greater than the sum of its components. A part transcends being merely a fraction of the
entirety." Thus, when we delve into the realm of systems, the outcome of adding two and two might well
surpass the simple sum of four. Similarly, when we dissect four into two pairs, the resulting pieces can
exceed two in size.[9]
Systems are intricate and organized structures. To comprehend these systems, we can begin with the most
fundamental structure: the human being. In the beginning Allah(C.C.) had the initial "wish" to create
humanity. In this first stage, humankind existed solely as an idea. Subsequently, God crafted human
beings in the most exquisite form, molding them from the earth. He breathed life into them with His own
spirit, completing the creation process. Thus, human beings came into existence in this world and had to
adapt to the conditions of the world. They learned to survive and make use of the world's blessings.
Ultimately, individuals age, pass away, and return to the earth. This cycle is a recurring pattern for every
person and can be divided into several stages. If we were to sort these stages [10];
• Conception
• Design
• Creation
• Transition to Use
• Use, Life
• Senility, aging
• Replacement
Page 807
SEMIT 2023
Indeed, there exists a parallel cycle in the world we inhabit. Could it be said that it all commences with
an idea and culminates with mortality?
These phases extend not only to the grand scope of existence but also resonate within the intricacies of
our own bodies. We observe a flawless mechanical system within us, one meticulously poised and
harmonized. It comprises cells, tissues, organs, organ systems, and ultimately the organism, which is our
own body. Furthermore, when individuals unite, they give rise to families and communities, which, in
turn, coalesce into tribes or nations. It becomes evident that each of these entities serves as a subsystem
of the next (as depicted in Figure 6). Consequently, we can delineate a hierarchical arrangement within
these systems.[11]
Fig. 33. Simple system model [11]

2.3. Systems Engineering
So, what exactly does "engineering of the system" entail? What do we signify when we mention "systems
engineering"? While numerous interpretations exist depending on the perspective and emphasis of those
who contemplate this field, there are some fundamental attributes that receive a general consensus. Before
crafting a concise definition, let's categorize these key characteristics:[12]
• Wholes
• Synthesis of the whole from complementary parts.
• Finding answers/solutions to ‘whole problems
• Analysis
• Design
• Complexity
• Discipline and science
• Integrity
• Planning
The catch-all definition of systems engineering is;
Systems engineering is the art and science of creating whole solutions to complex problems [12]
or
Systems engineering is synthesizing a complex system from less complex systems.
The central concept here revolves around the partitioning of tasks, establishing independent departments
that interact through a process of "decomposition and abstraction".
When you embark on designing a system to tackle complex challenges, as you implement what is
necessary and rectify errors, you gradually approach perfection, resembling it more closely. The pivotal
factor in this pursuit is your knowledge and competency in organization rather than the technology or
resources at your disposal.
In actuality, system design is an inherent aspect of our daily lives. Think about it – when we prepare a
meal, we follow a specific order of steps. Tasks such as cleaning, arranging the layout of our homes, and
Page 808
SEMIT 2023
managing work supplies can be seen as mini-systems in their own right. A casual glance around reveals
a multitude of designed or constructed systems, each with its unique attributes—some interactive, some
hierarchical, and others completely self-contained.[9]
Fig. 34. Cook algorithm

Intentionally or not, we use the systems engineering approach and algorithm structure in many areas of
our lives. For example, when we cook, we have to act in a certain order.
• Get the recipe
• Make ingredients list
• Provide them with correct amounts
• Use them in a certain order and way
• Prepare according to the cooking method (pre-heated oven, on fire, boiling etc.)
• Cook for a certain time
• Service
3. The Solution
3.1. SSADM
In the realm of system design, various methods and tools exist, and for this endeavor, we will be
employing the "Structured System Analysis and Design Method" (SSADM). While the complete
structure and sub-levels are not within the scope of this study, we will focus on revealing the four levels
of the proposed system for a state. Our primary objective is to articulate the critical aspects with utmost
clarity and comprehensiveness, thereby establishing a model that can be readily adapted to diverse
scenarios pertaining to product supply chains.
Now, let's delve into the first and highest level of our work:
Fig. 35. First Level of SSADM
Page 809
SEMIT 2023
There are 3 independent departments and 1 data store:

1. State Management (Government): At the pinnacle of our system, this department is overseen by
the Deputy President, who is answerable to the President. The President, given their manifold duties in
overseeing the entire state, delegates this crucial responsibility to the Deputy President. Within this board,
various entities are represented, including relevant government departments, ministries, governors, law
enforcement agencies, scientists, agricultural experts, the food sector, select small and medium-sized
enterprises (SMEs), and critical suppliers. This department is the central hub for managing and
responding to crises on a state level.
2. Ministries/Provincial Administration.
4. Directorate of Communication.
5. Data Storage. Central software and database warehouse. Every department *must* connect this
store and share information.
Inputs: Outputs:
(1, 3) General Policy (2, 4) Report
(5, 7) Public Information (6, 8) Public Information
(9, 11) Update (10, 12) Reports and Information
The second level of SSADM:
Fig. 36. The second level of the SSADM diagrams

Thin black lines indicate information flows, red lines indicate material flows. One of the most important
parts will be the supply of products and raw materials.
The aim of the local data store is to contribute to ensuring data security and create an alternative backup
center for local data. The data at the municipal scale will be collected and processed in this center, and
the gotten results and necessary details will be consolidated and sent to the central data store. In the event
of a possible interrupting between the central data store and the municipality, the local data store will be
able to continue to operate and be a resource without losing its function. Synchronization can be restored
by ensuring the connection, even at intervals.[7]
In the current landscape of our technological capabilities, it's feasible to track the activities of nearly every
citizen individually, often in real-time, allowing us to monitor their whereabouts and actions. While
ethical discussions surrounding this matter are pertinent, let's temporarily set them aside and explore how
we can harness these opportunities for the greater good.
Ideally, every citizen, regardless of their financial means, social standing, occupation, or title, should
have the ability to consume resources in line with their genuine needs, provided they do so willingly and
Page 810
SEMIT 2023
without coercion. However, in situations where compulsion is necessary, it becomes imperative to ensure
that resource allocation is carried out equitably and transparently under the oversight of a responsible
authority.[9]
At this level, material flow as well as information flow is shown. Especially critical materials can be
tracked at the provincial level. Both the governorship, municipality and non-governmental organizations
may have different suppliers. However, all of them should be in contact with the data storage. Here we
see a resource management center. Even if they come from different suppliers and sources, all materials
are collected here.
The third level:
Fig. 37. The third level of the diagrams
When we go down to the district level, especially food and vital materials need to be monitored in a more
controlled manner. There is a resource management center in the district as well as in the province.
More accurate and sensitive follow-up will be carried out here. Because district administrations are more
familiar with their demographic structure. They have information such as who lives in which
neighborhood, what is the average income level and which needs are at the forefront. It will be easier and
fairer for them to manage resource transfers by being in constant contact with neighborhood headmen.
Fig. 38. The last level

On this last level we see the hub points and the home gardens. In such harsh conditions, we generally
expect an environment where neighborliness and cooperation will be at the forefront. Although water
Page 811
SEMIT 2023
wells will be monitored at the district level, it would be good for each neighborhood to have its own water
source. These resources should be utilized effectively and all possible land should be used for agriculture.
Large gardens of detached houses and housing estates may be suitable for this purpose.
Planting activities should also be carried out in a disciplined manner rather than randomly. Governments
should decide, at least regionally, where to grow what and where not, and guide citizens. Both information
and appropriate equipment support should be provided.
3.2. The Database
We've created a relational database model (Figure 12) that serves as a repository for information related
to the supply chain of materials and goods.
We used MySQL to design the database and the chart prepared with MySQL Workbench application.
But it doesn't matter which database application is used. What we are trying to present here is a model.
Since the vast majority of databases work in a similar way, this model can be used for any other database.
In Figure 12, each box represents a database table, and the rows within these tables denote the stored
information, i.e., the columns. Dashed lines are employed to illustrate the connections or relationships
between these tables. In this particular design, the use of one-to-many relationships suffices. This implies
that there can be an unlimited number of records in the associated table linked to a single record in the
source table. For instance, a father may have multiple children, but each child is connected to just one
father. Similarly, an invoice comprises only one header record but can encompass an indefinite number
of details.
These relationships also play a vital role in maintaining data consistency. By establishing relationships
between tables, it becomes impossible to enter a detail record of an invoice without a corresponding
header record. Even when software controls fail, the database engine prevents such inconsistencies.
Likewise, record deletions that could lead to inconsistencies are precluded.[7]
Fig. 39. Relationship Diagram
Page 812
SEMIT 2023
The core of the database consists of three main objects: Person, Point, and Item. The other tables are
employed to delineate the relationships that exist between these core objects. "The Ration" is placed
centrally in the diagram, but its definition relies on the main objects. Without these core objects, we
cannot establish any entitlements or rights.
Ration, in this context, signifies a predetermined quantity of a commodity that is officially allotted to
each person during times of scarcity, such as in wartime or crisis situations. However, there's no need to
define rations for every region uniformly. Instead, it should be contingent upon the current state or
circumstances of the region. Some districts may not be obligated to store ration data depending on their
particular situation.[7]
Moreover, the system allows local farmers and other suppliers to register their products and services
within the system and deposit them at designated delivery hub points. This setup facilitates the
dissemination of information to individuals in need, enabling them to access essential resources directly,
without requiring intermediaries. Each delivery point is affiliated with a specific region or, in other words,
a neighborhood. Consequently, each region may boast multiple delivery points.
By employing an application, as elaborated in the subsequent section of this document, any individual
can conveniently access these resources, ensuring a more streamlined and efficient process.
3.3. Localization
Food is a perishable substance due to its nature. In particular, fresh fruits and vegetables should be
delivered to those in need as soon as possible after they are collected.
The distance that a person can reach on foot is important and should be taken into account mainly. Every
step of the supply chain for foods needs to be robust and, if possible, have alternatives. In addition, it
should be tried to ensure that the number of steps is small. Access to vital foods should be available as
quickly and over short distances as possible.
Urban agriculture is already producing about 15–20% of the world’s food supply, and this can play an
even more critical role in achieving food security during a crisis [13].
Studies of the urban model underscore the pivotal role of collaboration among various stakeholders,
including partners, citizens, corporations, and governmental bodies, in developing a sustainable business
model. This collaborative approach is instrumental in devising solutions that not only address local
district issues but also have the potential to impact and resolve challenges at the city-wide, national, and
even global levels [14]. But, there should be a good communication and well defined network between
the actors.
A home garden is an agricultural system that integrates various physical, social, and economic functions
on the land surrounding the family residence. Its primary purpose is to enhance the availability of fresh
food at the household level. From a logistical standpoint, home gardens offer convenient, day-to-day
access to fresh vegetables and fruits. This accessibility contributes to improved and well-rounded diets
by supplementing essential nutrients, including proteins, vitamins, and minerals. [15], [16]
Here is our general model map:[7]
Page 813
SEMIT 2023
Fig. 40. The Model Map

The basic idea of model in Figure 13 is based on “resource sharing”. As mentioned before, localization
is the key concept of our model.
A normal person can walk 1km in 12 minutes and 5km in 1 hour. Even if we have made our model on a
neighborhood basis, it has the potential to be applied modularly even on a district, provincial, regional
and even state scale. In addition, details such as food values, daily needs of families, household and street
needs can also be kept. The fact that the needs are scalable will allow the management of resources that
need to be found, the possible bottlenecks to be identified in advance. It can also be used to determine
priorities when it is possible to ensure product diversity or when it is decided to direct limited
opportunities.
Fig. 41. Accessibility standards for key services
Page 814
SEMIT 2023
4. Conclusion
In this research endeavors, we sought to devise a model proposal for a resource management system
utilizing the esteemed framework of Systems Engineering, specifically the "Structured System Analysis
and Design Method" (SSADM), complemented by the principles of relational database design. Our scope
extended from the state level down to the granular levels of provinces, districts, and neighborhoods. For
each administrative tier, we meticulously crafted SSADM graphical representations. Concurrently, we
delineated the entities pertinent to resource acquisition and management, elucidating their
interrelationships. Subsequently, we engineered a database tailored to the software in harmony with this
conceptual blueprint.
Famine periods pose among the most exigent trials that societies can confront. Yet, a proficient strategy
for food distribution and resource management stands poised to mitigate these challenges. Systematic
analysis and design tools like SSADM furnish robust mechanisms to refine food allocation amid famine
crises, thereby fostering a sustainable paradigm for resource governance. The adoption of such an
approach holds the potential for markedly enhanced outcomes in the battle against famine.
As a result of our inquiry, we could not find a comparable application. The crux of our rationale hinges
on the systematic resolution of issues. The application model showcased herein serves to illustrate the
system's functionality, yet it is paramount to recognize that the human element remains indispensable.
Ultimately, the linchpin of this endeavors lies in the unwavering virtues of "goodwill" and "honesty" [7].
References
[1] A. HÜR, “Anadolu’nun kuraklık tarihi: ‘Kaht-ü galâ, kahtzedegân, kıllet-i bârân, istiskâ,’” 2023.
[Online]. Available: https://kisadalga.net/yazar/kaht-u-gala-kahtzedegan-killet-i-baran-istiska-
anadolunun-kuraklik-tarihi_53219. [Accessed: 06-Sep-2023].
[2] İ. E. Bilici, Kıtlık ve Yokluklarıyla 1940’lar. 2015.
[3] G. Abdulkadir, “OSMANLI DEVLETİ’NDE KURAKLIK VE KITLIK (Erzurum Vilayeti
Örneği: 1892-1893 ve 1906-1908 Yılları).,” J. Int. Soc. Res., vol. 2, no. 9, pp. 144–158, Nov.
2009.
[4] A. Pilikoğlu, “COVID-19 pandemi sürecinde ve endüstri 4.0 çerçevesinde tedarik zinciri
uygulamalarının değerlendirilmesi: Lojistik sektöründe bir uygulama,” İstanbul Ticaret
Üniversitesi, 2021.
[5] A. AYDIN and A. GÜNER, “COVID-19 SALGINININ TARIM SEKTÖRÜ VE GIDA
GÜVENLİĞİ ÜZERİNE ETKİSİ: TÜRKİYE ÜZERİNE BİR DEĞERLENDİRME,” Artuklu
Kaime Uluslararası İktisadi ve İdari Araştırmalar Derg., vol. 3, no. 2, pp. 155–171, Nov. 2020.
[6] K. Benis and P. Ferrão, “Potential mitigation of the environmental impacts of food systems
through urban and peri-urban agriculture (UPA) – a life cycle assessment approach,” J. Clean.
Prod., vol. 140, pp. 784–795, Jan. 2017.
[7] M. ATA, “Masters Thesis,” Ankara Yıldırım Beyazıt University, 2022.
[8] C. M. Ashworth, “Structured systems analysis and design method (SSADM),” Inf. Softw.
Technol., vol. 30, no. 3, pp. 153–163, Apr. 1988.
[9] M. GÜNDOĞAN and M. ATA, “Resource Management in Crisis Situations,” Avrupa Bilim ve
Teknol. Derg., no. 27, pp. 935–941, Nov. 2021.
[10] D. K. Hitchins, “Youtube Video,” 2015. [Online]. Available:
https://youtu.be/BEvW9B64tsM?t=292. [Accessed: 20-Feb-2021].
[11] D. K. Hitchins, “System Models Website,” 2015. [Online]. Available:
http://systems.hitchins.net/systems/systems-models/index.html. [Accessed: 01-Mar-2021].
[12] D. K. Hitchins, Systems Engineering: A 21st Century Systems Methodology. Wiley, 2007.
[13] R. Lal, “Home gardening and urban agriculture for advancing food and nutritional security in
response to the COVID-19 pandemic.”
[14] M. R. Anandhika, F. R. Hassan, and Y. Nugraha, “Towards a collaborative framework for the
large-scale social collaboration: A case of Jakarta’s response to the Covid-19 pandemic,” in 7th
Page 815
SEMIT 2023
International Conference on ICT for Smart Society: AIoT for Smart Society, ICISS 2020 -
Proceeding, 2020, pp. 1–7.
[15] E. H. Gündeş, “Changing shopping habits in supermarket during Covid-19,” Sabancı University,
2021.
[16] O. Khan, “Beginning in January 2020, the Covid-19 pandemic inflicted massive disruption on
the world’s supply chains.,” Focus, pp. 50–51, 2020.
Page 816
SEMIT 2023
Determining the Location for Conquest in a Computer Game Using MCDM Methods
Zeynep Melike SEVER1 Babek Erdebilli2
1 Department of Industrial Engineering, TOBB ETU, 06520, Ankara, Turkey
2 Department of Industrial Engineering, Yıldırım Beyazıt University, 06010, Ankara, Turkey
Abstract
Within the scope of this project, research has been conducted on a computer game. The primary objective
of this study is to make a selection among the countries that will be conquered in the game. During the
process of making this selection, various Multiple Criteria Decision Making (MCDM) methods such as
SWARA, AHP, TOPSIS, ELECTRE, and VIKOR have been explained and applied. The game
developers meticulously examined historical details during the game's production phase in order to
determine relevant parameters that would accurately mirror reality. As a result, it can be confidently
asserted that the executed project maintains a high level of coherence. Moreover, owing to the inherent
structure of the game, which necessitates decision-making at each stage, it is exceptionally well-suited
for the application of MCDM methods. The result of this study reveal that Trebizond and Byzantium
emerge as the prominent alternatives for the conquest location. The alignment of the result with a decision
made in history adds a layer of realism to the issue, thereby contributing an original piece of work to the
existing literature. The decisions derived from the application of these methods within the game have
been meticulously compared and thoughtfully interpreted.
Keywords: MCDM, AHP, ELECTRE, TOPSIS, Location Problem.
1. Introduction
In decision-making processes like site selection, which inherently involve multiple criteria, Multi-Criteria
Decision Making (MCDM) methods are preferred. Problems of multi-criteria decision-making can be
examined under three headings: choice, classification, and ranking. These methods started to be used in
the 1960s. Nowadays, they have become widespread, diversified, and are applied in various decision-
making domains. The choice of method should be based on the type of decision to be made and the nature
of the problem.
Within the scope of this project, a realistic decision-making simulation has been designed using Multi-
Criteria Decision-Making methods alongside the computer game development process. The process of
selecting the country to conquer by players has been analyzed using prominent MCDM methods such as
SWARA, AHP, TOPSIS, ELECTRE, and VIKOR. These methods systematically address the complexity
of criteria and alternatives in the decision-making process.
The results of the project include an analysis where the decisions obtained using MCDM methods in the
game are compared and interpreted with real historical decisions.
This project represents a unique approach that combines the fields of computer games and decision
analysis, providing players with the opportunity to experience real historical scenarios. It is believed that
this study, which offers both enjoyable and educational experiences, will make a significant contribution
to the gaming industry and the field of decision analysis.
In this study, after the introduction section, the second section provides a general description and
application steps of MCDM methods. In the third section, a literature review is conducted and presented
in tabular form. As we move on to the fourth section, the flowchart of the problem is provided, an example
study is discussed, and the results of the methods are demonstrated. Finally, in the conclusion section, the
results are interpreted and compared.
2. Method Steps
2.1. SWARA Method
Page 817
SEMIT 2023
In this method, a step-by-step process is followed to evaluate the weights between criteria and
alternatives. It is used to determine the priority of criteria and alternatives, helping to assess their
importance and rank them accordingly.
2.1.1. The Steps of the SWARA Method
Step 1: All criteria are ranked based on experts' opinions, with the most important criterion considered
first. In this step, decision-makers identify the most important criterion according to their own perspective
and assign a value of "1.00" to this criterion. All criteria are given values between "0" and "1," in
increments of 0.05.
Step 2: In this step, decision-makers determine the relative importance values for each criterion other
than the criterion in the first place. For this purpose, the j-th criterion is compared with the j-1 criterion,
and the value of sj (the comparison ratio of the average) is obtained. The sj value helps determine how
important the j-th criterion is compared to the (j-1)-th criterion. Thus, all consecutive criteria are
compared, weight values are calculated, and their levels of importance are determined as percentages.
Step 3: The kj coefficient is calculated for all specified criteria using the 2.1.1.1 principle. The kj
coefficient for the first criterion is 1, and for subsequent criteria in the ranking, the kj coefficient is found
by adding the sj value (the comparison ratio of the average) calculated in the previous step to the kj
coefficient of the criterion that comes after j-th criterion in the ranking.
1, 𝑗 = 1
𝑘𝑗 = { (2.1.1.1)
𝑠𝑗 + 1, 𝑗 > 1
Step 4: The corrected weight values for all specified criteria are calculated using the qj equation.
1, 𝑗 = 1
𝑞𝑗 = {𝑞𝑗−1 (2.1.1.2)
, 𝑗>1
𝑘
Step 5: In the final stage, the relative weights of all criteria are calculated as shown in the equation below.
wj represents the relative weight of criterion j, and it is calculated by dividing the corrected weight value
of criterion j by the sum of the corrected weight values of all criteria.
𝑞𝑗
𝑤𝑗 = { 𝑛 (2.1.1.3)
∑𝑘=1 𝑞𝑘
2.2. AHP Method
This method aims to simplify the decision-making process by transforming a complex decision structure
into a hierarchical one. It is used to calculate the weights of criteria and the values of alternatives.
2.2.1. The Steps of the AHP Method
Step 1: Creating the Hierarchical Structure: It is a structure containing all the factors that assist in defining
the decision problem and influence the decision.
Step 2: Pairwise Comparison Matrix and Solution: After creating the hierarchical structure, decision-
makers evaluate the criteria at each level relative to each other. Pairwise comparisons start from the top
of the hierarchy, and comparisons are made between criteria at each level.
Step 3: Normalization and Relative Importance Weights: After comparisons are made for criteria at each
level, square matrices are normalized. The sum of each column is taken, and the row values are divided
by the column sums separately. In the normalized matrix, the relative importance weights are obtained
by averaging each row.
Step 4: Calculating the Consistency Ratio: The consistency ratio is calculated to demonstrate the accuracy
of the comparisons made. The result of the calculation showing a ratio less than 0.10 indicates the
accuracy of the comparisons made.
Step 5: Determining the Final Ranking: The relative weights of criteria at each level in the hierarchical
structure are used to find the priority values for the alternatives. (Eren vd., 2019).
Page 818
SEMIT 2023
2.3. TOPSIS Method

In this method, the similarity of alternatives to the ideal and anti-ideal solutions is measured to select
the most suitable alternative.
2.3.1. The Steps of the TOPSIS Method

Step 1: Formation of the Decision Matrix: The rows represent the decision points or options available to
the decision-makers, and the columns represent the criteria that influence this decision.
Step 2: Creation of the Weighted Standard Decision Matrix: After normalizing the decision matrix, this
step involves multiplying each criterion's weights, which indicate the proportion of influence on the
decision, with the decision matrix.
Step 3: Establishment of Ideal (A+) and Negative Ideal (A-) Solutions: To establish the ideal and negative
ideal solution sets, the highest and lowest values of the column entries in the weighted decision matrix
are considered.
Step 4: Calculation of Distinction Measures: Using the Euclidean Distance Approach, the distances of
each alternative to the positive ideal solution and negative ideal solution are calculated.
Step 5: Computation of Relative Closeness to the Ideal Solution: Calculating the relative closeness of
each decision point to the ideal solution provides insight into the absolute proximity of the decision point
to the ideal solution (Gelashvili, 2019).
2.4. ELECTRE Method
In this method, the extent to which alternatives meet specific criteria is evaluated. Alternatives are
grouped into a ranking or an acceptance/rejection list based on this evaluation.
2.4.1. The Steps of the ELECTRE Method
Stage 1: (A) Calculation of the Decision Matrix: The rows of the decision matrix represent the criteria of
the decision problem, and the columns represent the alternatives. The matrix A is the initial dataset
objectively created by decision-makers. The decision matrix A is expressed as follows:
𝑎11 ⋯ 𝑎1𝑛
𝑎21 𝑎22 𝑎2𝑛
𝐴𝑖𝑗 = [ ⋮ ⋮ ⋮ ]
𝑎𝑚1 ⋯ 𝑎𝑚𝑛
In the Aij matrix, m represents the number of alternatives, and n represents the number of criteria.
Stage 2: (X) Calculation of the Standard Decision Matrix: Each element of the X Standard Decision
Matrix can be found based on the elements of the A Decision Matrix using the formula given below
(Equation 2.4.1.1).
𝑎𝑖𝑗
x𝑖𝑗 = (2.4.1.1)
√∑𝑚 2
𝑘=1 𝑎 𝑘𝑗
For example, to find the x_11 element of the X standard decision matrix, the a11 element of the A decision
matrix is used in the numerator, and in the denominator, the square root of the sum of the squares of the
1st column elements of the A matrix is written. This division yields the result. The objective of this
process is to evaluate the relationship between the first alternative and the first criterion and determine
its weight with respect to other alternatives. As a result of these evaluations, the X standard decision
matrix can be expressed as follows.
𝑥11 𝑥12 ⋯ 𝑥1𝑛
𝑥21 𝑥22 … 𝑥2𝑛
𝑋𝑖𝑗 = ⋮ ⋮ ⋮ ⋮
⋮ ⋮ ⋮ ⋮
[𝑥𝑚1 𝑥𝑚2 ⋯ 𝑥𝑚𝑛 ]
Page 819
SEMIT 2023
Stage 3: (Y) Calculation of the Weighted Standard Decision Matrix: Each criterion in the decision
problem may have different importance for the decision maker. To incorporate these differences into the
solution process of the Electre method, the Y Weighted Standard Decision Matrix is created. Decision
makers first need to determine the weights (wi) of the criteria. These weights must sum to a value of 1.0
according to the formula provided below (2.4.1.2).
∑𝑛𝑖=1 𝑤𝑖 = 1 (2.4.1.2)
Next, each column element of the X Standard Decision Matrix is multiplied by its corresponding w_i
value, resulting in the Weighted Standard Decision Matrix (Y) matrix. The Weighted Standard Decision
Matrix (Y) is expressed below:
𝑤1 𝑥11 𝑤2 𝑥12 ⋯ 𝑤𝑛 𝑥1𝑛
𝑤1 𝑥21 𝑤2 𝑥22 … 𝑤𝑛 𝑥2𝑛
𝑌𝑖𝑗 = ⋮ ⋮ ⋮ ⋮
⋮ ⋮ ⋮ ⋮
[𝑤1 𝑥𝑚1 𝑤2 𝑥𝑚2 ⋯ 𝑤𝑛 𝑥𝑚𝑛 ]
Stage 4: Calculation of (Ckl) Concordance and (Dkl) Discordance Sets: To create the concordance and
discordance sets, the elements of the Weighted Standard Decision Matrix (Y) are utilized. Alternatives
are compared with each other considering the criteria. The concordance and discordance sets are
determined using the formula expressed below (2.4.1.3)
𝐶𝑘𝑙 = j, 𝑦𝑘𝑗  𝑦𝑙𝑗  (2.4.1.3)
For example, in the calculation of the C12 concordance set for k = 1 and l = 2, the 1st and 2nd row
elements of the Weighted Standard Decision Matrix (Y) need to be compared with each other. If there
are 3 criteria in the problem, the C12 concordance set will have a maximum of 3 elements. As a
characteristic of the Electre method, for each concordance set (Ckl), a discordance set (Dkl) is provided.
Hence, the numbers of concordance and discordance sets are equal.
In the Electre method, when determining the concordance sets, attention should be paid to the structures
of the criteria. If the relevant criterion is benefit-oriented, the formula (2.4.1.3) is used. However, if the
criterion is cost-oriented, the inequality
ykj < ylj should be utilized.
Stage 5: Calculation of (C) Concordance and (D) Discordance Matrices: In the calculation of the (C)
Concordance matrix, the concordance sets determined in the previous stage are utilized. The size of the
C matrix is (mxm). No value is written for the element where k=l. The elements of the C Concordance
Matrix can be calculated using the following formula (2.4.1.4).
𝐶𝑘𝑙 = ∑𝑗€𝐶𝑘𝑙 𝑤𝑗 (2.4.1.4)
For example, if 𝐶12 = 1,4, the value corresponding to the 𝑐12 element of the C Concordance matrix
can be calculated as 𝑐12 = 𝑤1 + 𝑤4 the sum of weights for w1 and w4. The C matrix created in this way
is expressed below.
− 𝑐12 ⋯ 𝑐1𝑚
𝑐21 − … 𝑐2𝑚
𝐶= ⋮ ⋮ ⋮ ⋮
⋮ ⋮ ⋮ ⋮
[𝑐𝑚1 𝑐𝑚2 ⋯ − ]
The elements of the (D) Discordance matrix (𝑑𝑘𝑙 ) can be calculated using the formula (2.4.1.5) provided
below;
𝑚𝑎𝑥|𝑦𝑘𝑗𝑗€𝐷 −𝑦𝑙𝑗 |
𝑘𝑙
𝑑𝑘𝑙 = (2.4.1.5)
𝑚𝑎𝑥|𝑦𝑘𝑗𝑗 −𝑦𝑙𝑗 |
Page 820
SEMIT 2023
Similarly, the dimension of the (D) Discordance matrix is also (m x m), and no value is written for k=l.
Thus, the (D) matrix is represented as follows:
− 𝑑12 ⋯ 𝑑1𝑚
𝑑21 − … 𝑑2𝑚
𝐷= ⋮ ⋮ ⋮ ⋮
⋮ ⋮ ⋮ ⋮
[𝑑𝑚1 𝑑𝑚2 ⋯ − ]
Stage 6: Calculation of concordance dominance (F) and Discordance Dominance (G) Matrices: The
dimension of the Concordance Dominance (F) Matrix is (m x m), and the elements of this matrix are
found by separately comparing the Concordance Matrix elements (𝑐𝑘𝑙 ) with the Concordance Threshold
(𝑐), which is the threshold of concordance. The Concordance Thresholdd (𝑐) is calculated using the
formula expressed below (2.4.1.6) :
1
(𝑐)=𝑚(𝑚−1) ∑𝑚 𝑚
𝑘=1 ∑𝑙=1 𝑐𝑘𝑙 (2.4.1.6)
The formula mentioned represents the alternative count denoted by m. The value of (𝑐) is the obtained by
1
multiplying the result of summing all the elements in the C Compatibility Matrix with 𝑚(𝑚−1) The
elements of the F Compatibility Superiority Matrix, denoted as 𝑓𝑘𝑙 , consist of values 1 or 0. No values
are assigned to the elements on the main diagonal of this matrix.
The size of the (G) Incompatibility Superiority Matrix is also (mxm). Before creating the G matrix, the
incompatibility threshold value needs to be calculated. This threshold value ( 𝑑 ) can be calculated using
the formula shown below (2.4.1.7).
1
(𝑑)=𝑚(𝑚−1) ∑𝑚 𝑚
𝑘=1 ∑𝑙=1 𝑑𝑘𝑙 (2.4.1.7)
1
So, the value of (𝑑) is equal to the product of and the sum of all elements in the D Incompatibility
𝑚(𝑚−1)
Matrix. The elements of the G Incompatibility Superiority Matrix ( 𝑔 𝑘𝑙 ) consist of values of 1 or 0. No
values are written for the elements on the main diagonal of the matrix.
Stage 7: Calculation of Total Dominance (E) Matrix: The elements ( 𝑒𝑘𝑙 ) of the Total Dominance (E)
Matrix are formed by the pairwise multiplication of elements from the Concordance Dominance (F) and
Discordance Dominance (G) Matrices. Since the elements of F and G Matrices are either 1 or 0, the
elements of the E Matrix will also be 1 or 0.
Stage 8: Calculation of Priority Order of Alternatives: The rows and columns of the E Total Dominance
Matrix represent the priorities of the alternatives. For example, if 𝑒21 = 1 and 𝑒31 = 1 this indicates the
superiority of the 2nd alternative over the 1st alternative and the 3rd alternative over the 1st alternative
(Arslan, 2018).
2.5. VIKOR Method
This method helps in selecting the most suitable alternative by considering the priorities and sensitivities
of the alternatives.
2.5.1. The Steps of the VIKOR Method
Step 1: The best ( 𝑓𝑖 ∗ ) and worst ( 𝑓𝑖 −
) values are determined for each evaluation criterion. For i =
1,2,…,n; the formula (2.5.1.1) is used.
𝑓𝑖 ∗ = 𝑚𝑎𝑥𝑗 𝑓𝑖𝑗 , 𝑓𝑖 − = 𝑚𝑖𝑛𝑗 𝑓𝑖𝑗 , (2.5.1.1)
Step 2: 𝑆𝑗 and 𝑅𝑗 values are calculated for each evaluation unit. Here, 𝑤𝑖 represents the criterion weights.
𝑆𝑗 = ∑𝑛𝑖=1 𝑤𝑖 (𝑓𝑖∗ − 𝑓𝑖𝑗 )/(𝑓𝑖∗ − 𝑓𝑖− ) (2.5.1.2)
𝑅𝑗 = max [𝑤𝑖 𝑓𝑖∗ − 𝑓𝑖𝑗 )/(𝑓𝑖∗ − 𝑓𝑖− )] (2.5.1.3)
Step 3: For each evaluation unit, 𝑄𝑗 values are calculated.
Page 821
SEMIT 2023
𝑄𝑗 = 𝑣(𝑆𝑗 − 𝑆 ∗) / (𝑆 − − 𝑆 ∗ ) + (1 − 𝑣)((𝑅𝑗 − 𝑅 ∗ )/( 𝑅 − − 𝑅 ∗ )) (3.6.1.4)

𝑆 ∗ = 𝑚𝑖𝑛 𝑗 𝑆𝑗 ; 𝑆 − =𝑚𝑎𝑥 𝑗 𝑆𝑗 ; 𝑅 ∗=𝑚𝑖𝑛 𝑗 𝑅𝑗 ; ; 𝑆 − = 𝑚𝑎𝑥 𝑗 𝑅𝑗 The 𝑣 value is used to represent the maximum
strategy weight in group benefit, while (1 − 𝑣) represents the weight of the minimum regret of those with
the opposing view (Opricovic ve Tzeng, 2007). In the literature, the 𝑣 value is commonly used as 𝑣 = 0.5
(Lixin vd., 2008); (Wu vd., 2009).
Step 4: Sorting 𝑄𝑗, 𝑆𝑗, 𝑅 values is performed. Having the smallest 𝑄𝑗 value is stated to be the best option
within the alternative group. There are two conditions to ensure the validity of the results. When these
conditions are met, the alternative with the minimum 𝑄 value can be considered as the most suitable (Eren
ve Yanık, 2017).
Condition 1 (C1) - Acceptable Advantage: It indicates a clear difference between the best and the closest
to the best option.
𝑄(𝑃2 ) − 𝑄(𝑃1 ) ≥ 𝐷(𝑄) (2.5.1.5)
In this inequality, 𝑃1 represents the first best alternative with the lowest 𝑄 value, and 𝑃2 represents the
second best alternative. 𝐷(𝑄) is expressed as 1/(𝑗 − 1), where 𝑗 represents the number of evaluation units.
When the number of evaluation units is less than 4, 𝐷(𝑄) is taken as 0.25 (Chen ve Wang, 2009).
Condition 2 (C2) - Acceptable stability: The first best alternative 𝑃1, with the lowest 𝑄 value, should
achieve the best score in one of the 𝑆 and 𝑅 values. If one of the specified two conditions is not met, a
compromise solution is reached: If the second condition is not met, alternatives 𝑃1 and 𝑃2 are considered.
If the first condition is not met, alternatives 𝑃1, 𝑃2, ……, 𝑃𝑀 are considered, expressed by the inequality
𝑄(𝑃𝑀) - 𝑄(𝑃1 ) ≥ 𝐷(𝑄) (Opricovic ve Tzeng, 2004).
3. Literature
Multi-criteria decision-making methods, initiated in the 1960s, were developed in response to the
perceived need for tools to aid decision-making processes (Saaty, 1980). There are numerous scientific
methods utilized for optimal site selection. Among these methods are mathematical modeling, simulation
techniques, financial methods, and Multi-Criteria Decision Making (MCDM) methods (Eleren, 2006).
Commonly used MCDM methods in location selection problems include AHP, ANP, TOPSIS,
PROMETHEE, ELECTRE, and VIKOR.
In the literature, it is observed that MCDM methods were not used on historical academic studies.
Therefore, since there were no location selection problems related to the subject, such problems were
investigated in different subjects. Studies on location selection problems and MCDM methods were
examined in the literature, and relevant studies are included in this section.
Duckstein, Gershon (1983) presented the theoretical basis of the ELECTRE method in their study and
demonstrated how alternatives for forest cover management can be compared with a multi-criteria
perspective using this method. The study addressed the transformation of forested watersheds into grass
areas with certain percentages over time and permanently, as management alternatives, evaluated based
on different criteria such as tree and forage production, agricultural water supply, flood, hydroelectric
power generation, and reservoir-based recreation.
Manoochehr Toshtzar (1988) proposed the Analytic Hierarchy Process (AHP) model as a suitable multi-
criteria decision-making framework for evaluating software considering the rapid growth of the computer
industry and the increasing number of publishers and products, making evaluation and selection processes
complex. The model was claimed to be flexible and comprehensive, allowing the evaluation of
quantitative and qualitative factors with different preference systems.
Godwin et al. (2004) reported in their study the application of multiple-criteria decision-making methods
PROMETHEE and GAIA to indoor and outdoor air quality data for the first time. Through these methods,
they evaluated and differentiated the ranking of homes and the parameters influencing the ranking,
especially focusing on quantitative indicators.
Page 822
SEMIT 2023
Abo-Sinna and Abou-El-Enien (2006) extended the Technique for Order of Preference by Similarity to
Ideal Solution (TOPSIS) method to solve large-scale multi-objective programming problems involving
fuzzy parameters. This approach provides an interactive fuzzy decision-making algorithm that produces
an alpha-Pareto optimal solution, where decision-makers are asked to specify the alpha degree and the
relative importance of the objectives.
Dagdeviren and Eraslan (2008) proposed a model based on the Analytic Network Process for evaluating
strategic energy policies. The model is used to prioritize Turkey's strategic energy policies. With this
model, different people are involved in the decision-making process, increasing the efficiency of the
process.
Kahraman et al. (2010) presented various approaches and techniques used in location selection problems.
In this context, they focused on fuzzy multi-criteria decision-making methods, particularly presenting
fuzzy AHP and fuzzy TOPSIS as the most commonly used ones in the literature. Additionally, a
framework based on fuzzy information axiom for facility location selection was proposed for the first
time.
Bottero and Ferretti (2010) emphasized the contribution of the ANP technique to the assessment of
sustainability of unwanted facilities and particularly highlighted the use of quantitative indicators in the
evaluation process. The main findings of the study demonstrated that the use of quantitative indicators as
nodes in the ANP-BOCR structure significantly increased the internal consistency of the model and made
the decision-making process more traceable and reliable.
Alimoradi et al. (2011) attempted to design a closed-loop supply chain by considering not only traditional
supply chain levels but also recycling options, including collection centers with reverse flows and
recycling facilities. They tried to determine the best location for a recycling facility by using the fuzzy
TOPSIS method due to uncertainties in decision parameters for a capacity-free facility layout problem.
Jahan et al. (2011) emphasized the compromise solution and proposed a new version of the VIKOR
method that covers all types of criteria. This comprehensive version improves the main drawback of
traditional VIKOR with a simpler approach. The proposed method may enhance the accuracy of material
selection results, especially in biomedical applications where materials should have similar properties to
human tissues. To explain and justify the proposed method, five examples were provided.
Diaby et al. (2011) established a solid decision analysis field for the selection of rubber tree clones. Today,
there is no known method for selecting the best clone to be planted for the highest return on investment.
The only known selection method is to benefit from the experience of different individuals moving in the
plantation. Therefore, there is a need for a tool that considers the critical criteria for achieving this goal.
This goal is achieved by using multi-criteria analysis methods for clone selection and ranking, especially
using the ELECTRE method.
Sadeghi et al. (2012) aimed to develop a model to assess the factors affecting the success of high-tech
SMEs in Iran. Group Analytic Hierarchy Process (AHP) method was applied, and Chang's extended
analysis of fuzzy data was used for the analysis of fuzzy data. Finally, a fuzzy TOPSIS model was used
in a fuzzy TOPSIS model to evaluate and rank the performance of high-tech SMEs in Iran to determine
and rank critical success factors (CSFs).
Avikal et al. (2013) proposed a heuristic method based on a multi-criteria decision-making technique to
balance and design disassembly lines efficiently where the disassembly of waste products containing soft
resources, intangible resources, i.e., intellectual resources, is performed, contributing to the presentation
of port performance that could contribute to the presentation of port service quality. The proposed
approach used the fuzzy TOPSIS-MCGP method for disassembly lines' efficient design and balancing,
which is one of the multi-criteria decision-making methods (MCDMs).
Alimardi et al. (2014) considered the concept of a flexible supply chain. In this article; A new hybrid
MADM method for the selection of a flexible supplier was proposed based on four criteria, performance,
cost, flexibility, and technology. Two MADM methods, Step-wise Weight Assessment Ratio Analysis
Page 823
SEMIT 2023
(SWARA) and Vise Kriterijumska Optimizacija I Kompromisno Resenje (VIKOR), were used in the
decision-making process.
Pak et al. (2015) emphasized that intangible resources, i.e., intellectual resources, which constitute soft
resources, are important for any company to gain sustainable competitive advantage, particularly for
ports, from a point of view of airlines. In this context, they aimed to identify intra-port intellectual
resources that could contribute to the provision of port service quality and proposed a fuzzy TOPSIS
approach focusing on intellectual resources for port selection problem.
Michal et al. (2016) compared the Weighted Sum Model (WSM) and the Analytic Hierarchy Process
(AHP) methods, popular multi-criteria decision-making methods. In their simulation study, they showed
that both methods can be considered efficient.
Emma et al. (2016) aimed to establish a new assessment method that aims to establish closer relationships
with sustainability, considering suitability, economic, social, and environmental criteria that are
traditionally evaluated in economic terms. In the article, they presented an experimental application and
comparison of five different multi-criteria decision-making (MCDM) approaches for assessing
sustainable housing suitability.
Srisawat and Payakpate (2016) explained a study focusing on the use of multi-criteria decision-making
(MCDM) techniques in the agricultural sector, especially in rubber plantations. They compared the
Analytic Hierarchy Process (AHP), TOPSIS, and SAW multi-criteria decision-making methods in
providing effective criteria for selecting crop plants among rubber trees in plantations.
Torf et al. (2016) presented a three-tier location and routing problem consisting of LRP, facilities, depots,
and customers. The problem solution consists of two stages, and Fuzzy MCDM methods are used to solve
the facility location problem to find the best location for a recycling facility due to uncertainties in
decision parameters for a capacity-free facility layout problem.
Franz et al. (2017) performed a real-world case study for an integrated rehabilitation management system,
analyzing the suitability and applicability of five different multi-criteria decision-making (MCDM)
methods for use in integrated asset management of water systems. To compare, they selected ELECTRE,
AHP, WSM, TOPSIS, and PROMETHEE methods.
Widianta et al. (2018) addressed the importance of making correct decisions in placing employees in
suitable positions within a company. For such a decision-making process, they compared AHP, TOPSIS,
SAW, and PROMETHEE, which are MCDM methods.
Dahooie et al. (2018) presented a competence framework consisting of five criteria to select the best IT
expert among five candidates. They used the Stepwise Weight Assessment Ratio Analysis (SWARA)
method to determine the importance of each criterion and calculate their weights.
Ic and Yurdakul (2020) examined whether reducing the number of criteria and processing centers made
the ranking results more sensitive to changes in criterion weights by using the Spearman rank correlation
test. The study found that reducing the number of criteria and processing centers made the ranking results
more sensitive to changes in criterion weights.
Korkusuz et al. (2020) developed a new occupational health and safety (OHS) performance measurement
method for the health sector by integrating MCDM methods into a structured framework. Initially, they
determined the OHS performance indicators and weighted them using the AHP method. To obtain
performance indexes, the PROMETHEE (Preference Ranking Organization Method for Enrichment
Evaluations) and GRA (Grey Relational Analysis) methods were used.
Quynh et al. (2020) discussed the challenges of distribution center selection, emphasizing that selecting
the best location among alternatives is a significant task for any company to reduce costs, improve
transportation flow efficiency, and ensure customer satisfaction. To select the best location for a
distribution center, they used the TOPSIS method aiming at the selection of the best location among
alternatives.
Page 824
SEMIT 2023
Fu et al. (2020) focused on the selection of hotels from an airline's perspective for the hotel-at-airport
(HAA) problem and how it is challenging in road infrastructure construction to choose the most suitable
mechanization. The study proposed the first study adopting the fuzzy TOPSIS-MCGP approach, which
is a fuzzy multi-criteria decision-making (MCDM) approach, in the HAA problem for addressing
uncertainty and to deal with various evaluation criteria.
Goswami et al. (2021) attempted to create two new hybrid MCDM systems by integrating the Additive
Ratio Assessment (ARAS) method with the Technique for Order of Preference by Similarity to Ideal
Solution (TOPSIS) and Complex Proportional Assessment (COPRAS) methods. They aimed to enhance
the efficiency of MCDM systems by combining these methods.
Egilmez et al. (2022) conducted a study using the circular economy model to analyze the challenges of
the fashion sector as a sustainability model where environmental impacts are reduced, and products are
effectively recycled. They analyzed the challenges of circular economy in the fashion sector using the
SWARA and BWM methods.
Jovanovic et al. (2023) focused on the selection of road construction machinery, one of the most important
challenges encountered in the selection of the most suitable mechanization for road infrastructure
construction, by using ARAS and SWARA methods to develop an intelligent Fuzzy Multi-Criteria
Decision Making model.
Peker and Gorener (2023) conducted a study to determine a new facility location using the
composite propeller manufacturing scenario and used the Fuzzy SWARA-COPRAS method.
Hinduja and Pandey (2023) compared the results produced by the classic and contemporary fuzzy Multi-
Criteria Decision Making (FMCDM) approach under increasing uncertainty using the random
uncontrolled Monte Carlo simulation approach and analyzed the results from the perspective of a reverse
ranking approach. Additionally, they examined the similarity between the rankings generated by each
pair of methods for the same decision problems."
Page 825
SEMIT 2023
No. Author(s) Title Year Journal Country Methods Used
Multicriterion analysis of a
Lucien Duckstein Applied Mathematical
1 vegetation management 1983 USA ELECTRE 2
Mark Gershon Modelling
problem using ELECTRE II
Multi-criteria decision
making approach to
computer software Mathematical and
2 Manoochehr Toshtzar 1988 Britain AHP
evaluation: Application of Computer Modelling
the analytical hierarchy
process
Application of Multicriteria
Godwın A.
DecisionMaking Methods
Ayoko
to Air Quality inthe Environmental Science GAIA
3 Lıdıa Morawska 2004 Austraila
Microenvironments of & Technology PROMETHEE
Serge Kokot
Residential Houses in
Anddale Gılbert
Brisbane,Australia
An interactive algorithm for
Mahmoud A. Abo large scale multiple
Sinna a objective programming Applied Mathematics
4 2006 Egypt TOPSIS
Tarek H.M. Abou-El- problems with fuzzy and Computation
Enien parameters through TOPSIS
approach
Priority determination in
Page
strategic energy policies in INTERNATIONAL

Metin Dagdeviren
826
5 Turkey using analytic 2008 JOURNAL OF Türkiye ANP

Ergun Eraslan
network process (ANP) ENERGY RESEARCH
with group decision making
SEMIT 2023
Cengiz Kahraman ProductıonEngıneerıng

Fuzzy Location Selection Fuzzy AHP
6 Selçuk Cebi 2010 And Management Türkiye
Techniques Fuzzy TOPSIS
Fatih Tuysuz Under Fuzzıness
An Analytic Network
Process-based Approach for JOURNAL OF
Marta Bottero Location Problems: The MULTI-CRITERIA
7 2010 Italy ANP
Valentina Ferretti Case of a New Waste DECISION
Incinerator Plant in the ANALYSIS
Province of Torino
Ali Alimoradi A hybrid model for INTERNATIONAL
Rosnah Mohd remanufacturing facility JOURNAL OF
8 2011 Malaysia Fuzzy TOPSIS
Yussuf location problem in a SUSTAINABLE
Norzima Zulkifli closed-loop supply chain ENGINEERING
Ali JAHAN
Faizal MUSTAPHA
A comprehensive VIKOR
Yusof ISMAİL MATERIALS &
9 method for material 2011 Malaysia VIKOR
S.M. SAPUAN DESIGN
selection
Marjan
BAHRAMİNASAB
Moussa DIABY
A Comprehensive Decision JOURNAL OF
Helene FERRER
Approach for Rubber Tree MULTI-CRITERIA
10 Fabrice VALOGNES 2011 France ELECTRE
Planting Management in DECISION
Andre CLEMENT-
Africa ANALYSIS
DEMANGE
Developing a fuzzy group WORLD
Arash SADEGHI AHP model for prioritizing CONFERENCE ON
AHP
11 Adel AZAR the factors affecting success 2012 BUSINESS, Iran
Fuzzy TOPSIS
Page
Ramin Sepehri RAD of High-Tech SME's in Iran: ECONOMICS AND

A case study MANAGEMENT
827
SEMIT 2023
Shwetank AVIKAL A PROMETHEE Method

Industrıal Engıneerıng
P.K MISHRA Based Heuristic for
12 2013 And Management India PROMETHEE
Rajeev JAIN Disassembly Line
Systems
H.C. YADAV Balancing Problem
Maryam
ALIMARDANI A NOVEL HYBRID
Zolfani SWARA AND VIKOR TECHNOLOGICAL
HASHEMKHANI METHODOLOGY FOR AND ECONOMIC SWARA
13 2013 Iran
Mohammad Hasan SUPPLIER SELECTION DEVELOPMENT OF VIKOR
AGHDAIE IN AN AGILE ECONOMY
Jolanta ENVIRONMENT
TAMOSAITIENE
Fuzzy MCDM Approach
Ji Yeong PAK ASIAN JOURNAL OF
for Evaluating Intangible South
14 Vinh THAI 2015 SHIPPING AND Fuzzy TOPSIS
Resources Affecting Port Korea
Gi Tae YEO LOGISTICS
Service Quality
COMPARISON OF
Michal ADAMCZAK MULTIPLE-CRITERIA
AHP
15 Roman DOMANSKI DECISION-MAKING 2016 LOGFORUM Poland
WSM
Norbert WAGENER METHODS-RESULTS OF
A SIMULATION STUDY
OMEGA- AHP
Comparative analysis of
Emma MULLINER INTERNATIONAL COPRAS
MCDM methods for the
16 Naglis MALYS 2016 JOURNAL OF England TOPSIS
assessment of sustainable
Vida MALIENE MANAGEMENT WPM
housing affordability
Page
SCIENCE WSM
Comparıson Of Mcdm
828
Chutiphon Journal Of Informatıon AHP

Methods For Intercrop
17 SRISAWAT 2016 And Communıcatıon Thailand SAW
Selectıon In Rubber
Janjira PAYAKPATE Technology-Malaysıa TOPSIS
Plantatıons
SEMIT 2023
Fatemeh TORFI Fuzzy MCDM for weight of APPLIED

18 Reza FARAHANI object's phrase in location 2016 MATHEMATICAL Iran Fuzzy MCDM
Iraj MAHDAVI routing problem MODELLING
Franz TSCHEIKNER-
AHP
GRATL Comparison of Multi-
ELECTRE
Patrick EGGER Criteria Decision Support
19 2017 WATER Austria PROMETHEE
Wolfgang RAUCH Methods for Integrated
TOPSIS
Manfred Rehabilitation Prioritization
WSM
KLEIDORFER
M.M.D. WIDIANTA Comparison of Multi- 2nd Internatıonal Joınt AHP
T. RIZALDI Criteria Decision Support Conference On Scıence PROMETHEE
20 2018 Indonesia
DPS SETYOHADI Methods for Employee And Technology (Ijcst) SAW
H.Y. RISKIAWAN Placement 2017 TOPSIS
Jalil DAHOOIE HUMAN FACTORS
Elham ABADI Competency-based IT AND ERGONOMICS
Hamid Reza personnel selection using a IN ARAS
21 2018 Iran
FIROOZFAR hybrid SWARA and ARAS- MANUFACTURING SWARA
Amir VANAKI G methodology & SERVICE
INDUSTRIES
Analysis of the effect of the
number of criteria and
alternatives on the ranking Journal of the Faculty
Tansel Yusuf IC results in applications of the of Engineering and
22 2020 Türkiye TOPSIS
Mustafa YURDAKUL multi criteria decision Architecture of Gazi
making approaches in University
machining center selection
problems
Page
829
SEMIT 2023
Occupational health and

Ammar Yasir
safety performance Journal of the Faculty
KORKUSUZ AHP
measurement in healthcare of Engineering and
23 Umut Hulusi INAN 2020 Türkiye GRA
sector using Architecture of Gazi
Yavuz OZDEMİR PROMETHEE
integrated multi criteria University
Huseyin BAŞLIGIL
decision making methods
M.P. QUYNH
Distribution center location
Thuy Luong THU
selection using a novel multi
Quynh Doan HUONG DECISION SCIENCE
24 criteria decision-making 2020 Vietnam TOPSIS
Anh VAN LETTERS
approach under interval
Hien VAN
neutrosophic complex sets
Nyugen VAN
The selection model for
horizontal alliances between
Yan-kai FU
hotels and airlines: an TOPSIS-
25 Weilun HUANG 2020 TOURISM REVIEW Taiwan
integrated application of MCGP
Chin-Nung LIAO
NGT, fuzzy TOPSIS and
MCGP methods
Shubhra GOSWAMI
Dhiren BEHERA
Abdul Razak Analysis of a Robot
KALADGI Selection Problem Using
ARAS-TOPSIS
Sher KHAN Two Newly Developed India-
26 2021 SYMMETRY-BASEL ARAS-
Ram SUBBIAH Hybrid MCDM Models of Malaysia
COPRAS
Mohammad ASIF TOPSIS-ARAS and
Asif AFZAL COPRAS-ARAS
Rajendran
Page
PARVATHY
830
SEMIT 2023
Gözde KOCA JOURNAL OF

Analysıs Of Drıvers And
Özüm EGİLMEZ MEHMET AKIF
Challenges In Cırcular
Ezgi DEMİR ERSOY UNIVERSITY BWM
27 Economy Wıth Swara And 2022 Türkiye
Çağlar KARAMASA ECONOMICS AND SWARA
Bwm Methods In Clothıng
Halil İbrahim ADMINISTRATIVE
Sector
GOKCAN SCIENCES FACULTY
Stanislav
JOVANOVIC
Edmundas An Intelligent Fuzzy
ZAVADSKAS MCDM Model Based on D
Saudi Fuzzy ARAS
28 Zeljko STEVIC and Z Numbers for Paver 2023 AXIOMS
Arabia SWARA
Milan MARINKOVIC Selection: IMF D-SWARA-
Adel F. Fuzzy ARAS-Z Model
ALRASHEEDI
Ibrahim BADI
Facility location selection
with improved fuzzy Journal of Turkish Fuzzy
Bilge Nur PEKER
29 SWARA and fuzzy CODAS 2023 Operations Türkiye COPRAS
Ali GORENER
methods: An application in Management Fuzzy SWARA
the manufacturing industry
Analysis and Comparison of
State-of-the-Art Fuzzy VISION-THE
Akshay HINDUJA Multi-criteria Decision- JOURNAL OF AHP
30 2023 India
Manju PANDEY making Methods Under BUSINESS DEMATEL
Different Levels of PERSPECTIVE
Uncertainty
Page
831
SEMIT 2023
4. Sample Application
Problem Statement
Identification of Selecting one of the

Alternatives alternatives /
Comparing Method
Results
Determination of
Criteria
Performing
Sensitivity Analysis
Solution of the
Problem Using AHP
Weighting Criteria
Solution of the Using SWARA
Problem Using
TOPSIS
Solution of the Solution of the

Problem Using Problem Using
Page
ELECTRE VIKOR
832
SEMIT 2023
In the flowchart provided above, the initial stage of the computer game involves the player making
specific decisions. Based on the decisions they make at the beginning, the course and shape of the game
will change and take form. At the starting stage, based on the decisions made, the player can face various
problems based on cause and effect relationships.
To provide examples of these problems, if the player displays an aggressive attitude, they might face
reactions from other countries. On the other hand, if they adopt a defensive stance, they might receive
reactions from people within their own religion and fellow countrymen. Considering these possibilities,
it can be said that every decision made in the game is crucial. When examining the rules and mechanics
of the game, numerous criteria and alternatives are encountered.
In the project, while determining the player's next move, the goal was to use MCDM methods to select
the alternatives in the best possible way based on criteria.
4.1. Problem Definitio
In the game Europa Universalis IV, at the starting stage, the selected nation is the Ottoman Empire. The
decision has been made regarding which territories the first conquest with the Ottoman Empire will
be.During this decision-making stage, there are many criteria and alternatives to consider. The player
makes these decisions based on their own intuition and encounters positive or negative consequences. In
the project phase, important criteria for the game and considering the historical context of the game,
alternatives have been identified. These identified alternatives and criteria will be explained in detail in
the next section.
4.2. Alternatives
In the conducted study, the alternatives were determined as Byzantium, Karaman, Trebizond, Hungary,
and Dulkadir. The reasons for determining these alternatives are as follows:
The Ottoman Empire being a neighboring country with the alternatives.
The absence of any peace treaties between the chosen countries and the Ottoman Empire.
The geopolitical positions of the alternatives being crucial points for the Ottoman Empire.
The reason for choosing Byzantium is that Constantinople is the trade center of the region.
The reason for choosing Hungary is the Ottoman Empire's desire to expand into Europe.
The reason for choosing Trebizond is the Ottoman Empire's wish to have full authority in the Black Sea.
The reason for choosing Karaman and Dulkadir is the Ottoman Empire's desire to dominate Anatolia.
4.3. Criteria
In the mentioned computer game, the gameplay was analyzed, and the necessary criteria for the problem
at hand and its solution were identified.
Criterion 1: Military Strength
Description: Military strength, a crucial factor in selecting the country for conquest, represents the
number of soldiers the country will face in a possible war. The best value for this criterion is the minimum
value.
Criterion 2: Ally Military Support
Description: It signifies the military support that will come from the allies of the country to be conquered.
The highest value for this criterion is considered the worst value for us.
Criterion 3: Country Development Level
Description: Having a high level of development in the territories acquired in case of winning a war is
crucial in many aspects. In this criterion, the highest value is considered the best value.
Criterion 4: Income to Be Gained
Page 833
SEMIT 2023
Description: Countries incur material and spiritual expenses during war. Naturally, countries aim to cover
these expenses incurred after the war. Therefore, the income to be gained after the war is a factor that
significantly influences the country's decision to engage in war. Hence, in this criterion, the highest value
is considered the best value.
Criterion 5: Reaction Points from Other Countries
Description: This criterion is the most important one to consider. The reason for this is the importance
of the reaction that will come from other countries after conquering the selected country. The lowest
value in this criterion is considered the best value for us.
Criterion 6: Potential Number of Soldiers the Countries Can Produce
Description: This criterion represents the maximum military power potential of the selected countries.
Since reaching the level of potential high countries takes time, this criterion is not very important. The
highest value in this criterion is considered the worst value.
Criterion 7: Prestige to Be Gained
Description: It represents the prestige gain of the country to be selected. Having high prestige to be gained
is important for us, and the highest value represents the best value.
Criterion 8: Military Loss
Description: It is the number of soldiers that will be lost in case of war. The lowest value for this criterion
is considered the best value for us.
Criterion 9: Share in Countries' Trading Centers
Description: In the world, there are specific trading centers. Having a high share in these trading centers
for the selected country is crucial and effective.
Criterion 10: Distance to Countries
Description: It represents the distance to the countries to be conquered. Minimizing the distance as much
as possible is a desired condition.
4.4. Solution of the Problem
4.4.1. AHP Solution
The comparison of the criteria identified as the first step of the AHP has been made. Later, the matrix of
this comparison was created. The comparison matrix illustrates the importance levels of the criteria in
relation to each other in a specific logic. Using the 1-9 comparison scale suggested by Saaty, the matrix
was created, and the importance levels were determined by our decision-makers. Upon examining the
matrix, it is seen that the most important criterion is "Expected Reaction Score from Other Countries."
Table 1 shows the created matrix.
Table 5: Comparıson Matrıx Of Crıterıa
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
C1 1 1 2 3 0,5 7 4 8 5 6
C2 1 1 2 3 0,5 7 4 8 5 6
C3 0,5 0,5 1 2 0,33 6 3 7 4 5
C4 0,33 0,33 0,5 1 0,25 5 2 6 3 4
C5 2 2 3 4 1 8 5 9 6 7
C6 0,14 0,14 0,17 0,2 0,12 1 0,25 2 0,33 0,5
C7 0,25 0,25 0,33 0,5 0,2 4 1 5 2 3
C8 0,12 0,12 0,14 0,17 0,11 0,5 0,2 1 0,25 0,33
C9 0,2 0,2 0,25 0,33 0,17 3 0,5 4 1 2
C10 0,17 0,17 0,2 0,25 0,14 2 0,33 3 0,5 1
Page 834
SEMIT 2023
A normalized matrix has been created by dividing each cell in the comparison matrix of criteria by the
sum of its respective column. Table 2 shows the normalized values.
Table 6: Normalızed Matrıx
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
C1 0,18 0,18 0,21 0,21 0,15 0,16 0,20 0,15 0,19 0,17
C2 0,18 0,18 0,21 0,21 0,15 0,16 0,20 0,15 0,19 0,17
C3 0,09 0,09 0,10 0,14 0,10 0,14 0,15 0,13 0,15 0,14
C4 0,06 0,06 0,05 0,07 0,08 0,12 0,10 0,11 0,11 0,12
C5 0,35 0,35 0,31 0,28 0,30 0,18 0,25 0,17 0,22 0,20
C6 0,03 0,03 0,02 0,01 0,04 0,02 0,01 0,04 0,01 0,01
C7 0,04 0,04 0,03 0,04 0,06 0,09 0,05 0,09 0,07 0,09
C8 0,02 0,02 0,02 0,01 0,03 0,01 0,01 0,02 0,01 0,01
C9 0,04 0,04 0,03 0,02 0,05 0,07 0,03 0,08 0,04 0,06
C10 0,03 0,03 0,02 0,02 0,04 0,05 0,02 0,06 0,02 0,03
Eigenvector is obtained by taking the arithmetic mean of each row in the normalized matrix table. The
eigenvector is shown in Table 3.
Table 7: Eigenvalues of the Normalized Matrix
Eigenvalues
0,178
0,178
0,123
0,086
0,261
0,022
0,061
0,016
0,043
0,031
The comparison matrix and the eigenvector column of the normalized matrix were multiplied to obtain
new values. These values are shown in Table 4.
Table 8: The Comparison Matrix * Eigenvector values
The Comparison Matrix *
Eigenvector values
1,917096966
1,917096966
1,312377408
0,910186323
2,786387852
0,218076545
0,631635282
0,163463064
0,439201699
0,307250922
Page 835
SEMIT 2023
The values in Table 4 are divided by the corresponding eigenvector values for each row to obtain new
values. These values are shown in Table 5.
Table 9: The Comparison Matrix * Eigenvector values/Eigenvector
Eigenvector values/Eigenvector
10,75189
10,75189
10,70094
10,52694
10,65874
10,08016
10,31392
10,18729
10,12907
10,04426
The average of the values in Table 5 is taken, and the λmax value is calculated, resulting in a λmax value
of 10.4145. Finally, the CI and CR values are calculated. The formulas for these values are provided
below (4.1.1.1 - 4.1.1.2). A CR value below 0.10 indicates that the comparisons made by the decision
maker are consistent. If the CR value is greater than 0.10, it indicates either a calculation error in AHP
or inconsistency in the decision maker's comparisons. Since our CR value is 0.0391, we can say that our
comparison matrix is consistent. The obtained values are given in Table 6.
𝜆𝑚𝑎𝑥−𝑛 𝐶𝐼
𝐶𝐼 = 𝑛−1
(4.1.1.1) 𝐶𝑅 = 𝑅𝐼 (4.1.1.2)
Table 10: λmax, CI and CR Value
λmax CI CR
10,41451 0,046057 0,03091
In the next step, a comparison matrix for alternatives is created for each criterion. The comparison matrix
for the military power criterion is shown in Table 7.
Table 11: Comparison Matrix for Military Power Criterion
Comparison Matrix Byzantium Karaman Trebizond Hungary Dulkadirogulları
for Military Power
Criterion
Byzantium 1 0,5 0,25 2 0,33
Karaman 2 1 0,33 3 0,5
Trebizond 4 3 1 5 2
Hungary 0,5 0,33 0,2 1 0,25
Dulkadirogulları 3 2 0,5 4 1
Sum 10,5 6,83 2,28 15 4,08
The normalized values of the comparison matrix created for the first criterion are given in Table 8.
Table 12: The Normalized Matrix for the Military Strength Criterion Comparison Matrix
Normalized Byzantium Karaman Trebizond Hungary Dulkadirogullar
Matrix ı
Page 836
SEMIT 2023
Byzantium 0,0952381 0,073206442 0,109649123 0,13333333 0,080882353

Karaman 0,19047619 0,146412884 0,144736842 0,2 0,12254902
Trebizond 0,38095238 0,439238653 0,438596491 0,33333333 0,490196078
Hungary 0,04761905 0,048316252 0,087719298 0,06666667 0,06127451
Dulkadirogulları 0,28571429 0,292825769 0,219298246 0,26666667 0,245098039
The eigenvector matrix was created by finding the arithmetic mean of the rows of the normalized matrix.
The eigenvectors are shown in Table 9.
Table 13: Eigenvector
Eigenvector
0,098461869
0,160834987
0,416463387
0,062319155
0,261920601
The matrix multiplication of the comparison matrix and the eigenvectors is performed. The resulting
values are then divided by the eigenvectors again. These values are presented in Table 10 and Table 11.
Table 14: The matrix product of the comparison matrix and eigenvectors.
Eigenvector
0,494067
0,813109
2,128253
0,313398
1,336484
Table 15: The division of the values in Table 10 by the eigenvectors
The Comparison Matrix * Eigenvector /
Eigenvector
5,017854
5,055551
5,1103
5,028927
5,102632
Finally, the arithmetic mean of the values in Table 11 is calculated, and λmax value is obtained. CI and
CR values are calculated to observe whether the matrix is consistent or not. Looking at the values in
Table 12, it can be observed that the comparison matrix we created for the military power criterion is
consistent.
Table 16: λmax, CI ve CR Value
λmax CI CR
5,063053 0,015763 0,014074
Page 837
SEMIT 2023
For the other 9 criteria, the same steps are followed. In our study, all CR values turned out to be less than
0.10. In this case, it can be said that a consistent comparison has been made. Table 13 shows the CR
values obtained for all criteria.
Table 17: CR Values for Criteria
Military Power 0,0140
Ally Military Support 0,0022
Country Development Level 0,0440
Expected Income 0,0140
Expected Reaction from Other Countries 0,0060
Number of Troops Countries Can Mobilize 0,0140
Prestige Gain 0,0140
Military Losses 0,0292
Share of Countries in Trade Centers 0,0010
Distance to Other Countries 0,0119
In the final stage, the eigenvector matrix is created. Cities are listed in rows, criteria in columns, and the
eigenvectors obtained for each criterion are entered into the matrix. This matrix is then multiplied by the
eigenvectors obtained from the comparison of criteria in the first stage. The resulting values are sorted
in descending order to find the outcome. The largest value indicates the country we should choose. The
resulting matrix and the ranking are shown in Table 15.
Table 18: Matrix of Results (Eigenvectors)

MP AMS CDL EI EROC NTCM PG MS SCTC DOC
Byzantiu 0,09846 0,13181 0,19681 0,26192 0,21478 0,09846 0,26192 0,1044 0,31324 0,48395
m 1869 1094 2763 0601 8095 1869 0601 9815 8945 5837
Karaman 0,16083 0,21598 0,15137 0,16083 0,12128 0,16083 0,09846 0,2604 0,09843 0,22209
4987 6145 3553 4987 7905 4987 1869 3383 8819 3023
Trebizond 0,41646 0,52179 0,04847 0,06231 0,21478 0,41646 0,16083 0,1598 0,17662 0,07882
3387 6067 8799 9155 8095 3387 4987 5372 4473 6051
Hungary 0,06231 0,05063 0,52843 0,41646 0,07385 0,06231 0,41646 0,0617 0,31324 0,13629
9155 2267 2094 3387 8359 9155 3387 2445 8945 9037
Dulkadiro 0,26192 0,07977 0,07490 0,09846 0,37527 0,26192 0,06231 0,4134 0,09843 0,07882
gulları 0601 4427 279 1869 7547 0601 9155 8985 8819 6051
Table 19: AHP Results

0,20021 BYZANTIUM
0,15728 KARAMAN
0,26797 TREBIZOND
0,17213 HUNGARY
0,20242 DULKADIROGULLARI
As seen from Table 15, Trebizond is selected with a value of 0.266272.
4.4.2. TOPSIS Solution
The decision matrix created in the first stage of the TOPSIS method is presented in Table 16. The scores
for each criterion in the decision matrix are individually divided by the square root of the sum of the
Page 838
SEMIT 2023
squares of the values in the respective columns to obtain the standardized decision matrix. These values
are shown in Table 17.
Table 20: Decision Matrix
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Byzantium 6000 11000 49 5,06 49 11000 25 4000 20,80% 237

Karaman 5000 4000 43 3,62 138 10000 10 2000 6,10% 942
Trebizond 3000 2000 12 1,94 49 7000 15 3000 16,20% 1300
Hungary 16000 21000 166 12,71 173 28000 40 15000 22% 1037
Dulkadirogulları 4000 14000 23 2,28 39 8000 5 1500 6,50% 1356
SUM 34000 52000 293 25,61 448 64000 95 25500 0,713 4872
𝒓 𝒂𝒊𝒋 (4.4.2.1)
𝒊𝒋=
𝒎
√∑𝒌=𝟏 𝒂𝟐 𝒌𝒋
Table 21: TOPSIS Standard Decision Matrix (R)

R C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Byzantium 0,324442 0,3943 0,2718 0,3498 0,2083 0,3289 0,4926 0,2498 0,5893 0,100622
842 691 91 284 767 817 646 78 912 311
Karaman 0,270369 0,1434 0,2385 0,2502 0,5868 0,2990 0,1970 0,1249 0,1728 0,399941
035 069 982 725 569 743 659 39 503 845
Trebizond 0,162221 0,0717 0,0665 0,1341 0,2083 0,2093 0,2955 0,1874 0,4590 0,551936
421 035 855 239 767 52 988 085 451 729
Hungary 0,865180 0,7528 0,9211 0,8787 0,7356 0,8374 0,7882 0,9370 0,6148 0,440275
913 865 191 974 08 634 426 937 683
Dulkadirog 0,216295 0,5019 0,1276 0,1576 0,1658 0,2392 0,0985 0,0937 0,1841 0,575712
ulları 228 243 223 302 509 594 329 043 848 465
Table 22: Weighted Standard Decison Matrix (V)

V C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Byzantium 0,058 0,070 0,033 0,030 0,054 0,007 0,030 0,004 0,026 0,003
Karaman 0,048 0,026 0,029 0,022 0,153 0,006 0,012 0,002 0,007 0,012
Trebizond 0,029 0,013 0,008 0,012 0,054 0,005 0,018 0,003 0,020 0,017
Hungary 0,154 0,134 0,113 0,076 0,192 0,018 0,048 0,015 0,027 0,013
Dulkadirogulları 0,039 0,089 0,016 0,014 0,043 0,005 0,006 0,002 0,008 0,018
After obtaining the standardized decision matrix, the weighted standardized decision matrix is calculated
by multiplying the weights of the criteria with the standardized decision matrix. Table 18 shows the
weighted standardized decision matrix. The weights of the criteria, obtained using the AHP method, are
presented in Table 19.
Table 19: Criterion weights
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
wi 0,18 0,18 0,12 0,09 0,26 0,02 0,06 0,02 0,04 0,03
Page 839
SEMIT 2023
In the next step, assuming that the criteria exhibit a monotonically increasing or decreasing trend, the
maximum (A*) and minimum (A-) values from the weighted standardized decision matrix are identified.
These values are shown in Table 20.
Table 20: A* ve A- Values
A* 0,029 0,013 0,113 0,076 0,043 0,005 0,048 0,002 0,027 0,003
A- 0,154 0,134 0,008 0,012 0,192 0,018 0,006 0,015 0,007 0,018
Next, using the formulas below, the distances of each decision point in the matrix to the criterion values
of the ideal and negative ideal solutions are calculated. The calculated values are shown in Table 21 and
Table 22.
𝑺∗𝒊 = √∑𝒏𝒋=𝟏(𝒗𝒊𝒋 − 𝒗∗𝒋 )𝟐 (4.4.2.2)
𝑺− 𝒏 − 𝟐
𝒊 = √∑𝒋=𝟏(𝒗𝒊𝒋 − 𝒗𝒋 ) (4.4.2.3)
Table 21: S* Values

S1* 0,114204142
S2* 0,155849738
*
S3 0,127315886
*
S4 0,230263436
*
S5 0,146518859
Table 22: S- Values
S1- 0,185791457
-
S2 0,159560094
-
S3 0,223816625
S4- 0,131449981
-
S5 0,194912779
After finding the values of S* and S- , the relative closeness to the ideal solution is calculated using the
separation criteria with the following formula. The calculated C* values are shown in table 23.
𝑺−
𝑪∗𝒊 = 𝑺−+𝑺
𝒊
∗ (4.4.2.4)
𝒊 𝒊
Table 23: C* Values

C1*(Byzantium) 0,619313941
*
C2 (Karaman) 0,505881802
C3*(Trebizond) 0,637413564
*
C4 (Hungary) 0,363409193
C5 (Dulkadirogulları)
*
0,570869121
Finally, the obtained C* values are sorted in descending order. The highest C* alue gives us the result.
Trebizond is chosen with a value of 0.6374.
4.4.3. ELECTRE Solution
The first three steps of the ELECTRE method are identical to the first three steps of the TOPSIS method.
The decision matrix, standard decision matrix, and weighted standard decision matrix are the same as
provided in Table 16, 17, and 18 in the TOPSIS section.
Page 840
SEMIT 2023
Using the weighted standard decision matrix, concordance and discordance sets are calculated in the
ELECTRE method. The maximum value in the concordance and discordance sets is found using the
formula m*(m-1) where m represents the number of alternatives. The maximum number of elements in
the sets should be equal to the number of criteria. Comparisons are made in pairs for each criterion within
the rows, and the dominant ones are placed in the concordance set, while the non-dominant ones are
placed in the discordance set. Table 24 shows the sets.
Table 24: Compliance and Non-compliance Sets

C12 3-4-5-6-7-9-10 D12 1-2-8
C13 3-4-5-7-9-10 D13 1-2-6-8
C14 1-2-5-6-8-9-10 D14 3-4-7-
C15 2-3-4-6-7-9-10 D15 1-5-8-
C23 3-4-8-10 D23 1-2-5-6-7-9
C24 1-2-5-6-8-10 D24 3-4-7-9
C25 2-3-4-6-7-9-10 D25 1-5-8
C34 1-2-5-6-8- D34 3-4-7-9-10
C35 1-2-4-6-7-9-10 D35 3-5-8-
C45 3-4-7-9-10 D45 1-2-5-6-8
C21 1-2-3-6-8 D21 4-5-7-9-10
C31 1-2-5-6-8 D31 3-4-7-9-10
C41 3-4-7-9 D41 1-2-5-6-8-10
C51 1-5-6-8 D51 2-3-4-7-9-10
C32 1-2-5-6-7-9 D32 3-4-8-10
C42 3-4-7-9-10 D42 1-2-5-6-8
C52 1-5-6-7-8-9 D52 2-3-4-10
C43 3-4-7-9-10 D43 1-2-5-6-8
C53 3-4-5-8-10 D53 1-2-6-7-9
C54 1-2-5-6-8 D54 3-4-7-9-10
In the next step, a concordance and discordance matrix is created. In the concordance matrix, the weights
of the criteria found for each cell value in the concordance set are summed to create the matrix. When
creating the discordance matrix, the differences between the criteria in the discordance set are calculated
(absolute differences). Then the maximum value is selected, and that value is divided by the largest
difference value among the compared alternatives. The matrices are shown in Table 25 and Table 26.
Table 25: Compatibility Matrix
Byzantium Karaman Trebizond Hungary Dulkadirogulları
Byzantium 0,63 0,61 0,73 0,54
Karaman 0,52 0,26 0,69 0,54
Trebizond 0,66 0,74 0,66 0,60
Hungary 0,31 0,34 0,34 0,34
Dulkadirogulları 0,48 0,58 0,52 0,66
Table 26: Incompatibility Matrix
Byzantium 0.4 1 0,57 1
Karaman 1 1 0,8 0,11
Trebizond 0,33 0,2 0,71 0,125
Hungary 1 1 1 1
Dulkadirogulları 1 0,54 1 0,60
Page 841
SEMIT 2023
Next, the concordance dominance and discordance dominance tables are multiplied to obtain the total
dominance matrix. This matrix is shown in Table 27.
Table 27: Total Dominance Matrix
Byzantium 1 1
Karaman 1
Trebizond 1
Hungary
Dulkadirogulları 1
Upon reviewing the results, it can be observed that Byzantium dominates over Trebizond and Dulkadir.
In the final result, Byzantium is ranked first.
4.4.4. VIKOR Solution
In the first step of the VIKOR method, the initial matrix used in other methods is utilized.
Table 28: VIKOR Initial Matrix
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Byzantium 6000 11000 49 5,06 49 11000 25 4000 20,8% 237
Karaman 5000 4000 43 3,62 138 10000 10 2000 6,1% 942
Trebizond 3000 2000 12 1,94 49 7000 15 3000 16,2% 1300
Hungary 16000 21000 166 12,71 173 28000 40 15000 21,7% 1037
Dulkadirogulları 4000 14000 23 2,28 39 8000 5 1500 6,5% 1356
Next, the fj* and fj- values for each column are calculated, representing the maximum and minimum
values of the columns, respectively. These values are shown in Table 29.
Table 29: fj* ve fj- values
fi* 3000 2000 166 12,71 39 7000 40 1500 21,70% 237
fj- 16000 21000 12 1,94 173 28000 5 15000 6,10% 1356
Along with these calculated values, the weighted decision matrix is constructed using the formula
provided below for each cell.
The weighted standard decision matrix is shown in Table 30.
Table 30: Weighted Standard Decision Matrix
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10
Byzantium 0,041 0,084 0,093 0,061 0,019 0,004 0,026 0,004 0,002 0,000
Karaman 0,027 0,019 0,098 0,073 0,192 0,003 0,051 0,001 0,040 0,019
Trebizond 0,000 0,000 0,123 0,086 0,019 0,000 0,043 0,002 0,014 0,028
Hungary 0,178 0,178 0,000 0,000 0,260 0,020 0,000 0,020 0,000 0,021
Dulkadirogulları 0,014 0,113 0,114 0,084 0,000 0,001 0,060 0,000 0,039 0,030
Once the weighted standard decision matrix is obtained, Si values are calculated. Si values are computed
by summing each row of the weighted standard decision matrix. Subsequently, Ri values are calculated
by taking the maximum value of each row. Si and Ri values are shown in Table 31.
Page 842
SEMIT 2023
Table 31:Si ve Ri Values

Si Ri
Byzantium 0,33514 0,0932
Karaman 0,52315 0,1921
Trebizond 0,31619 0,1226
Hungary 0,67805 0,2600
Dulkadirogulları 0,45387 0,1139
Once this table is created, S* and S- values are determined. These values are, respectively, the maximum
and minimum of the Si values.
Similarly, R* and R- values are also calculated. Additionally, in the methodology, the v value is taken as
0.5. The relevant values are presented in Table 32.
Table 32: S*, S-, R*, R- and v Values
S* 0.31618
S- 0.67805
R* 0.09317
R- 0.26
v 0.5
Once these values are obtained, Si and Ri values are sorted in ascending order. This ordering will be used
to test the consistency of the method in subsequent steps. To determine the DQ value needed for the
consistency test, the operation 1/ (Number of Alternatives - 1) is performed. After obtaining this result,
Q''-Q' value is calculated.
In this calculation, the smallest second value in the Q ordering is subtracted from the smallest value. The
method is considered consistent when Q''-Q' value is greater than the DQ value.
When the consistency of the method is ensured, the Qi values obtained are sorted in ascending order.
The ordering is done based on the values corresponding to the relevant row. The resulting ordering is
shown in Table 33.
Table 33: Qi values sorted from smallest to largest
Byzantium 0,0261
Trebizond 0,1677
Dulkadirogulları 0,3081
Karaman 0,8492
Hungary 1,45
With this ordering, it can be observed that the Byzantium alternative is the best alternative.
4.4.5. SWARA Solution
SWARA (Step-wise Weight Assessment Ratio Analysis); It is used to determine the weights of criteria,
not alternatives. Therefore, it will be included as a method in the study, and the steps will be followed
to show the weights of the criteria in the result matrix. Since SWARA is not seen as an effective
weighting method, the weights obtained have not been used in any other method.
In the first step, the criteria were ranked according to their importance. Normally, the military power and
ally military support criteria are considered equal, but for the method to provide more accurate results,
military power was placed in 2nd place, and ally military support was placed in 3rd place.
Page 843
SEMIT 2023
Table 34: SWARA Step1

CRITERIA RANK
Expected Reaction from Other Countries 1
Military Power 2
Ally Military Support 3
Country Development Level 4
Expected Income 5
Gained Prestige 6
Share in Countries' Trade Centers 7
Distance to Other Countries 8
Number of Soldiers Countries Can Produce 9
Military Loss 10
Moving on to the second step, it is necessary to find the sj values. The sj cell value for the first row
(criterion) should be empty, and for the remaining criteria, weighting should be done sequentially to find
the sj value. In this weighting process, multiples of 5 and 5 must be used, and when comparing, each
criterion should only be compared with the criterion after it. The matrix resulting from the second step
is shown in Table 35.
Table 35: SWARA Step 2
CRITERIA RANK Sj
Expected Reaction from Other Countries 1
Military Power 2 0,10
Ally Military Support 3 0,05
Country Development Level 4 0,10
Expected Income 5 0,15
Gained Prestige 6 0,15
Share in Countries' Trade Centers 7 0,25
Distance to Other Countries 8 0,30
Number of Soldiers Countries Can Produce 9 0,10
Military Loss 10 0,05
In the 3rd and 4th steps, the kj and qj values need to be calculated. When calculating kj values, 1 is added
to each sj value, and the kj value of the criterion in the 1st row is set to 1. When calculating the qj value,
the criterion in the 1st row takes the value of 1. When finding the results for other criteria, the qj value
of the previous criterion is divided by the kj value of the respective row. The matrices resulting from
steps 3 and 4 are given in Table 36 and Table 37.
CRITERIA RANK Sj Kj
Expected Reaction from Other Countries 1 1
Military Power 2 0,10 1+0,10
Ally Military Support 3 0,05 1+0,05
Country Development Level 4 0,10 1+0,10
Expected Income 5 0,15 1+0,15
Gained Prestige 6 0,15 1+0,15
Share in Countries' Trade Centers 7 0,25 1+0,25
Distance to Other Countries 8 0,30 1+0,30
Number of Soldiers Countries Can Produce 9 0,10 1+0,10
Military Loss 10 0,05 1+0,05
Page 844
SEMIT 2023

CRITERIA RANK Sj Kj Qj
Expected Reaction from Other Countries 1 1 1
Military Power 2 0,10 1,10 1/1,10=0,91
Ally Military Support 3 0,05 1,05 0,91/1,05=0,87
Country Development Level 4 0,10 1,10 0,87/1,10=0,79
Expected Income 5 0,15 1,15 0,69
Gained Prestige 6 0,15 1,15 0,6
Share in Countries' Trade Centers 7 0,25 1,25 0,48
Distance to Other Countries 8 0,30 1,30 0,37
Number of Soldiers Countries Can 9 0,10 1,10 0,34
Produce
Military Loss 10 0,05 1,05 0,32
In the final step, the final weights (wj) are calculated. When determining the weights, the column sum of
qj values is computed. Then, each qj value is divided by the total qj value to find the wj values. The result
matrix and the final criterion weights are provided in Table 38.
Table 38: SWARA Result Matrix
CRITERIA RANK Sj Kj Qj Wj
Expected Reaction from Other Countries 1 1 1 1/6,37=0,16
Military Power 2 0,10 1,10 0,91 0,91/6,37=0,14
Ally Military Support 3 0,05 1,05 0,87 0,87/6,37=0,13
Country Development Level 4 0,10 1,10 0,79 0,12
Expected Income 5 0,15 1,15 0,69 0,11
Gained Prestige 6 0,15 1,15 0,6 0,10
Share in Countries' Trade Centers 7 0,25 1,25 0,48 0,08
Distance to Other Countries 8 0,30 1,30 0,37 0,07
Number of Soldiers Countries Can 9 0,10 1,10 0,34 0,05
Produce
Military Loss 10 0,05 1,05 0,32 0,04
SUM Qj = 6,37 1.0
When we arrange the wj values from Table 38, we can see that we reach the ranking matrix obtained in
the first step. This indicates that we haven't made any processing errors.
4.4.6. Comparison of Solution Methods
The rankings determined according to the applied methods are shown in Table 39.
Table 39: Methods and Rankings
METHOD RANK
AHP Trebizond-Dulkadirogulları-Byzantium-Hungary-Karaman
TOPSIS Trebizond-Byzantium-Dulkadirogulları-Karaman-Hungary
ELECTRE Byzantium-Trebizond-Dulkadirogulları-Karaman- Hungary
VIKOR Byzantium-Trebizond-Dulkadirogulları-Karaman- Hungary
Page 845
SEMIT 2023
When we look at the methods and rankings, it can be observed that the alternatives Trebizond and
Byzantium rank first in both methods, while the alternatives Karaman and Hungary generally rank last
in the methods. Based on these results, it can be stated that Trebizond and Byzantium are the best
alternatives, and Karaman and Hungary are the worst alternatives.
Looking at the actual data, for Trebizond, it can be seen that it ranks best in important criteria such as
military power, allied military support, expected reactions from other countries, and the number of
soldiers it can produce. In other criteria, it maintains an average ranking. For the Karaman alternative, it
ranks worst in criteria such as expected reactions from other countries, gain of prestige, and share in trade
centers. When other criteria are examined, it can also be concluded that Karaman ranks towards the
bottom.
Comparing the rankings of the methods, it is expected that the results of TOPSIS, ELECTRE, and
VIKOR methods will be the same. This is because the weights of the criteria are taken into account.
However, when examining the methods in detail, it can be said that the TOPSIS and VIKOR methods
look at the maximum and minimum values, while the ELECTRE method does not look at values higher
or lower than the specified criteria values and tries to find average values. Therefore, it is thought that
there might be slight differences in the results of the methods.
In the AHP method, it is considered that expecting as consistent results as the other methods might be
wrong, as the AHP method involves evaluations based on the discretion of the decision-maker without
relying on too many formulations. However, it is observed that AHP behaves very consistently during
the scoring process, and therefore the results are almost the same as the results of the other methods.
Additionally, no inconsistency was encountered in the CR values calculated for each criterion in the AHP
method during the implementation phase. Therefore, it is assumed that the AHP method is also quite
consistent.
4.4.7 Sensitivity Analysis
Sensitivity analysis helps determine which risks or uncertainties could potentially have a greater impact
on the project. It is used to analyze the relationship between uncertainties in the project and project
objectives, and how they influence them. In this project, it is very clear that the importance levels of the
criteria will affect the outcome. However, in the initial stage, consistency analyses were conducted based
on the results obtained from the Analytic Hierarchy Process (AHP), and the outcomes demonstrated the
correctness of the decisions made. Nevertheless, not only AHP but also the Swara method were used for
the criterion weighting. Both equal weights (10%) and weights determined by Swara were used to see if
the results would change, and the effect of the weights on the choice is shown in Table 40 and Table 41.
Table 40: Results Obtained When Criteria Weights Are Equal

TOPSIS Byzantium>Trebizond>Karaman>Dulkadir>
Hungary
ELECTRE Trebizond> Byzantium
>Karaman>Dulkadir> Hungary
VIKOR Byzantium
>Trebizond>Karaman>Dulkadir> Hungary
Table 41: Results Obtained When SWARA Weights Are Accepted

TOPSIS Trebizond> Byzantium > Hungary
>Dulkadir>Karaman
ELECTRE Byzantium >Trebizond> Hungary
>Dulkadir>Karaman
VIKOR Byzantium >Trebizond> Hungary
>Dulkadir>Karaman
In addition to these results; when the criteria weights are set as 0.20 for the Country Development Level
criterion, 0.05 for the Ally Military Support and Military Loss criteria, Hungary is chosen in all three
Page 846
SEMIT 2023
methods. In sensitivity analysis, the most important criterion is assigned a value of 1. The original value
is divided by 1 minus the most important weight value to obtain a value. When the original value is
divided by this new value, we obtain the delta x range. This range shows us the values within which the
result will not be affected when changes are made. For our problem, the calculated range is determined
as follows:
-0.18 ≤ Δx ≤ 0.85.
5. Results and Recommendations

In this study, the selection of the country to be conquered by the Ottoman Empire in the Europa
Universalis IV game has been made. In making this selection, various MCDA methods including AHP,
TOPSIS, ELECTRE, VIKOR, and SWARA (SWARA is not a method for alternative selection, it is only
used for criterion weighting) have been applied. The results were compared, and in three out of four
methods (excluding SWARA as it is not used for alternative selection), Trebizond emerged as the best
alternative. During the application phase, the importance of criteria was determined based on the steps
to succeed in the game, considering the rules of the game. Looking back in history, it is seen that Mehmed
the Conqueror decided to conquer Byzantium during the relevant period. Despite Trebizond appearing
as the best alternative using modern methods, the Byzantium option ranks first in one method and second
with a very slight difference in others. Based on this result, it can be said that the parameters considered
in the game are quite consistent.
The project topic is based on a subject that is very close to reality but not entirely real. In future studies,
real data can be used to address real-world problems and find solutions to these problems using MCDM
methods. Additionally, it is considered appropriate to use more scientific methods like TOPSIS and
VIKOR rather than decision-maker-centric techniques such as AHP and SWARA among the MCDM
methods used.
References
Abo-Sinna, M. A., & Abou-El-Enien, T. H. (2006). An interactive algorithm for large scale
multiple objective programming problems with fuzzy parameters through TOPSIS approach.
Applied Mathematics and Computation, 177(2), 515-527.
Adamczak, M., Domański, R., & Wagener, N. (2016). Comparison of multiple-criteria decision-
making methods-results of a simulation study. LogForum, 12(4), 237-246.
Alimardani, M., Hashemkhani Zolfani, S., Aghdaie, M. H., & Tamošaitienė, J. (2013). A
novel hybrid SWARA and VIKOR methodology for supplier selection in an agile environment.
Technological and economic development of economy, 19(3), 533-548.
Alimoradi, A., Yussuf, R. M., & Zulkifli, N. (2011). A hybrid model for remanufacturing
facility location problem in a closed-loop supply chain. International Journal of Sustainable
Engineering, 4(01), 16-23.
Anık, Z. (2007). Nesne yönelimli yazılım dillerinin analitik hiyerarşi ve analitik network prosesi
ile karşılaştırılması ve değerlendirilmesi. Fen Bilimleri Enstitüsü, Ankara.
ARSLAN, H. M. (2018). ELECTRE VE MAKSİMUM KAPSAMA MODELİ
YÖNTEMLERİ İLE BİLİM MERKEZLERİNİN OPTİMUM TESİS YERİ SEÇİMİ. Elektronik
Sosyal Bilimler Dergisi, 17(65), 337-355.
Avikal, S., Mishra, P. K., Jain, R., & Yadav, H. C. (2013). A PROMETHEE method based
heuristic for disassembly line balancing problem. Industrial Engineering & Management Systems,
12(3), 254-263.
Page 847
SEMIT 2023
Ayoko, G. A., Morawska, L., Kokot, S., & Gilbert, D. (2004). Application of multicriteria
decision making methods to air quality in the microenvironments of residential houses in Brisbane,
Australia. Environmental science & technology, 38(9), 2609-2616.
Bottero, M., & Ferretti, V. (2010). An analytic network process‐based approach for location
problems: the case of a new waste incinerator plant in the Province of Torino (Italy). Journal of
Multi‐Criteria Decision Analysis, 17(3‐4), 63-84.
Chater, N., Oaksford, M., Nakisa, R., & Redington, M. (2003). Fast, frugal, and rational:
How rational norms explain behavior. Organizational behavior and human decision processes,
90(1), 63-86.
San Cristóbal, J. R. (2012). Contractor selection using multicriteria decision -making
methods. Journal of Construction Engineering and Management, 138(6), 751-758.
Çakir, S., & Perçin, S. (2013). Çok kriterli karar verme teknikleriyle lojistik firmalarinda
performans ölçümü/Performance measurement of logistics firms with multi-criteria decision
making methods. Ege Akademik Bakis, 13(4), 449.
Dagdeviren, M. (Metin), & Eraslan, E. (Erguen). (2008). Priority determination in strategic
energy policies in Turkey using analytic network process (ANP) with group decision making.
Energy Reviews, 32(11), 1047-1057.
Dağdeviren, M. (2005). Performans Değerlendirme sürecinin çok ölçütlü karar verme
Yöntemleri ile Bütünleşik Modellenmesi. Fen Bilimleri Enstitüsü, Ankara.
Diaby, M., Ferrer, H., Valognes, F., & Clement‐Demange, A. (2011). A comprehensive
decision approach for rubber tree planting management in Africa. Journal of Multi‐Criteria
Decision Analysis, 18(3-4), 187-201.
Duckstein, L., & Gershon, M. (1983). Multicriterion analysis of a vegetation management
problem using ELECTRE II. Applied Mathematical Modelling, 7(4), 254-261.
Eleren, A. (2010). Kuruluş yeri seçiminin analitik hiyerarşi süreci yöntemi ile belirlenmesi;
deri sektörü örneği. Atatürk Üniversitesi İktisadi ve İdari Bilimler Dergisi, 20(2), 405 -416.
Fu, Y. K., Huang, W., & Liao, C. N. (2020). The selection model for horizontal alliances
between hotels and airlines: an integrated application of NGT, fuzzy TOPSIS and MCGP
methods. Tourism Review, 75(4), 681-698.
Gelashvili, T. (2019). Çok kriterli karar verme yöntemleri ile performans değerlendirmesi:
AHP, TOPSIS ve PROMETHEE yöntemlerinin karşılaştırılması (Master's thesis, Sosyal
Bilimler Enstitüsü).
Goswami, S. S., Behera, D. K., Afzal, A., Razak Kaladgi, A., Khan, S. A., Rajendran, P., ...
& Asif, M. (2021). Analysis of a robot selection problem using two newly developed hybrid
MCDM models of TOPSIS-ARAS and COPRAS-ARAS. Symmetry, 13(8), 1331.
Gözde, K. O. C. A., EĞİLMEZ, Ö., DEMİR, E., KARAMAŞA, Ç., & GÖKCAN, H. (2022).
Analysis of Drivers and Challenges in Circular Economy with SWARA and BWM Methods in
Clothing Sector. Mehmet Akif Ersoy Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi ,
9(2), 763-787.
Heidary Dahooie, J., Beheshti Jazan Abadi, E., Vanaki, A. S., & Firoozfar, H. R. (2018).
Competency‐based IT personnel selection using a hybrid SWARA and ARAS‐G methodology.
Human Factors and Ergonomics in Manufacturing & Service Industries, 28(1), 5 -16.
Hinduja, A., & Pandey, M. (2023). Analysis and Comparison of State-of-the-Art Fuzzy
Multi-criteria Decision-making Methods Under Different Levels of Uncertainty. Vision, 27(1),
93-109.
İç, Y. T., & Yurdakul, M. (2019). İşleme merkezi seçim problemlerinde kullanılan çok kriterli
karar verme yöntemi uygulamalarında kriter ve alternatif sayısının sıralama sonuçlarına
Page 848
SEMIT 2023
etkisinin incelenmesi. Gazi Üniversitesi Mühendislik Mimarlık Fakültesi Dergisi, 35(2), 991 -1002.
Jahan, A., Mustapha, F., Ismail, M. Y., Sapuan, S. M., & Bahraminasab, M. (2011). A
comprehensive VIKOR method for material selection. Materials & Design, 32(3), 1215 -1221.
Jovanović, S., Zavadskas, E. K., Stević, Ž., Marinković, M., Alrasheedi, A. F., & Badi, I. (2023).
An Intelligent Fuzzy MCDM Model Based on D and Z Numbers for Paver Selection: IMF D -
SWARA—Fuzzy ARAS-Z Model. Axioms, 12(6), 573.
Kahraman, C., Cebi, S., & Tuysuz, F. (2010). Fuzzy location selection techniques. In Production
engineering and management under fuzziness (pp. 329-358). Berlin, Heidelberg: Springer Berlin
Heidelberg.
Korkusuz, A., Inan, U., Ozdemir, Y., & BAŞLIGİL, H. (2020). Occupational health and safety
performance measurement in healthcare sector using integrated multi criteria decision making
methods. Journal of the Faculty of Engineering and Architecture of Gazi University, 35(1).
Lixin, D., Ying, L., & Zhiguang, Z. (2008, April). Selection of logistics service provider based
on analytic network process and VIKOR algorithm. In 2008 IEEE International Conference on
Networking, Sensing and Control (pp. 1207-1210). IEEE.
Mulliner, E., Malys, N., & Maliene, V. (2016). Comparative analysis of MCDM methods for the
assessment of sustainable housing affordability. Omega, 59, 146-156.
Opricovic, S., & Tzeng, G. H. (2004). Compromise solution by MCDM methods: A comparative
analysis of VIKOR and TOPSIS. European journal of operational research, 156(2), 445 -455.
Opricovic, S., & Tzeng, G. H. (2007). Extended VIKOR method in comparison with outranking
methods. European journal of operational research, 178(2), 514-529.
Özcan, E. C., Ünlüsoy, S., & Tamer, E. R. E. N. (2017). ANP ve TOPSIS yöntemleriyle
türkiye'de yenilenebilir enerji yatirim alternatiflerinin değerlendirilmesi. Selçuk Üniversitesi
Mühendislik, Bilim ve Teknoloji Dergisi, 5(2), 204-219.
Pak, J. Y., Thai, V. V., & Yeo, G. T. (2015). Fuzzy MCDM approach for evaluating intangible
resources affecting port service quality. The Asian Journal of Shipping and Logistics, 31(4), 459 -
468.
Peker, B. N., & Görener, A. (2023). Geliştirilmiş Bulanık SWARA ve Bulanık CODAS
Yöntemleriyle Tesis Yeri Seçimi: İmalat Sektöründe Bir Uygulama. Journal of Turkish Operations
Management, 7(1), 1493-1512.
Quynh, M. P., Thu, T. L., Huong, Q. D., Van, A. P. T., Van, H. N., & Van, D. N. (2020).
Distribution center location selection using a novel multi criteria decision -making approach under
interval neutrosophic complex sets. Infinite Study.
Saaty, L.T. (1980), The Analytic Hierarchy Process, McGraw-Hill comp. U.S.A
Saaty, T. L. (1994). Fundamentals of decision making and priority theory with the analytic
hierarchy process. RWS publications.
Sadeghi, A., Azar, A., & Rad, R. S. (2012). Developing a fuzzy group AHP model for
prioritizing the factors affecting success of high-Tech SME's in Iran: A Case Study. Procedia-Social
and Behavioral Sciences, 62, 957-961.
SARIÇALI, G., & KUNDAKCI, N. (2016). AHP ve COPRAS yöntemleri ile otel
alternatiflerinin değerlendirilmesi. International Review of Economics and Management, 4(1), 45 -
66.
Srisawat, C., & Payakpate, J. (2016). Comparison of MCDM methods for intercrop selection in
rubber plantations. Journal of Information and Communication Technology, 15(1), 165 -182.
Torfi, F., Farahani, R. Z., & Mahdavi, I. (2016). Fuzzy MCDM for weight of object’s phrase in
location routing problem. Applied Mathematical Modelling, 40(1), 526-541.
Page 849
SEMIT 2023
Toshtzar, M. (1988). Multi-criteria decision making approach to computer software

evaluation: Application of the analytical hierarchy process. Mathematical and Computer
Modelling, 11, 276-281.
Triantaphyllou, E., & Triantaphyllou, E. (2000). Multi-criteria decision making methods (pp.
5-21). Springer US.
Tscheikner-Gratl, F., Egger, P., Rauch, W., & Kleidorfer, M. (2017). Comparison of multi -
criteria decision support methods for integrated rehabilitation prioritization. Water, 9(2), 68.
Tuskan, Y., & Basari, E. AHP-VIKOR BÜTÜNLEŞİK YAKLAŞIMI İLE ŞEV STABİLİTE
PROBLEMLERİNDE GEOTEKNİK DAYANMA YAPISI SEÇİMİ1. Mühendislikte Güncel
Araştırmalar..
Wang, T. C., Lee, H. D., & Wu, C. C. (2007, April). A fuzzy TOPSIS approach with
subjective weights and objective weights. In 6th WSEAS International Conference on Applied
Computer Science (pp. 1-6).
Widianta, M. M. D., Rizaldi, T., Setyohadi, D. P. S., & Riskiawan, H. Y. (2018, January).
Comparison of multi-criteria decision support methods (AHP, TOPSIS, SAW & PROMETHEE)
for employee placement. In Journal of Physics: Conference Series (Vol. 953, p. 01 2116). IOP
Publishing.
Wu, H. Y., Tzeng, G. H., & Chen, Y. H. (2009). A fuzzy MCDM approach for evaluating
banking performance based on Balanced Scorecard. Expert systems with applications, 36(6),
10135-10147.
Yang, J., & Lee, H. (1997). An AHP decision model for facility location selection. Facilities,
15(9/10), 241-254.
YANIK, L., & Tamer, E. R. E. N. (2017). Borsa İstanbul’da işlem gören otomotiv imalat
sektörü firmalarının finansal performanslarının AHP, TOPSIS, ELECTRE ve VIKOR
yöntemleri ile analizi. Yalova Sosyal Bilimler Dergisi, 7(13), 165-188.
YEŞİLYURT, B., KARAKUŞ, K., Şeyda, G. Ü. R., & Tamer, E. R. E. N. (2019). Çok ölçütlü
karar verme yöntemleri ile hastane bilgi yönetim sistemleri için paket programı seçimi. Başkent
Üniversitesi Ticari Bilimler Fakültesi Dergisi, 3(1), 1-21.
Yurdoğlu, H. ve Kundakcı, N. (2017). SWARA ve WASPAS yöntemlerı̇ ile sunucu seçı̇ mı̇ .
Balıkesir Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 20(38), 253–270.
Zolfani, H., S., Esfahani, M. H., Bitarafan, M., Zavadskas, E. K. and Arefi, S.L. (2013).
Developing a new hybrid mcdm method for selection of the optimal alternative of mechanical
longitudinal ventilation of tunnel pollutants during automobile accidents. Transport, 28(1), 89–
96.
Page 850
SEMIT 2023
MCDM
Yavuz Selim Özdemir, Babek Erdebilli, Ender Sevinç
Abstract
Industrial, economic, and financial concepts are affected by multiple factors. Analytical techniques that
must lead to logical conclusions are frequently used to solve such multi-criteria problems. Due to the
COVID-19 pandemic process, customers’ behaviors were affected. To evaluate a bank's dependability, a
set of evaluation criteria from the perspective of its clients can be used. Using multi-criteria evaluation
methods to regulate their importance to the review purpose, we can associate multidimensional and
contradictory accurate criteria into a single universalizing value. The main objective of this study is to
examine public banks from the perspective of their customers using the fuzzy multi-criteria decision-
making approaches. In this study, the ranking of banks from the customer's viewpoint is determined using
the PROMETHEE, while the weight of the banks is inferred using the Fuzzy AHP method with a group
of experts. These techniques are favored due to their quick computation times and straightforward
calculations. The bank data set was provided from June 2020 to June 2022. The number of bank clients,
ATMs, branches, deposit rates, and ten more criteria were used to rank the banks from the customer's
perspective.
Keywords: bank performance evaluation, Fuzzy AHP, PROMETHEE, group decision making.
I. Introduction
The banking industry is crucial to the economy because it facilitates commercializing the deposits it
receives from the general public. The ability of banks to draw in capital and directly offer loans and other
financial services impact economic growth and development. As a result, the banking industry's growth
directly correlates with how well it is doing financially and developing steadily. The stability of the
banking industry's structure also lessens its susceptibility to unfavorable shocks like crises.
The COVID-19 pandemic had profound effects on the banking industry. The restrictions reduced the
activities of businesses, and there were difficulties in paying debts. Banks took measures against
increasing risks in loans and tightened their credit rating standards. They cut costs with branch closures
and staff reductions. At the same time, there was a rapid transition to digital banking. The use of Internet
banking and mobile applications has increased. Customers gained access to banks with online
transactions. The pandemic has led banks to reshape their customer service and adopt remote working
methods. In summary, the pandemic has caused difficulties and accelerated digitalization in the banking
sector.
Banks have been ranked several times concerning their dependability. Analysis of Banks is detailed since
they have many criteria evaluating the bank customers’ perspective. The trustworthiness of banks from
the customer's perspective is a complex, multifaceted incident. For this reason, this set of criteria defining
the research object is ensured. The matter is confused in point of fact that these criteria have many
dimensions. Also, their significance may change in opposite aspects, including that in some instances,
the more significant values of the criteria may point out ahead of status. According to some other cases,
the smaller values signify identifying it. In addition, some criteria are essential for some customers, while
others may be more valuable for other customers. In this study, it is not possible to associate them with a
single value. To analyze this set of problems and multi-criteria evaluation methods, finding a public
divisor for multidimensional criteria and evaluating their significance may be impressive. One
comprehensive result track is that the industry has run through much recruitment and improvement so
soon and bitter that it was, the environment of the banking sector is so far excessively advanced and has
an organizing substructure.
In the second part, a brief literature review has been made. In the third part, the methodologies used in
the paper were explained. The application part is given in the fourth part of the paper. The final section
provides the conclusions, and future studies are presented.
Page 851
SEMIT 2023
II. Literature Review

To better comprehend client retention, the intensity or amount of consumption of services or products
over time, cross-buying or add-on purchases, and word-of-mouth, Blattberg (2001) aimed to grasp the
length, depth, and breadth of relationships. According to Hawkins and Mothersbaugh (2010), the study
of people or organizations and the methods they employ to choose, offer, and construct products, services,
or ideas is at the heart of the customer perspective field. This is done to acknowledge these incidents'
impact on various facets of consumer behavior. Wang., Chi R. and Yang Y. (2004) held that performance
should be evaluated from the customer's perspective since they represent most of a company's current
clients and can potentially increase both the current and future revenue streams. According to Wang, as
customers' primary goals are to generate a steady source of income and maximize their lifetime value or
equity, their perspectives on potential revenue streams are now strategically important.
According to Figuera J., Greco S., and Ehrgott M. (2005), deciding requires comprehension of the issue,
the decision's goal, the sub-criteria and groups under consideration, and the optional actions to be taken.
The next step is to determine the optimum option for supply assignment, considering the options' ability
to support the appropriate portion of the facilities. Today, making decisions is a quantitative science that
formalizes the thinking process and makes it clear from all angles what one must consider to reach a
sensible conclusion. Das and Teng (1999) consulted on the organizational method observed as the
caution. They conceive that a structured decision-making process, such as standard operating processing,
can take away the thinking piece of decision-making. They worried decision-making is made at the cost
of novelty in this environment. Bodily (1985) studied decision-making as a sight that can be established
in many different kinds of studies. So that have a purpose overview of the guide of decision making,
seven phases can be detected. These phases are a conventional way of decision-making. These seven
phases are Case description, requirement determination, goal establishment, options recognition, criteria
determination, selection of decision-making tool, assessment of options against criteria, and approval of
solutions against issue explanation.
Brans and Mareschal (2005) researched that the description issue needs to be clear and specific so that
every stakeholder in the operation knows the problem to be answered. The necessity specification sets
the circumstances the needed compound must meet to become a feasible solution. These requirements
are the restriction of the model. Criteria need to be defined to detect the different alternatives regarding
conformity. These criteria can change into many other appearances if they conduce to a thorough problem
assessment. The next step is one of the most complicated steps of the decision-making process. It is
selecting the instrument that conducts you through the rest of the decision-making process. Huang E.,
Keisler J., and Linkov J. (2011) defined outranking approximations, which are methods that require
holding various votes across measurements. Different measurements of actions are checked to derive
alternatives that can be sectional across dimensions. These models search to find out the whole order of
the alternatives.
According to Ramanathan (2001), AHP is a compensative Multi-Criteria Decision Analysis (MCDA)
technique that confesses commerce among the different elements of the model. Gülfem Tuzkaya and
Cengiz Kahraman (2014) applied with Fuzzy PROMETHEE Method to a supplier selection problem. In
this way, it helps to find solutions in the manufacturing and service sectors. Ayşegül Şahin and Cenk
Akkaya (2013) created a portfolio with PROMETHEE Ranking Method that can help investors. For this
reason, they used the ISE-50 index to rank firms for creating portfolios.
Hung-Yi Wu, Gwo-Hshiung Tzeng, and Yi-Hsuan Chen (2009) utilized Fuzzy AHP Method Multi-
Criteria Decision Analysis (MCDA) access for evaluating banking performance based on the balanced
scorecard. Masoud Goli and Shahab Bayati (2009) advanced a model for electing a convenient Semantic
Web Service combination using PROMETHEE Method and PROMETHEE Among the approaches used
for making multi-criteria judgments, Gaia Methodology is regarded as the most effective and user-
friendly. The Decision Lab is the name of the recently created software available to individuals. The
software was developed in collaboration with the Canadian business Visual Decision for all queries to
aid decision-making. PROMETHEE is a method for multi-criteria decision-making that can be utilized
Page 852
SEMIT 2023
to achieve the best results. The authors believed this study would be appropriate for assessing the banking
performance model for the Taiwan banking sector.
Tolga Kaya and Cengiz Kahraman (2011) defined the Fuzzy Approach for e-banking websites’ quality
evaluation based on customers and integrated using the AHP method. For this reason, the main aim of
this study is to reduce the costs of building e-banking websites. Mei Fang Chen and Tzungi Tang (2005)
developed Fuzzy Multi-Criteria Decision Making (MCDM) to evaluate expatriate assignments, which is
used to vice a solution for human resource departments. Also, this study was developed to focus on human
resource managers with concerns about how the expatriate candidates evaluate the expatriate
assignments.
Cem Koçak (2013) specified a new high-order Fuzzy Arma Time Series for forecasting Methods using
neural networks. It was created for the deficiency of forecasting ARMA models, which tried to prove that
forecasting performance towards the future can be significantly increased with Fuzzy Arma Models.
III. Methodology
Analytic Hierarchy Process (AHP) Theory
Triantaphyllou and Mann (1995) considered n elements to be compared, C1… Cn denotes the relative
weight or significance of Ci concerning Cj by aij and forms a square matrix A= (aij) of order n with the
constraints that aij = 1/ aij for i ≠ j, and aii = 1, all i. Such a matrix is said to be a reciprocal matrix. The
weights are consistent if they are transitive, that aik = aij .ajk for all i, j, and k. Such a matrix might exist if
the aij is calculated from precisely measured data. Then find a vector ω of order n such that Aω = λω. For
such a matrix, ω is said to be a vector of order n, and λ is an Eigenvalue. For a consistent matrix, λ = n.
For matrices involving human judgment, the condition aik = aij ajk does not hold, as human decisions are
inconsistent to a greater or lesser degree. In such a case, the ω vector satisfies the equation Aω= λ max ω
and λmax ≥ n. The difference, if any, between λmax and n indicates the inconsistency of the judgments. If
λmax = n, then the decisions have become consistent. In the end, (λmax-n)/ (n-1) can be used to calculate a
Consistency Index. Saaty (1980) computed huge samples of random matrices of increasing order and the
Consistency Indices of those matrices to compare against judgments made wholly at random. By reducing
the consistency index for the group of decisions by the index for the corresponding random matrix, an
actual consistency ratio (CR) is determined. Using paired comparisons of the criteria and alternatives or
rating measurements of the alternatives based on the criteria, Saaty's (1996) AHP can be used to make
relative measurements.
A pairwise comparison of the goal-related criteria is included in the rating model. The Likert Scale is one
of the rating scales that is later prescribed for grading. The rating levels of each criterion are then
compared pairwise to produce a set of weights for those levels. Priorities for each rating level are split by
the maximum rating weight for that criterion, and each level's strength or intensity is linear. The ultimate
score of an option is then calculated across all criteria as the product of the scaled weight for that criterion
and the product of the criterion weights.
Fuzzy AHP (FAHP)
Using the Fuzzy Analytic Hierarchy Process (FAHP), Saaty's basic Analytic Hierarchy Process (AHP) is
combined with the fuzzy theory. AHP is a frequently used decision-making tool in various multi-criteria
decision-making problems. The responses on the inquiry form determine the relevant Triangular Fuzzy
Numbers (TFNs) for the linguistic variables. A certain hierarchy level is used to generate the pairwise
comparison matrix. The 𝐴̃𝑘 matrix representation is given in Equation 1.
1 𝑑̃𝑘 12 ⋯ 𝑑̃1𝑛
𝑘
𝐴̃𝑘 = 𝑑̃21
𝑘
1 ⋱ 𝑑̃2𝑛
𝑘
(1)
⋮ ⋮ ⋱ ⋮
𝑑̃𝑛1
𝑘
𝑎
̃ … 1 )
( 𝑛2
Page 853
SEMIT 2023
The Fuzzy AHP is a set of values that incorporates the uncertainty of the decision-maker. The fuzzy
conversion scale for this method is shown in Table 1. In the technique of fuzzy prioritization, this scale
has been used. The AHP approach, based on fuzzy numbers, was used to set up the TFNs. Each expert
compares the decision criteria in pairs and assigns relative scores.
Table 1 Fuzzy function of the linguistic scale
Linguistic Scale Crisp Number TFNs Inverse of TFNs
Equally Important 1 (1, 1, 1) (1,1,1)
Moderately More Important 3 (2, 3, 4) (1/4, 1/3, 1/2)
Strongly More Important 5 (4, 5, 6) (1/6, 1/5, 1/4)
Very Strongly More Important 7 (6, 7, 8) (1/7, 1/7, 1/6)
Absolutely More Important 9 (8, 9, 9) (1/8, 1/9, 1/9)
Buckley (1985) used the geometric mean method to compute fuzzy weights and performance scores. This
approach is employed because it is simple to apply to the fuzzy case and ensures that the reciprocal
comparison matrix will have a unique answer. The weight-assessing method using the geometric mean
was chosen since it is straightforward and straightforward in the fuzzy case. Equation 2 is used to find
each row's geometric mean.
1⁄
𝑛
𝑟̃𝑖 = (∏𝑛𝑗=1 𝑑̃𝑖𝑗 ) where 𝑖 = 1, 2, … , n (2)
The fuzzy priority value's fuzzy geometric mean is calculated with normalization priorities for factors
using Equation 3.
𝑟̃ 𝑖
𝑟̃𝑖 = ∑𝑛 (3)
𝑟=1 𝑟̃𝑛
In this final stage, the weights are defuzzified.

The PROMETHEE Method
The scales and deviations of the criteria were examined by Brans, J.P., and Vincke (1985); each criterion
has a corresponding preference function. To achieve this, a preference function Pi (Aj, Ak) is created,
which shows how much alternative Aj is preferred over alternative Ak for criterion Ci.
In most practical cases, pi (Aj, Ak) is the function of the deviation d= aij-aik, i.e., pi (Aj, Ak) = pi (aij-aik),
where pi is a non-decreasing function, pi (d)=0 for d ≤ 0, and 0 ≤ pi pi (d) ≤ 1 for d > 0. It was suggested
to use a set of six common preference functions. The simplicity of these preference functions (each having
no more than two parameters with distinct economic implications) is their key selling point. Additionally,
this index accepts values between 0 and 1, which show the global strength of preference for various
options. The precedence flows defined in Equations 4 and 5 below are used to order the alternatives.
Positive Outranking flow:
1
Φ+ = n−1 ∑ π(x, A) (4)
Negative Outranking flow:
1
Φ− = n−1 ∑ π(x, A) (5)
The PROMETHEE I Partial Ranking
When Φ+ (Aj) ≥ Φ + (Ak), Φ - (Aj) ≤ Φ - (Ak),, and at least one of the inequalities holds as a strict
inequality, Aj is preferred to Ak. When Φ+ (Aj) = Φ + (Ak), Φ - (Aj) = Φ - (Ak), and otherwise, Aj and Ak
are incomparable, Aj and Ak are indifferent. Some alternative couple pairs in this partial ranking are
comparable, while others are not. Decision-making can benefit from these details in practical
applications.
The PROMETHEE II Complete Partial Ranking
The net outranking flow can be considered if the decision-maker demands an exhaustive ranking of the
alternatives while avoiding incomparability. Equation 6 provides a definition.
Page 854
SEMIT 2023
Φ (Aj) = Φ + (Aj) – Φ - (Aj). (6)

Preference evaluation types could be listed as; Type-I Usual, Type-II U-Shape, Type-III V-Shape, Type-
IV Level, Type-V Linear preference function, and Type-VI Gaussian.
IV. Application and Results

In this study ranking of nineteen Turkish Banks in terms of fourteen criteria is completed using Fuzzy
AHP and PROMETHEE Methods. The Fuzzy AHP (FAHP) is particularly interested in separation from
consistency. For this reason, the measurement of this consistency used its most expansive applications
with Multi-Criteria Decision Making (MCDM). The hieratical structure of bank selection criteria is given
in Table 2.
Table 2 Hieratical Structure of Bank Selection Criteria
Goal Main Criteria Sub Criteria
Physical Service The Number of Bank Clients
Network and The Number of ATMs
Awareness of Bank The Number of Branches of Bank
One-Year Fixed Deposit Interest Rate
One-Year Fixed Deposit Foreign Currency Interest Rate
Loan, Deposit, and
One-Year Fixed Deposit Gold Interest Rate
Bank Service Rates
New Auto Loan Monthly Interest Rate (48 Months)
Performance
Mortgage Loan Monthly Interest Rate (120 Months)
Evaluation
Annual Credit Card Fees
Commissions
Commission Charges on Payments
Cash Dividend per Shareholders of Bank
Total Assets
Bank Assets
Total Equity
Net Profit/Loss
As a result of the literature review, sectoral banking data, and meetings with experts, four main and
fourteen sub-criteria were proposed for the bank performance evaluation. The criteria weights were
scored by the FAHP method. The main criteria used in this paper are; physical service network and
awareness of bank, loan, deposit, service rates, commissions, and bank assets, respectively. Expert
questionnaires define comparisons of fuzzy pairwise matrices; each expert must appoint linguistic terms
by TFNs, given in Table 1, to the pairwise comparisons for all criteria in a hierarchy system. The results
of the comparisons are established as fuzzy pairwise comparison matrices.
Table 3 Defuzzified Global and Local Weights for Sub-criteria

Defuzzified Local Defuzzified
Criteria
Weights Global Weights
The Number of Bank Clients 0,253435 0,062298
The Number of ATMs 0,431047 0,107118
The Number of Branches of Bank 0,314429 0,078138
One-Year Fixed Deposit Interest Rate 0,345879 0,080758
One-Year Fixed Deposit Foreign Currency Interest Rate 0,171112 0,039952
One-Year Fixed Deposit Gold Interest Rate 0,190523 0,044485
New Auto Loan Monthly Interest Rate (48 Months) 0,094706 0,022113
Mortgage Loan Monthly Interest Rate (120 Months) 0,197779 0,046179
Annual Credit Card Fees 0,544690 0,086029
Commission Charges on Payments 0,455310 0,071912
Cash Dividend per Shareholders of Bank 0,265021 0,095425
Total Assets 0,194208 0,069928
Total Equity 0,169222 0,060931
Net Profit/Loss 0,371549 0,133782
Page 855
SEMIT 2023
Total Weight 1,000000
Fourteen sub-criteria weights were calculated as fuzzy numbers. In the last step of weight calculation,
the results were defuzzified and given in Table 3. These weights were used in the performance
evaluation during the Covid-19 conditions for the nineteen Turkish Banks in terms of customer
perspective by the PROMETHEE approach.
Table 4 Bank Comparison from 2022 to 2022
2020 2021 2022
Bank No Phi Rank Phi Rank Phi Rank
Bank 1 0,5930 1 0,6116 1 0,5262 1
Bank 2 0,4018 2 0,4541 2 0,4573 2
Bank 3 0,3707 3 0,3667 3 0,3149 3
Bank 4 0,266 4 0,2177 5 0,2138 7
Bank 5 0,2386 5 0,0840 6 0,2256 6
Bank 6 0,2326 6 0,2374 4 0,2289 4
Bank 7 0,0704 7 0,0809 7 0,1032 9
Bank 8 0,0608 8 0,0463 8 0,2289 5
Bank 9 -0,0123 9 0,0007 9 0,1124 8
Bank 10 -0,0297 10 -0,0434 11 -0,0926 12
Bank 11 -0,0479 11 -0,1668 13 0,0241 10
Bank 12 -0,1037 12 -0,0173 10 -0,0464 11
Bank 13 -0,1509 13 -0,0652 12 -0,1168 13
Bank 14 -0,2167 14 -0,2354 14 -0,3412 16
Bank 15 -0,2407 15 -0,2726 16 -0,2155 14
Bank 16 -0,2977 16 -0,2419 15 -0,2723 15
Bank 17 -0,3502 17 -0,3288 17 -0,3819 18
Bank 18 -0,3553 18 -0,3292 18 -0,3683 17
Bank 19 -0,4288 19 -0,3990 19 -0,4154 19
As shown in Table 4, the removal of negative superiority values from positive superiority values was
taken by finding net superiority values in PROMETHEE II.
IV. Conclusion
Performance evaluation is an essential process from the customers' perspective. As a result, this complex,
multifaceted incident may have various dimensions and even be oppositely proportional to preference.
There are many criteria to consider in this evaluation process. It is easier to choose the most discussed
criterion than it is difficult for decision-making under different criteria. Consequently, the bank ranking
from the customers' perspective should be reviewed through a rational decision process, and analytical
methods should be preferred. Various techniques have been developed in multi-Criteria decisions,
including the PROMETHEE Method. This study compares values from 2020 to 2022, three-year period
data. In the performance evaluation process, PROMETHEE Method was used for bank ranking from the
customers’ perspective. According to the application results, Bank Number 1 succeeded in first place for
three years by meeting customer needs and fulfilling customer expectations. Respectively, Bank 2 and 3
stood the same as second and third in the same period.
Page 856
SEMIT 2023
Future work for this study is to enhance the results of the PROMETHEE Method by changing the
parameters used in calculations. Especially parameter is a good candidate for future studies since it has a
prominent effect on decision-making.
References
Ayşegül Şahin, G. Cenk Akkaya (2013) PROMETHEE sıralama yöntemi ile portföy oluşturma üzerine
bir uygulama Journal of Economics and Management Research Vol: 2, No: 2.
Brans, J.P., Mareschal, B. (2005) PROMETHEE Methods, Multiple Criteria Decision Analysis, Chapter
5, 1-33.
Brans, J.P., Vincke, PH. (1985) A Preference Ranking Organization Method the PROMETHEE Method
for Multiple Criteria Decision Making. Management Science, 31(6), 647-656.
Brans, J.P. Bodily, (1998) Combining multi-criteria decision aid and system dynamics for the control of
socio-economic processes. European Journal of Operational Research, Vol. 109, pp 428-441.
Buckley, J. J. (1985) Fuzzy Hierarchical Analysis, Fuzzy Sets and Systems, 17(3), 233–247.
Blattberg (2001) Want to build your business? Grow your customer equity. Harvard Management Update,
pp. 4-.
Bodily SE (1985) Modern decision making: a guide to modeling with decision support systems. McGraw-
Hill Book Company, New York.
Constanta (2005) Resources Planning using the AHP and PROMETHEE Multi-criteria methods The 7th
Balkan Conference on Operational Research in May 2005.
Cem Kocak (2013) First Order Arma type fuzzy time series method based on logic relation tables,
Mathematical Problems in Engineering, vol. 2013, Article ID 769125, 12 pages, 201.
Das, T. K., Teng, B. (1999) Cognitive Biases and Strategic Decision Processes: An Integrative
Perspective. Journal of Management Studies, 36: 757-778.
Engel, J., Blackwell, R. D., and Miniard, (1986) Consumer Psychology in Behavioral Perspective, 5th
ed., Dryden.
Figuera J., Greco, S., Ehrgott M., (2005) Multiple Criteria Decision Analysis: State of the Art Surveys,
Springer, and New York, NY.
Feinberg, R.A. (1986), Credit Cards as Spending Facilitating Stimuli: A Conditioning Interpretation,
Journal of Consumer Research, 13, December, pp. 348-356.
Harry Barton (2009) A PROMETHEE based uncertainty analysis of UK police force performance rank
improvement international police executive symposium Geneva Centre for the Democratic Control of
Armed Forces working paper no: 14.
Hawkins, D. I., Mothersbaugh, D. L. (2010) Customer Behavior: Building marketing strategy. Boston:
McGraw-Hill Irwin.
Huang I.E., Keisler J., Linkov J. (2011) Multi-criteria decision analysis in environmental sciences: Ten
years of applications and trends. Science of the Total Environment, Vol. 409. Issue 19. Pages 3578–3594.
The United States.
Hung-Yi Wu, Gwo-Hshiung Tzeng, Yi-Hsuan Chen (2009) A Fuzzy MCDM approach for evaluating
banking performance based on Balanced Scorecard Expert Systems with Applications 36 (2009) 10135–
10147.
Jankowski P., (1995) Integrating geographical information systems and multiple criteria decision-making
methods. Int J Geogr Inf Syst 9(3):251–273.
Keyser, W.D., Peeters P. (1996) A note on the use of PROMETHEE Multi-Criteria Methods, European
Journal of Operational Research, 89, 457-461.
Page 857
SEMIT 2023
Keeney, R.., H. Raffia. (1976) Decisions with Multiple Objectives: Preferences and Value Tradeoffs.
Wiley, New York.
Koutroumanidis, J. Papathanassiou, B. Manos (2011) Multicriteria Analysis of Productivity of
Agricultural Regions of Greece Operational Research. An International Journal. Vol.2, No.3 (2002),
pp.339-346 339.
Lewis (1995) Religiosity and Happiness: No Evidence for an Association among Undergraduates.
Personality and Individual Differences, 1997, Vol. 22. 119 – 121.
Masoud Goli, Shahab Bayati (2009) Selecting Suitable Semantic Web Service Composition, Using
PROMETHEE Method and Non-Functional Parameters International Conference on Computer
Engineering and Applications IPCSIT vol.2(2011) IACSIT Press, Singapore.
Marsden, D. and Littler, D. (1998), Positioning Alternative Perspectives of Customer Behavior, Journal
of Marketing Management, 14, pp. 3-28.
Mei Fang Chen, Tzungi Tang (2005) A Fuzzy MCDM Approach for Evaluation of Expatriate
Assignments International Journal of Information Technology & Decision Making Vol. 4, No. 2 (2005)
277–296.
Mitroff, I., Linstone, H. (1993): The Unbounded Mind, Oxford University Press.
Podvezko, V., Podviezko, A. (2010) PROMETHEE I Method Application for identification of the best
alternative, Verslas: teorija ir praktika business theory and practice 10(2): 84–92 in Lithuanian
doi:10.3846/1648-0627.2009.10.84-92.
Prakash, T. N. (2003) Land Suitability Analysis for Agricultural Crops: A Fuzzy Multi-Criteria Decision-
Making Approach. MS Thesis, ITC, Institute.
Ramanathan R. (2001) a note on the use of the analytic hierarchy process for environmental impact
assessment. Journal of Environmental Management, Volume 63, Issue 1. Pages 27-35. India.
Saaty, T. L. (1980) the Analytic Hierarchy Process. New York: McGraw-Hill.
Saaty, T. L. (1996) Fundamentals of decision making and priority theory with the Analytic Hierarchy
Process. Pittsburgh, PA: RWS Publications.
Shankar (2002) A developer Toolkit for web service composition in proceedings of the Eleventh
International World Wide Web Conference, 2002.
Slavica Prvulovic, Dragisa Tolmac (2008) Multi-Criteria Decision in the choice of advertising tools.
Mechanical Engineering Vol. 6, No 1, 2008, pp. 91 – 100.
Steuer, R. E. (1986) Multiple Criteria Optimization: Theory, Computation and Application, Wiley Series
in Probability and Mathematical Statistics, John Wiley, New York, 546 pp.
Solomon M.R. (1995) the world of products as social stimuli: A Symbolic Interactions Perspective,
Journal of Customer Research. 10, 319-329.
Tolga Kaya, Cengiz Kahraman (2011) A Fuzzy Approach to e-banking website quality assessment based
on an integrated AHP Electre Method Technological and
Economic Development of Economy 17:2, DOI: 10.3846/20294913.2011.583727.
Triantaphyllou, E., A. Sanchez, (1997), A Sensitivity Analysis Approach for Some Deterministic Multi-
Criteria Decision-Making Methods. Decision Sciences, Vol. 28, No. 1, pp. 151-194.
Triantaphyllou, E., Mann, S.H. (1995) Using the Analytic Hierarchy Process for decision making in
engineering applications, some challenges, Journal of Industrial Engineering Applications and Practice,
Vol. 2, pp. 35-44.
Wang Y., Chi R., Yang Y. (2004): An integrated framework for customer perspective and customer
relationship management performance: a customer-based perspective from China. Managing Service
Quality, Vol. 14, No. 2/3, pp. 169-182.
Page 858
SEMIT 2023
Westing, H.J., Albaum, G. (1975) Modern Marketing Thought, 3rd Ed., Collier Macmillan Publishers.
Page 859
SEMIT 2023
Supplier Selection in the Context of Industry 4.0 Using Hybrid DEA-SMART Method
Parmida Bahreini1, Babek Erdebilli2
Abstract
In the era of Industry 4.0, choosing suppliers for online commerce is of utmost importance and calls for
the application of efficient, data-centric techniques. Businesses are under increasing pressure to improve
their supply chain management strategies and select the best suppliers in the online commerce
environment of Industry 4.0. Traditional approaches, however, sometimes don't include a thorough
assessment of suppliers across several dimensions. In order to close this gap, this paper proposes a new
approach that combines Data Envelopment Analysis (DEA) with the Simple Multi-Attribute Rating
Technique (SMART).The first phase is applying DEA to determine how effective suppliers are using the
data that has been gathered. DEA offers a quantitative indicator of how efficiently providers convert their
inputs into outputs. This combination score enables rating suppliers while simultaneously considering
multi-attribute evaluation and quantifying efficiency assessment, then using a number of different criteria,
providers are evaluated using the SMART approach. The findings of this analysis help to improve
supplier selection procedures in the context of online commerce, which falls under the purview of
Industry 4.0.
Keywords: Supplier selection, Industry 4.0, Data Envelopment Analysis, Simple Multi-Attribute Rating
Technique
1. Introduction
Industry, a part of an economy, produces highly automated and mechanized materials. Since the
beginning of industrial development, technological developments have led to changes in attitudes that are
currently referred to as "industrial revolutions": mechanization (known as the first industrial revolution),
a heavy dependence on electrical energy (known as the second industrial revolution), and the wide
adoption of digitalization (referred to as the third industrial revolution). A new fundamental paradigm
change in industrial production appears to be brought about by the combination of Internet technologies
and future-oriented technologies in the field of "smart" objects (machines and goods), which is based on
a highly advanced digitalization within factories [1]. The fourth industrial revolution, often known as
Industry 4.0, and the digitization of business are taking place as the 21st century gets underway [2]. The
implementation of technological advances is essential to Industry 4.0 because they allow for the real-time
collection and analysis of data, which gives the industrial system vital information. This was made
possible by the development of the Internet of Things (IoT), cloud services, big data analytics, and the
cyber-physical system concept of Industry 4.0 [3]. The foundation for Industry 4.0 supply chains (SCs)
has been laid by developments in communication and information technology, which provide multiple
opportunities for supply chain intelligence and autonomy. The earlier research has given considerable
emphasis to the assessment and choice of sustainable suppliers as a crucial SC decision. This process has
not yet been accomplished by industry 4.0 SCs, where decentralized of supply chain participants and real-
time information transparency, technical assistance, and connectivity are believed to be the core
architectural elements. Actions for supplier evaluation and selection have an impact on almost every
choice that needs to be made in the management of supply chains and networks. Although researchers
and business leaders have focused a lot of attention on the significance of sustainability and
environmental issues and their incorporation in SCs, there haven't been many practical attempts to
incorporate sustainability concerns within the framework of Industry 4.0 SCs, and more primarily
sustainable supplier evaluation and selection within this context. To clarify the research directions in this
field, further research efforts are needed [4]. The manufacturing industries, in specific, are changing to
prepare for the fourth Industrial Revolution, or Industry 4.0, in today's fiercely competitive and
innovation-driven corporate climate. Supply chain strategy is vital for achieving an edge over the
competition, and it does offer operational and strategic advantages to firms, regions, and nations,
Page 860
SEMIT 2023
according to business leaders, academics, and policymakers. Businesses today must optimize their
company operations and support the performance of their whole supply chain to compete in the market.
Different factors should be defined and assessed considering various providers' features to choose the
best source. As a result, this issue can be categorized as a multi-criteria decision-making problem [5]. E-
commerce has flourished since the introduction of the Internet. Because client needs are changing so
quickly in the e-commerce era, the complexity and ease of dealing with procurement have become
increasingly noticeable. There is consensus in the pertinent literature that choosing the right suppliers is
a challenging task due to the sheer volume of criteria that must be considered. Therefore, there is an
urgent need for a more structured and open approach to making purchasing decisions, particularly in
supplier selection. [6]. The supplier selection problem (SSP), which involves many criteria for evaluation,
is fundamentally a multi-criteria decision-making problem. The Simple Multi Attribute Rating Technique
(SMART) is one of the multi-criteria decision-making approaches that is particularly well suited for
modeling quantitative criteria. It has found widespread application in a variety of fields, including
selection, evaluation, planning and development, decision making forecasting, and other areas.
The Simple Multi Attribute Rating Technique (SMART) is a method of multiple criteria decision making
developed by Edwards in 1971. It is a compensatory method that aims to provide an easy way to
implement the basic principles of Multi-Attribute Utility Theory (MAUT) [7]. The Simple Multi Attribute
Rating Technique (SMART) is a popular multi-attribute decision-making (MADM) technique used to
evaluate options based on multiple attributes or criteria. In this method, the decision-maker (DM)
identifies the relevant attributes, assigns weights to each attribute based on their relative importance,
scores each alternative on each attribute, and then chooses the alternative with the highest total score as
the preferred option. The Simple Multi Attribute Rating Technique (SMART) employs a linear additive
model, where the total value of an option is determined by the sum of the weighted performance scores
for each criterion. This implies that the value of each attribute is multiplied by its respective weight and
then added up to calculate the overall value of the alternative [8]. The SMART uses weights assigned to
each attribute by the decision-maker to determine the hierarchy. In other words, the decision maker
assigns weights to each criterion based on their individual judgment [9]. Despite the large number of
providers, the selection procedure is very difficult because many of them do not satisfy all the
requirements established by the business. Based on this, a determining model that aids in simplifying and
improving the selection process is needed. To provide a proper assessment procedure, this research
concentrates on the many criteria obtained from earlier studies. Additionally, this study used both the
Simple Multi Attribute Rating Technique (SMART) and Data Envelopment Analysis
(DEA) methodologies for its computations. This study uses the DEA Method to assess the effectiveness
of suppliers. The efficiency of the provider's input variables is considered, and the best supplier is chosen
using hybrid DEA-SMART approach. The effectiveness of our method is proven by offering a numerical
example of a manufacturing business that identifies and selects suppliers in the context of an electronic
marketplace and provides decision support services on this basis.
Industry 4.0 has changed how businesses produce, enhance, and distribute their goods. The Internet of
Things (IoT), cloud computing, analytics, AI, and machine learning are among the cutting-edge
technologies that manufacturers are incorporating into their manufacturing processes. These "smart"
businesses have cutting-edge gauges, software that is embedded, and robotics that gather data, analyze it,
and help with decision-making. When operational data from ERP, supply chain, customer service, and
other corporate systems is linked with data from production processes, even greater value from previously
segregated information is produced. Javaid, M. et al. (2022) figure out the barriers and solutions to the
use of current technology in production processes to line with Industry 4.0 objectives. 22 barriers and 14
solutions were identified by the researchers using a thorough literature analysis. These were then ranked
using the BWM and combined compromise solution (CoCoSo) approaches. The paper suggests a
framework to assist managers in manufacturing firms in comprehending the challenges and strategically
implementing the solutions to convert their systems into Industry 4.0-based systems [10]. Also, Kumar,
R et al. (2016) intend to suggest an effective technique for choosing a cloud service provider that satisfies
the user's Quality of Service (QoS) needs in the context of Industry 4.0. To deal with the complexity and
Page 861
SEMIT 2023
ambiguity involved in the selection process, the study uses a fuzzy-based Analytic Hierarchy Process
(fuzzy-AHP) as a decision-making tool [11]. Meanwhile, Trung, N. Q. et al. (2022) intend to suggest a
fuzzy linguistic MCDM model for the assessment and choice of industry 4.0-compliant digital marketing
technologies. The methodology utilized in this research entails identifying the evaluation criteria and
alternatives and employing two MCDM methods, SF-AHP and TOPSIS, to support the decision-making
process [12]. Besides, Naveed, Q. N. et al. (2021) attempt to define the critical success factors (CSFs) of
Cloud Enterprise Resource Planning (CERP), a key technology for reaching Industry 4.0. The study offers
significant knowledge into how businesses may effectively execute CERP as part of their Industry 4.0
strategy using the Analytic Hierarchy Process (AHP) and Fuzzy Analytic Hierarchy Process (FAHP)
approaches to figure out the ranking of the CSFs accountable for the achievement of the information
management system (IMS) [13]. Likewise, In the framework of Industry 4.0, Medi, N. et al. (2018) seek
to find the most suitable innovations in organization for manufacturing firms in developing nations. The
researchers utilized the European Manufacturing Survey data to determine the weights of the criterion
and the Fuzzy Analytic Hierarchy Process (FAHP) to rank the organizational innovations. Additionally,
they ranked the organizational innovations using the Preference Ranking Organization Method for
Enrichment Evaluations (PROMETHEE) method [14]. Moreover, Kumar, V., et al. (2021) use a hybrid
MCDM technique to investigate strategies for overcoming the obstacles to Industry 4.0. The SWARA-
WASPAS method was adjusted for the study's evaluation of Industry 4.0's obstacles and prioritization of
solutions for overcoming them [15].
Besides, Erdogan, M. et al. (2018) want to present an extensive strategy for choosing the most suitable
Industry 4.0 implementation plan, which can assist firms in improving their effectiveness and
competitiveness in today's competitive business environment. The methods used includes a methodical
approach to identifying different Industry 4.0 implementation strategies, laying out criteria for choosing
the best one, and using multi-criteria decision-making (MCDM) methods based on AHP-VIKOR
methodologies to evaluate and contrast various implementation strategies. To deal with uncertainty in the
selection process, fuzzy set theory is often used [16]. Likewise, A. Jamwal et al. (2021) propose to provide
a framework for sustainability in Industry 4.0 for (Micro Small Medium enterprises) sector (MSMEs) in
India. For the construction of the framework, the study employs a hybrid MCDM methodology based on
the F-AHP and DEMATEL techniques. The enablers of each enabler group are compared pairwise using
F-AHP, and the relationships between the enablers are determined using DEMATEL [17]. In addition,
Raj, A., et al. (2020) seek to identify and study the obstacles to the deployment of Industry 4.0 in the
manufacturing sector in various economic scenarios. The study makes use of a suggested approach that
has four parts, including the identification of obstacles, the examination of how obstacles relate to one
another, expert confirmation of the findings, and sensitivity analysis. The Grey-DEMATEL approach is
used in the study to examine the connections between the obstacles [18].
Medić, N. et al. (2018) explore the deployment of advanced digital technologies in manufacturing firms
and assess their impact in the context of Industry 4.0 in transition nations (i.e., Slovenia, Croatia, and
Serbia). The approaches utilized in this study include Fuzzy Analytic Hierarchy Process (FAHP) to
generate criteria weights and Preference Ranking Organization Method for Enrichment Evaluations
(PROMETHEE) to rank sophisticated digital technologies [19]. Additionally, Ucal Sari, I., & Ak, U.
(2022) use fuzzy data envelopment analysis to assess machine efficiency in Industry 4.0. The Data
Envelopment Analysis (DEA), Cronbach Alpha reliability analysis, and traditional DEA methods were
used in this study's approach. The study also employs the Interval DEA (IDEA) technique as the way to
measure efficiency [20]. In the era of Industry 4.0, Pishdar, M. et al. (2021) evaluate the sustainability
performance of third-party logistics service providers (3PLs) that are contemplating circular economy
methods. The study compares the sustainability performance of 17 3PLs in terms of various factors using
an interval type-2 fuzzy super-slack-based measure network DEA technique [21]. Similarly, Azadi, M.
et al. (2021) investigate a novel DEA model to assess the viability of CSPs (Cloud Service Providers) in
the context of Industry 4.0. Data Envelopment Analysis (DEA), a nonparametric methodology that
considers numerous inputs and multiple outputs, was utilized in this work to analyze the performance of
a group of peer entities [22]. Also, Azadi, M. et al. (2021) compare internal finance to external funding
and examine the significance of financing for manufacturers investing in Industry 4.0 technology. The
researchers provide a novel data envelopment analysis (DEA) model that considers economic,
environmental, and social elements to analyze the sustainability of financial resources [23]. Also, Arora,
Page 862
SEMIT 2023
C. et al. (2022) evaluate, and rate Industry 4.0 technologies based on their relative efficiencies in dealing
with hurdles in the agricultural supply chain. Analytic Hierarchy Process (AHP) and Data Envelopment
Analysis (DEA), two multi-criteria decision-making methodologies, are combined in the AHP-DEA
framework used in this study. While DEA is used to assess the relative effectiveness of the alternatives,
AHP is used to create a hierarchy of pair-wise comparison and resolve conflicts while achieving the
desired aim [24].
Table 30. The studies examined in the literature

Authors Methodology Purpose
Evaluating the barriers and solutions to the use of current
Javaid, M. et al. BWM &
technology in production processes to line with Industry 4.0
(2022) CoCoSo
objectives.
Suggesting an effective technique for choosing a cloud service
Kumar, R et al.
Fuzzy - AHP provider that satisfies the user's Quality of Service needs in the
(2016)
context of Industry 4.0.
Trung, N. Q. et al. SF-AHP & Providing a fuzzy linguistic MCDM model for the assessment and
(2022) TOPSIS choice of industry 4.0-compliant digital marketing technologies.
Table 31 (Cont.) The studies examined in the literature

Authors Methodology Purpose
Attempting to define the critical success factors of Cloud
Naveed, Q. N. et
AHP & FAHP Enterprise Resource Planning, a key technology for reaching
al. (2021)
Industry 4.0.
Medi, N. et al. FAHP & seek to find the most suitable innovations in organization for
(2018) PROMETHEE manufacturing firms in developing nations.
Kumar, V., et al. SWARA- Investigating the strategies for overcoming the obstacles to
(2021) WASPAS Industry 4.0.
Presenting an extensive strategy for choosing the most suitable
Erdogan, M. et al. Industry 4.0 implementation plan, which can assist firms in
AHP-VIKOR
(2018) improving their effectiveness and competitiveness in today's
competitive business environment.
Jamwal et al. FAHP- Propose to provide a framework for sustainability in Industry 4.0
(2021) DAMATEL for Micro Small Medium enterprises sector in India.
Seek to identify and study the obstacles to the deployment of
Raj, A., et al. Grey-
Industry 4.0 in the manufacturing sector in various economic
(2020) DAMATEL
scenarios.
Ucal Sari, I., & Use fuzzy data envelopment analysis to assess machine efficiency
DEA
Ak, U. (2022) in Industry 4.0.
Evaluating the sustainability performance of third-party logistics
Pishdar, M. et al.
DEA service providers (3PLs) that are contemplating circular economy
(2021)
methods in the context of industry 4.0.
Azadi, M. et al. investigate a novel DEA model to assess the viability of Cloud
DEA
(2021) Service Providers in the context of Industry 4.0.
Compare internal finance to external funding and examine the
Azadi, M. et al.
DEA significance of financing for manufacturers investing in Industry
(2021)
4.0 technology.
Evaluate and rate Industry 4.0 technologies based on their relative
Arora, C. et al.
AHP-DEA efficiencies in dealing with hurdles in the agricultural supply
(2022)
chain.
The literature study demonstrates a deficit in our knowledge of supplier selection in the context of
Industry 4.0, particularly in e-commerce. There is a lack of precise advice on choosing suppliers who are
prepared for this period of digital transformation, even while current research provides insightful analyses
of Industry 4.0's larger landscape. This study fills this knowledge void by emphasizing industry 4.0's
Page 863
SEMIT 2023
supplier selection techniques, notably in e-commerce. Our research intends to offer helpful tips for supply
chain experts, stakeholders, and e-commerce companies navigating this challenging market.
3. Methodology
3.1 Simple Multi Attribute Rating Technique (SMART)
The Simple Multi-Attribute Rating Technique (SMART) is a method of multiple criteria decision making.
The SMART uses weights assigned to each attribute by the decision-maker to determine the hierarchy.
In other words, the decision maker assigns weights to each criterion based on their individual judgment.
The weights reflect the relative importance of each attribute in the decision-making process. The
alternatives are then evaluated on each attribute using a predetermined scale, and the scores are combined
using the assigned weights to determine the overall value of each alternative. By using this approach,
SMART simplifies the decision-making process by breaking down the problem into smaller, more
manageable pieces.
• Process involved in SMART:
STEP 1: Define the decision problem and identify the criteria that are relevant to the decision.
STEP 2: Assign weights to each criterion by allocating values in the range of 1 to 100 for each criterion
based on their level of importance.
STEP 3: Normalize each criterion by dividing its weight by the sum of weights of all criteria using the
given formula.
𝒘𝒋
Normalization =
∑𝒘𝒋
• Where wj is weight of criteria and ∑ wj is total wight for all criteria.

STEP 4: Find the utility value for converting the criterion value of each attribute to the raw data value
by utilizing the formula provided:
𝐶𝑜𝑢𝑡− 𝐶𝑚𝑖𝑛
Ui (ai) =
𝐶max − 𝐶𝑚𝑖𝑛
• Where Ui (ai) is utility value of first criteria to the last criteria, Cmax is value of maximum criteria
and Cmin is value of minimum criteria, Cout is value criterion to the last criteria.
STEP 5: Calculating the final value of each criterion by combining the normalized value of the raw data
criteria with the weight-normalized value criteria, followed by multiplying them using the formula below:
U(ai) = ∑𝑚
𝑗=1 𝑤𝑗 𝑢𝑖 (𝑎𝑖 ), i= 1, 2,…,m
3.2 Data Envelopment Analysis (DEA)

DEA is a non-parametric method used to evaluate the relative efficiency of decision-making units
(DMUs) based on their inputs and outputs. The main idea behind DEA is to compare the performance of
DMUs that have similar inputs and outputs but operate under different conditions or constraints. DMUs
can be any entity that consumes resources (inputs) to produce goods or services (outputs), such as
hospitals, banks, schools, or manufacturing plants. DEA evaluates the efficiency of DMUs by comparing
their input-output relationships to the best performing DMU (the "efficient frontier"). The efficient
frontier is a set of DMUs that represent the best combination of inputs and outputs, given the constraints
or conditions faced by each DMU. DEA assumes that each DMU has a unique set of inputs and outputs,
and that the performance of each DMU can be represented by a set of input-output ratios. So, DEA then
determines the relative efficiency of each DMU by comparing its input-output ratios to those of the
efficient frontier. The DMUs that have a ratio of 1 are referred to as efficient and the units that have a
ratio less than 1 are less efficient compared to the most efficient units.
Page 864
SEMIT 2023
• Process involved in DEA:

STEP 1: Choose the decision-making units (DMUs) that desire to compute efficiency scores for.
STEP 2: Determining the data inputs and outputs for DMUs.
STEP 3: After determining inputs and outputs, calculate DMUs efficiency score using DEA formula
provided below:
𝑝
Ek= Max ∑𝑟=1 𝑢r yrj
s.t
∑𝑧𝑖=1 𝑣ixrj = 1
∑𝑝𝑟=1 𝑢ryrj – ∑𝑧𝑖=1 𝑣ixrj ≤ 0
ur ≥ 0, r=1,…,p
vi ≥ 0, i=1,…,z
• which Ek is efficiency,score, yrj is,output r for,DMUj , xrj is,input i for,DMUj , ur is,weight devoted
to output r, vi is,weight devoted to input i, n is # of DMUs, p is # of,outputs and z is # of,inputs.
STEP 4: After obtaining formula and implementing it for each of DMU’s, using LINGO software obtain
efficiency score.
STEP 5: Rank the DMUs based on efficiency scores.
3.3 Hybrid DEA-SMART Approach for Suppliers Selection
Step 1: Choose the,decision-making units (DMUs) that desire to compute efficiency scores for.
Step 2: Determining the data inputs and,outputs for DMUs.
Step 3: After determining inputs and outputs, Calculate DMU’s efficiency score using DEA
mathematical formula. After obtaining formula and implementing it for each of DMU’s, using LINGO
software obtain efficiency score.
𝑝
Ek= Max ∑𝑟=1 𝑢r yrj
s.t
∑𝑧𝑖=1 𝑣ixrj = 1
∑𝑝𝑟=1 𝑢ryrj – ∑𝑧𝑖=1 𝑣ixrj ≤ 0 (3)
ur ≥ 0, r=1,…,p
vi ≥ 0, i=1,…,z
Step 4: To make a hybrid model, Assess the criteria for DMU’s (attributes) , attributes in rows as Ai
and criteria in columns as Cj to evaluate decision matrix.
Ai (i=1,2,…,z)
Cj (j=1,2,..,p)
𝑦11 ⋯ 𝑦1𝑝
Y= [ ⋮ ⋱ ⋮ ]
𝑦𝑧1 ⋯ 𝑦𝑧𝑝
Step 5: Using SMART approach, find the raw data.
𝐶𝑜𝑢𝑡− 𝐶𝑚𝑖𝑛
Ui (ai) = 𝐶 (3.1)
max − 𝐶𝑚𝑖𝑛
Step 6: assign weights to each criterion by allocating values in the range of 1 to 100 for each criterion.
Page 865
SEMIT 2023
Step 7: Normalize each criterion weight by dividing its weight by the sum of weights of all criteria using
the given formula:
𝑤𝑗
Normalization = (3.2)
∑𝑤𝑗
Step 8: Calculating the final weighted value of each criterion by combining the raw value,of criteria with
the,weight-normalized value criteria, followed by multiplying them using the formula below:
U(ai) = ∑𝑚
𝑗=1 𝑤𝑗 𝑢𝑖 (𝑎𝑖 ), i= 1, 2,…,m (3.3)
Step 9: Normalizing the final weighted value.
𝑤𝑗
U(ai)’ = ∑𝑤𝑗
(3.4)
Step 10: To evaluate a hybrid model, multiplying the,efficiency scores of all DMU’s with the final
weighted value of the criteria, a hybridized efficiency score is produced.
Ek’ = Ek × U(ai)’ (3.5)
Step 11: Rank the DMU’s based on their hybridized,efficiency score.
3.4 Case Study
This section provides supplier selection in an B2B e-commerce setting in the context of industry 4.0. The
study was carried out using 4 criteria and 4 decision-making units, as shown in Fig.2. For businesses
involved in B2B manufacturing e-commerce, selecting suppliers is a crucial and difficult issue. In this
study, we provide an integrated data envelopment analysis and a simple multi-attribute ranking
technique for the selection of suppliers based on several criteria. The efficiency score of each supplier is
determined using the data envelopment analysis (DEA) in the proposed DEA-SMART approach, and the
weights of each criterion are then computed using Simple Multi Attribute Rating Technique (SMART).
A numerical example illustrates how to use the suggested way to compare the ranking order changes
between the DEA method and the hybrid DEA-SMART method.
Figure 42. Alternatives and criteria’s hierarchical chart

A brief description and clarification of all criteria are provided in Table 3.
Page 866
SEMIT 2023
Table 32. Description of criteria

Criteria Description
Price Monetary value or cost assigned to a product or service.
The actual distance that needs to be covered for goods or services to be transported
Distance
from the supplier's location to the company's location.
The time taken from the placement of the order to the receipt of the products or
Delivery time
completion of the requested services.
The level of excellence or how well the supplier's offerings meet the company's
Quality
requirements or standards.
The following data is acquired through investigation and with the support of findings from additional
pertinent studies, as indicated in the study, and is based on specified criteria and Decision-Making Units
(DMUs) as detailed in Table 4.
Table 33. Collected data according to criteria and DMU's
DMU’s Input Variable Output Variable
Suppliers Price (I1) Distance (I2) Delivery Time (O1) Quality (O2)
S1 53 39 0.84 0.92
S2 56 35 0.68 0.85
S3 46 47 0.93 0.74
S4 49 57 0.83 0.89
After collecting data, the mathematical model using the data in Table 4 was formulated, then using
LINGO software efficiency score of each formulated mathematical model was calculated using equation
(3) for each supplier. Efficiency score calculation result using Data Envelopment Analysis (DEA) method
is shown in Table 5.
E1 = Max 0.84u1 + 0.92u2 E2 = Max 0.68u1 + 0.85u2

s.t. s.t.
53v1 + 39v2 =1 56v1 + 35v2 =1
0.84u1 + 0.92u2 - 53v1 - 39v2 <= 0 0.84u1 + 0.92u2 - 53v1 - 39v2 <= 0
0.68u1 + 0.85u2 - 56v1 - 35v2 <= 0 0.68u1 + 0.85u2 - 56v1 - 35v2 <= 0
0.93u1 + 0.74u2 - 46v1 - 47v2 <= 0 0.93u1 + 0.74u2 - 46v1 - 47v2 <= 0
0.83u1 + 0.89u2 - 49v1 - 57v2 <= 0 0.83u1 + 0.89u2 - 49v1 - 57v2 <= 0
U1 >= 0 U1 >= 0
U2 >= 0 U2 >= 0
V1 >= 0 V1 >= 0
V2 >= 0 V2 >= 0
E3 = Max 0.93u1 + 0.74u2 E4 = Max 0.83u1 + 0.89u2

s.t. s.t.
46v1 + 47v2 =1 49v1 + 57v2 =1
0.84u1 + 0.92u2 - 53v1 - 39v2 <= 0 0.84u1 + 0.92u2 - 53v1 - 39v2 <= 0
0.68u1 + 0.85u2 - 56v1 - 35v2 <= 0 0.68u1 + 0.85u2 - 56v1 - 35v2 <= 0
0.93u1 + 0.74u2 - 46v1 - 47v2 <= 0 0.93u1 + 0.74u2 - 46v1 - 47v2 <= 0
0.83u1 + 0.89u2 - 49v1 - 57v2 <= 0 0.83u1 + 0.89u2 - 49v1 - 57v2 <= 0
U1 >= 0 U1 >= 0
U2 >= 0 U2 >= 0
V1 >= 0 V1 >= 0
V2 >= 0 V2 >= 0
Page 867
SEMIT 2023
Following the formulation of the mathematical model using Equation (3), the subsequent equations,
denoted as E1, E2, E3, and E4, were systematically solved using Lingo software. This comprehensive
analysis resulted in the determination of the efficiency scores for each equation, as presented in Table 5.
Table 34. Evaluation of efficiency score

Delivery Time Efficiency
Suppliers Price (I1) Distance (I2) Quality (O2)
(O1) Score
S1 53 39 0.84 0.92 1.00
S2 56 35 0.68 0.85 1.00
S3 46 47 0.93 0.74 1.00
S4 49 57 0.83 0.89 1.00
Following the calculation of efficiency scores, the SMART approach is used to find the best provider,
assuming that all suppliers obtained a computed score of 1.00. The first critical step in starting this process
is standardizing the data, which is currently available at multiple scales. This standardization technique
starts with determining the lowest and highest values for each criterion in the dataset. Following that, we
calculate the raw data as shown in Table 6 for each provider and criterion using Equation (3.1).
Table 35. Evaluation of raw data
Suppliers Price (I1) Distance (I2) Delivery Time Quality (O2)
(O1)
S1 53 39 0.84 0.92
S2 56 35 0.68 0.85
S3 46 47 0.93 0.74
S4 49 57 0.83 0.89
Min Value 46 35 0.68 0.74
Max Value 56 57 0.93 0.92
Max-Min Value 10 22 0.25 0.18
S1 0.70 0.18 0.64 1
S2 1.00 0 0.00 0.61
S3 0 0.55 1.00 0
S4 0.3 1.00 0.60 0.83
As shown in Table 7, allocating weights to each criterion is performed by assigning values ranging from
1 to 100 using the Simple Multi Attribute Rating Technique. Furthermore, Equation (3.2) is used to
achieve weight normalization. It is vital to highlight that the decision maker's judgment is used to
determine the weight allocated to each criterion.
Table 36. Weight and normalized weight of each criterion
Criteria Weight Normalized-Weight
C1 40 0.4
C2 20 0.2
C3 10 0.1
C4 30 0.3
Sum 100
The eventual weighted value for each criterion is calculated by multiplying the raw value of the criteria
by their weighted values. The values derived from equations (3.3) and (3.4) are then computed to produce
the final scores. Table 8 displays the total cumulative values of the criteria as calculated by the Simple
Multi Attribute Rating Technique.
Table 37. Evaluation of final score using SMART method
DMU’s Criteria Criteria Criteria 3 Criteria 4 Score Total Score
1 2
Page 868
SEMIT 2023
S1 0.70 0.18 0.64 1 0.68 68.04

S2 1.00 0 0.00 0.61 0.58 58.33
S3 0 0.55 1.00 0 0.21 20.91
S4 0.3 1.00 0.60 0.83 0.63 63.00
Normalized
0.4 0.2 0.1 0.3
weight
In the ultimate phase, a hybrid score known as 'Hybrid DEA-SMART,' calculated using Equation (3.5),
was constructed to determine the ranking of suppliers and identify the best performance. This hybrid
score is obtained by multiplying each supplier's efficiency score by the overall score acquired using the
Simple Multi Attribute Rating Technique. Table 9 displays the resulting hybrid scores.
Table 38. Evaluation of hybridized score
Hybridized
DMU’s Criteria 1 Criteria 2 Criteria 3 Criteria 4 Efficiency
Score
S1 0.70 0.18 0.64 1 15.65
S2 1.00 0 0.00 0.61 14.58
S3 0 0.55 1.00 0 5.23
S4 0.3 1.00 0.60 0.83 15.75
In the course of this study, our first DEA evaluation resulted in an efficiency score of 1.00 for each of the
providers, which, while indicative of a high degree of efficiency, presented a hurdle in identifying the top
supplier based purely on these uniform results. We used a unique hybrid approach in our research since
we were aware of this restriction. The main goal was to set up a ranking system based on hybrid efficiency
ratings and to distinguish between the providers. Upon applying this hybrid algorithm, the ranking of the
suppliers emerged as follows: with a hybrid efficiency value of 15.75, Supplier 4 took the top spot, closely
followed by Supplier 1, who scored 15.65. Supplier 2 achieved the third spot with a respectable hybrid
efficiency score of 14.58. Lastly, among the efficient providers, Supplier 3 came in last with a score of
5.23. This ranking not only enables a more detailed evaluation of supplier performance, but it also offers
important insights into the relative degrees of efficiency of each provider within the framework of our
research. Table 10 shows the ranking of suppliers according to their hybridized efficiency score.
Table 39. Suppliers final ranking
Ranking DMU’s Hybridized Efficiency Score
1 Supplier 4 15.75
2 Supplier 1 15.65
3 Supplier 2 14.58
4 Supplier 3 5.23
4. Conclusion
In summary, when we apply hybrid DEA-SMART model, applying Simple Multi Attribute Rating
Technique gives weight to each criterion and this results in changing the order of ranking. DEA-SMART
supplier selection in B2B e-commerce within the context of Industry 4.0 delivers considerable benefits
and boosts corporate productivity and competitiveness. Data Envelopment Analysis (DEA) and the
Simple Multi-Attribute Rating Technique (SMART) strengths are combined in this hybrid technique to
give a thorough and quantitative evaluation of providers based on several criteria. Businesses may assess
supplier effectiveness using the DEA-SMART approach, compare performance, and give priority to
suppliers that provide the highest value. It makes data-driven decision-making easier, lessens personal
biases, and makes supplier selection procedures more objective. Businesses may take advantage of
Industry 4.0 to improve their procurement procedures, optimize supply chain operations, and boost their
operational effectiveness by utilizing digital technology and data analytics. Businesses are better
equipped to choose and work with the most effective and dependable suppliers in the B2B e-commerce
market when they implement the DEA-SMART approach for supplier selection. Consequently, there is
Page 869
SEMIT 2023
an increase in customer happiness, timely delivery, cost optimization, and product quality. The technique
also aids in the supply chain's entire digital transformation, which is in line with Industry 4.0's objectives
and tenets. The DEA-SMART approach offers a useful framework for firms to adapt and succeed in the
rapidly evolving electronic marketplace as B2B e-commerce and Industry 4.0 technologies develop.
Organizations may arrive at knowledgeable supplier selection decisions that have an advantageous effect
on their operational efficiency, competitiveness, and overall success in the B2B e-commerce landscape
of Industry 4.0 by utilizing the strength of quantitative analysis and integrating it with the thorough
evaluation of numerous attributes. This methodology provides a trustworthy outcome and may be
expanded for similar businesses. Using knowledge-driven approaches like fuzzy logic, the model may
be expanded, and a decision support system may be created.
Reference
1. Lasi, H., Fettke, P., Kemper, H. G., Feld, T., & Hoffmann, M. (2014). Industry 4.0. Business &
information systems engineering, 6, 239-242.
2. Ghobakhloo, M. (2020). Industry 4.0, digitization, and opportunities for sustainability. Journal of
cleaner production, 252, 119869.
3. Frank, A. G., Dalenogare, L. S., & Ayala, N. F. (2019). Industry 4.0 technologies: Implementation
patterns in manufacturing companies. International Journal of Production Economics, 210, 15-26.
4. Ghadimi, P., Wang, C., Lim, M. K., & Heavey, C. (2019). Intelligent sustainable supplier selection
using multi-agent technology: Theory and application for Industry 4.0 supply chains. Computers &
Industrial Engineering, 127, 588-600.
5. Sachdeva, N., Shrivastava, A. K., & Chauhan, A. (2021). Modeling supplier selection in the era of
Industry 4.0. Benchmarking: An International Journal, 28(5), 1809-1836.
6. Pan, X. L., & Tian, Y. (2011). Supplier selection in B2B manufacturing commerce using AHP-DEA.
In Advanced Materials Research (Vol. 323, pp. 23-27). Trans Tech Publications Ltd.
7. Patel, M. R., Vashi, M. P., & Bhatt, B. V. (2017). SMART-Multi-criteria decision-making technique
for use in planning activities. New Horizons in Civil Engineering (NHCE 2017), 1-6.
8. Siregar, D., Arisandi, D., Usman, A., Irwan, D., & Rahim, R. (2017, December). Research of simple
multi-attribute rating technique for decision support. In Journal of Physics: Conference Series (Vol.
930, No. 1, p. 012015). IOP Publishing.
9. Yoon, K. P., & Hwang, C. L. (1981). Multiple attribute decision making: Methods and applications:
A state-of-the-art survey. Springer.
10. Javaid, M., Khan, S., Haleem, A., & Rab, S. (2022). Adoption of modern technologies for
implementing industry 4.0: an integrated MCDM approach. Benchmarking: An International
Journal, (ahead-of-print).
11. Kumar, R. R., & Kumar, C. (2016, December). An evaluation system for cloud service selection
using fuzzy AHP. In 2016 11th International Conference on Industrial and Information Systems
(ICIIS) (pp. 821-826). IEEE.
12. Trung, N. Q., & Thanh, N. V. (2022). Evaluation of Digital Marketing Technologies with Fuzzy
Linguistic MCDM Methods. Axioms, 11(5), 230.
13. Naveed, Q. N., Islam, S., Qureshi, M. R. N. M., Aseere, A. M., Rasheed, M. A. A., & Fatima, S.
(2021). Evaluating and ranking of critical success factors of cloud enterprise resource planning
adoption using MCDM approach. IEEE Access, 9, 156880-156893.
14. Medić, N., Marjanović, U., Zivlak, N., Anišić, Z., & Lalić, B. (2018, March). Hybrid fuzzy MCDM
method for selection of organizational innovations in manufacturing companies. In 2018 IEEE
International Symposium on Innovation and Entrepreneurship (TEMS-ISIE) (pp. 1-8). IEEE.
Page 870
SEMIT 2023
15. Kumar, V., Vrat, P., & Shankar, R. (2021). Prioritization of strategies to overcome the barriers in
Industry 4.0: a hybrid MCDM approach. Opsearch, 1-40.
16. Erdogan, M., Ozkan, B., Karasan, A., & Kaya, I. (2018). Selecting the best strategy for industry 4.0
applications with a case study. In Industrial Engineering in the Industry 4.0 Era: Selected papers
from the Global Joint Conference on Industrial Engineering and Its Application Areas, GJCIE 2017,
July 20–21, Vienna, Austria (pp. 109-119). Springer International Publishing.
17. Jamwal, A., Agrawal, R., Sharma, M., Kumar, V., & Kumar, S. (2021). Developing A sustainability
framework for Industry 4.0. Procedia CIRP, 98, 430-435.
18. Raj, A., Dwivedi, G., Sharma, A., de Sousa Jabbour, A. B. L., & Rajak, S. (2020). Barriers to the
adoption of industry 4.0 technologies in the manufacturing sector: An inter-country comparative
perspective. International Journal of Production Economics, 224, 107546.
19. Medić, N., Marjanović, U., Prester, J., Palcic, I., & Lalić, B. (2018). Evaluation of advanced digital
technologies in manufacturing companies: Hybrid fuzzy MCDM approach. In 25th International
EurOMA Conference (pp. 1-10).
20. Ucal Sari, I., & Ak, U. (2022). Machine efficiency measurement in industry 4.0 using fuzzy data
envelopment analysis. Journal of fuzzy extension and applications, 3(2), 177-191.
21. Pishdar, M., Danesh Shakib, M., Antucheviciene, J., & Vilkonis, A. (2021). Interval type-2 fuzzy
super sbm network dea for assessing sustainability performance of third-party logistics service
providers considering circular economy strategies in the era of industry 4.0. Sustainability, 13(11),
6497.
22. Azadi, M., Moghaddas, Z., Cheng, T. C. E., & Farzipoor Saen, R. (2021). Assessing the sustainability
of cloud computing service providers for Industry 4.0: a state-of-the-art analytical
approach. International Journal of Production Research, 1-18.
23. Azadi, M., Moghaddas, Z., Farzipoor Saen, R., & Hussain, F. K. (2021). Financing manufacturers
for investing in Industry 4.0 technologies: internal financing vs. External financing. International
Journal of Production Research, 1-17.
24. Arora, C., Kamat, A., Shanker, S., & Barve, A. (2022). Integrating agriculture and industry 4.0 under
“agri-food 4.0” to analyze suitable technologies to overcome agronomical barriers. British food
journal.
Page 871
SEMIT 2023
SP23: Smart Marketing
Investigating the social and conceptual structure of AI in marketing: a bibliometric analysis

Mostafa Esmaeili Mahyari, PhD of Business management, Department of business Administration,
College of Farabi, University of Tehran, Qom, Iran
Hamid Reza Irani*, Associate professor, Department of business Administration, College of Farabi,
University of Tehran, Qom, Iran hamidrezairani@ut.ac.ir
Vahid Nourmandi Pour, Master student of business management, College of Farabi, University of
Tehran, Qom, Iran
Hamed Nozari, Islamic Azad University, Dubai Branch, UAE
Abstract
The role of AI in marketing in becoming more and more important and is being widely applied in various
sector of marketing. This popularity encouraged many scholars to conduct research in this filed. The aim
of this study is investigating the social and conceptual structure of AI in marketing. To reach this goal,
along with descriptive analyses, keywords co occurrence and co authorship networks were portrayed
through VOS viewer software. Descriptive analyses introduces the highly cited papers, authors, countries,
universities and journals completed with the co authorship analysis showing the social structure of this
filed. Keywords co occurrence analysis also show the conceptual structure of this filed in particular
through temporal status highlights the recent research topics. Findings show that AI is becoming an
important tool for businesses to identify and understand the behavioral patterns and needs of customers
specifically through customer journey and customer experience. These technologies not only assisted
businesses but also helped customers in their decision making.
Keywords: artificial intelligence, AI, marketing, bibliometrics, machine learning, internet of things
Introduction
Artificial intelligence derives from information technology. It is often used interchangeably with
concepts such as automation or even robotization. It also is usually being used mistakenly with machine
learning or algorithm application(Jarek and Mazurek 2019). Artificial intelligence (AI) has recently
turned into a very popular subject in the area of management and marketing sciences. Technological
advancements has continuously caused a shift in the evolution of marketing and has seriously proved that
marketing can cooperate closely with artificial intelligence (AI) to make a difference (Siau, 2017,Wirth
2018). In general, contemporary marketing is becoming more and more data driven, automated, and
intelligent(Kumar et al., 2019,Chintalapati and Pandey 2022).
Along this, research on digital and technological evolution in marketing has developed considerably with
the fast space (Crittenden et al., 2019). In particular via scholars seeking to investigate the ways in which
technological advancements impact the firms and businesses, especially pertaining managing customer
needs and delivering offerings (Kumar et al., 2019).
With constantly growing number of studies and as a result, body of literature, providing an overview and
comprehensive image of a scientific field through review studies becomes more and more significant. In
particular, as traditional reviews suffer limitation in terms of scope and subjectivity because of huge
concentration on researchers’ mental tendency resulting in mental biases and also focusing on a particular
part of the field. Therefore, those traditional reviews are not able to portray the comprehensive status of
Page 872
SEMIT 2023
the knowledge development and structures of a knowledge field instead, quantitative reviews including
bibliometric review are the appropriate alternative for traditional reviews(Fan, et al. 2022) because they
have reliable (Donthu, et al. 2021, Kumar, et al. 2022) and can provide overview of the field such as
evolution also underlying structures and current and future research streams by pinpointing the theoretical
gaps in the literature (Chakraborty, et al. 2021), In spite of traditional reviews(e.g. Chintalapati and
Kumar pandy), which provide useful insights but literature needs an updated bibliometric review
providing overview with today status quo of artificial intelligence in tourism.
Therefore, the aim of this study is to present an overview of studies in the field of artificial intelligence
in marketing due to its huge impact on marketing specifically in recent years as expanding dramatically
and in particular revolving service sector (Robinson et al., 2020, Flavián, et al. 2022).
Literature review
Artificial intelligence in marketing
Various definitions of AI show different aspects and attributes of this technology. In service literature,
AI is referred to as “machines that exhibit aspects of human intelligence” (Huang & Rust, 2018; p. 155).
The development of Artificial intelligence as a phenomenon encouraged scholars to work enthusiastically
push AI concepts forward over the few decades. This effort caused some major innovations like big data
analytics and machine learning applications in myriad sectors and context.(Verma, et al. 2021).
The technology which is based on artificial intelligence is able to imitate the cognitive actions and
functions which usually are being executed by human mind, including the abilities to think, learn and
solve problems (Syam, Sharma, 2018).
Technological solutions driven from Artificial intelligence can be used for business actions and
improving their performance In fact, AI is being used to develop numerous marketing strategies and other
marketing-related activities, such as product promotion, pricing, planning, development, and distribution
(Mogaji et al., 2020) through AI application in business functions such as social media, data mining,
customer growth, consumer engagement, collecting of data and segregating relevant data for further
decision making, enhancing activities of logistics, and increasing efficiencies of various functions of the
business(Bhalerao, et al. 2022). Emerging technologies provide a competitive advantage by facilitating
the customers’ product and service offerings (Verma, et al. 2021). In sum AI is now affecting relatively
all kinds of functions in marketing (Pitt et al., 2018, Hildebrand, 2019), From customer experience to
marketing performances and to business decision making. In particular advances in computing power and
the emergence of Big Data, paved the way for integration of AI In digital marketing which revolutionized
the way of advertizing and creation the online channels for businesses (Haenlein and Kaplan,
2019,Mogaji and Nguyen 2022).
Research method
In the current research, which is descriptive-analytical, bibliometric method has been used. Bibliometric
analysis is a quantitative method that can analyze and examine a specific research field. it can help us
significantly by illustrating and presenting various structures of a field or research topic (Chen, et al.
2022; Liu, et al. 2022). Two visualization analyzes were used in this study: keyword co-occurrence
analysis, which displays interesting and important topics for researchers based on the repetition of words
and their co-occurrence. And co-authorship analysis that shows the cooperation of authors, countries and
universities in knowledge creation in a specific field (Koseoglu, et al. 2016; Singh, et al. 2021).
To start the bibliometric analysis, the database was selected and the criteria for data retrieval from it were
determined. The web of science database and all its subsets were selected for data retrieval, because this
database is a reliable and standard source for bibliometric research (Meho and Yang 2007). This database
includes high quality magazines that publish scientific articles in various fields (Hassan Shah, et al. 2022).
Then the related keywords were identified and searched for retrieving articles in this database, which
were: "Artificial intelligence", "AI", "marketing"
Page 873
SEMIT 2023
Temporal limit was applied to the ending point excluding to end of 2022 but no starting points was
considered. In the initial search, 1066 articles were retrieved from the database. The aim of the current
research was to investigate the global trend of artificial intelligence research in marketing, so as the first
filter the articles were limited to English language articles. By applying this filter, 353 articles were
removed. Then, the second filter related to the type of research was applied and the output set was limited
only to research and review articles, because they have been accepted by more accurate referees. Finally,
698 articles remained in the output and entered the analysis process. In the next step, the full bibliometric
information of these articles, including title, abstract, organizational affiliation, sources, etc., were stored
and then analyzed in the VOS Viewer software. This software is a practical tool for illustrating
bibliometric data that has been used in various researches, and by using its illustrations and the networks
it displays, it is possible to provide more comprehensive and efficient analyses about a research field
(Bragge et al., 2019; Van Eck et al. 2014). By using this software, two widely used analyses were carried
out: keyword co-occurrence which makes it possible to understand the evolution and conceptual structure
of a research field (Bajaj, et al. 2022, Pahrudin, et al. 2022), and the co authorship which highlights the
collaborations among countries, authors etc (Chakraborty, et al. 2021; León-Gómez, et al. 2021).
The steps of this research, which is derived from Pahrudin, et al. (2022), are shown in the figure below:
(1) Identification of the term

(2) Database selection: web of science and all its
subsets
(3) search criteria and filters
Time limit: beginning till 2022
Article type: research and review
Language: English
(4) Finding the result
Mapping the result
(5) Analysis of the result
Discussion
Figure1: the steps of the study
Findings
Data retrieved from the Web of Science database show significant growth in this research area since
2017. The upward trend of publishing articles in this field, which has continued until now, has set a new
record in 2022 with the publication of 230 articles (which is about 33% of all articles in this field). Also,
the trend of citing articles in this field, which started in 1992, has been on the rise and reached its peak in
2022 with 5656 citations and in the current year (2023), 617 citations have been made to articles in this
field until now. Figure 1 show the graph of the publication statistics of articles in the field of artificial
intelligence (AI) and marketing research and the citations made to them. In Table 1, statistics presented
in detail by year.
Page 874
SEMIT 2023
Figure2: quantitative trend of publications of AI in marketing

Table 1: detail of paper publication of AI in marketing
Year Publications Citations Year Publications Citations Year Publications Citations
2023 - 617 2012 3 120 2001 1 11
2022 230 5656 2011 8 92 2000 3 11
2021 182 3697 2010 4 71 1999 1 13
2020 110 1551 2009 9 72 1998 1 12
2019 45 868 2008 2 38 1997 4 7
2018 24 554 2007 1 48 1996 2 7
2017 12 47 2006 1 41 1995 3 3
2016 9 300 2005 5 39 1994 2 8
2015 11 214 2004 4 37 1993 3 2
2014 5 176 2003 3 22 1992 1 1
2013 6 118 2002 3 15 1991 3 -
Based on the results of the search in the database, the most cited articles in the field of artificial
intelligence and marketing research were identified. Accordingly, the article "Machine learning: trends,
perspectives, and prospects" published by Jordan, MI and Mitchell, TM in 2015 published in the
"Science" journal, was recognized as the most cited article in this field with 2613 citations. The article
by Wu, YY; Kosinski, M and Stillwell, D entitled "Computer-based personality judgments are more
accurate than those made by humans" in Proceeding of the national academy of science of the united
states of America with 451 citations is in the second place. In addition, the third most cited article is
Davenport, T; Guha, A; Grewal, D; Bressgott, T's article titled "How artificial intelligence will change
the future of marketing" with 357 citations, which was published in Journal of the academy of marketing
science. A list of the top 10 most cited articles in the field of artificial intelligence and marketing research
with details are shown in Table 2.
Page 875
SEMIT 2023
Table2: highly cited papers of AI in marketing

Pub. Per
Rank Article Authors Year Journal Citations
year
Machine learning: Jordan, MI
1 trends, perspectives, andand Mitchell, 2015 Science 2613 290.33
prospects TM
Computer-based Wu, YY;
Proceeding of the national
personality judgments Kosinski, M
2 2015 academy of science of the 451 50.11
are more accurate than and Stillwell,
united states of America
those made by humans D
Davenport,
How artificial
T; Guha, A; Journal of the academy of
3 intelligence will change 2020 357 89.25
Grewal, D; marketing science
the future of marketing
Bressgott, T
Effects of COVID-19 on
International journal of
hotel marketing and Jiang, YY
4 2020 contemporary hospitality 328 82
management: a and Wen, J
management
perspective article
Setting the future of
Dwivedi,
digital and social media
YK;
marketing research: International journal of
5 Ismagilova, 2021 302 100.67
Perspectives information management
E; (...);
and research
Wang, YC
propositions
Adaptive
Submodularity: Theory Golovin, D
Journal of artificial
6 and Applications in and Krause, 2011 270 20.77
intelligence research
Active Learning and A
Stochastic Optimization
Handling class Burez, J and
Expert systems with
7 imbalance in customer Van den 2009 262 17.47
applications
churn prediction Poel, D
User generated content:
the use of blogs for
8 2009 Service business 246 16.4
tourism organisations Akehurst, G
and tourism consumers
Real-time co-creation
Buhalis, D
and nowness service: Journal of travel & tourism
9 and Sinarta, 2019 203 40.6
lessons from tourism marketing
Y
and hospitality
Poria, S;
Sentic patterns:
Cambria, E;
Dependency-based rules
10 Winterstein, 2014 Knowledge-based systems 201 20.1
for concept-level
G; Huang,
sentiment analysis
GB
The findings of also identify the top authors in the field of artificial intelligence and marketing research
in terms of the number of citations made to their articles. Therefore, Jordan, M. I. and Mitchell, T. M. are
the most cited authors in this field by publishing an article in this field that their article has been cited
2613 times, and also known as the most cited article in this field. After them, Grewal is in the next rank
by publishing 3 articles which had caught 433 citations. A list of the top 10 authors in the field of artificial
intelligence and marketing research is presented in Table 3.
Page 876
SEMIT 2023
Table3: highly cited authors

Number of Total link
Author Citation
doc strength
1 Jordan, m. i. 1 2613 46
2 Mitchell, t. m. 1 2613 46
3 Grewal, Dhruv 3 433 273
4 Guha, Abhijit 2 375 265
5 Dwivedi, yogesh k. 4 365 198
6 Bressgott, Timna 1 357 251
7 Davenport, Thomas 1 357 251
8 Jiang, Yangyang 1 328 3
9 Wen, Jun 1 328 3
Rauschnabel,
10 2 323 95
Philipp A.
Based on the findings, the "Science" journal is identified as the most cited journal in the field of artificial
intelligence and marketing by publishing only one article that has reached 2613 citations. The second
most cited journal is the "journal of the academy of marketing science" with 836 citations and the third
one is the "industrial marketing management" with 442 citations made to their articles. In Table 4, a list
of the top 10 journals in the field of artificial intelligence and marketing research is presented along with
the number of published articles, the number of citations made to them, and the total link strength.
Table4: highly cited journals
Number Total link
source Citation
of doc strength
1 Science 1 2613 17
2 journal of the academy of marketing science 12 836 143
3 industrial marketing management 22 442 113
4 journal of business research 21 419 129
5 international journal of contemporary hospitality management 3 360 9
6 journal of interactive marketing 8 352 60
7 international journal of information management 2 306 18
8 expert systems with applications 4 291 2
9 journal of artificial intelligence research 1 270 0
10 service business 1 246 0
Highly cited universities of this research field were also indentified by examining the data extracted from
the database. Therefore, University of California, Berkeley is the most cited university in this field with
2632 citations. Moreover, Carnegie Mellon University is in second place with 2613 citations and Babson
College is in third place with 433 citations. The 10 most cited universities of this research field are listed
in Table 6 along with the number of articles, the number of citations and the total link strength.
Table5: highly cited universities
Number Total link
organization Citation
of doc strength
1 University of California, Berkeley 2 2632 54
2 Carnegie Mellon University 2 2613 40
3 Babson College 3 433 232
4 University of Bradford 3 393 198
5 University of South Carolina 4 382 240
6 Edith Cowan University Joondalup Campus 5 369 27
7 Maastricht University 2 366 209
8 Swansea University 4 365 172
9 Newcastle University 3 354 141
10 Ghent University 3 332 9
Page 877
SEMIT 2023
The findings show that United States of America, with 6459 citations to its articles, is the most cited
country in the field of artificial intelligence and marketing research. After USA, the Australia ranks
second with 1435 citations and England ranks third with 1381 citations. A list of the top ten countries in
this research area is shown in Table 6.
Table 6: highly cited countries
Country Number of doc Citation Total link strength
United states of
1 145 6459 1274
America
2 Australia 53 1435 508
3 England 57 1381 741
4 Netherland 20 1117 337
5 China 85 1098 287
6 Germany 35 1061 368
7 India 62 883 534
8 Canada 26 800 347
9 France 27 799 310
10 Spain 25 559 244
Co-occurrence analysis
Key words co occurrence analysis shows the highly used words related to the artificial intelligence in
marketing top five words based on occurrence are artificial intelligence(291), machine learning(91),
marketing(67), big data(28) and chtbot(23) table 7 and figure 3 show the top keywords in this filed.
Table 7 : keywords co occurrence

No keywords occurrence Total link strength
1 artificial intelligence 291 360
2 machine learning 91 149
3 marketing 68 108
4 big data 28 72
5 chatbot 23 33
6 digital marketing 22 39
7 deep learning 19 35
8 natural language processing 14 32
9 social media 13 20
10 anthropomorphism 10 13
11 customer experience 10 14
12 sentiment analysis 10 21
13 Tourism 10 25
14 Trust 10 16
15 data mining 9 17
16 digitalization 9 18
17 e-commerce 9 15
18 robots 9 25
19 advertising 8 14
20 blockchain 8 16
Keywords co-occurrence analysis is also able to show the temporal order of keywords indicating
frequency of words by authors and also ongoing topics. The analysis reveals that in recent years topics
have changed from customer loyalty to the customer engagement in relation to the artificial intelligence
in marketing, this trend is in the line with general trend of marketing which the consumer centricity is the
Page 878
SEMIT 2023
dominant theory (e.g. Lemon and Verhoef 2016). It is also evident that along with the general research
stream of marketing which recently highlights the role of customer journey in customer experience,
artificial intelligence also addressed these concepts with the focus of AI role in customer experience and
customer journey.
Figure3: top keywords of AI studies in marketing

Findings also show that within artificial intelligence there are various technologies assisting marketing
such as machine learning(Syam and Sharma 2018), internet of things ( e.g. Hacid, et al. 2023),
chatbots(Luo, et al. 2019), etc. figure 4 illustrates the temporal statue of these keywords.
Figure 4: temporal statue of keywords of AI studies in marketing
Co authorship analysis which envisage the status quo of collaboration among countries in a special field,
indicate that in addition to possessing of the highest number of articles with 145, with 32 links has the
highest collaboration with other countries in scientific publications following by England, India and
Germany. The analysis revealed 6 clusters with the close topic covering in the same cluster. Figure 5
illustrates the collaboration networks among countries
Page 879
SEMIT 2023
Figure 5: collaboration network among countries
Discussion and conclusion

In addition to introducing the most productive authors, countries, universities and journals also highly
cited papers, collaboration among countries indicate the top countries in this field such as USA and China
contributing significantly in this filed. Finding arise from conceptual structure also highlights the role of
technologies in particular artificial intelligence in enormous changes in marketing. Although in recent
years development of marketing has taken place with emergence of some theories such as Service
dominant logic (Vargo and Lusch 2008) leading to some concepts and theories such as consumer
centricity, co creation and etc, but not only technologies contributed to this shift also accelerated this
development. Technology breakthrough in particular artificial intelligence, and driven technologies,
facilitated the customer participation which ultimately empower them to co create (Sthapit, et al. 2019)
and even further personalize what they want to consume and also choose the channel and their own
customer journey(Sugathan and Ranjan 2019). It also helps managers to understand customer needs and
preferences more precisely (Neuhofer, et al. 2015). These technologies also paved the way for customer
engagement(Gretzel, et al. 2015, Moriuchi 2021).
As findings of this study confirm, customer experience management and customer journey is one of the
most important application of AI in marketing. Analyzing customer behavior through AI can provide
information for understanding various patterns of consumer behavior and then predicting their needs
result in better planning for creating greater experience(Daqar and Smoudy 2019). For instance Chabots
as automated customer support tools also can create dialogue with customer to improve their
experience(Torres and Delgado 2023) which ultimately can create customer loyalty (Chen, et al. 2023).
The same thing is true with IOT which is the enabler of big data, provide information for customers
moreover assist businesses for knowing customers much more precisely(Elkhwesky and Elkhwesky
2022).
References
Bajaj, V., et al. (2022). "Linkage dynamics of sovereign credit risk and financial markets: A bibliometric
analysis." Research in International Business and Finance 59: 101566.
Bhalerao, K., et al. (2022). "A Study of Barriers and Benefits of Artificial Intelligence Adoption in Small
and Medium Enterprise." Academy of Marketing Studies Journal 26: 1-6.
Chakraborty, K., et al. (2021). "A systematic literature review and bibliometric analysis based on pricing
related decisions in remanufacturing." Journal of Cleaner Production 310: 127265.
Page 880
SEMIT 2023
Chen, Q., et al. (2023). "Can AI chatbots help retain customers? Impact of AI service quality on customer
loyalty." Internet Research(ahead-of-print).
Chintalapati, S. and S. K. Pandey (2022). "Artificial intelligence in marketing: A systematic literature
review." International Journal of Market Research 64(1): 38-68.
Crittenden, W. F., et al. (2019). "Embracing digitalization: Student learning and new technologies."
Journal of Marketing Education 41(1): 5-14.
Daqar, M. A. A. and A. K. Smoudy (2019). "The role of artificial intelligence on enhancing customer
experience." International Review of management and Marketing 9(4): 22.
Donthu, N., et al. (2021). "How to conduct a bibliometric analysis: An overview and guidelines." Journal
of business research 133: 285-296.
Elkhwesky, Z. and E. F. Y. Elkhwesky (2022). "A systematic and critical review of Internet of Things in
contemporary hospitality: a roadmap and avenues for future research." International Journal of
Contemporary Hospitality Management(ahead-of-print).
Fan, D., et al. (2022). "Advancing literature review methodology through rigour, generativity, scope and
transparency." International Journal of Management Reviews 24(2): 171-180.
Flavián, C., et al. (2022). "Intention to use analytical artificial intelligence (AI) in services–the effect of
technology readiness and awareness." Journal of Service Management 33(2): 293-320.
Gretzel, U., et al. (2015). "Smart tourism: foundations and developments." Electronic Markets 25(3): 179-
188.
Hacid, H., et al. (2023). "An exploratory study of rethinking user experience in the internet of things."
Computing 105(1): 1-12.
Hildebrand, C. (2019). "The machine age of marketing: How artificial intelligence changes the way
people think, act, and decide." NIM Marketing Intelligence Review 11(2): 10-17.
Huang, M.-H. and R. T. Rust (2018). "Artificial intelligence in service." Journal of Service Research
21(2): 155-172.
Jarek, K. and G. Mazurek (2019). "Marketing and Artificial Intelligence." Central European Business
Review 8(2).
Kumar, S., et al. (2022). "Fuzzy-set qualitative comparative analysis (fsQCA) in business and
management research: A contemporary overview." Technological Forecasting and Social Change 178:
121599.
Kumar, V., et al. (2019). "Understanding the role of artificial intelligence in personalized engagement
marketing." California management review 61(4): 135-155.
Luo, X., et al. (2019). "Frontiers: Machines vs. humans: The impact of artificial intelligence chatbot
disclosure on customer purchases." Marketing science 38(6): 937-947.
Mogaji, E. and N. P. Nguyen (2022). "Managers' understanding of artificial intelligence in relation to
marketing financial services: insights from a cross-country study." International Journal of Bank
Marketing 40(6): 1272-1298.
Mogaji, E., et al. (2020). "Using AI to personalise emotionally appealing advertisement." Digital and
Social Media Marketing: Emerging Applications and Theoretical Development: 137-150.
Moriuchi, E. (2021). "An empirical study on anthropomorphism and engagement with disembodied AIs
and consumers' re‐use behavior." Psychology & Marketing 38(1): 21-42.
Neuhofer, B., et al. (2015). "Smart technologies for personalized experiences: a case study in the
hospitality domain." Electronic Markets 25(3): 243-254.
Sthapit, E., et al. (2019). "Extending the memorable tourism experience construct: an investigation of
memories of local food experiences." Scandinavian Journal of Hospitality and Tourism 19(4-5): 333-353.
Page 881
SEMIT 2023
Sugathan, P. and K. R. Ranjan (2019). "Co-creating the tourism experience." Journal of business research
100: 207-217.
Syam, N. and A. Sharma (2018). "Waiting for a sales renaissance in the fourth industrial revolution:
Machine learning and artificial intelligence in sales research and practice." Industrial Marketing
Management 69: 135-146.
Torres, A. I. and C. J. Delgado (2023). Assessing Customer Interactions With Chatbots in Online
Shopping Experiences: An Empirical Study. Promoting Organizational Performance Through 5G and
Agile Marketing, IGI Global: 203-223.
Vargo, S. L. and R. F. Lusch (2008). "Service-dominant logic: continuing the evolution." Journal of the
academy of marketing science 36(1): 1-10.
Verma, S., et al. (2021). "Artificial intelligence in marketing: Systematic review and future research
direction." International Journal of Information Management Data Insights 1(1): 100002.
Wirth, N. (2018). "Hello marketing, what can artificial intelligence help you with?" International Journal
of Market Research 60(5): 435-438.
Page 882
SEMIT 2023
Examining the Impact of Entrepreneurial Culture on SME Performance: The Role of Marketing
Capabilities and Social Media
Mohsen Mazaheriasad
PhD. candidate in IT Management – Business Intelligence, Islamic Azad University, South Tehran
Branch, Iran
m.mazaheriasad@gmail.com
Maryam Jamali
MSc. in Entrepreneurship Management, University of Tehran, Iran
mjamalim072@gmail.com
Abstract
This research investigates the influence of entrepreneurial culture on the performance of small and
medium-sized enterprises (SMEs) in Tehran Province, Iran, focusing on the mediating effect of marketing
capacities and social media usage. A questionnaire of 25 questions, separated into four sections—
"Entrepreneurial culture," "SMEs' performance," "Marketing skills," and "Social media usage"—was
used to collect data. The replies of 80 SME owners or managers were collected by non-probability
sampling, and the validity and dependability of the data were confirmed. For data analysis, the study
applied structural equation modeling (SEM) with the partial least squares (PLS) technique and SmartPLS
4 software. The findings indicated that entrepreneurial culture has a favorable and substantial effect on
the performance, marketing capacities, and social media usage of SMEs. In addition, marketing
capabilities and social media usage have a positive and substantial effect on the performance of SMEs.
Additionally, social media utilization has a good impact on marketing capabilities. The study
demonstrated the indirect influence of entrepreneurial culture on the performance of SMEs through the
mediating roles of marketing capabilities and social media usage. The article concludes with evidence-
based, actionable recommendations based on the research findings.
Keywords: Entrepreneurial Culture, SMEs’ performance, Marketing Capabilities, Social Media Usage
1. Introduction
SMEs are essential for fostering economic growth in both developed and developing countries, including
Iran. Their effects are seen in the growth of the GDP, the creation of jobs, rising incomes, and the
establishment of new businesses (Sopha et al., 2021). As a result, research on SME-related challenges
has gained popularity in a number of sectors (Omar et al., 2020). It is now even more important to
understand entrepreneurship and the use of digital technologies, particularly social media, in order to
improve the performance and productivity of SMEs after the COVID-19 pandemic had such a devastating
impact on them (Papadopoulos et al., 2020). In the post-pandemic era, Ratten (2020) underlines the value
of entrepreneurial expertise. SMEs in difficult circumstances may need to take entrepreneurial actions,
such as refocusing innovation resources on obtaining practical insights, selecting the best options for the
production or service delivery process, realigning business models with the changing needs and
expectations of their customers, and implementing technologies to enhance their operations (Susanto et
al., 2023). Like in any other nation, SMEs are the foundation of the Iranian economy. The success of
SMEs is vital to the economic development of nations since they are acknowledged as the engine of
economic growth in both developed and developing nations (Singh et al., 2020). Even though they have
many advantages, SMEs confront a number of common problems, including lower productivity, less
legitimacy, and higher capital expenses. Regardless of the business climate they operate in, these
challenges pose a danger to the viability and success of SMEs, leading to an alarming failure rate during
the first few years of operation (Sadeghi, 2018). Long-term performance for even those that do survive is
frequently only mediocre, which is worrying for both the economy and enterprises. Social and
demographic changes are continually changing the entrepreneurial environment, and SMEs struggle to
operate in a commercial world that is more controlled by technology. Iran's developing economy presents
particular difficulties that call for reconsideration and a better understanding of Iranian business culture
Page 883
SEMIT 2023
and performance. A definition of entrepreneurial culture (EC) is a setting that fosters creativity and
invention and is marked by a set of values, precepts, and anticipated behaviors. As a result, this culture
encourages employees at entrepreneurial firms to be proactive and free-thinking when developing and
implementing novel ideas. It also encourages a willingness to embrace uncertainty, take chances, and
start new projects. As a result, these companies exhibit highly creative workplace behavior. In the quickly
changing environment of the fourth industrial revolution, where innovation dynamics are bursting
because of the open innovation paradox, understanding the function of EC in fostering inventive work
behavior is extremely critical (Yun et al., 2020). This study also answers Obschonka's (2017) demand for
additional research on EC in order to capitalize on unparalleled prospects for entrepreneurial activities
while addressing twenty-first-century concerns (Nguyen et al., 2023). In recent years, the marketing
paradigm has shifted significantly as a result of the development of technology and the increased use of
social media in marketing (Qalati et al., 2021). Consequently, it is considered that marketing capabilities
(MCs) and social media usage (SMU) may influence the connection between EC and the performance of
SMEs (Pascal and Shin, 2015). However, research on the relevance of these characteristics in the context
of SMEs, particularly in developing nations, has been sparse (Trainor, 2012). MCs and SMU should be
examined for two reasons. Initially, MCs can strengthen a company's competitive position and customer
relationships, resulting in enhanced performance (Omar et al., 2020). MC refers to a company's capacity
to comprehend and predict the demands of its customers better than its competitors and to engage them
successfully. Therefore, it is proposed that SMEs need marketing skills to develop a strong relationship
between their EC and performance. In addition, MCs may serve as both a mediator and an influencer on
EC performance. This dual role of MCs has not been investigated in the context of emerging markets
(Wiwoho et al., 2020), broadening our understanding of the responsibilities of MCs within the EC-SMEs’
performance relationship. Second, SMU platforms such as Facebook, Instagram, WhatsApp, LINE, and
WeChat offer multiple chances for SMEs to increase their competitiveness in the social media
environment (Gillin and Schwartzman, 2010). SMEs can use social media to identify new possibilities,
strengthen customer relationships, increase collaboration within the business and with other entities, and
advertise their products and services. In addition, social media plays a significant role in establishing
consumer interactions, making it an essential part of marketing (Appel et al., 2020). The association
between EC and SME performance may therefore be moderated by the SMU. The prospective marketing
activities via social media show that it can establish the link between EC and MC, and ultimately, EC and
the performance of SMEs. Previous studies have neglected the possibility of a serial or double mediator
between EC and SMEs' performance (Wales et al., 2021). In order to examine the relationship between
EC and SMEs' performance, the study has five objectives:
• Analyzing the connection between EC and SMEs’ performance.
• Examining the role of MC as a mediator between EC and SME performance.
• Examining SMU as a mediator between EC, SMEs’ performance, and MC.
• Examining SMU as a mediator between EC and the SMEs’ performance.
• Examining the serial mediation of EC and SME performance via SMU and MC.
These objectives suggest that the study intends to investigate the relationship between EC and SME
success, as well as the role that social media and MC may play in this relationship. In order to examine
the relationship between EC and SME success, the study has five objectives:
• Analyzing the connection between EC and the SMEs’ performance.
• Examining the role of MC as a mediator between EC and SME performance.
• Examining the SMU as a mediator between EC, SME performance, and MC.
• Examining the SMU as a mediator between EC and the SMEs’ performance.
• Examining the serial mediation of EC and SME performance via SMU and MC.
These objectives suggest that the study intends to investigate the relationship between EC and SME
success, as well as the role that SMU and MC may play in this relationship. It also says that the study
would investigate how SMU may moderate the connection between EC and SME performance. It also
says that the study would investigate how SMU may moderate the connection between EC and SME
performance.
Page 884
SEMIT 2023
2.1. Entrepreneurial Culture (EC)
EC is the combination of personal values, management expertise, experiences, and behaviors that
determine an entrepreneur's spirit of initiative, risk-taking, innovative capability, and ability to manage a
company's economic ties (Minguzzi & Passaro, 2001). This concept is multifaceted and comprises
numerous facets. Despite the fact that multiple features of EC have been found in the relevant literature
(Kanter, 1985; Hornsby et al., 1999; Ireland et al., 2003), there is a lack of uniformity in how this term is
defined and quantified across different disciplines. Leadership is likely the most frequently mentioned
aspect of EC, since it entails managers encouraging their people to fulfill their full potential and fostering
an entrepreneurial mindset among them. The second factor relates to the characteristics of EC, such as
risk-taking and a tolerance for failure, as well as the adoption of innovative techniques by both the firm
and its board of directors, as well as human resource (HR) policies and initiatives (Nguyen et al., 2023).
Bau and Wagner (2015) highlighted four aspects of EC that encompass the aforementioned themes. These
are "leadership quality and effectiveness" (LQE), "collaboration, information, and innovation" (CII),
"product and market expertise," and "tasks and responsibilities." In addition, this concept incorporates
some aspects that were not previously included in research, such as product and market expertise,
teamwork, and information exchange (Nguyen et al., 2023).
In an entrepreneurial organization, the "LQE" component refers to the significance of competent
leadership. This can appear in several ways, such as developing self-governing teams through clear
instructions and objectives for all members (as advocated by Bell and Kozlowski in 2002) or mentoring
and guiding people to nurture innovative thought and the ability to see new opportunities (as proposed by
Hisrich and Peters in 1986).
The "CII" dimension refers to HR systems that empower employees to act independently and demonstrate
entrepreneurial behavior through fostering collaboration, creativity, and information exchange (Bau and
Wagner, 2015). This includes fostering a supportive workplace culture in which employees are
encouraged to collaborate, adopting effective information technologies to allow communication and
knowledge exchange, and providing an atmosphere that fosters innovation.
The "product and market expertise" dimension is a cultural quality that emphasizes the significance of
having a thorough understanding of a company's goods, customer needs, and industry trends (Bau and
Wagner, 2015). This includes a thorough understanding of the company's services, an awareness of the
changing needs and desires of clients, and an awareness of the most recent industry advances and
innovations.
The "tasks and responsibility" factor relates to the clarity with which employees comprehend their job
responsibilities. This includes having a full understanding of their jobs and responsibilities, as well as
their role within the business. Such clarity enables employees to align their own ambitions with the
organization's objectives, thus assisting the organization in achieving its goals (Bau and Wagner, 2015).
Based on this perspective, the following hypothesis is proposed:
• H1: EC is positively associated with SMEs’ performance.
2.2. Social Media Usage (SMU)
Social media is a readily accessible external resource that any SME can utilize. The amount of benefit
derived from social media mostly depends on the talents, methods, habits, and decision-making abilities
of SME owners (Susanto et al., 2023). The rise of social media as a meaningful and innovative tool for
corporate growth and sustainability has been the subject of extensive academic and managerial discussion
(Jabbour et al., 2019). Consequently, entrepreneurship researchers have acknowledged the growing
significance of social media, especially for SMEs (Wamba and Carter, 2016). Utilizing their real-time
market knowledge, resource-matching abilities, networking, social media customer relationship
management (CRM), co-creation of products and services, and overall MCs, social media can provide
valuable market intelligence to bridge the resource gap and reduce uncertainty for SMEs (Omar et al.,
2020). As noted by Sigala (2012), the usefulness of social media in increasing SME performance is
contingent on the company's total utilization of multiple technologies and the synergy produced between
Page 885
SEMIT 2023
these technologies and the company's internal resources. In addition to utilizing corporate IT resources,
social media provides opportunities for networking, knowledge and information sharing between
businesses and customers, and other advantages, as mentioned by Trainor et al. (2014). According to this
perspective, SMU can strengthen the impact of EC on SMEs.
Effective two-way communication with consumers and the development of long-term relationships are
essential to a company's success because they result in enhanced customer happiness, business goodwill,
and profitability. By leveraging SMU, SMEs may effectively communicate with customers and collect
information about their requirements and problems, enabling them to better comprehend real-time
expectations and fluctuating product and service demand. Furthermore, SMU enables SMEs to access
wider client groups and demographics, which can increase overall performance. The favorable mediation
effects of social media marketing on organizational and brand performance discovered by Wang and Kim
(2017) and Odoom and Mensah (2019), respectively, suggest that SMU likely influences the relationship
between EC and SME success. According to Mason et al. (2021), social media marketing has been
actively exploited throughout the COVID-19 epidemic.
Therefore, our hypotheses in this stance are as follows:
• H2: EC is positively associated with SMU.
• H3: SMU is positively associated with SMEs’ performance.
• H4: EC has an indirect positive effect on SMEs’ performance through SMU.
2.3. Marketing Capabilities (MCs)
MCs are the abilities and skills that enable firms to identify and comprehend the needs of their customers
and the market through interactions. By exploiting these competencies, businesses can proactively adapt
their approach to market conditions and achieve specific performance goals (Elsharnouby and Elbanna,
2021). As a result, MCs have been identified as a crucial company performance determinant (Martin and
Javalgi, 2016).
MCs are regarded as an internal asset because it is the obligation of the decision-maker to decide when
to engage, publicize, and react. SMEs require an entrepreneurial mindset to produce exceptional results
through their MCs. These skills permit the integration of improved services that suit consumer needs,
ultimately resulting in stronger customer connections (Kohtamaki et al., 2015). Utilizing MCs enables
SMEs to maximize their potential profit and gain a competitive advantage over other organizations.
Consequently, incorporating the market capability element into the interaction between EC and SME is
essential for strengthening the model and gaining a thorough grasp of its dynamics. In addition, a number
of research papers, including those by Murray et al. (2011), Martin and Javalgi (2016), and Su et al.
(2013), have examined the relationship between MCs and corporate performance. These studies have
collectively established a strong and positive association between MCs and organizational performance.
Therefore, it is logical to conclude that SMEs could improve their performance by enhancing their MCs.
Khan et al. (2021) found a positive correlation between entrepreneurial competencies and the
performance of enterprises during the COVID-19 pandemic.
In addition, a number of research papers, including those by Murray et al. (2011), Martin and Javalgi
(2016), and Su et al. (2013), have examined the relationship between MCs and corporate performance.
These studies have collectively established a strong and positive association between MCs and
organizational performance. Therefore, it is logical to conclude that SMEs could improve their
performance by enhancing their MCs. Khan et al. (2021) found a positive correlation between
entrepreneurial competencies and the performance of enterprises during the COVID-19 pandemic.
Consequently, the hypotheses for this position are as follows:
• H5: EC is positively associated with MC.
• H6: MC is positively associated with SMEs’ performance.
• H7: SMU is positively associated with MC.
• H8: EC has an indirect positive effect on SMEs’ performance through MC.
After building a theoretical basis and assumptions, we suggest the research framework represented in
Figure 1.
Page 886
SEMIT 2023
H2 Social Media H3
Usage
H7
H4
Entrepreneurial SMEs’
Culture H1 performance
H8
Marketing
H5 H6
Capabilities
Figure 43. Research Conceptual Model
3. Research Method
This study design adopts a quantitative methodology with closed-ended survey questions. The
organizational level will serve as the unit of analysis for this investigation of SME proprietors and
managers in Tehran Province, Iran. Due to a dearth of data, this study employed the technique of
purposive sampling based on government-defined criteria for SMEs. The research team circulated 90
questionnaires via research assistants and enumerators and gathered 80 completed surveys, which met
Hair et al.'s (2014) sample requirements. PLS-SEM (partial least squares structural equation modeling)
was used in this study to test the conceptual framework and hypotheses. Given the mediation hypothesis,
Hair et al. (2019) indicated that PLS-SEM was an appropriate approach. For the PLS-SEM analysis,
SmartPLS 4 software was utilized.
This study utilized a five-point scale that varied from "strongly disagree" (1) to "strongly agree" (5) for
measuring responses. The choice of a five-point scale was made to encourage respondent engagement
and reduce the amount of time needed to complete the questionnaire, as compared to an open-ended
approach (Churchill, 1979). To verify the validity and reliability of the metrics, each item was adapted
from earlier SME research and field-tested in emerging economies.
We used the measurement items for "EC" from Nguyen et al. (2023) due to their effective implementation
and validation in previous studies on SMEs in developing countries. The "EC" variable consisted of a
total of 9 items, including 2 items pertaining to leadership and quality of exchange, 3 items pertaining to
creativity, innovation, and risk-taking, 2 items pertaining to product and market knowledge, and 2 items
pertaining to task management and responsibility. This variable contains the example "My employer
encourages and supports my career development."
To measure "MC", we utilized the measuring items created and validated by Susanto et al. (2023),
consisting of a total of 4 items. In accordance with their research, we also evaluated a multidimensional
construct of "SMU" comprising three dimensions: information search, marketing and branding, and
customer relations. Each dimension was measured with two items, for a total of six items derived from
Susanto et al. (2023).
To assess "SME performance" (PER), we utilized the measurement items created by Susanto et al. (2023),
which consist of a total of six measures that are centered on efficiency, growth, and profitability.
As a result, a questionnaire with 25 questions was distributed. Prior to dissemination, we did a pilot test
of the survey with a small sample of SME owner-managers to check that there were no phrasing,
grammar, or readability errors. The pilot study included 30 managers or owners of SMEs, which is
regarded as an appropriate quantity according to Hertzog (2007) as the minimal criterion for a pilot study.
No problems were discovered during the pilot research.
Page 887
SEMIT 2023
4. Results
4.1. Demographic Statistics
84 out of 90 surveys were returned, for a response rate of 93.3 percent. Eighty percent of the submitted
surveys were useful, while the rest could not be utilized due to unanswered questions. Following the
recommendations of Hair et al. (2014), we are sure that our sample size and response rate are enough for
conducting the analysis.
On the basis of a questionnaire given to SME owners and managers, the demographic characteristics of
the respondents were identified. These statistics indicate that 56 percent of respondents are male and 44
percent are female. In addition, the ages of the respondents are as follows: 11% are between the ages of
25 and 34, 59% are between the ages of 35 and 44, 21% are between the ages of 45 and 54, and 9% are
over the age of 54. Regarding marital status, 63% of respondents are married, while 27% are single. 35
percent of respondents own a bachelor's degree, 55 percent possess a master's degree, and 10 percent
possess a doctorate.
In conclusion, the questionnaire issued to SME owners and managers found that the majority of
respondents were male, between the ages of 35 and 44, married, and master's degree-educated. However,
there was also significant participation from other age groups and educational backgrounds, showing the
existence of a broad community of SME owners and managers.
4.2. Measurement Model Evaluation
In accordance with Hair et al. (2019), we utilized SEM-PLS to develop a reflecting measurement model
(algorithm) and a structural model (bootstrapping). The initial phase entails assessing the quality of
measurement, which requires a review of validity and reliability measures to ensure that criteria are met.
According to the measurement test and quality standards established by Hair et al. (2019), it has been
confirmed that each item has a loading factor greater than 0.70. We used the PLS-SEM algorithm to
assess the measurement model's internal consistency (Cronbach's alpha, composite reliability),
convergent validity (loading factors, average variance extracted), and discriminant validity (e.g., Fornell-
Lacker Criterion and Heterotrait-Monotrait ratio/HTMT) in accordance with Hair et al.'s (2019)
recommendations.
Table 1 provides proof that the measuring model meets the necessary standards. Cronbach's alpha and
composite reliability for each construct, which both surpass 0.70, indicate that the internal consistency is
dependable. In addition, the average variance extracted (AVE) exceeds 0.50, indicating adequate
convergent validity.
Table 1. the status of convergent validity and internal consistency reliability

Construct AVE CR Cronbach's alpha
EC 0.574 0.892 0.866
MC 0.711 0.873 0.863
PER 0.573 0.872 0.840
SMU 0.717 0.926 0.921
Table 2 demonstrates that the square roots of the AVE values are bigger than the correlation between
constructs in terms of discriminant validity. Further, Table 3 demonstrates that the HTMT ratio falls
below the crucial value of 0.90, suggesting that discriminant validity is obtained. These outcomes
permitted us to proceed with the evaluation of the structural model.
Page 888
SEMIT 2023
Table 2. the status of discriminant validity (Fornell-Lacker criterion)

EC MC PER SMU
EC 0.888
MC 0.856 0.843
PER 0.850 0.818 0.857
SMU 0.743 0.800 0.841 0.847
Note: The correlation of the latent construct is lower than the square roots of the AVE values (italicized).
Table 3. the status of discriminant validity (Heterotrait-Monotrait ratio (HTMT))
EC MC PER SMU
EC -
MC 0.824 -
PER 0.873 0.861 -
SMU 0.765 0.855 0.884 -
4.3. Structural Model Evaluation

According to Hair et al. (2019), the evaluation of the structural model consists of a set of consecutive
processes. Among these are measuring collinearity, identifying the significance of route coefficients,
calculating the R-squared value (R2), and checking predictive significance (Q2). Prior to doing the study,
it is essential to test for collinearity using variance inflation factor (VIF) values. The findings of the
internal VIF suggest that each construct (EC, MC, and SMU) has a value less than 3 (EC = 2.23, MC =
1.000, and SMU = 2.33), showing that collinearity is not a concern (Hair et al., 2019).
The coefficient of determination (R2) reflects the predictive accuracy of MCs (0.845) and SME
performance (0.897), which both exhibit high levels of accuracy, while SMU (0.552) demonstrates
moderate levels of accuracy. According to Hair et al. (2019) and Henseler et al. (2015), R2 values of 0.19,
0.33, and 0.67 imply weak, moderate, and strong levels of in-sample predictive power (Sarstedt et al.,
2019). Our blindfold tests utilizing the cross-validated redundancy method (Hair et al., 2019) indicated
Q2 values for exogenous constructs (EC) to be greater than zero, suggesting their predictive validity for
endogenous constructs (SME performance and MC).
These outcomes suggest that the structural model evaluation requirements have been satisfied. We
utilized PLS-SEM bootstrapping with a subsample of 5,000 and used bias-corrected accelerated (BCa)
bootstrapping and two-tailed to determine path coefficient significance. We established a significance
criterion of 5 percent with a p-value less than or equal to 0.05. The path coefficients can be classified as
direct or indirect (mediation) coefficients, which are reported in Table 4.
Figure 2 displays the coefficients and t-values of the model's relationships.
Page 889
SEMIT 2023
Figure 44. Result of the structural model including path coefficients (t-values)
Table 4. Results for direct effects, mediating effect
Hypotheses Relationships β STDEV t-value p-values Decisions
H1 EC → PER 0.230 0.067 3.419 0.001 Supported
H2 EC → SMU 0.743 0.046 16.219 0.000 Supported
H3 SMU → PER 0.381 0.080 4.786 0.000 Supported
H4 EC → SMU → PER 0.283 0.059 4.779 0.000 Supported
H5 EC → MC 0.485 0.055 8.870 0.000 Supported
H6 MC → PER 0.393 0.088 4.494 0.000 Supported
H7 SMU → MC 0.500 0.057 8.802 0.000 Supported
H8 EC → MC → PER 0.191 0.056 3.418 0.001 Supported
5. Discussion
The study suggests that the presence of an EC can have a significant and positive effect on the
performance of SMEs. The statistical analysis demonstrates that the hypothesis H1 is supported, as
evidenced by the statistically significant effect size (β = 0.230, p-value = 0.001). This means that SMEs
with a greater degree of EC will likely do better. The study underlines the positive impact of EC on the
success of SMEs, which may be promoted through a set of values, beliefs, attitudes, and behaviors that
encourage innovation, risk-taking, and proactive decision-making. As SMEs are viewed as economic
development, job creation, and innovation engines, this conclusion is especially crucial for policymakers,
entrepreneurs, and researchers. Thus, organizations should seek to foster an environment that supports
innovation and risk-taking, as well as provide the resources necessary to pursue new ideas. In addition to
providing financial help, policymakers can assist SMEs by fostering an entrepreneurial mindset. Overall,
EC contributes to improved organizational performance, which fosters economic expansion and
innovation. This finding is similar to prior research, such as those of Wardi et al. (2018), Omar et al.
(2016), and Nguyen et al. (2023), which discovered that businesses with high entrepreneurial behavior
are likely to enjoy more success. In addition, this study contradicts previous studies, which revealed only
a moderate association between EC and firm performance (Dimitratos et al., 2004; Lumpkin and Dess,
2001).
Page 890
SEMIT 2023
The study demonstrates that hypothesis H2 is supported, as EC has a significant and positive effect on
SMU (β = 0.743, p-value = 0.000). This result suggests that an increase in EC will likely result in a wider
SMU for business purposes. Additionally, Amoah et al. (2021) have investigated the SMU for business
reasons. With an estimated effect size of β = 0.743 and a statistically significant p-value of 0.000, the
finding implies that there is a strong association between EC and SMU. This suggests that individuals or
firms with a better entrepreneurial spirit are more likely to use SMU platforms for marketing, networking,
and consumer communication. The study also underlines the significance of developing an EC in the
digital age in order to capitalize on the potential benefits of SMU. This research provides organizations
and governments with useful insights to consider when building policies to promote innovation, growth,
and engagement via social media channels.
It has been determined that hypothesis H3 is supported since SME performance is positively affected by
SMU (β = 0.0381, p-value = 0.000). This result is consistent with earlier studies that have proved the
significance of social media in increasing the marketing efforts and consumer involvement of businesses.
By exploiting social media channels to communicate with customers and advertise their products or
services, SME performance and market competitiveness can potentially be enhanced. These findings
emphasize the need for SMEs to incorporate social media strategies into their business planning in order
to fully enjoy the benefits of these communication channels. It is essential to remember, however, that
effective SMU necessitates careful planning and execution, as well as continuous evaluation of its impact
on business outcomes. Several past studies have corroborated the positive effects of SMU on the
performance of businesses, as have our findings. In particular, Susanto et al. (2023), Dodokh et al. (2019),
Tajudeen et al. (2018), Parveen et al. (2016), and Ainin et al. (2015) have shown comparable outcomes.
In addition, Amoah et al. (2021) have backed this view by acknowledging that corporations have been
utilizing social media to conduct specific operations.
Table 4 confirms that EC has a favorable indirect influence on SME performance via SMU (β = 0.283,
p-value = 0.000). The result that EC has a positive indirect effect on the performance of SMEs through
the SMU underscores the significance of fostering a strong, innovative, and risk-taking culture in SMEs.
Such an EC encourages people to take risks, be innovative, and investigate new business ideas. This may
result in greater SMU channels for brand promotion and customer outreach. SMU can improve the
exposure and reach of SMEs, which has a positive effect on their financial success. Therefore,
entrepreneurs should seek to cultivate a culture that fosters innovation and risk-taking while embracing
social media platforms to improve the performance of their SME.
The supported hypothesis H5 suggests an association between EC and MC (β = 0.485, p-value = 0.000).
This conclusion emphasizes the significance of developing a culture of innovation and risk-taking in
SMEs. When SMEs encourage employees to think creatively, take risks, and investigate new business
models, they are more likely to develop robust MCs. This allows them to effectively identify their target
market, craft a persuasive message, and design customer-resonating promotional techniques.
Entrepreneurs should attempt to construct a culture that supports creativity, innovation, and risk-taking
while investing in developing excellent MCs. This can ultimately result in an increase in consumer
engagement, brand recognition, and financial performance for the SME.
By accepting hypothesis H6, it was determined that MCs are positively related to SME performance (β =
0.393, p-value = 0.000). This discovery is a significant advance in our understanding of what makes
SMEs successful. This shows that SMEs that invest in enhancing their MCs will likely perform better
than those that do not. This is consistent with prior studies that have shown how good marketing tactics
may assist organizations in achieving their objectives and gaining a competitive advantage in their
respective marketplaces. This conclusion emphasizes the need for SMEs to prioritize and invest in the
development of their MCs in order to improve their overall success. Multiple studies have investigated
how the MCs of a company or manager can facilitate the appraisal of market needs and the achievement
of both financial and non-financial goals (Krasnikov and Jayachandran, 2008; Morgan et al., 2009; Martin
and Javalgi, 2016). Khan et al. (2021), who examined the effects of COVID-19, and our own findings are
consistent. In addition, the positive influence of MCs on SME performance gives additional support for
prior studies, such as Susanto et al. (2023), Martin and Javalgi (2016), Su et al. (2013), Murray et al.
(2011), and Krasnikov and Jayachandran, among others (2008).
Page 891
SEMIT 2023
According to Table 4, hypothesis H7 is supported since SMU is positively linked with MC (β = 0.50, p
= 0.000). The result that SMU is positively linked with MC is not surprising. With the rising significance
of social media platforms in the digital age of today, it has become essential for businesses to develop a
strong online presence and utilize social media platforms for marketing reasons. When businesses
actively connect with customers on social media, not only can they enhance brand exposure, but they
may also get vital insights about customer preferences and behavior. This allows businesses to better
customize their marketing tactics and offers to their target audience's requirements and expectations.
Therefore, it is essential for businesses to cultivate their social media community managers in order to
remain competitive in the market and promote development. This conclusion is consistent with those of
Mason et al. (2021) and Susanto et al. (2023), who found that marketing via social media platforms
attracted more attention.
Finally, hypothesis H8 results show that EC has an indirect positive effect on SMEs’ performance through
MC (β = 0.191, p-value = 0.001). This discovery provides a fresh perspective on how MCs mediation
impacts the relationship between EC and SME performance, thereby enriching the research findings of
previous studies such as Ahmed et al. (2014), Ju et al. (2018), Dirgiatmo et al. (2019), Wiwoho et al.
(2020), and Susanto et al. (2023). This finding highlights the importance of nurturing a culture of
entrepreneurship in SMEs. The study suggests that firms with a strong EC are better equipped to develop
effective MCs, which in turn leads to improved business performance. This indicates that investing in
developing an entrepreneurial mindset among employees and fostering a culture of innovation can yield
significant benefits for SMEs. By recognizing the mediating role of MCs, policymakers and managers
can focus on building a robust marketing function within their organizations and provide opportunities
for employees to enhance their entrepreneurial skills. Overall, this finding underscores the need for SMEs
to prioritize the development of an EC as a critical foundation for achieving sustained growth and success.
6. Conclusion
SEM-PLS was utilized to analyze the association between EC and the performance of Iranian SMEs. The
study also considered the mediating effects of MC and SMU. The primary result of this study is that the
European Union has a positive effect on the performance of Iran's SMEs. In addition, the SMU and MC
impact the interaction between the two as intermediary factors.
SMEs that develop an atmosphere that supports innovation, risk-taking, and creativity are more likely to
achieve greater levels of company success, according to the research. In addition, this study emphasizes
the significance of social media and management communication in mediating the link between an EC
and SME performance. By effectively utilizing social media platforms and implementing effective
marketing strategies, SMEs can increase their reach and visibility, attract new customers, and ultimately
improve their bottom line. These findings have important implications for policymakers and business
leaders in Iran. Not only can encouraging the growth and development of entrepreneurship assist
individual enterprises, but it may also contribute to the economic growth and development of the nation
as a whole. Governments can support this by providing funding and resources for entrepreneurial training
programs and incentives for starting small businesses.
For SME owners and managers, this research underscores the need to focus not only on product or service
quality but also on creating a supportive environment that fosters innovation and creativity. Additionally,
investing in MCs and building a strong online presence can be crucial for achieving long-term success.
This study contributes to the expanding body of knowledge regarding the importance of entrepreneurship
and innovation for economic progress and growth. By understanding the factors that facilitate SME
success, policymakers and business leaders can make informed decisions that promote sustainable
business growth and create a thriving economy.
Overall, this study indicates that SMEs should concentrate on enhancing their MCs and social media
abilities in order to boost sales, profitability, and sustainability. Social media may give real-time market
intelligence, enabling proactive analysis and the implementation of suitable measures to establish
customer connections and enhance corporate performance. Additionally, it might be advantageous to
encourage entrepreneurs to create company concepts and provide social marketing training to those who
lack the necessary abilities. Social media networks such as Facebook, Twitter, and Instagram may connect
clients and expedite business transactions, particularly when paired with a home delivery service, due to
Page 892
SEMIT 2023
the prevalence and widespread usage of smartphones and Internet-based applications. Therefore, SMU
and MC are crucial for both consumer connections and commercial transactions.
According to the results of this study, here are some practical recommendations for SMEs in Iran to
improve their performance:
• Foster an EC: It is important for SMEs to cultivate a culture that encourages innovation, risk-
taking, and creativity. This can be achieved by promoting a work environment that values ideas
and encourages employees to take initiative.
• Improve social media presence: Since SMU and MCs serve as intermediate variables between an
EC and a SME's success, it is suggested that SMEs concentrate on enhancing their social media
presence and MCs. They can do this by creating engaging content, using relevant hashtags,
collaborating with influencers, and analyzing their metrics to determine what works best.
• Ensure work-life balance: It is essential for small business owners to consider their own and their
workers' health. This can be done by establishing clear boundaries between work and personal
life and providing flexible working arrangements.
• Seek policy support: SMEs can benefit from policy support that provides them with access to
resources and facilitates their growth. Therefore, it is recommended that SME owners seek out
policies and initiatives that support the development of SMEs in Iran.
Future research recommendations are as follows:
• Consider niche categories: Future study can investigate particular elements, such as cost issues,
that may be critical to the sustainability of each SME group. In order to make decisions that will
contribute to their company's growth and long-term viability, SME owners need conduct research
and gain a thorough awareness of their industry's distinctive qualities.
• Longitudinal design testing: Since the current study is based on cross-sectional data, it is
suggested that future research employ a longitudinal design to examine the mode. This will result
in a more accurate grasp of the model's genuine dynamic effects.
• Compare and contrast emerging markets: The market in Iran is the topic of this investigation.
However, future research may concentrate on other emerging areas in order to simplify the
comparison and contrast of SME behavior in these markets. This will enable SME owners to
acquire insight into the methods that worked for their competitors in other marketplaces and
adopt them within their own enterprises.
References
Ahmed, M.U., Kristal, M.M. and Pagell, M. (2014), “Impact of operational and marketing capabilities on
firm performance: evidence from economic growth and downturns”, International Journal of
Production Economics, Vol. 154, pp. 59-71.
Ainin, S., Parveen, F., Moghavvemi, S., Jaafar, N.I. and Mohd Shuib, N.L. (2015), “Factors influencing
the use of social media by SMEs and its performance outcomes”, Industrial Management and Data
Systems, Vol. 115 No. 3, pp. 570-588.
Amoah, J., Belas, J., Khan, K.A. and Metzker, Z. (2021), “Antecedents of sustainable SMEs in the social
media space: a partial least square-structural equation modeling (PLS-SEM) approach”, Management
and Marketing. Challenges for the Knowledge Society, Vol. 16 No. 1, pp. 26-46
Appel, G., Grewal, L., Hadi, R. and Stephen, A.T. (2020), “The future of social media in marketing”,
Journal of the Academy of Marketing Science, Vol. 48 No. 1, pp. 79-95.
Bau, F. and Wagner, K. (2015), “Measuring corporate entrepreneurship culture”, International Journal of
Entrepreneurship and Small Business, Vol. 25 No. 2, pp. 231-244.
Bell, B.S. and Kozlowski, S.W. (2002), “A typology of virtual teams: implications for effective
leadership”, Group and Organization Management, Vol. 27 No. 1, pp. 14-49.
Page 893
SEMIT 2023
Churchill, G.A. Jr (1979), “A paradigm for developing better measures of marketing constructs”, Journal
of Marketing Research, Vol. 16 No. 1, pp. 64-73.
Dimitratos, P., Lioukas, S. and Carter, S. (2004), “The relationship between entrepreneurship and
international performance: the importance of domestic environment”, International Business Review,
Vol. 13 No. 1, pp. 19-41.
Dirgiatmo, Y., Abdullah, Z. and Ali, R.H.R.M. (2019), “The role of entrepreneurial orientation in
intervening the relationship between social media usage and performance enhancement of exporter
SMEs in Indonesia”, International Journal of Trade and Global Markets, Vol. 12 No. 2, pp. 97-129.
Dodokh, A. and Al-Maaitah, M.A. (2019), “Impact of social media usage on organizational performance
in the Jordanian dead sea cosmetic sector”, European Journal of Business and Management, Vol. 11
No. 2, pp. 75-91.
Elsharnouby, T.H. and Elbanna, S. (2021), “Change or perish: examining the role of human capital and
dynamic marketing capabilities in the hospitality sector”, Tourism Management, Vol. 82, p. 104184.
Gillin, P. and Schwartzman, E. (2010), Social Marketing to the Business Customer: Listen to Your B2B
Market, GenerateMajor Account Leads, and Build Client Relationships, 1st ed., JohnWiley and Sons.
Hair, J.F., Black, W.C., Babin, B.J. and Anderson, R.E. (2014), Multivariate Data Analysis, 7th ed.,
Pearson Prentice Hall, New Jersey.
Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M. (2019), “When to use and how to report the results
of PLS-SEM”, European Business Review, Vol. 31 No. 1, pp. 2-24.
Henseler, J., Ringle, C.M. and Sarstedt, M. (2015), “A new criterion for assessing discriminant validity
in variance-based structural equation modeling”, Journal of the Academy of Marketing Science, Vol.
43 No. 1, pp. 115-135.
Hertzog, M.A. (2007), “Considerations in determining sample size for pilot studies”, Research in Nursing
and Health, Vol. 31 No. 2, pp. 180-191.
Hisrich, R.D. and Peters, M.P. (1986), “Establishing a new business venture unit within a firm”, Journal
of Business Venturing, Vol. 1 No. 3, pp. 307-322.
Hornsby, J.S., Kuratko, D.F. and Montagno, R.V. (1999), “Perception of internal factors for corporate
entrepreneurship: a comparison of Canadian and US managers”, Entrepreneurship Theory and
Practice, Vol. 24 No. 2, pp. 9-24.
Ireland, R.D., Hitt, M.A. and Sirmon, D.G. (2003), “A model of strategic entrepreneurship: the construct
and its dimensions”, Journal of Management, Vol. 29 No. 6, pp. 963-989.
Jabbour, C.J.C., de Sousa Jabbour, A.B.L., Sarkis, J. and Godinho Filho, M. (2019), “Unlocking the
circular economy through new business models based on large-scale data: an integrative framework
and research agenda”, Technological Forecasting and Social Change, Vol. 144, pp. 546-552.
Ju, M., Jin, J.L. and Zhou, K.Z. (2018), “How can international ventures utilize marketing capability in
emerging markets? Its contingent effect on new product development”, Journal of International
Marketing, Vol. 26 No. 4, pp. 1-17.
Kanter, R.M. (1985), “Supporting innovation and venture development in established companies”,
Journal of Business Venturing, Vol. 1 No. 1, pp. 47-60.
Khan, M.A., Zubair, S.S., Rathore, K., Ijaz, M., Khalil, S. and Khalil, M. (2021), “Impact of
entrepreneurial orientation dimensions on performance of small enterprises: do entrepreneurial
competencies matter?”, Cogent Business and Management, Vol. 8 No. 1, p. 1943241.
Kohtamaki, M., Hakala, H., Partanen, J., Parida, V. and Wincent, J. (2015), “The performance impact of
industrial services and service orientation on manufacturing companies”, Journal of Service Theory
and Practice, Vol. 25 No. 4, pp. 463-485.
Page 894
SEMIT 2023
Krasnikov, A. and Jayachandran, S. (2008), “The relative impact of marketing, research-and-

development, and operations capabilities on firm performance”, Journal of Marketing, Vol. 72 No. 4,
pp. 1-11.
Lumpkin, G.T. and Dess, G.G. (2001), “Linking two dimensions of entrepreneurial orientation to firm
performance: the moderating role of environment and industry life cycle”, Journal of Business
Venturing, Vol. 16 No. 5, pp. 429-451.
Martin, S.L. and Javalgi, R.R.G. (2016), “Entrepreneurial orientation, marketing capabilities and
performance: the moderating role of competitive intensity on Latin American international new
ventures”, Journal of Business Research, Vol. 69 No. 6, pp. 2040-2051.
Mason, A.N., Narcum, J. and Mason, K. (2021), “Social media marketing gains importance after covid-
19”, Cogent Business and Management, Vol. 8 No. 1, p. 1870797.
Minguzzi, A. and Passaro, R. (2001), “The network of relationships between the economic environment
and the entrepreneurial culture in small firms”, Journal of Business Venturing, Vol. 16 No. 2, pp. 181-
207.
Morgan, N.A., Slotegraaf, R.J. and Vorhies, D.W. (2009), “Linking marketing capabilities with profit
growth”, International Journal of Research in Marketing, Vol. 26 No. 4, pp. 284-293 Murray
Murray, J.Y., Gao, G.Y. and Kotabe, M. (2011), “Market orientation and performance of export ventures:
the process through marketing capabilities and competitive advantages”, Journal of the Academy of
Marketing Science, Vol. 39 No. 2, pp. 252-269.
Nguyen, H. T. N., Nguyen, H. T. T., Truong, A. T. L., Nguyen, T. T. P., & Nguyen, A. V. (2023).
Entrepreneurial culture and innovative work behaviour: the mediating effect of psychological
empowerment. Journal of Entrepreneurship in Emerging Economies, 15(2), 254-277.
Obschonka, M. (2017), “The quest for the entrepreneurial culture: psychological big data in
entrepreneurship research”, Current Opinion in Behavioral Sciences, Vol. 18, pp. 69-74.
Odoom, R. and Mensah, P. (2019), “Brand orientation and brand performance in SMEs: the moderating
effects of social media and innovation capabilities”, Management Research Review, Vol. 42 No. 1,
pp. 155-171.
Omar, N.A., Aris, H.M. and Nazri, M.A. (2016), “The effect of entrepreneurial orientation, innovation
capability and knowledge creation on firm performance: a perspective on small scale entrepreneurs”,
Jurnal Pengurusan, Vol. 48, pp. 187-200.
Omar, N.A., Kassim, A.S., Shah, N.U., Shah Alam, S. and Che Wel, C.A. (2020), “The influence of
customer value co-creation behavior on SME brand equity: an empirical analysis”, Iranian Journal of
Management Studies, Vol. 13 No. 2, pp. 165-196.
Papadopoulos, T., Baltas, K.N. and Balta, M.E. (2020), “The use of digital technologies by small and
medium enterprises during COVID-19: implications for theory and practice”, International Journal of
Information Management, Vol. 55, p. 102192.
Parveen, F., Jaafar, N.I. and Ainin, S. (2016), “Social media’s impact on organizational performance and
entrepreneurial orientation in organizations”, Management Decision, Vol. 54 No. 9, pp. 2208-2234.
Pascal, V.J. and Shin, S. (2015), “The role of marketing capabilities and entrepreneurial orientation on
SME performance”, Journal of International Marketing Strategy, Vol. 3 No. 1, pp. 37-54.
Qalati, S.A., Yuan, L.W., Khan, M.A.S. and Anwar, F. (2021), “A mediated model on the adoption of
social media and SMEs’ performance in developing countries”, Technology in Society, Vol. 64, p.
101513.
Ratten, V. (2020), “Coronavirus (covid-19) and entrepreneurship: changing life and work landscape”,
Journal of Small Business and Entrepreneurship, Vol. 32 No. 5, pp. 503-516.
Page 895
SEMIT 2023
Sadeghi, A. (2018). Success factors of high-tech SMEs in Iran: A fuzzy MCDM approach. The Journal
of High Technology Management Research, 29(1), 71-87.
Sarstedt, M., Hair, J.F. Jr., Cheah, J.H., Becker, J.M. and Ringle, C.M. (2019), “How to specify, estimate,
and validate higher-order constructs in PLS-SEM”, Australasian Marketing Journal, Vol. 27 No. 3,
pp. 197-211.
Sigala, M. (2012), “Web 2.0 and customer involvement in new service development: a framework, cases
and implications in tourism”, In Web, Vol. 2, pp. 25-38.
Singh, N., Bamel, U. and Vohra, V. (2020), “The mediating effect of meaningful work between human
resource practices and innovative work behavior: a study of emerging market”, Employee Relations:
The International Journal, Vol. 43 No. 2, pp. 459-478.
Sopha, B.M., Jie, F. and Himadhani, M. (2021), “Analysis of the uncertainty sources and SMEs’
performance”, Journal of Small Business and Entrepreneurship, Vol. 33 No. 1, pp. 1-27.
Su, Z., Ahlstrom, D., Li, J. and Cheng, D. (2013), “Knowledge creation capability, absorptive capacity,
and product innovativeness”, R&D Management, Vol. 43 No. 5, pp. 473-485.
Susanto, P., Hoque, M. E., Shah, N. U., Candra, A. H., Hashim, N. M. H. N., & Abdullah, N. L. (2023).
Entrepreneurial orientation and performance of SMEs: the roles of marketing capabilities and social
media usage. Journal of Entrepreneurship in Emerging Economies, 15(2), 379-403.
Tajudeen, F.P., Jaafar, N.I. and Ainin, S. (2018), “Understanding the impact of social media usage among
organizations”, Information and Management, Vol. 55 No. 3, pp. 308-321.
Trainor, K.J. (2012), “Relating social media technologies to performance: a capabilities-based
perspective”, Journal of Personal Selling and Sales Management, Vol. 32 No. 3, pp. 317-331.
Trainor, K.J., Andzulis, J.M., Rapp, A. and Agnihotri, R. (2014), “Social media technology usage and
customer relationship performance: a capabilities-based examination of social CRM”, Journal of
Business Research, Vol. 67 No. 6, pp. 1201-1208.
Wales, W.J., Kraus, S., Filser, M., Stöckmann, C. and Covin, J.G. (2021), “The status quo of research on
entrepreneurial orientation: conversational landmarks and theoretical scaffolding”, Journal of
Business Research, Vol. 128, pp. 564-577.
Wamba, S.F. and Carter, L. (2016), “Social media tools adoption and use by SMEs: an empirical study”,
Social Media and Networking: Concepts, Methodologies, Tools, and Applications, IGI Global, pp.
791-806.
Wang, Z. and Kim, H.G. (2017), “Can social media marketing improve customer relationship capabilities
and firm performance? Dynamic capability perspective”, Journal of Interactive Marketing, Vol. 39,
pp. 15-26.
Wardi, Y., Susanto, P., Abror, A. and Abdullah, N.L. (2018), “Impact of entrepreneurial proclivity on
firm performance: the role of market and technology turbulence”, Pertanika Journal of Social Sciences
and Humanities, Vol. 26, pp. 241-250.
Wiwoho, G., Suroso, A. and Wulandari, S. (2020), “Linking adaptive capability, product innovation and
marketing performance: results from Indonesian SMEs”, Management Science Letters, Vol. 10 No.
10, pp. 2379-2384.
Yun, J.J., Zhao, X., Jung, K. and Yigitcanlar, T. (2020), The Culture for Open Innovation Dynamics,
Multidisciplinary Digital Publishing Institute.
Page 896
SEMIT 2023
E-Shopping Sites Preference Analysis with Moosra Method

Vuslat Erat1 Babek Erdebilli2
1
TOBB University of Economics and Technology, Söğütözü 43, 06510 Çankaya,Ankara
2
Department of Industrial Engineering, Ankara Yıldırım Beyazıt University, 06010, Ankara, Turkey
verat@etu.edu.tr
Abstract
With the developing technology, the use of the internet has become widespread. From this point
of view, in this study, Turkey's five most popular e-shopping sites (Trendyol, Hepsiburada,
Amazon, Getir, Morhipo) were selected as alternatives. Seven criteria were used in the study:
site design, product variety, reliability, detailed filtering, service quality, ease of use of the site,
and price. As a method, Multi-Objective Optimization on the basis of Simple Ratio Analysis
(MOOSRA) from Multi-Criteria Decision Making (MCDM) method was selected. While
determining the criterion weights, the Pairwise Comparison method was used. As a result of the
study, it has been seen that the most important criterion is the price, and the least important
criterion is the site design. When the MOOSRA method was applied, it was concluded that
Trendyol was the best e-shopping site.
Keywords: Multiple Criteria Decision Making (MCDM), MOOSRA, e-shopping, Pairwise
Comparison
1.Introduction
With the rapid development of the internet, e-shopping has become more preferred instead of
traditional shopping. Many criteria play a role in customers choosing an e-shopping site. Among
these criteria, seven criteria were selected to be used in the study: site design, product variety,
reliability, detailed filtering, service quality, ease of site use and price. Alternatives were
determined as five popular e-shopping sites in Turkey; Trendyol, Hepsiburada, Amazon, Getir
and Morhipo.
The second part of this study, located after the introduction, includes a literature review about
online shopping. In the third part, method used is included. In the 4th part of the study, there is
application of the method. In the discussion and conclusion section, the result of the method is
discussed.
In order to weight the criteria and evaluate the alternatives, the data to be used was obtained by
survey method. Necessary questions were asked to 52 e-shopping site users belonging to
different age groups, and the answers given were used in the application.
This section includes literature research on online shopping, Azhar, Mulyani, Hutahaean and
Mayhaky (2022) conducted a best e-commerce company decision analysis using the MOOSRA
method. In the study, 8 alternatives were determined and 4 criteria were used. The criteria are
ease of use, ease of access, reliability and price.
Guru, Nenavani, Patel and Bhatt (2020) ranked India's best B2C e-commerce brands using the
AHP method. As criteria, three risks perceived by shoppers were determined. These are
performance risk, financial risk and time loss risk.
Page 897
SEMIT 2023
Mardhia and Normawati (2018) determined the 3 most popular shopping sites in Indonesia and
used 5 criteria. These criteria are price, seller location, seller reputation, number of products sold
and transportation support.
3. Method
The MOOSRA method was developed by Das, Sarkar and Ray in 2012.
The steps of the method are as follows;
Step 1: Creating the decision matrix
Step 2: Normalizing the decision matrix
Step 3: Finding the total performance score (𝑆i ) of the alternatives
Step 4: The determined 𝑆i values are sorted from largest to smallest. The one with the highest
score is chosen as the best alternative.
4. Application
This section includes the application of MOOSRA method
According to the results obtained from the survey, the decision matrix in Table 1 was created.
Table 23: Decision Matrix

Criteria
Benefit Benefit Benefit Benefit Benefit Benefit Cost
(max) (max) (max) (max) (max) (max) (min)
Criterion 0,03 0,04 0,18 0,05 0,23 0,10 0,37
Weights
Alternatives SD PV R DF SQ ESU P
Trendyol 6,50 7,69 7,04 7,27 7,12 7,50 3,65
Hepsiburada 5,65 6,50 6,46 6,35 6,27 6,04 4,04
Amazon 5,88 6,42 6,46 6,23 6,50 6,23 3,96
Getir 6,31 5,85 6,92 6,00 6,58 7,50 4,58
Morhipo 4,54 4,81 5,73 5,77 5,62 5,46 4,88
Table 24: Normalized Decision Matrix

Criteria
Criterion 0,03 0,04 0,18 0,05 0,23 0,10 0,37
Weights
Trendyol 0,4997 0,5438 0,4815 0,5124 0,4947 0,5084 0,3845
Hepsiburada 0,4344 0,4597 0,4419 0,4476 0,4356 0,4095 0,4256
Amazon 0,4521 0,4540 0,4419 0,4391 0,4516 0,4223 0,4172
Getir 0,4851 0,4137 0,4733 0,4229 0,4572 0,5084 0,4825
Morhipo 0,3491 0,3402 0,3919 0,4067 0,3905 0,3701 0,5141
Page 898
SEMIT 2023
Table 25: Weighted Normalized Matrix

Criteria
Criterion 0,03 0,04 0,18 0,05 0,23 0,10 0,37
Weights
Trendyol 0,0150 0,0218 0,0867 0,0256 0,1138 0,0508 0,1423
Hepsiburada 0,0130 0,0184 0,0795 0,0224 0,1002 0,0409 0,1575
Amazon 0,0136 0,0182 0,0795 0,0220 0,1039 0,0422 0,1543
Getir 0,0146 0,0165 0,0852 0,0211 0,1052 0,0508 0,1785
Morhipo 0,0105 0,0136 0,0705 0,0203 0,0898 0,0370 0,1902
Tablo 26: Performance Score of Alternatives

Alternatives Sum of Sum of Cost 𝑆i Ranking
Benefit Criteria
Criteria
Trendyol 0,3137 0,1423 2,2048 1
Hepsiburada 0,2745 0,1575 1,7431 3
Amazon 0,2793 0,1543 1,8097 2
Getir 0,2934 0,1785 1,6438 4
Morhipo 0,2418 0,1902 1,2712 5
According to the MOOSRA method, the ranking of the alternatives is determined as Trendyol-
Amazon- Hepsiburada- Getir-Morhipo.
5. Discussion and Conclusion

In the study, an attempt was made to analyze the decisions of five e-shopping sites used in
Turkey. In order to access the data, questions were asked to 52 e-shopping site users with
different gender and age distributions. The obtained answers were used in the Pairwise
Comparison method to perform criterion weighting. For this purpose, MOOSRA was selected
from the Multi-Criteria Decision Making methods.
In this study, the order in which five e-shopping sites widely used in Turkey are preferred by
customers, in line with the determined criteria, was determined as Trendyol-Amazon-
Hepsiburada-Getir-Morhipo.
References
Azhar, Z., Mulyanı̇ , N., Hutahaean, J. ve Mayhaky, A. (2022). Sistem pendukung keputusan
pemilihan e-commerce terbaik menggunakan metode moosra . Jurnal Medı̇ a İnformatı̇ ka
Budı̇ darma .
Dorfeshan, Y., Mousavı̇ , S. M., Mohagheghı̇ , V. ve Vahdanı̇ , B. (2018). Selecting project-critical
path by a new interval type-2 fuzzy decision methodology based on multımoora, moosra
and tpop methods. Computers & İndustrı̇ al Engı̇ neerı̇ ng .
Page 899
SEMIT 2023
Fan, L., Yager, R. R., Mesı̇ ar, R. ve Jı̇ n, L. (2020). Two-level multi-criteria comprehensive
evaluation for preference vectors in online shopping platform evaluation. Journal Of
İntellı̇ gent & Fuzzy Systems .
Febransyah, A. ve Goni J. I. C. (2022). Measuring the supply chain competitiveness of e-
commerce industry in ındonesia. Competı̇ tı̇ veness Revı̇ ew .
Guru, S., Nenavanı̇ , J., Patel, V. ve Bhatt, N. (2020). Ranking of perceived risks in online
shopping. Decı̇ sı̇ on.
Page 900
SEMIT 2023
Category Sort wıth TODIM Method, an Applıcatıon for Retaıl Sector E-Commerce Sıte
Ayşe Hümeyra Akar1 Babek Erdebilli2
1
TOBB University of Economics and Technology, Söğütözü 43, 06510 Çankaya,Ankara
2
Department of Industrial Engineering, Ankara Yıldırım Beyazıt University, 06010, Ankara, Turkey
aysehumeyraakar@etu.edu.tr
Abstract
In recent years, classical multi-criteria decision-making methods, including the widely used TODIM
(Topsis-based multi-criteria decision-making), have been frequently used to address various real-world
challenges. Based on prospect theory, TODIM is one of the effective methods that can be used to solve
ranking problems.
Considering the increase in online sales, evolving global dynamics and the impact of the epidemic, e-
commerce has become indispensable for businesses. In this context, establishing an online presence is
crucial. The easiest way to step into the digital world is through successful e-commerce sites. Therefore,
it is important to strategically organize product categories in a way that improves user navigation and
improves the overall shopping experience. Our goal is to rank product categories on e-commerce
platforms with the TODIM method, optimize user experience and increase sales performance.
Keywords: TODIM, e-commerce, ranking problems, MCDM
1. Introduction
E-commerce, one of the rapidly growing and developing sectors, is adopting various strategies to improve
customer experience and gain competitive advantage as consumers' online shopping habits increase. One
of these strategies is to organize the product categories offered on the home page of e-commerce sites.
E-commerce sites have a wide range of options to offer various user experiences to attract users' attention
and enable them to easily access their products. These experiences include many options such as sorting
products by gender, color options, and displaying customer-based ads. However, no studies have been
done on how to rank product categories and how this ranking can directly affect user experience and
conversion rates. Correctly sorting product categories allows users to find the products they are looking
for more easily and improves site navigation. Therefore, it is important for e-commerce sites to have a
user-friendly interface.
As a result, this study will enable the product categories on the e-commerce site to be ranked using the
TODIM method and it will be possible to determine which category should be highlighted for the home
page layout.
This study will be based on the TODIM method, which is one of the multi-criteria decision-making
methods. In this context, there are many studies conducted in the past. The TODIM method is a method
based on the "Expectation Theory" first put forward by Kahneman and Tverski in the 1970s. Unlike other
multi-criteria decision-making methods, in the TODIM method, the measurement value calculated by
applying the "Expectation Theory" paradigm is used instead of the highest value [1].
Page 901
SEMIT 2023
Figure 6. Expectation Theory
Although there is no direct study on the category ranking of e-commerce sites, there are many studies on
product ranking. The volume of online product reviews has increased in recent years [2]. Some studies in
this field have investigated product rankings according to customer feedback [3] by obtaining information
about emotional tendencies [4]. There are also studies in the literature to sort monthly sales according to
categories [5].
3. TODIM Method
The main purpose of the TODIM method is to determine the degree of dominance of each alternative
over the others by using the potential theory-based utility function [29].
The mathematical formulations can be systematically summarized as follows:
Step 1: Creation of the decision matrix.
Step 2: Normalization of decision matrix. These values calculated separately for beneficial criteria and
unbeneficial criteria.
Step 3: The reference criterion is selected. This selected criterion has the highest degree of importance
and relative weights are calculated
Step 4: Calculation of degreeof superlatives.
Step 5: Calculation of the global value of ith alternative.
Step 6: Listing and evaluating of the results.
4. Case Study
This research involves a process in which the data used to evaluate the performance of the e-commerce
site is obtained through SQL queries and then these data are processed and analyzed.
Steps for category sorting with the TODIM method are shown below.
Table 1. Decision Matrix
Beneficial Non Beneficial

Category Name Number of Average Quantity The amount of Age range Number of Product
Customers Amount (TL) stock Types
Baby 3562 756702,34 7034,00 578487 3 4498
Meat-Chicken-Fish 17270 3427899,45 51823,97 15173466,36 4 34015
and Seafood
Page 902
SEMIT 2023
Food 36993 8052296,25 323430,6 168337155,9 3 205862

Non-Food Products 33606 530364,83 218354 36487704 1 47572
Beverages 33003 4115302,59 233169,3 92831677,84 2 100182
Paper- Cosmetics 24865 3433588,18 92050 17125405 2 63330
Greengrocer 27218 3339204,27 336314,6 255660284 5 214803
Healthy Living 351 18740,08 602 123746 3 424
Products
Salad Dressings 2047 47457,12 3148 72746 2 2285
Milk, Breakfast 34380 6817943,35 166307,4 142174481,7 4 114340
Cleaning 24080 3473696,39 89499,36 13460561 2 70066
Bakery products 15144 506058,8 58660,23 49123770 1 34834
Weights 0,268865 0,20682 0,17235 0,156681 0,104454 0,09083
The weights were determined by an expert by the SWARA method in this study.
Table 2. Normalized Matris
0,014105869 0,021921167 0,00445079 7,25E-02 5,85E-02 7,04E-02

0,068390893 0,099303985 0,032791815 2,77E-03 4,39E-02 9,32E-03
0,146495907 0,233269709 0,204651914 2,49E-04 5,85E-02 1,54E-03
0,133083055 0,015364319 0,138164326 1,15E-03 1,75E-01 6,66E-03
0,130695116 0,1192176 0,14753877 4,52E-04 8,77E-02 3,16E-03
0,098467838 0,099468784 0,058244988 2,45E-03 8,77E-02 5,00E-03
0,107785949 0,096734544 0,212804358 1,64E-04 3,51E-02 1,48E-03
0,001389994 0,000542888 0,000380918 3,39E-01 5,85E-02 7,47E-01
0,008106321 0,001374801 0,001991909 5,77E-01 8,77E-02 1,39E-01
0,136148171 0,19751132 0,105231667 2,95E-04 4,39E-02 2,77E-03
0,095359161 0,100630692 0,056631061 3,12E-03 8,77E-02 4,52E-03
0,059971725 0,01466019 0,037117484 8,54E-04 1,75E-01 9,10E-03
Then, the weights of the criteria:
Wcr 1 0,76923363 0,641028 0,582749707 0,3885 0,337827534
Table 4. Degrees of Distinction
Baby -25,81
Meat-Chicken-Fish and Seafood -23,4
Food -10,86
Non-Food Products -13,32
Beverages -11,69
Paper- Cosmetics -16,45
Greengrocer -18,69
Healthy Living Products -19,8
Salad Dressings -18,58
Milk, Breakfast -16,02
Cleaning -16,71
Bakery products -21,2
Page 903
SEMIT 2023
Table 5. Recommended Category Ranking
Food
Beverages
Non-Food Products
Milk, Breakfast
Paper- Cosmetics
Cleaning
Salad Dressings
Greengrocer
Healthy Living Products
Bakery products
Meat-Chicken-Fish and Seafood
Baby
5. Conclusion and Future Research

In this study, the TODIM method was applied to solve the problem of ranking the categories of websites
and to suggest a more effective ranking. Sales data of the e-commerce site was used to sort between
categories. According to the results obtained from the TODIM method, the most visible category should
be the food category.
In future studies, in addition to home page editing in the field of e-commerce, many areas such as gender-
based home page editing, color editing, and advertising campaigns can be studied. Getting customers to
browse websites designed for them can increase customer loyalty and also bring in new customers.
References
1. Gomes L, Rangel L. “An Application of the TODIM method to the multicriteria rental evaluation of
residential properties”. European Journal of Operational Research, 193(1), 204-211, 2009.
2. Zha, Z.J., Yu, J., Tang, J., Wang, M. and Chua, T.S. (2014), “Product aspect ranking and its
applications”, IEEE Transactions on Knowledge & Data Engineering, Vol. 26 No. 5, pp. 1211-1224.
3. J. Santhana Preethi, A.M. Abirami, A. Askarunisa, G. Sathya Priya, E. Sankaragomathy, Applying
MCDM Techniques for Ranking Products Based on Online Customer Feedback, International Journal
of Knowledge Based Computer Systems Volume 3, Issue 2, December 2015.
4. Chong Wu, Dong Zhang Ranking products with IF-based sentiment word framework and TODIM
method. School of Management, Harbin Institute of Technology, Harbin, China.
Jia-Wei Gong, Hu-Chen Liu, Xiao-Yue You, Linsen Yin, An integrated multi-criteria decision making
approach with linguistic hesitant fuzzy sets for E-learning website evaluation and selection,
Applied S
Page 904
SEMIT 2023
SP24: Advancement and Future of Technologies
Dissecting the Economic Feasibility and Life Cycle Assessment of Battery Electric and Internal
Combustion Engine Vehicles: A Case Study of India
Amrut P. Bhosaleab, Sachin A. Mastudb, Muzammil Beparic, Ketaki Bhosaled
a
Rajarambapu Institute of Technology, Rajaramnagar, Uran Islampur, Maharashtra, India
b
Veermata Jijabai Technological Institute, Matunga, Mumbai, India.
c
Bharati Vidyapeeth's College of Engineering, Kolhapur, India.
d
D. Y. Patil College of Engineering and Technology, Kolhapur, India.
Abstract:
Fuel supplies for conventional vehicles are vulnerable to scarcity, which could ultimately lead to an increase
in fuel prices. There has been a realization regarding national energy security as a result of these high
gasoline costs, which further increase the overall cost of ownership. Additionally, the emissions from
burning conventional fuels make consideration of the already pressing environmental issues necessary. On
the other hand, because they have low running costs and no tailpipe emissions, electric vehicles are being
considered as a viable alternative to conventional automobiles. But when a vehicle's whole life cycle is
taken into account, the common-sense belief that electric vehicles are cheaper and emit no emissions may
be misleading. Therefore, it is necessary to consider both the economic and environmental elements of
whether electric vehicles are a viable alternative to conventional automobiles. In this article, a life cycle
analysis—both economic and environmental—between battery-electric and conventional automobiles is
presented in the context of India. For financial analysis, a Total Cost of Ownership (TCO) model is created
to show how compatible battery-electric vehicles are. OpenLCA software, which is based on the ReCePi
2016 technique, is used to conduct the environmental analysis for all impact categories at both the mid-
point and end-point levels. According to the findings, electric vehicles are more expensive than
conventional automobiles in India based on current data and regulations. However, it is shown that electric
vehicles have cost parity and can sometimes even become more inexpensive than conventional automobiles
by using specific optimizing factors in sensitivity analysis. The results of environmental studies show that
battery electric vehicles emit fewer greenhouse gases (GHGs) than do conventional automobiles. Battery
electric vehicles, however, had less of an impact in ten of the eighteen impact categories that were
examined, and they even have a lower impact score at the end-point level.
Keywords: Life Cycle Analysis, Economic Compatibility, Total cost of Ownership, Electric vehicles in
India.
1 Introduction
Researchers from all around the world are being pushed to develop more sustainable transportation
solutions as the drive to convert to carbon-neutral transportation becomes more intense. Consequently, the
global vehicle fleet is witnessing an abundance of electric vehicles. The main factors driving the adoption
of electric vehicles are the avoidance of exhaust emissions from tailpipes and the achievement of national
energy security through the use of an affordable fuel substitute [1]. Among the EV research community,
there is currently no agreement on the sustainability, viability, techno-economic, and environmental aspects
of electric vehicles (EVs). It's still uncertain if electric cars would work as a reliable substitute and, if so, in
what circumstances. According to Gautam [2], fully electrified vehicles won't be practical by 2040, but
Page 905
SEMIT 2023
electric-hybrid cars will be a preferable choice over traditional fuel-powered automobiles. Conversely, for
a variety of reasons, including technical, financial, and environmental, experts Hauschild et al. [3] support
electric vehicles as the superior choice.
The inconsistent adoption of battery-powered cars around the world is shown in the Global EV Outlook
report [4]. Similarly, a variety of literary works highlight the obstacles to the widespread adoption of electric
vehicles, as well as the public's perceptions and regulations aimed at promoting their use. Lieven et al. [5],
Heidrich et al. [6]. These publications also emphasize that the absence of reliable facilities, financial
concerns, and environmental issues are the main obstacles to the widespread use of electric vehicles. These
obstacles become increasingly important for nations such as India. Given that the Indian auto industry is
price-sensitive, a comprehensive cost study is crucial. Furthermore, it is evident from the regional energy
mix—which primarily consists of coal—that environmental analysis is also necessary in order to power
electric vehicles.
The sluggish adoption of electrically powered cars is also intimately related to India's national energy
security. India is the third-most populous crude oil importer because less than 20% of the country's crude
oil is derived from domestic sources [7]. The fiscal year 2022 saw a nearly twofold increase in the crude
oil import cost, amounting to $120 billion [8]. Furthermore, India is ranked second in Asia for CO2
emissions, with a significant majority of these emissions coming from the transportation sector, primarily
from road transportation [9]. To increase the market share and accelerate the deployment of electric
vehicles, these complex barriers need to be carefully considered. But even after the Indian government
implemented programs like the National Electric Mobility Mission Plan (NEMMP 2020) and Faster
Adoption and Manufacturing of (hybrid &) Electric vehicles (FAME I & II), there was a gap among the
plan and its implementation, which undermined the mission and made it necessary to think more deeply
about the problem. The worldwide market contribution of battery-powered two-wheelers is consistent with
the worldwide marketplace, but the numbers for electric four-wheelers are significantly worse [10]. In order
to determine the sustainability implications, this article compares the economic viability of battery electric
vehicles (BEVs) with internal combustion engine vehicles (ICEVs). Additionally, an environmental
evaluation is conducted to compare the two vehicle types.
2 Literature Review
The literature on the environmental and economic aspects of battery-electric cars is highlighted in this
section. According to J. Seixas et al.'s literature [11], one of the main obstacles to the widespread use of
electric vehicles is their somewhat greater cost than internal combustion engines. On the other hand, electric
cars may be more affordable than internal combustion engines (ICEVs) when taking into account their
entire life cycle. Persuading potential buyers of the long-term advantages of electric cars can increase sales
of the type and entice them to change their minds. et al. Bhosale [12]. Numerous academic works disclose
that variations in local rules, exemptions, and limits can cause the Total Cost of Ownership (TCO) to
fluctuate across various parts of the world. Even after accounting for incentives, battery electric vehicles
seem to be more expensive in China than ICEVs Zhao et al. [13]. However, when certain intangible costs
are taken into consideration, ICEVs become far more costly than BEVs, according to a related study done
in China Diao et al. [14]. Even in Singapore and Australia, battery powered cars have higher total costs
than internal combustion engines (ICEVs) for many reasons. Electric vehicles with batteries have a higher
total cost of ownership (TCO) in Singapore due to high local and customs duties as well as excise taxes
[12], while in Australia, peak and off-peak electricity prices cause the TCO of battery electric vehicles to
be significantly lower. Kara and others [15].
Norway is at the top of the European economic chart, which is explained by the combination of clean and
renewable energy. Lévay et al al. [16]. Following Norway on the list are France, the UK, and the
Netherlands, whose TCOs are cheaper than ICEVs but still somewhat higher than Norway's. In contrast to
ICEV, other nations fare poorly when the entire cost of ownership is taken into account. When compared
Page 906
SEMIT 2023
to the TCO of BEVs in Japan, roaming incentives are an elixir for BEVs in the US and the UK, according
to comprehensive research by Palmer et al. [17]. Potk'any et al. [18] reported similar findings in Slovakia,
where providing subsidies had a significant impact on lowering the TCO of BEVs. While some studies
categorize TCO comparisons based on geographic regions, others differentiate TCO based on governing
characteristics such gasoline prices, incentives, and annual kilometers traveled (AKT). Many governments
use incentives as their main policy instrument, but how effective is this strategy has to be examined?
According to Levey et al. [16], incentives are determined to be adequate in some areas, such as Norway
and other European and American nations. Lieven et al. [19] observed similar things in roughly 20 other
countries, demonstrating that the monetary grant is the most valued incentive. However, incentives don't
seem to work all that well in various parts of the world. In Germany, it is determined that a 4,000 Euro
incentive is insufficient. According to Bubeck et al. [20], battery-electric cars are still 5% more expensive
than gas-powered cars. In Germany, it is determined that a 4,000 Euro incentive is insufficient. According
to Bubeck et al. [20], battery-electric cars are still 5% more expensive than gas-powered cars. After
receiving government grants et al. Tseng [21]. A year AKT, or kilometers traveled, is a metric that also
influences Wu et al.'s TCO on Battery Electric Vehicles [22]. AKT has been referenced in literature which
is upto 20,000 kms in order to cover the entire expenses between the automobiles Mitropoulos et al. [23]
recommend at least 23,000 kms AKT for BEVs to surpass gasoline in efficiency and Italian cars using
diesel fuel. Some literatures have placed emphasis on the battery price, battery replacement cost, and its
prognosis, in addition to incentives and yearly kilometers traveled. When battery prices for battery electric
vehicles drop to less than $300/kWh and €240/kWh for various parts of the world, the vehicles become
economically competitive. Newbery, David, and others [24]. In addition to the cost of the batteries, high
depreciation rates are also thought to contribute to the total cost of ownership (TCO), making battery-
electric vehicles more expensive than internal combustion engines. The literature shows that while battery
electric vehicles (BEVs) were initially welcomed with gusto, buyers are becoming hesitant to purchase
BEVs due to their higher total cost of ownership. One of the most challenging problems is minimizing the
total cost of ownership (TCO), which is determined by a complex interplay of characteristics that needs to
be addressed giving priority.
When it comes to battery-electric vehicles, environmental damage is the main worry in addition to the
barrier of economic compatibility. The battery-electric cars produce no emissions is a frequent trick that
should be watched out for as the tail pipe's emissions are being transferred to Kalghatgi is another website
[4]. But we can't ignore it. The traditional automobiles' emissions have been a significant problem since
there are more cars on the road Lucas [25] but examining the emissions of both kinds of vehicles grounds
will be used to assess these cars' prospects. When driving a battery-electric car, greenhouse emissions come
from a variety of sources, including material extraction, transit, the amount of energy used in manufacturing
and its nature Ma et al. (energy-mix utilized) [26]. However, rather than focusing on just one effect category
(greenhouse gases), a whole life cycle study should be carried out taking into account all of the impact
categories in order to obtain a panoramic emission impact. It makes sense that, if a clean energy source is
utilized to charge the BEVs, emissions from battery electric vehicles might be substantial during the
manufacture phase of the two phases (use and production). Although the battery pack is thought to
contribute significantly to overall emissions, differences in emissions from various battery technologies
also need to be taken into consideration. According to research by Premrudee et al. [27], Li-Ion batteries
have the least number of pollutants when compared to Lead Acid batteries. According to Held et al. [28],
the primary cause of the twice-as-high global warming potential impact category for battery-electric
vehicles compared to equivalent internal combustion engines is battery production. Remarkably, with
Belgium's existing energy mix, electric cars are discovered to be releasing fewer emissions. While gasoline
and diesel vehicles emit more than 200 g/km CO2eq, battery powered cars emit less than 52 g/km CO2eq
[29]. Picirelli de Souza et al.'s [30] observations from Brazil, where they estimated the emissions from
BEVs compared with ethanol mixed gasoline, show that while BEVs perform well overall in terms of
emissions, ethanol blended fuel has less of an influence on human toxicity and global warming. In the
global warming category, BEV emissions exceed 140 g/km CO2eq, while ICEVs using an ethanol mix fuel
Page 907
SEMIT 2023
emit less than 100 g/km CO2eq. The average greenhouse gas emissions from battery electric cars (BEVs)
in China are 210 g/km CO2eq, according to Zhou et al. [31]. The carbon footprints from BEVs vary across
China, ranging from 160 to 245 g/km CO2eq. Comparably, Qiao et al. [32] propose that by choosing the
battery recycling option, BEV GHG emissions can be lowered below 50% of those of ICEVs, while current
evidence shows that BEV emissions are 18% lower than those of equivalent ICEVs. In tandem with the
development of BEVs, laws pertaining to their emissions must be upheld. As the technology behind electric
cars develops, regulations (such as ICEV emission standards, such as EURO 2, 3...) will become stricter,
hence it will be crucial to monitor BEV emissions lest the whole transformation in transportation modes be
jeopardized.
3 Motivation and Objectives

In order to replace the whole fleet of vehicles with 100% electric vehicles by 2030, the government created
the National Electric Mobility Mission Plan 2020 (NEMMP). Regrettably, taking into account the present
pace of the market, the goal has been reverted to electrifying thirty percent of the whole fleet of electric
vehicles by 2030. The original impetus for the study was these "Unachieved Targets". Second, the "Current
Market Status" shows that conditions for electric four-wheelers in India are still unfriendly, even with the
introduction of incentive schemes like FAME I & II. The third factor is "Anxiety," which arises from the
fact that the Indian market is extremely receptive to cost. This leads to erroneous concerns such as "Are
electric vehicles in India cost-competitive with traditional automobiles?" Therefore, the primary goal of
this study is to expedite the launch of electric vehicles in India, allay public anxiety (deception questions),
pursue the target as soon as possible, and revitalize the electric 4-wheeler market in India.
In the context of India, this article examines how battery electric vehicles compare economically and
environmentally against vehicles with similar internal combustion engines. to conduct life cycle
environmental and economic analyses while taking into account more accurate data and Indian conditions
as opposed to depending solely on general information. Additionally, conduct sensitivity analysis taking
into account various regulating factors and recommend appropriate inputs to policy drafters to increase the
acceptance of BEVs.
4 Methodoloy
4.1 Economic Analysis
Life cycle economics accounts for the Total Cost of Ownership (TCO), which includes all costs from the
time a vehicle is built to the end of its useful life. In order to conduct a fair comparison, the pairing
approach—as recommended by Gilmore et al. [33]—is chosen to estimate total cost of ownership. With the
matching vehicle approach, automobiles with nearly comparable dimensions and attributes are compared
in order to confirm that the estimations are produced using the right datum. Figure 1 shows the TCO block
diagram, several phases, and significant obstacles. As shown in Figure 1, the TCO is divided into three
phases: acquisition phase, utilization phase, and end-life phase.
The total cost of ownership (TCO) for the base case is first evaluated. Afterwards, optimization is chosen
with the aid of sensitivity analysis to determine the optimum alternative that may be recommended to the
strategy drafters. Evaluations of sensitivity take into account a number of factors, including Annual
Kilometer Traveled (AKT), Battery Replacement Cost, Incentives/Subsidies, Finance Interest Rates, and
EV-PV Integration (Battery electric vehicles utilized in combination with solar energy alternative).
Page 908
SEMIT 2023
Figure 1: Approaching the Total Cost of Ownership

The TCO/km is computed by Equation 1:
Here, IC is the possession cost, RC is Operating cost, PVC and PSS are the associated with the solar energy
(cost and sale respectively), RS is salvage cost, and I is subsidies/Incentives/ other exemptions and B is
loading principal balance, D is annual distance travelled in kilometers, n is number of years vehicle used.
The acronyms Finance TCO (FTCO) and Purchase TCO (PTCO) are used to calculate the effects of
financing and not financing on the TCO. The total cost of acquisition is the amount the owner must pay to
purchase the car. Ex-showroom costs, government taxes, road taxes, registration fees, additional ancillary
charges from the vehicle's manufacturer company, and interest if the car is financed are all included. During
the vehicle's use phase, expenses include fuel, upkeep, parking fees, municipal entry taxes, tire replacement,
and other parts. Estimating the fuel cost (Petrol/Diesel and Electricity) involves taking the timeline's
inflation rates into consideration. General inflation rates, or around 3.98%, are applied whenever necessary
to account for inflation. According to literature, electric vehicles typically require 30% less maintenance
than internal combustion engines (ICEVs), and their tires have a 50,000 km lifespan. Ultimately, the salvage
phase includes the car's market value for that year, which is calculated using the car's depreciation. In line
with the trend in the literature, which is reported by Messagie et al. [34], battery-electric cars lose value
faster than internal combustion engines.
Page 909
SEMIT 2023
Figure 2: Various Phases in an Environmental Life Cycle Analysis of an automobile application.

The assumptions and usage statistics for the vehicles can be summed up as follows: Two sets of cars are
compared in this TCO model; pair A comprises a TATA pair that contrasts the TATA Nexon EV and TATA
Nexon Diesel/Petrol. Hyundai Kona EV and Hyundai Creta are contrasted for pair B. To complete the
paired vehicle technique, the appropriate model variant is chosen. For the basic case, with a loan tenure of
five years and an annual kilometer traveled of 15,000 km, the financing interest rate of 9.7% is taken into
account for FTCO computation. According to FAME I & II standards, pair A BEV is eligible for an
incentive of ₹115,000, and pair B BEV is presumed to have an incentive of ₹ 400,000. The battery
replacement is done after 8 years/ 160,000 km.
4.2 Environmental Analysis
The Life Cycle Environmental Analysis, which follows the ISO 14040 and ISO 14044 European standard
series, assesses a product's emissions during its whole life. Here, the consequences and emissions from
ICEVs and BEVs that have been modelled are contrasted in relation to the Indian setting. The extraction of
materials, transportation, the production and consumption of energy, the usage phase, and lastly the end of
life phase are all included in the emissions.
The comparison of emissions from both internal combustion engines (ICEVs) and BEVs from birth to death
is known as "cradle-to-grave" analysis.
An examination that solely considers the route of fuel from production to combustion or consumption is
referred to as a "Well to Wheel" analysis; well-to-tank and tank-to-wheel analyses are also included in this
phrase. Environmental analysis is divided into three parts, much like economic analysis is, as Figure 2
shows. Pollution from the mining and processing of raw materials to the construction of the vehicle and
other essential parts like battery packs are included in the manufacturing phase. Emissions from the
extraction of fuel, the production of power, and other incidental elements like tire and brake wear are all
taken into account throughout the use phase. The vehicle's disassembly and recycling of its required pieces
are the last steps in the disposal phase.
Page 910
SEMIT 2023
Figure 3: Interlinking of various processes for BEV production

With the aid of the software OpenLCA, which estimates emissions for each effect category, the LCA is
carried out. Ecoinvent-3 is the compatible database that is utilized for data inventory. The modeling process
incorporates the essential product flows, systems, and processes before being compared to a real project.
Figure 3 depicts the prototype product system for battery-electric vehicles. It shows the main processes,
with over 9000 processes or flows interacting with the end flow diagram. ReCePi 2016 is the impact
assessment approach that was employed; all 16 effect categories are taken into consideration in the
evaluation, both at the mid- and end-points.
5.Results and Discussion

5.1 Economic Analysis results
5.1.1 Base Case
With a base case of 15,000 km and a 15-year life span, the established methodology is used to assess the
Total Cost of Ownership. The loan has a five-year term and a 9.75% financing interest rate. The replacement
of the battery is taken care of until the eighth year. The FTCO and PTCO findings for the base case for both
couples are shown in Figures 4a and 4b. In this case, FBEV stands for the financed TCO of a BEV, and
P-BEV for the purchased TCO of a BEV. In a similar vein, F-ICEV D/P stands for financed TCO and P-
ICEV D/P for purchased TCO associated with ICEVs (diesel/petrol).
Figures 4a and 4b's results indicate that for parings in this study, the financed overall cost (TCO) of the
BEV (F-BEV) is higher than the corresponding TCO of the ICEV (F-ICEV D & F-ICEV P). In a similar
vein, for both pairs, the purchased TCO of the BEV is higher than the comparable ICEV. Nonetheless, there
is a slight difference in the pair A (TATA pair) purchase TCO between the I.C.V. purchases TCO. However,
even pair A's financed TCO is significantly lower than ICEV TCOs. Sadly, pair B (the Hyundai pair) has
dismal performance for both purchased and financed choices, highlighting the TCO discrepancies more
Page 911
SEMIT 2023
clearly. An additional intriguing discovery indicates a significant increase in the TCOs of BEVs in both
pairings at the eighth year. This sudden rise is due to the high battery replacement cost.
Figure 4: TCO for TATA pair Figure 5: TCO for Hyundai pair
5.1.2 Sensitivity Analysis

According to the base case study, no pair's BEV becomes more cost-effective than an ICEV or reaches cost
equality. In order to accomplish the economic compatibility of BEVs, it becomes necessary to examine the
TCO in combination with a few extra variables. Aspects of the sensitivity analysis include Policy Governed:
offering incentives (discussed in the preceding section), Technology governed: a 50% decrease in battery
replacement costs; Extended Policy governed: a reduction in financing interest rates (ROI to 6% from
9.75%) Consumer-Governed: using BEVs with EV+PV integration and imposing a higher Annual
Kilometer Traveled (AKT) of 20,000 AKT rather than 15,000 AKT.
According to the findings of the sensitivity analysis, pair A (TATA)'s purchased TCO of a BEV is lower
than that of an ICEV for each of the previously mentioned parameters. When the "reduction battery
replacement cost" feature is taken into account, the funded TCO for pair A BEV is also equated with the
TCO of an ICEV. When examining every variable separately, the funded TCO of pair A BEV is still higher
than the TCO of an ICEV. In every variable, outcomes for pair B don't appear sides the electric vehicles.
In every sensitivity characteristic listed above, the TCO of the BEV in pair B (the Hyundai) is still greater
than that of a comparable ICEV.
It only happens once the BEV's purchase TCO in pair B approaches the ICEV's buy TCO for the "reduction
battery replacement cost" criterion just a little bit. The sensitivity analysis reveals that, in the majority of
circumstances, the individual variables are insufficient to make the pair's costs competitive; this can only
be accomplished by combining the aforementioned characteristics.
As seen in Figure 6, a "compatibility wheel" is created for this. This wheel shows the many variable
configurations that were used, as well as how both couples performed under 12 different sets of situations.
The planes indicate finance and purchase TCO for the corresponding pairs, while the orbits represent a
collection of conditions and parameters.
The wheel consists of 12 different sets of variables (called "orbits") that range from O to K and are listed
in the sensitivity analysis. For example, Orbit "A," also known as "Policy Governed," has a set of settings
consisting of 15,000 AKT, incentives, a finance interest rate of 9.75%, and no integration of EV and PV.
Compare the financed/purchased TCO of a BEV with the financed/purchased TCO of an ICEV within a
Page 912
SEMIT 2023
pair (for example, Planet 1 compares the FTCO of a Hyundai pair of BEVs with the FTCO of an ICEV).
These comparisons are known as the planets.
Page 913
SEMIT 2023
Figure 6: The Economic Compatibility Wheel for the TCO analysis of battery electric vehicles compared to ICEVs.
Page 914
SEMIT 2023
The following is an explanation of the economic compatibility wheel using an example: The parameters
considered are 15000 AKT with incentives, 6% interest rate, and no integration of PV and EV. An orbit
"B" is observed. For the previously described orbit (set of conditions), Planet 1 (the FTCO of the Hyundai
pair BEV compared to the FTCO of the ICEV) is situated in a costly zone or quadrant, meaning that the
FTCO of the Hyundai pair BEV is more than 10% more expensive than the FTCO of the Hyundai pair
ICEV. Planet 2, on the other hand, is in the moderately expensive range when compared to the PTCO of
the ICEV. This implies that the PTCO of the Hyundai pair BEV is 2–10% more expensive than the PTCO
of the Hyundai pair ICEV.
The FTOC and PTCO of TATA pair BEVs are situated in the economic and near cost parity zones,
respectively, in contrast to ICEV. In a nutshell, the economic compatibility wheel shows how each pair's
TCO functions in different scenarios in a simple and visual manner.
The graphic shows that the Hyundai pair BEV's FTCO (planet 1) mainly falls in the expensive zone when
compared to ICEV FTCO. Almost all of the orbits for the TATA pair, with the exception of the base orbit
"O," put BEV's PTCO (planet 4) in the economic zone. Known as orbits 'G' and 'K,' we have found two
intriguing orbits (set of conditions) in which every planet is in the economic zone. This implies that the
sponsored and acquired TCOs (FTCO and PTCO) of the TATA pair and the Hyundai pair are more
economical than the TCOs of ICEV. We also find that in orbit 'J', the Hyundai pair FTCO reaches the
cost parity zone, which is around 2% economical. In conclusion, the Hyundai dual BEV financing option
may also prove to be economical when the proper arrangements are made.
5.2 Environmental Analysis results
Figure 7 (a-g) shows the climatic implications for the simulated ICEV and BEV in the OpenLCA for the
Indian setting. To further aid in mitigating the problem, the total impacts are further divided into the pre-
use and usage phases. This allows for a clearer knowledge of the primary emitter source.
The results' primary effect categories—which are frequently discussed in the literature—are displayed in
Figure 7(a–g).
(a) (b)
Page 915
SEMIT 2023
(c) (d)
(e) (f)
(g)
Figure 7: Impact assessment of BEV in comparison to ICEV for a) Global warming b) fine
particulate matter c) human toxicity (carcinogenic) d) human toxicity (non-carcinogenic) e) Fossil
resource scarcity f) Mineral resource scarcity g) Ionizing radiation.
Page 916
SEMIT 2023
The global warming gases (GHG) emissions are the subject of the most frequent plot literature discussion;
it is discovered that BEVs emit fewer GHGs than ICEVs.
Wu et al.'s findings are also consistent with other literature [40]. Though in a distributed study, the GHG
emissions from BEVs during the use phase are somewhat higher than those from ICEVs, the GHG
emissions from BEVs are still only 15% of those from ICEVs. This could be explained by the fact that
coal-powered facilities already make up a sizable portion of India's electricity mix.
Beyond the production of greenhouse gases, BEVs do well in impact categories like resource scarcity
and ionizing radiation. In ionizing radiation impact categories, where emissions from BEVs are nearly
half those of ICEVs, the emissions gap is observed rather clearly. When there is a shortage of mineral
resources, the BEV has very high emissions during the production phase, but these emissions are closed
during use, resulting in lower overall emissions than ICEVs. However, when taking into account all the
impact areas, ICEV do not always end out negatively. ICV emit equivalent levels of pollution that are
lower in categories including human toxicity and finite particle matter. In contrast with BEVs, IEVs'
emissions in the human carcinogenic toxicity impact category during use are essentially insignificant.
The findings show that neither the BEV nor the ICEV benefit from the emissions in the various effect
categories. Figure 8's relative graph, which displays all of the impact categories for both automobiles,
can be used to get deeper understanding. It is discovered that the BEV emits less in nearly ten of the
eighteen categories; the category with the least variation in emissions each category is terrestrial
acidification, at roughly 7%. The impact category of freshwater eutrophication has the largest range, with
nearly over 73% of the total.
Figure 8: LCA relative results.

Last but not least, a single point score is obtained from the end-point analysis, showing that battery
electric vehicles are more cost-effective when the total score—0.59 kpt—is taken into account, compared
to the traditional vehicle's 2.12 kpt score.
6. Conclusion
For the Indian setting, a life cycle analysis comparison between a modeled or chosen battery electric
vehicle and an internal combustion engine is conducted, taking into account both environmental and
economic factors. This page assists end users in addressing the highly complex question of whether BEVs
Page 917
SEMIT 2023
are environmentally and economically viable in India. Regarding economic compatibility, the base case
taken into consideration here is significantly closer to actual data of the current situation and finds that
BEVs are not financially compatible with ICEV for both financed and purchased options in both pairings.
In contrast, pair A's (TATA) total cost of ownership (TCO) gap is lower than pair B's (Hyundai).
Additionally, a sensitivity analysis was conducted with various parameters considered in order to reduce
the total cost of ownership (TCO) of BEVs in both pairs. The examination of sensitivity clarifies that
aside from the component handled by technology (a 50% decrease in the cost of replacing batteries) not
another each of the parameters alone can determine the BEVs' TCO. To be more affordable than ICEVs
or almost cost parity. It's just with the parameter combinations from the sensitivity analysis the BEV
attains cost parity and even gains enough efficiencies that the similar ICEV that is seen in the
compatibility diagram. The compatibility wheel shows the financial performance of cars with twelve
distinct parameter configurations, two of which (in orbits G and K) allow the BEV's total cost of
ownership to enter the profitable range. With an annual mileage of roughly 20,000 km, the usage of
EV+PV integration, 6% loan ROI, and incentives, battery electric vehicles seem to be the answer.
Environmental friendliness analysis was used to resolve the disturbed greenhouse gas emissions, and the
result was an agreement that, in the Indian context, the greenhouse gas emissions from BEVs are lower
than those from ICEVs. BEVs do not, however, perform better than ICEVs in all impact categories when
it comes to emissions. Although not in the vast majority of cases, ICEVs have less emissions in 8 out of
the 18 impact areas. The class with the least amount of emissions difference between BEVs and ICEVs
is the one that affects terrestrial acidification the least, while freshwater eutrophication varies the most.
By dividing the overall emissions into the pre-use and usage phases, one may identify the real significant
source of emissions, which in turn helps to reduce or remove the source of concern. The BEV has a
significantly lower overall emission score (0.59 kpt) than the ICEV (2.12 kpt) based on the end-point
study. Lastly, switching to renewable energy sources can make this battle for environmental compatibility
worse. Recycling the products and using correct production techniques will also assist to carefully reduce
the emissions from battery-electric vehicles.
References
[1] K. Petrauskiene˙ , M. Skvarnavicˇi_ute˙ , J. Dvarioniene˙ , Comparative environmental life cycle
assessment of electric and conventional vehicles in Lithuania, J. Clean.Prod. 246.
[2] Kalghatgi, G., 2018. Is it really the end of internal combustion engines and petroleum in transport?
Appl. Energy 225 (May), 965.
[3] Hauschild, M.Z., Rosenbaum, R.K. and Olsen, S.I., 2018. Life cycle assessment. Springer
International Publishing, Cham.
[4] Global EV Outlook 2022, available at https://www.iea.org/reports/global-ev-outlook-2022.
[5] Lieven, T., Mühlmeier, S., Henkel, S., Waller, J.F., 2011. Who will buy electric cars? An empirical
study in Germany. Transport. Res. Transport Environ. 16 (3), 236–243.
[6] O. Heidrich, G. A. Hill, M. Neaimeh, Y. Huebner, P. T. Blythe, and R. J. Dawson, “How do cities
support electric vehicles and what difference does it make?,” Technol. Forecast. Soc. Change, vol. 123,
no. May, pp. 17–23, 2017
[7] India’s crude imports from Russia up 7.2 times in April-May 2022. Available at :
https://www.newindianexpress.com/business/2022/jul/14/indiascrude-imports-from-russia-up-72-
timesin-april-may-2022-
2476374.html#:~:text=India%20is%20the%20world's%20third,and%2014%25%20from%20the%20US
[8] India's oil import bill doubles to $119 billion in FY22. Available at:
https://economictimes.indiatimes.com/industry/energy/oilgas/indias-oil-import-bill-doubles-to-usd-119-
bn-infy22/articleshow/91049349.cms
[9] Total CO2 emissions-India. available at: https://www.iea.org/countries/india
Page 918
SEMIT 2023
[10] Sales of electric vehicles across India from financial year 2020 to 2022, by type, available at:
https://www.statista.com/statistics/1234761/india-electric-vehiclesales-by-type/
[11] Seixas, J., Simões, S., Dias, L., Kanudia, A., Fortes, P., Gargiulo, M., 2015. Assessing the cost-
effectiveness of electric vehicles in European countries using integrated modeling. Energy Pol. 80, 165–
176.
[12] A. P. Bhosale, S. Sharma, and S. A. Mastud, “Characterizing the economic competitiveness of
battery electric vehicles in India,” Asian Transp. Stud., vol. 8, no. 274, p. 100069, 2022.
[13] Zhao, X., Doering, O.C., Tyner, W.E., 2015. The economic competitiveness and emissions of battery
electric vehicles in China. Appl. Energy 156, 666–675.
[14] Diao, Q., Sun, W., Yuan, X., Li, L., Zheng, Z., 2016. Life-cycle private-cost-based competitiveness
analysis of electric vehicles in China considering the intangible cost of traffic policies. Appl.Energy 178,
567–578.
[15] Kara, S., Li, W., Sadjiva, N., 2017. Life cycle cost analysis of electrical vehicles in Australia.
Procedia CIRP 61, 767–772.
[16] Lévay, P.Z., Drossinos, Y., Thiel, C., 2017. The effect of fiscal incentives on market penetration of
electric vehicles: a pairwise comparison of total cost of ownership. Energy Pol. 105 (October2016), 524–
533.
[17] Palmer, K., Tate, J.E., Wadud, Z., Nellthorp, J., 2018. Total cost of ownership and market share for
hybrid and electric vehicles in the UK, US and Japan. Appl. Energy 209 (November 2017), 108–119.
[18] Potkány, M., Lesníková, P., 2019. The amount of subsidy for the electric vehicle in Slovakia
through a strategic cost calculation.Transport. Res. Proc. 40, 1168–1175.
[19] Lieven, T., Mühlmeier, S., Henkel, S., Waller, J.F., 2011. Who will buy electric cars? An empirical
study in Germany. Transport. Res. Transport Environ. 16 (3), 236–243.
[20] Bubeck, S., Tomaschek, J., Fahl, U., 2016. Perspectives of electric mobility: total cost of ownership
of electric vehicles in Germany.Transport Pol. 50, 63–77.
[21] Tseng, H.K., Wu, J.S., Liu, X., 2013. Affordability of electric vehicles for a sustainable transport
system: an economic and environmental analysis. Energy Pol. 61, 441–447.
[22] Wu, L., Liu, S., Liu, D., Fang, Z., Xu, H., 2015a. Modelling and forecasting CO2 emissions in the
BRICS (Brazil, Russia, India, China, and South Africa) countries using a novel multi-variable grey
model. Energy 79 (C), 489–495
[23] Mitropoulos, L.K., Prevedouros, P.D., Kopelias, P., 2017. Total cost of ownership and externalities
of conventional, hybrid and electric vehicle. Transport. Res. Proc. 24, 267–274.
[24] Newbery, D., Strbac, G., 2016. What is needed for battery electric vehicles to become socially cost
competitive? Econ. Transport. 5.
[25] Lucas A, Civa CA, Neto RC. Life cycle analysis of energy supply infrastructure for conventional
and electric vehicles. Energy Pol 2012;41:537–47
[26] Ma H, Balthasar F, Tait N, Riera-Palou X, Harrison A. A new comparison between the life cycle
greenhouse gas emissions of battery electric vehicles and internal combustion vehicles. Energy Pol
2012;44:160–73
[27] K. Premrudee, U. Jantima, A. Kittinan, L. Naruetep, K. Kittiwan, B. Sudkla, Life cycle assessment
of lead acid battery. case study for Thailand, Environ. Prot. Eng. 39 (1) (2013) 101–114,
[28] M. Held and M. Baumann, “Towards Life Cycle Sustainability Management,” Towar. Life Cycle
Sustain. Manag., pp. 535–546, 2011, doi: 10.1007/978-94-007-1899-9.
[29] J. Van Mierlo, M. Messagie, and S. Rangaraju, “Comparative environmental assessment of
alternative fueled vehicles using a life cycle assessment,” Transp. Res. Procedia, vol. 25, pp. 3435–3445.
Page 919
SEMIT 2023
[30] L. L. P. de Souza, E. E. S. Lora, J. C. E. Palacio, M. H. Rocha, M., “Comparative environmental life

cycle assessment of conventional vehicles with different fuel options, plug-in hybrid and electric vehicles
for a sustainable transportation system in Brazil,” J.Clean. Prod., vol. 203, pp. 444–468, 2018.
[31] G. Zhou, X. Ou, and X. Zhang, “Development of electric vehicles use in China: A study from the
perspective of life-cycle energy consumption and greenhouse gas emissions,” Energy Policy, vol. 59, no.
2013, pp. 875–884, 2013.
[32] Q. Qiao, F. Zhao, Z. Liu, and H. Hao, “Electric vehicle recycling in China: Economic and
environmental benefits,” Resour. Conserv. Recycl., vol. 140, no. July 2018, pp. 45–53, 2019.
[33] Gilmore, E.A., Patwardhan, A., 2016. Passenger vehicles that minimize the costs of ownership and
environmental damages in the Indian market. Appl. Energy 184, 863–872.
[34] Messagie, M., Lebeau, K., Coosemans, T., Macharis, C., van Mierlo, J., 2013. Environmental and
financial evaluation of passenger vehicle technologies in Belgium. Sustainability 5 (12),5020–5033.
Page 920
SEMIT 2023
Optimization of Controller Parameter Multilevel Converter Based STATCOM for Reactive

Power Compensatio
Hamed Atyia Soodi 1, a), Yousif A. Hamad 2,3, b *), and Ardam H. Mohammed 1, c)
1
2
Artificial Intelligence Laboratory, Siberian Federal University, Krasnoyarsk, 660074 Krasnoyarsk,
Russia,
3
Department of Computer Science, Collage of Computer Science and Information Technology,
University of Kirkuk, Kirkuk, Iraq
b)
Corresponding author: y.albayati@uokirkuk.edu.iq
a)
hamid.attia@sadiq.edu.iq
c)
ardam.haseeb@sadiq.edu.iq
Abstract
Static Synchronous Compensator (STATCOM) is a self-tuning proportional-integral (PI) controller that
has been developed to handle the voltage and reactive power of Cascaded H-Bridges (CHB-
STATCOM) in order to improve the power system. The purpose of optimizing controller parameters
for a Multilevel Converter-based STATCOM (Static Synchronous Compensator) in the context of
reactive power compensation is to enhance the performance and efficiency of the STATCOM system
in regulating the power factor and voltage of an electrical system. This controller was introduced in this
article. This research presented that the STATCOM should have the capability to be operate both in
inductive and capacitive mode of operation and absorb or injects the desired amount of reactive power
into the grid. The reactionary power that is received or supplied contributes to the improvement of the
power system. The benefits of the self-tuning PI controller have been optimized to their full potential
through the use of Accelerated Particle Swarm Optimization (APSO) strategies. The STATCOM is
implemented into the electrical system, and its reaction to varying reactive and voltage levels is
analyzed as part of the process of determining whether or not the optimization techniques are successful.
Two controllers are now in operation at the STATCOM. The first controller is in charge of managing
the DC voltage (which is employ me to produce reference signal depended on differences in phase
angle), and the second controller is in charge of controlling the reactive power. Both controllers are in
use at the same time. (Which is what is utilized to produce the modulation index). The output of the
controller is then transferred to the phase change carrier, which is followed by pulse width modulation
in a sinusoidal pattern. Because of this, the power converter switch may now function properly.
Keywords: APSO; Reactive power compensation; STATCOM; PI control; Phase shift pulse width
modulation.
Introduction
Brought on by non-linear loads have a deleterious effect on both power quality and grid
dependability over time. Because total system power is determined by adding together the system's
active and reactive powers, an increase in reactive power is inextricably related to a similar increase in
total power. In [1, 2], several strategies for increasing the grid's response power are described. Because
of how successfully they control the overall system's reactive power, shunt Flexible AC Transmission
System (FACTS) devices are widely used. A coordinated power source consisting completely of solid-
state components is used to power STATCOM and other FACTS systems. Connected in series with the
power supply, it is an excellent replacement for the cumbersome reactive components of conventional
static VAR compensators (SVC). It also beats SVC in terms of the dependability of the nation's
electrical grid. [3]. STATCOM's primary role is to control voltage at the point of common coupling
(PCC) during a power system changeover. The STATCOM obtains these reactive currents from the
power infrastructure in a controlled fashion. Intricate design challenges for these layered STATCOM
[4] stem from the need for a large number of instruments to track DC voltage and uneven voltage across
the DC-link capacitor.[5] A conventional two-loop management system was proposed. The PCC
voltage is controlled by the outer loop, and the inverter current is controlled by the interior loop with a
Page 921
SEMIT 2023
steady-state inaccuracy of zero. Capacitor banks were among the first components of electrical
apparatus to be installed in an effort to regulate power supply. However, with the help of modern
technology, system managers have access to more flexible and effective tools that can manage huge
quantities of power. Flexible AC transmission systems are a specific type of these tools. (FACTS).
These devices are carefully installed on selected electric buses, and they are used to quickly alter various
system settings in order to boost system efficiency, control power flow, and enhance system reliability
[6]. They are large-scale regulators that modify system-wide parameters like reactive power, voltage,
and phase angle. [1].
Numerous researches have determined PI benefits through trial and error to reach an optimal
working point, power system under STATCOM control exhibits significantly non-linear behavior.
Linear controllers are simple to implement but difficult to fine-tune in this architecture. Transient
performance requirements must be addressed by fine tweaking of controller settings. In this research,
we use the non-linear and non-gradient Simplex algorithm to optimize the control system parameters
of a STATCOM built from two converters based on six-pulse voltage sources. The performance index
serves as the basis for the objective function of the system, which is written in terms of the actual and
ideal values of the system's variables. Several simulation tests examining step-response performance
have validated the optimized performance of the STATCOM control system, which consists of two PI
controllers and two compensators. With the right settings in place, the system responds quickly and
precisely [2, 7]. This is not a good option in practice due to the radical changes in system features
brought about by the inclusion of STATCOM devices or the improvement of the transmission
architecture. [8].
In the realm of heuristics for parameter optimization, accelerated particle swarm optimization
(APSO) [9] is a well-known method. It has been used to modify controller gains [10,11]. It is possible
for the gains of an APSO-based PI scheduler to auto-tune in response to variations in the system and
demand. This will result in a significantly more adaptable PI system. The consensus time of APSO is
low (A Moreover, it outperforms SVC in terms of power system reliability [3]) because it does not
necessitate offline training like an artificial neural network. During a transition in the electricity system,
STATCOM's main function is to regulate the voltage at the point of common coupling (PCC). The most
important development challenges for these layered STATCOM [7] involve extracting controlled
reactive currents from the electrical infrastructure. are the many instruments used to track DC voltage
and imbalanced voltage at the DC link capacitor. It also ignores the reasoning principles of the
approximate control, which points to PSO's superior performance. [12, 25].
Optimizing the controller parameters for a Multilevel Converter-Based Static Synchronous
Compensator for reactive power compensation can be a challenging task, and there are several potential
problems and issues that may arise during the optimization process, Multilevel converters are inherently
complex due to the multiple voltage levels they can generate. Optimizing the controller parameters for
such converters requires a deep understanding of their operation and the interaction between different
voltage levels. There may be trade-offs between different control objectives, such as voltage regulation,
harmonic mitigation, and response time. Finding the right balance among these objectives can be
challenging. Power systems are subject to variability due to changes in load, generation, and network
conditions. The optimized parameters that work well under one set of conditions may not perform
optimally under different conditions. The accuracy of the system model used for optimization is crucial.
Inaccurate models can lead to suboptimal parameter values or instability in the real system. Over-
optimization, where parameters are tuned too precisely for specific operating conditions, can lead to
instability or poor performance when conditions deviate from the optimized scenario. Changes to
controller parameters can impact the safety and reliability of the power system. Safety measures and
risk assessments must be considered during the optimization process.
A Multilevel Converter-Based STATCOM presents various technical and practical challenges. It
requires a combination of expertise in control theory, power systems, and optimization techniques,
along with careful consideration of system complexity and constraints. Successful optimization should
aim to strike a balance between different control objectives while accounting for the dynamic and
variable nature of power systems.
Page 922
SEMIT 2023
In this investigation, we use Accelerated Particle Swarm Optimization to design a self-tuning PI

controller for CHB-STATCOM [13]. (APSO). By tweaking the original PSO algorithm, we get
accelerated PSO. The primary goal of this research is to develop a robust CHB-STATCOM scheduler
that can effectively account for disturbances within a few cycle time windows of their occurrence.
STATCOM
To sum up, STATCOM is an apparatus that counteracts the reactive power of an electrical grid. It
may behave as both a generator and a sink of reactive energy. To complement the tiny synchronous
condenser compensator (STATCOM) that controls the PCC's power factor (PF), the proposed inverter
does double duty by transforming DC power from the DC link into AC power for the mains supply and
syphoning off enough reactive power from the grid to increase the PCC's PF [14]. The major power line
and the primary power line servicing the PCC are both illustrated in Fig.1 as part of the basic
STATCOM arrangement. Transformer, capacitor, coupling reactor, and regulated power supply make
up the voltage source converter (VSC) in the centre of a STATCOM.
In addition to its more common name, "voltage source converter," or "VSC," can also refer to a
device that converts direct current (DC) to alternating current (AC). It can produce asymmetrical AC
power with user-defined phase and angle of rotation. A VSC can be charged while the DC battery stores
energy. By using coupling reactance to link the sources to the system and regulate the current, the
converter output can be joined to the power system. The coupling transformer regulates the input
voltage of the STATCOM converter and the output voltage of the PCC.
FIGURE 1. The grid connection of STATCOM
System software A STATCOM can be used either in VAR mode or with voltage control turned on.
The STATCOM aids the PCC in keeping the voltage stable during dynamic disruptions, allowing the
PCC to provide superior transient properties. By switching into this state, a STATCOM will either:
• When the system voltage is high, the STATCOM will absorb reactive power.
• When the system voltage is low, the STATCOM will generate and inject reactive power.
Page 923
SEMIT 2023
The STATCOM reactive power production is either held steady or adjusted based on VAR mode
monitoring of the reactive power and comparison to a standard number.
Multi-level simulation of a single-phase cascaded H-bridge converter. Multilayer cascaded H-bridge
inverters have a basic layout shown in Fig.2. This N-level converter offers the possibility of 2N+1
different values for the phase voltage at the output. Because of the power requirements of each H-
bridge, the previously noted voltage discrepancy problem may be avoided. However, to prevent voltage
fluctuations of the DC-bus capacitor, it is necessary to adjust the active power provided to each bridge.
[4]. There are a number of methods outlined in [5] for dampening DC-link voltage variations.
Many applications also take reactive power into account, which is where the cascaded H-bridge
architecture shines. Since high-power electrical switches have their limits, we have to settle with a low-
frequency modulation strategy. [6]- [8]. When compared to a regular power supply, a multilayer
converter offers several advantages. Since the output pattern has fewer spurious harmonics, cheaper,
smaller output filters may be employed. Since there are more voltage steps between the highest and
lowest values in the staircase pattern, the voltage burden on semiconductor components is lessened.
Switching losses may be minimized in multi-level inverters due to their lower switching frequency [1].
Multiple converters often use SPWM as the primary management technique. By contrasting a triangle
waveform with a sine wave, PSPWM generates control signals for converter switches.
FIGURE 2. Model of power-stage CHB multilevel converter
Overview of Accelerated Particle Swarm Optimization (Apso)

The optimization method known as particle swarm optimization [14-18] is inspired by natural
swarms like those of fish and birds. The PSO algorithm gets the best answer by first producing a random
sample of candidates, or particles, and then offering incentives to the candidates still in the running, or
particles. Particles may be created via either collaborative or adversarial interactions. The equation
allows for a point in D-dimensional space to represent any given particle I. 𝑋𝑖 = (𝑥𝑖1, 𝑥𝑖2, . . . , 𝑥𝑖𝐷).
The optimal answers in the full collection fall into two distinct categories: 𝑃𝑏𝑒𝑠𝑡, or cognitive
learning, is the particle's best response across rounds and is defined as 𝑃𝑏𝑒𝑠𝑡 = (𝑝𝑖1, 𝑝𝑖2, . . . , 𝑝𝑖𝐷),
while Gbest, or the swarm's best solution, is defined as the solution with the highest fitness. (Called
Page 924
SEMIT 2023
social learning). Every repetition k brings the particle located at x(k) one step closer to the 𝑃𝑏𝑒𝑠𝑡 and
𝐺𝑏𝑒𝑠𝑡, and this motion is indicated by its velocity (written v(k)). The speed is determined by combining
the knowledge of both individual particles and the collective swarm, and can be written in D-
dimensional space as 𝑉𝑖 =. (𝑣𝑖1, 𝑣𝑖2, . . . , 𝑣𝑖𝐷). Each particle then uses the following calculation to
make an approximation of its motion for the next cycle:
𝑣𝑖 (𝑘 + 1) = 𝑤𝑖. 𝑣𝑖 (𝑘) + 𝑐1 . 𝑟𝑎𝑛𝑑. (𝑃𝑏𝑒𝑠𝑡 − 𝑥𝑖 (𝑘)) + 𝑐2 . 𝑟𝑎𝑛𝑑. (𝐺𝑏𝑒𝑠𝑡 − 𝑥𝑖 (𝑘)) (1)
With the help of the gravitational constant (w), the cognitive constant (c1), and the societal constant
(c2), the rand () algorithm generates a random integer between 0 and 1. The rate of convergence is
affected by all these factors. Clerc's optimal numbers were c1=1.494, c2=1.494, and w=0.729 [22-25].
PSO's cost function must be adjusted to reflect the intended system behavior under these conditions.
The PSO keeps running its procedures for the given amount of time to find the best answer. The next
iteration's particle position is calculated using the following algorithm. 'k+1':
𝑋𝑖 (𝑘 + 1) = 𝑣𝑖 (𝑘 + 1) + 𝑋𝑖 (𝑘) (2)
After making some educated guesses, we can figure out what size community to use to initiate PSO.
In spite of the fact that the optimum population is context-dependent, the PSO developers have stated
that starting with a population of 20-100 particles typically yields the same outcomes [16, 26]. The
Quick Optimization of Particle Swarms (APSO) While standard PSO takes into account both global
and local best particle sites when determining particle motion for the next cycle, accelerated PSO only
takes into account global best particle locations [17]. In PSO, utilizing local best serves to increase the
size of the search area; in APSO, sampling is used to achieve the same effect. Unless the problem is
very complex, this quantity causes the strategy to "accelerate" towards optimality [26]. The simplified
new algorithm is as follows:
𝑣𝑖 (𝑘 + 1) = 𝑤𝑖. 𝑣𝑖 (𝑘) + 𝑐1 . 𝑟𝑎𝑛𝑑
+ 𝑐2 . 𝑟𝑎𝑛𝑑. (𝐺𝑏𝑒𝑠𝑡 − 𝑥𝑖 (𝑘)) (3)
Implementation Strategy
Design Objectives
The STATCOM employed in this study has two major purposes:
• To give reactive power to the system when it is operating in capacitive mode and
• To remove reactive power while it is running in inductive mode.
System Design
STATCOM used in this study consists of nine-level with four similar CHB connected in series,
each with its own power supply. (DC Capacitor). Each H-bridge can produce a positive, neutral, or
negative voltage at its output. When these energies are added together, a sine pulse is produced. The
phase magnitude of an AC output is represented as follows [18-22]:
𝑉𝑐𝑎 = 𝑉𝑐𝑎1 + 𝑉𝑐𝑎2 + 𝑉𝑐𝑎3 + … . . . +𝑉𝑐𝑎𝑀 (4)
M denotes the number of inverters per phase. The flexible design of a cascading multilevel inverter
(CMI) makes adding more similar H-bridge inverters to the circuit a straightforward way to boost
output. Energy from the connection inductance is corrected and supplied through the H-bridge to the
DC capacitors attached to the other side of the bridge. Capacitance can be calculated as follows:
𝑖𝑐
𝐶= (5)
𝑓𝑟𝑝 ∗∆𝑉𝐷𝐶
𝑓𝑟𝑝 : frequency of ripples
Page 925
SEMIT 2023
𝑖𝑐 : 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑟𝑚𝑠 𝑐𝑢𝑟𝑟𝑒𝑛𝑡

𝑄
𝑖𝑐 = 𝑉 (6)
𝑝𝑐𝑐
Q : the maximum reactive power can be generate or absorb by STATCOM . As a reactive power
coupling between the PCC and the STATCOM, the inductor's capacitance is affected by the current
harmonics and the CMI. To calculate inductance, use the following formula:
𝑣𝑃𝐶𝐶
𝐿𝑓𝑚𝑎𝑥 = (7)
2∗𝜋∗𝑓∗𝑖𝑐
F: System fundamental frequency

𝑣𝑝𝑐𝑐 : Voltage at PCC
The above solution gives the highest allowable inductance capacity. Due to its massive scale, this
inductor effectively dampens current waves and overtones at the PCC Transformer. Due to the need to
limit the residual of voltage power networks, the STATCOM cannot be hardwired to the grid. In order
to bring the power down to a level that the equipment can manage, a converter is required. However,
there may be power deficits associated with using a converter, such as increased compensatory currents
on the STATCOM side. Procedure for controlling depending on the PI controller.
The proposed control approach involves the use of PI controllers, one to determine the phase angle
from the DC voltage and another to determine the modulation index from the difference between the
targeted and measured reactive power. The modulator receives inputs such as the toggling components'
on/off states, the current, and the voltages across the capacitors. As shown in Fig.3, the control
architecture's primary block layout is laid out in logical fashion.
FIGURE 3. Schematic of proposed control strategy using PI controller

The reactive power that the STATCOM supplies to the line is given by the equation below:
𝐸𝑐 2 −𝐸𝑐 𝐸𝑠 𝑐𝑜𝑠𝛿
𝑄𝑠 = 𝑋
(8)
Where:
x : the coupling reactance
𝐸𝑠 : RMS value of PCC Voltage
𝐸𝑐 : STATCOM output voltage
Page 926
SEMIT 2023
The STATCOM's can adjust the reaction strength by adjusting the frequency it operates at. The
following is a broad description of the connection between the DC link voltage and the STATCOM
output voltage.
𝑉𝑜𝑢𝑡
=𝑚 (9)
𝐸𝐷𝐶
Therefore, the reactive power equation changes to

𝐸𝑠 2 −𝑚𝐸𝐷𝐶 𝐸𝐿 𝑐𝑜𝑠𝛿
𝑄𝑠 = 𝑋
(10)
The solution above illustrates how the STATCOM operates. In the suggested setup, the modulation
index regulates the reaction strength of the STATCOM. The primary regulator of the STATCOM. The
phase locked loop (PLL) component provides not only the phase but also the phase angle (𝑤𝑡) of a
voltage signal for verification purposes.
Simulation And Results
The whole form of the modeling software’s depiction of the system is shown in Fig. 4.
FIGURE 4. Simulink block diagram of complete system
To cope with a nonlinear load, the H-bridge circuit in Fig.5 uses a perfect IGBT and a reactive power
reference. The scenario is executed without STATCOM, with a power factor of 0.7 for all loads, and
with inductive/capacitive load powers varying from 0 to 14 seconds and capacitive load powers ranging
from 14 to 28 seconds. Table 1 contains the model's characteristics.
Table 1. System parameters

Type Parameter Value
Source PCC Voltage (L- 311.12 V
L)
System Frequency (Hz) 50 Hz
Load Voltage (V) 220Vrms
Real Power (Watts) 450 W
Reactive Power Inductive 450 VAR
Reactive Power Capacitive 450 VAR
Dc Link volt DC Link reference Voltage (Volts) 27 V
Coupling Nominal Power (VA) 5000 VA
Transformer 220/110 V
Page 927
SEMIT 2023
FIGURE 5. Reactive power reference signal for STATCOM

The PI-based APSO management method is advocated by the computer program. Fig.6 displays the
layered inverter's output voltage's thirteen discrete steps. The reactionary power of the STACOM is
depicted in Fig.7, and the APSO technology is what gives the PI control its durability and versatility.
FIGURE 6. STATCOM-output voltage and reference signal
FIGURE 7. STATCOM- reactive power with APSO
The effectiveness of the APSO method can be further examined by plotting the cost functions for
both controllers in addition to the PI controller gains k p and k I. Fig.8 displays the gains achieved by
the reactive power converter over the course of 50 trials. The advantage in k I increases to 6.1, then
drops to the optimal number of 0.1. The k p trajectory, on the other hand, shows a negligible decrease
from its initial 0.6 to its ultimate 0.01035.
Page 928
SEMIT 2023
FIGURE 8. PI parameters for reactive power controller iterations using APSO

The Fig.9 depicts the reduction in IAE (cost function) caused by a reactive power converter. After
running through APSO's rounds, the cost naturally drops from a high of 1.1 to a low of 0.17, which is
an acceptable margin of error.
FIGURE 9. IAE for reactive power controller iterations using APSO

Another controller that provides comparable trajectory options is the APSO-optimized DC-Link PI
controller. The Fig.10 can be seen , the K_P gain number begins at 11.8 and increases gradually to 3.1.
The rise has been larger in recent versions than in previous ones. In a similar vein, the K_I gain, which
was originally fixed at 013.9, steadily declines to 0.355 after 50 cycles. Fig.11 shows the iterative
readings of the IAE. The APSO method fails after 50 rounds, with an initial error of 0.67 and a final
error of 0.091.
FIGURE 10. PI parameters for DC-Link Voltage controller with iterations using APSO
Page 929
SEMIT 2023
FIGURE 11. Trajectory of IAE for DC-Link Voltage controller with iterations using APSO
Optimizing the controller parameters of a Multilevel Converter-Based Static Synchronous
Compensator for reactive power compensation can offer several advantages in power systems.
STATCOMs are a type of Flexible AC Transmission System (FACTS) device used to control and
stabilize the voltage and reactive power in electrical grids. Optimization of controller parameters allows
for precise control of the STATCOM's output voltage. This helps in maintaining grid voltage within
specified limits, reducing voltage fluctuations, and ensuring a stable and reliable power supply. By
optimizing controller parameters, the STATCOM can provide efficient reactive power compensation.
This helps in minimizing voltage drops, improving power factor, and reducing line losses. Properly
tuned controller parameters enable the STATCOM to respond quickly to changes in the grid's reactive
power requirements. This rapid response time ensures that voltage and power factor are kept within
desired limits, even during transient events. Multilevel converters used in STATCOMs can generate
harmonics if not controlled properly. Parameter optimization can mitigate harmonics, ensuring a cleaner
and more stable power supply with minimal distortion. A well-tuned STATCOM contributes to grid
stability by damping out oscillations, reducing voltage flicker, and preventing voltage sags and swells.
This, in turn, enhances the overall reliability of the power system. Optimizing the controller parameters
can lead to more efficient operation of the STATCOM. This means that the device consumes less
reactive power from the grid, reducing energy losses and improving the overall energy efficiency of the
system. Improved voltage regulation and reduced losses result in cost savings for utilities and end-users.
With optimized STATCOM control, there is less need for additional equipment to maintain grid quality.
STATCOMs are increasingly used in conjunction with renewable energy sources like wind and solar
power. Proper parameter optimization ensures seamless integration and improved grid stability in
renewable energy systems. Optimization allows for the development of control strategies tailored to
specific grid conditions and requirements. This adaptability is crucial in addressing varying grid
challenges. Many countries have stringent grid codes that require power quality standards to be met.
Optimized STATCOM controllers can help meet these regulatory requirements, avoiding penalties and
ensuring compliance.
Conclusion
the Adaptive Particle Swarm Optimization (APSO) algorithm is a powerful optimization approach
rooted in the principles of swarm intelligence, drawing inspiration from the collective behaviors of
natural swarms like fish or birds [14]. Voltage optimization plays a pivotal role in maintaining the
operational efficiency and reliability of electrical distribution systems by ensuring voltage levels remain
within acceptable parameters. Leveraging the STATCOM (Static Synchronous Compensator) in
conjunction with APSO offers a viable strategy for achieving this optimization goal. A STATCOM, as
a component of Flexible AC Transmission Systems (FACTS), possesses the capability to control
reactive power flow within the electrical grid. This device, consisting of a voltage source converter
(VSC) driven by power electronics, can dynamically generate or absorb reactive power as required,
providing essential voltage support to the system. APSO, an evolution of the Particle Swarm
Page 930
SEMIT 2023
Optimization (PSO) algorithm, harnesses the innate behaviors of swarming creatures to fine-tune the
settings of STATCOM and effectively regulate the system's voltage in the context of voltage
optimization. PSO, a metaheuristic optimization technique, orchestrates a population of potential
solutions, referred to as particles, in a dynamic and purposeful manner, much like the cooperative or
competitive interactions observed in natural swarms. Each particle, representing a point in a
multidimensional space, is guided towards the optimal solution through iterative adjustments,
ultimately yielding the best configuration for achieving voltage optimization. The successful
optimization of controller parameters not only enhances the STATCOM's ability to provide reactive
power compensation but also contributes to the overall stability and reliability of the power system. It
allows for better control of voltage levels and mitigation of power quality issues such as voltage sags
and swells, thereby improving the quality and efficiency of electricity distribution.
Additional, the optimization of controller parameters for Multilevel Converter-based STATCOMs
represents a critical step towards achieving optimal reactive power compensation and improving the
overall performance of power systems. This ongoing research and development effort continue to play
a vital role in ensuring a stable, reliable, and high-quality supply of electricity to meet the demands of
modern society.
References
1. J. Rodriguez, J. S. Lai and F. Z. Peng, “Multilevel inverters: survey of topologies, controls, and
applications,” IEEE Trans. on Industry Applications, Vol. 49, No. 4, pp. 724-738, 2002 .
2. S. Panda and N. P. Padhy, “location and controller design of STATCOM for power system
stability improvement using PSO,” Journal of the Franklin Institute, (2007).
3. Saman Babaei “A Control Structure for Line- Frequency-Switched STATCOMs under System
Faults”, Department of Electrical and Computer Engineering, 978-1-4799-0336-8/13 IEEE, Sept
2013
4. Verma Amitkumar, Kamani PL, Kapadia RR. A review on gridpower quality improvement in
wind energy system using STATCOM with BESS. J Emerg Technol Innov Res
(JETIR);2(1):26e31,(2015).
5. Neyshabouri, Y., Iman-Eini, H., Miranbeigi, M.: ‘State feedback control strategy and voltage
balancing scheme for a transformer-less STATic synchronous. COMpensator based on cascaded
H-bridge converter’, IET Power Electron., 8, (6), pp. 906–917, 2015.
6. Zotin, A., Hamad, Y., Simonov, K., Kurako, M., & Kents, A. (2020, June). Processing of CT lung
images as a part of radiomics. In International Conference on Intelligent Decision Technologies
(pp. 243-252). Singapore: Springer Singapore.
7. Hamad, Y., Mohammed, O. K. J., & Simonov, K. (2019, March). Evaluating of tissue germination
and growth rate of ROI on implants of electron scanning microscopy images. In Proceedings of
the 9th international conference on information systems and technologies (pp. 1-7).
8. C. Tareila, P. Sotoodeh, R. D. Miller., “Design and control of a singlephase D-STATCOM
inverter for wind application”, Power Electronics and Machines in Wind Application PEMWA
2012, Denver, Co, 2012.
9. Sun, Y., Zhao, J., Ji, Z., et al.: ‘An improved DC capacitor voltage balancing strategy for PWM
cascaded H-bridge converter-based STATCOM’. Proc. IEEE 10th Int. Conf. Power Electronics
Drive Syst., Kitakyushu, Japan ,pp. 1245–1250, April 2013.
10. Vural, A. M. (2016). Self-capacitor voltage balancing method for optimally hybrid modulated
cascaded H-bridge D-STATCOM. IET Power Electronics, 9(14), 2731-2740.
Page 931
SEMIT 2023
11. Taha, O.A., Pacas, M.: ‘Hardware implementation of balance control for three-phase grid
connection 5-level cascaded H-Bridge converter using DSP’.Proc. IEEE 23rd Int. Symp.
Industrial Electronics, Istanbul, Turkey ,pp. 1366–1371, June 2014.
12. Sotoodeh, Pedram, and Ruth Douglas Miller. "A single-phase 5-level inverter with FACTS
capability using modular multi-level converter (MMC) topology." Electric Machines & Drives
Conference (IEMDC), 2013 IEEE International. IEEE, 2013.
13. Neyshabouri, Y., Iman-Eini, H., Miranbeigi, M.: ‘State feedback control strategy and voltage
balancing scheme for a transformer-less STATic synchronous. COMpensator based on cascaded
H-bridge converter’, IET Power Electron., 8, (6), pp. 906–917, 2015.
14. Behrouzian, Ehsan, Massimo Bongiorno, and Remus Teodorescu. "Impact of switching
harmonics on capacitor cells balancing in phase-shifted PWM-based cascaded H-bridge
STATCOM." IEEE Transactions on Power Electronics 32.1 (2017): 815-824.
15. S. U. Shinde, M. Sharmila, R. S. Patil and D. V. Malkhede, "Performance comparison of PI &
ANN based STATCOM for 132 KV transmission line," in International Conference on Electrical,
Electronics, and Optimization Techniques (ICEEOT), Chennai, India, 2016.
16. Debnath, Suman, et al. "Operation, control, and applications of the modular multilevel converter:
A review." IEEE transactions on power electronics 30.1 (2014): 37-53.
17. Xu, Jianzhong, Penghao Zhao, and Chengyong Zhao. "Reliability analysis and redundancy
configuration of MMC with hybrid submodule topologies." IEEE Transactions on Power
Electronics 31.4 (2015): 2720-2729.
18. Leon, Jose I., Sergio Vazquez, and Leopoldo G. Franquelo. "Multilevel converters: Control and
modulation techniques for their operation and industrial applications." Proceedings of the
IEEE 105.11 (2017): 2066-2081.
19. M. Moazzami, G. B. Gharehpetian, H. Shahinzadeh, and S.H. Hosseinian, “Optimal locating and
sizing of DG and DSTATCOM using Modified Shuffled Frog Leaping Algorithm,” in
Proceedings of the 2nd Conference on Swarm Intelligence and Evolutionary Computation, CSIEC
2017, pp. 54–59, Iran, March 2017.
20. Safonova, A., Hamad, Y., Dmitriev, E., Georgiev, G., Trenkin, V., Georgieva, M., ... & Iliev, M.
(2021). Individual tree crown delineation for the species classification and assessment of vital
status of forest stands from UAV images. Drones, 5(3), 77.
21. Kabaev, E., Hamad, Y., Simonov, K., & Zotin, A. (2020). Visualization and Analysis of the
Shoulder Joint Biomechanics in Postoperative Rehabilitation. In SibDATA (pp. 34-41).
22. Seno, M. E., Abed, A. A., Hamad, Y. A., Bhatt, U. M., Babu, B. R., & Bansal, S. (2022,
December). Cloud Based Smart Kitchen Automation and Monitoring. In 2022 5th International
Conference on Contemporary Computing and Informatics (IC3I) (pp. 1544-1550). IEEE.
23. Seno, M. E., Abed, A. A., Hamad, Y. A., Pundir, S., Srivastava, A. P., & Ali, S. (2022, December).
Forwarding Factor Routing Protocol for Opportunistic Internet of Things Networks. In 2022 5th
International Conference on Contemporary Computing and Informatics (IC3I) (pp. 988-995).
IEEE.
24. Safonova, A., Hamad, Y., Alekhina, A., & Kaplun, D. (2022). Detection of norway spruce trees
(Picea Abies) infested by bark beetle in UAV images using YOLOs architectures. IEEE Access,
10, 10384-10392.
25. Mustafa, M. O., & Abbas, N. M. (2021). Harmonic Suppression Compensation of Photovoltaic
Generation Using Cascaded Active Power Filter. Eastern-European Journal of Enterprise
Technologies, 6(8), 114.
26. Zainel, Q. M., Darwish, S. M., & Khorsheed, M. B. (2022). Employing quantum fruit fly
optimization algorithm for solving three-dimensional chaotic equations. Mathematics, 10(21),
4147.
Page 932
SEMIT 2023
Numerical Investigation of The Effect of the Use Of Multiple Propeller Systems On Flight
Performance in Unmanned Aerial Vehicles.
Kasım Şahin DÖNMEZ1[0009-0001-0305-2282] and Metin UZUN 2[0000-0002-0744-3491]
1
department of aeronautical science and technologies, graduate student, Iskenderun Technical
University, 31200 Hatay, Türkiye
2
Department of Airframe and Powerplant Maintenance, Faculty of Aeronautics and Astronautics,
Iskenderun Technical University, 31200 Hatay, Türkiye
Abstract
The aim of this study is to try to solve the need for Fixed Propeller Unmanned Aerial Vehicles by using
multiple propellers in the horizontal and vertical directions without changing the electronic equipment
and propeller type with the changes in the total take-off weight. In unmanned aerial vehicles, energy
needs are met by using electrical power effectively with the help of a battery, which is stored as a fuel
system in many of them. There will be changes in all of the engine, battery and propeller systems that
the aircraft will need as a result of flight altitude, flight altitude, body weight and useful load
information. The weight and cost of electronic equipment do not increase proportionally to the amount
of increase in aircraft take-off weight. For example, if you want the battery capacity to increase by 2
times, the weight and cost of the battery will increase more than 2 times. However, the required thrust,
that is, the weight increase in the aircraft, will cause the change of all electronic equipment. When the
compatibility between these working equipments is ensured, the problem can be solved by changing
the number of electronic equipment even if the weight of the aircraft changes. In this study, with the
help of Ansys Fluent, a certain distance was left between the propellers placed in horizontal and vertical
directions as double and triple of equal diameter, and the minimum distance of this distance and whether
there was a loss of efficiency in the use of multiple propellers were explained in detail. Here, multiple
propeller pairs were operated both in the same direction and in opposite directions, and distance and
direction analyzes were also carried out. In addition, the compatibility of the propeller pairs is shown
graphically and visually whether it is similar or not in different revolutions.
Keywords: Propeller, Unmanned Aerial Vehicles, Efficiency.
1. Introduction
Propellers generate thrust using the same principle that airplane wings do to create lift. If you look at a
section of a propeller blade, you will notice that it looks like a blade. Air approaches the wing section
(airfoil) at an angle of attack that causes a pressure change on the airfoil and provides lift. This lift
produces a force to propel the aircraft in the direction of flight. In recent years, multirotor unmanned
aerial vehicles (UAVs) are gaining more and more importance in both civil and military fields.
However, the endurance time of electric multicopters is still too short for many mission requirements,
and this is a major factor hindering further development of the multicopter field. Optimizing the
efficiency of the propulsion system, which is limited to battery technology, is a practical and effective
way to increase vehicle durability multicopters. Accordingly, a typical propulsion system usually
consists of a battery pack and several propulsion units [including a propeller, brushless direct current
(BLDC) motor, and an electronic speed control (ESC)]. Of these components, the engine and propeller
are the most important components that determine the effectiveness of a propulsion system [1-6].
2. Propeller Systems Types
Page 933
SEMIT 2023
Although many different propeller designs have been available since the beginning of powered flight,
not all of them have been adapted to widespread use. Detailed information about some propeller types
and their working principles will be given in this section.
2.1. Fixed pitch propeller
The most preferred propellers in aviation are fixed pitch propellers due to their simple design. They can
be made of wood, aluminum or composite materials, and are considered one piece because the pitch
angle does not change. Fixed pitch propellers are preferred in mini UAVs, quadcopters and light fixed
wing aircraft.
2.2. Ground adjustable propeller
Since the pitch angle is not changed during flight, it is similar to a fixed-pitch propeller. Although the
pitch angle is not actively changed during flight, the pitch angle can be adjusted passively to the desired
level on the ground. The pitch angle on the ground is changed depending on the take-off weight of the
aircraft and the airport runway condition. This type of propellers, which can passively change the pitch
angle, are widely used in light experimental aircraft.
2.3. The two position propeller
It is designed to allow the pilot to choose between two wing angles during flight; It allows the use of a
low wing angle for takeoff and a high wing angle for cruise. This is similar to a two-speed transmission
in a car. It is actively used in today's aircraft.
2.4. The controllable pitch propeller
It is designed to provide the pilot with the ability to adjust the blades at any angle within the range of
the propeller during flight, giving the pilot greater control over the propeller than a two-position design.
This style of propeller was popular on light aircraft in the 1940s due to its low weight and simple
mechanism, and is still produced today.
The automatic pitch changing propeller is designed to operate independently of the pilot. Rather than
being controlled by the pilot from the cockpit, it will theoretically adjust the propeller blades to the most
efficient angle by reacting to the forces created by engine loads and airspeed. Although designed for
engines up to 450 hp, most aircraft using these propellers today have engines of 200 hp or less.
Constant speed propellers are actively used in most of today's modern aircraft. By controlling all
conditions that may change the propeller thrust, the propeller speed can be adjusted via a regulator in
fixed adjustable propellers. We can talk about it as adjusting the pitch angle for the same thrust in
variable situations in order to keep the propeller thrust constant [7-8].
3. Propeller Theory
It is possible to estimate the preliminary performance of the propeller with the help of dimensionless
coefficients. The performance of a propeller is highly dependent on the relationship between the
airspeed V0 and the angular velocity of the propeller. Push forward speed is a function of V0, RPM, and
diameter. This function is expressed as the feed rate (J), which measures the ratio of the free-flowing
fluid velocity (aircraft velocity) to the propeller tip velocity [9-19]. The advance rate is defined as
follows;
V0 60.V0
J= = (Eq 1)
nD RRM .D
Power and thrust coefficients combined with the advance rate to calculate propeller efficiency;
Page 934
SEMIT 2023
P
CP =
 n3 D 5
T
CT = 2 4
n D ( Eq 2)
Q ( Eq3)
CQ =
 n2 D5 ( Eq 4)
Thus, the propeller efficiency is:
TV0 C
= =J T (Eq 5)
P CP
3.1.Materyal and Metod
Inlet
Rotating
Domain
Outlet
Figure 3.1. Boundary Conditions of CAD Propeller
Figüre 3.2. Mesh Images of Horizontally Positioned Propellers

A grid convergence study is mandatory when performing CFD Analysis with limited resources. Grid
leads to convergence analysis to a final output with minimal variation within the expected range We
can call this acceptable. In this study, with speed as changing input parameter and in combination with
other parameters Grid convergence analysis was performed in the same way. Variation the amount of
torque produced is the criterion observed during analysis [20-26].
Table 3.1. Specification of The Mesh Structure.
Page 935
SEMIT 2023
Minimum Element Size 0,00005 m Skewness (max) 0,72599

Number of Elements 2120212 Growth Rate 1,12
Maximum Size 0,1 m Curvature Normal Angle 18o
Orthogonal Quality 0,2702 Mesh Method Poly-hexcore
Figure 3.3. The Mesh Independence Test Results
4. Result
Figüre 4.1. Vertical And Horizontal Propeller Positioning on Quadcopter Arms

The figure shows the horizontal and vertical positions of the propellers placed on the quadcopter arms
to generate thrust. Here it is shown how to position a horizontal double blade or a vertical double blade
instead of a large single blade.
Table 4.1. Single Blade Aerodynamic Parameters
20000 4000 6000 8000 10000
Thrust (N) 0,381 1,52 3,454 6,173 9,678
Power (W) 1,35 11,0144 37,5095 88,7962 174,849
Produced thrust power and consumed energy information regarding the rotation of the propeller of APC
9*6 dimensions at different revolutions are given. Although the amount of energy consumed per thrust
Page 936
SEMIT 2023
produced is expressed as efficiency, the efficiency value decreases as the number of revolutions
increases. As the number of revolutions increases, the energy consumed increases faster than the thrust
produced. It can produce 9.6 N thrust at the highest speed applied in this study.
Figüre 4.2. Thrust information of 2 blades placed in the horizontal direction and rotating in the same
direction
In the above graphic, it gives information about the effect of the thrust values of the 2 pallets placed in
the horizontal direction on each other and whether there is a loss of efficiency as a result of this effect.
In the light of this information, detailed thrust force information is given for the blades rotating in the
same direction and horizontally in the table below.
Table 4.2. Horizontally placed cv-cv 2 pal Aerodynamic Parameters
20000 4000 6000 8000 10000
Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2
35 cm (Thrust) N 0,35 0,36 1,42 1,449 3,21 3,279 5,73 5,85 8,98 9,162
35 cm (Power) W 1,37522 1,3919 10,302 11,2657 37,5 37,698 88,796 90,471 173,8 176,94
40 cm (Thrust) N 0,363 0,362 1,46 1,46 3,316 3,3 5,954 5,924 9,203 9,246
40 cm (Power) W 1,4 1,4 11,3 11,3 38,95 38,95 92,147 92,147 179,03 177,99
45 cm (Thrust) N 0,369 0,369 1,502 1,502 3,4 3,403 6,066 6,072 9,498 9,5
45 cm (Power) W 1,414 1,4 11,3 11,47 38,95 38,95 92,147 92,147 180,08 179,037
50 cm (Thrust) N 0,381 0,38 1,467 1,454 3,482 3,481 6,204 6,204 9,71 9,71
50 cm (Power) W 1,419 1,41 11,45 11,34 38,64 39,7 90,47 90,47 179,246 18,34
55 cm (Thrust) N 0,383 0,384 1,54 1,548 3,492 3,509 6,234 6,263 9,768 9,811
55 cm (Power) W 1,4212 1,425 11,3 11,517 38,95 39,2 92,98 92,147 182,178 181,131
When the cv-cv 2 pal parameter information placed on the quadcopter arms in the horizontal direction
is examined, a very low loss is observed compared to the single blade values. When the blades are
examined relative to each other, it is seen that they have very close thrust and power consumption. In
this case, when the blades are placed in the horizontal direction, it is seen that they do not affect each
other much. Even if the distance between the central positions of the blades is increased in the horizontal
direction, there is no significant change in thrust value and efficiency. it can be said that there is a small
increase in efficiency as the distance between the paller centers increases, but since Decently increasing
the distance will be a challenge for control abilities, in this case the increase can be said to be negligible.
Table 4.3. Horizontally placed cv-ccv 2 pal Aerodynamic Parameters
20000 4000 6000 8000 10000
Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2
35 cm (Thrust) N 0,366 0,366 1,472 1,479 3,411 3,364 6,094 6 9,553 9,418
35 cm (Power) W 1,3376 1,3585 11,056 11,098 37,069 37,69 91,3 89,63 174,8 175,89
40 cm (Thrust) N 0,364 0,365 1,487 1,487 3,381 3,381 6,125 6,1325 9,5986 9,586
40 cm (Power) W 1,3585 1,356 11,05 11,05 37,698 37,069 88,79 88,79 174,84 174,84
45 cm (Thrust) N 0,366 0,369 1,5 1,5 3,4 3,44 6,1354 6,149 9,61 9,64
45 cm (Power) W 1,356 1,358 11,05 11,05 37,635 37,698 89,63 89,88 174,84 175,89
Page 937
SEMIT 2023
50 cm (Thrust) N 0,37 0,37 1,516 1,517 3,443 3,445 6,152 6,157 9,646 9,652
50 cm (Power) W 1,358 1,356 11,09 11,05 37,69 37,63 89,63 89,63 175,89 175,68
55 cm (Thrust) N 0,372 0,37 1,521 1,519 3,452 3,432 6,169 6,167 9,671 9,675
55 cm (Power) W 1,358 1,348 11,09 10,97 37,69 37,321 89,63 88,79 175,89 173,8
Parameter information for 2 blades placed as cv-ccv in the horizontal direction is given in the table
above. When compared with the blades rotating in the same direction, there is no significant difference.
Similarly, as the distance between the centers of the blades increases, there is a partial increase in the
efficiency and thrust values, but there is no significant difference.
Figüre 4.3. Pressure and velocity images formed as a result of the variation of the blades placed in the
horizontal direction with the distance
Figüre 4.4. Thrust values of each of the cv-cv blades placed in the vertical direction
When the simulation image is examined, it is seen that there is a difference between the thrust values
for the blades placed as cv-cv in the vertical direction. From the blades placed in the vertical direction,
Page 938
SEMIT 2023
it is understood that the thrust value of the blade in the upper region is higher and affects the efficiency
of the blade placed under it.
Table 4.4. Vertical placed cv-cv 2 pal Aerodynamic Parameters

20000 4000 6000 8000 10000
Distance between Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2
diameters
5 cm (Thrust) 0,214 0,354 0,899 1,425 2,052 3,217 3,674 5,731 5,765 8,968
5 cm (Power) 0,999 1,316 8,208 10,763 27,959 36,441 66,59 86,7 129,82 169,4
10 cm (Thrust) 0,225 0,36 0,929 1,478 2,11 3,355 3,773 5,993 5,916 9,396
10 cm (Power) 1,045 1,337 8,376 10,972 28,9 37,258 68,69 87,958 134,016 172,75
20 cm (Thrust) 0,232 0,385 0,962 1,595 2,187 3,63 3,909 6,492 6,128 10,184
20 cm (Power) 1,067 1,356 8,376 11,056 29,78 37,69 70,953 89,63 138,204 1744,849
50 cm (Thrust) 0,377 0,381 1,52 1,53 3,335 3,396 6,136 6,194 10 10,247
50 cm (Power) 1,3585 1,3585 11,01 11,056 37,44 37,57 89,382 89,633 175,89 174,84
The thrust and power consumption information obtained as a result of the rotation of the 2 blade
propellers placed in the vertical direction in the same direction at different speeds and according to the
change in the distance between their diameters are given in the table. It has been observed that the
distance between the centers of the propellers placed in the upper and lower positions is very short, and
the positions such as 5 cm and 10 cm have different thrust and power consumption. As the distance
between the diameters decreases, the thrust loss and efficiency decrease increase as the effect of the
propellers on each other increases. however, as the distance between the diameters decreases, the
difference in thrust power produced at increasing revolution numbers becomes greater. When the
distance between the diameters reaches 50 cm, it reaches the thrust produced by a single propeller and
the difference between them reaches a size that is almost non-existent.
Table 4.5. Vertical placed cv-ccv 2 pal Aerodynamic Parameters
20000 4000 6000 8000 10000
Distance between Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2 Pal 1 Pal 2
diameters
5 cm (Thrust) 0,306 0,339 1,312 1,261 3,018 3,013 5,425 5,351 8,073 8,984
5 cm (Power) 1,356 1,289 11,977 10 39,834 34,55 95,497 82,01 176,94 165,426
10 cm (Thrust) 0,3085 0,345 1,319 1,268 3,025 3,02 5,432 5,358 8,08 8,991
10 cm (Power) 1,378 1,311 12,01 10,05 39,897 34,619 95,581 82,094 177,047 1665,53
20 cm (Thrust) 0,311 0,351 1,327 1,276 3,033 3,028 5,44 5,366 8,088 8,99
20 cm (Power) 1,4 1,33 12,06 10,09 39,959 34,682 95,665 82,178 177,152 1665,634
50 cm (Thrust) 0,367 0,371 1,5 1,529 3,426 3,473 6,124 6,2 9,6 9,727
50 cm (Power) 1,3522 1,35641 11,05 11,056 37,635 37,635 89,633 89,633 174,84 174,84
Thrust and consumed power information of 2 propellers placed in the vertical direction and rotating in
the opposite direction at different speeds and at different distances between diameters are given. When
we compare it with the propeller pair rotating in the same direction, the difference in thrust produced
between the propellers is very low in the propellers rotating in the opposite direction.
Page 939
SEMIT 2023
Figüre 4.5. Pressure and velocity images formed as a result of the variation of the blades placed in the
vertical direction with the distance
The pressure and velocity images of the propeller pair placed in the vertical direction have been added
in the figure. Here, images are given for the shortest distance between diameters of 5 cm and the longest
distance of 50 cm. As the distance between the diameters increases, it is clearly seen that the interaction
of the propeller pairs decreases.
Table 4.6. Horizontally placed cv-cv-cv 3 pal Aerodynamic Parameters
2000 4000 6000 8000 10000
Distance Thrust Power Thrust Power Thrust Power Thrust Power Thrust Power
between (N) (W) (N) (W) (N) (W) (N) (W) (N) (W)
diameters
35 cm Pal 1 0,362 1,40866 1,458 11,3278 3,286 38,4964 5,852 91,233 9,154 179,037
Pal 2 0,361 1,40866 1,446 11,286 3,256 37,68 5,794 90,396 9,062 176,943
Pal 3 0,351 1,32924 1,39 10,5754 3,124 35,9216 5,558 85,374 8,692 166,473
40 cm Pal 1 0,366 1,41493 1,512 11,3069 3,343 38,7162 5,9635 91,6515 9,332 179,5605
Pal 2 0,3645 1,41911 1,5065 11,4114 3,3205 38,3394 5,9145 91,233 9,2535 179,037
Pal 3 0,3595 1,37522 1,474 11,077 3,256 37,5544 5,7985 89,1405 9,069 174,3255
45 cm Pal 1 0,368 1,418065 1,539 11,29645 3,3715 38,8261 6,01925 91,86075 9,421 179,8223
Pal 2 0,36625 1,424335 1,53675 11,4741 3,35275 38,6691 5,97475 91,6515 9,34925 180,084
Pal 3 0,36375 1,39821 1,516 11,3278 3,322 38,3708 5,91875 91,02375 9,2575 178,2518
50 cm Pal 1 0,37 1,4212 1,566 11,286 3,4 38,936 6,075 92,07 9,51 180,084
Pal 2 0,368 1,42956 1,567 11,5368 3,385 38,9988 6,035 92,07 9,445 181,131
Pal 3 0,368 1,4212 1,558 11,5786 3,388 39,1872 6,039 92,907 9,446 182,178
55 cm Pal 1 0,3774 1,449624 1,59732 11,51172 3,468 39,71472 6,1965 93,9114 9,7002 183,6857
Pal 2 0,37536 1,458151 1,59834 11,76754 3,4527 39,77878 6,1557 93,9114 9,6339 184,7536
Pal 3 0,37536 1,449624 1,58916 11,81017 3,45576 39,97094 6,15978 94,76514 9,63492 185,8216
Page 940
SEMIT 2023
In this section, information about the aerodynamic parameters resulting from the rotation of 3
horizontally placed propellers of the same diameter and characteristics in the same direction at different
speeds is given. In table 4.6, the center diameter distance between 3 blades on the vertical axis and the
thrust and power information values obtained for each blade are given. On the horizontal axis of the
table, the number of revolutions applied in the study and the thrust information resulting from this
process are given numerically in N, and the consumed power is given numerically in W. We have given
the parameters resulting from the placement of 2 propellers in the horizontal direction in the previous
tables. In order to find out whether increasing the number of propellers in the horizontal direction will
make a difference in efficiency or thrust production, aerodynamic parameters were determined by
placing 3 propellers in the horizontal direction. It is clearly seen in Table 4.6 that the thrust values
produced and the power consumed for different propellers placed horizontally are quite close. This
means that no matter how many propellers are added on the horizontal axis, it acts as if there was a
single propeller and there is no loss of thrust. Similarly, changing the distance between the propellers
on the horizontal axis does not affect the propeller efficiency and the thrust value produced much.
Table 4.7. Horizontally placed cv-ccv-cv 3 pal Aerodynamic Parameters
2000 4000 6000 8000 10000
Distance thrust power thrust power thrust power thrust power thrust power
between
diameters
35 cm Pal 1 0,366 1,3585 1,467 11,1188 3,339 37,68 5,999 89,559 9,371 175,896
Pal 2 0,365 1,35641 1,447 11,077 3,297 37,68 5,919 89,559 9,269 175,896
Pal 3 0,362 1,36059 1,484 11,077 3,382 37,68 6,08 89,9775 9,515 176,314
40cm Pal 1 0,366 1,386715 1,4895 11,21285 3,341 38,198 5,98125 90,6052 9,3515 177,728
Pal 2 0,36475 1,38776 1,47675 11,2442 3,3087 38,009 5,91675 90,396 9,26125 177,466
Pal 3 0,36075 1,367905 1,479 11,077 3,319 37,6172 5,93925 89,559 9,292 175,320
45cm Pal 1 0,367 1,388283 1,503 11,20763 3,3552 38,2530 6,009125 90,7098 9,396 177,851
Pal 2 0,365625 1,390373 1,491875 11,27555 3,32487 38,1745 5,946875 90,6052 9,30912 177,99
Pal 3 0,362875 1,3794 1,5 11,2024 3,352 38,0254 5,999375 90,50063 9,38625 177,283
50cm Pal 1 0,368 1,38985 1,5165 11,2024 3,3695 38,308 6,037 90,8145 9,4405 177,99
Pal 2 0,3665 1,392985 1,507 11,3069 3,341 38,3394 5,977 90,8145 9,357 178,513
Pal 3 0,365 1,390895 1,521 11,3278 3,385 38,4336 6,0595 91,44225 9,4805 179,246
55cm Pal 1 0,3717 1,404062 1,53216 11,31526 3,4035 38,6973 6,09775 91,7352 9,5356 179,790
Pal 2 0,37018 1,407281 1,52267 11,42227 3,37485 38,7293 6,03735 91,7352 9,45145 180,324
Pal 3 0,36868 1,405107 1,53658 11,44359 3,41888 38,8254 6,11989 92,37132 9,57496 181,068
In this section, information is given about the aerodynamic parameters that arise as a result of three
horizontally placed propellers of the same diameter and characteristics rotating in opposite directions
at different speeds. Table 4.6 gives the center diameter distance between 3 wings on the vertical axis
and the thrust and power information values obtained for each wing. On the horizontal axis of the table,
the number of revolutions applied in the work and the thrust information resulting from this process are
given numerically in N, and the consumed power is given numerically in W. Among the propellers
placed horizontally in Table 4.7, the propeller closest to the hub on the quadcopter arm is named blade
1, and towards the end of the arm it is named blade 2 and blade 3, respectively. The aerodynamic
parameter values resulting from blade 1 rotating clockwise, blade 2 counterclockwise, and blade 3
Page 941
SEMIT 2023
rotating clockwise are given. In table 4.6, the parameter values resulting from the rotation of all
propellers in the same location and in the same direction are given. In 4.7, the parameters resulting from
the rotation of the propellers in the opposite direction are given. When the data in the table is examined,
whether they rotate in the same direction or in the opposite direction does not affect the result.
Table 4.8. Vertically placed cv-cv-cv 3 pal Aerodynamic Parameters
2000 4000 6000 8000 10000
Distance thrust Power thrust Power Thrust Power Thrust Power Thrust Power
between
(N) (W) (N) (W) (N) (W) (N) (W) (N) (W)
diameters
5 cm Pal 1 0,152 0,75867 0,6 6,1864 1,365 21,038 2,443 49,9689 3,832 97,8945
Pal 2 0,24 1,07426 0,947 8,778 2,143 29,7044 3,818 70,5591 5,975 137,157
Pal 3 0,374 1,35223 1,5 11,077 3,397 37,68 6,061 88,722 9,494 174,849
10 cm Pal 1 0,1598 0,797667 0,63084 6,504 1,43516 22,1194 2,56857 52,5374 4,028972 102,9265
Pal 2 0,2523 1,129479 0,99567 9,229 2,25315 31,23127 4,01425 74,18598 6,282127 144,2071
Pal 3 0,3932 1,421737 1,57710 11,646 3,57161 39,61683 6,37254 93,28249 9,982011 183,8366
20 cm Pal 1 0,1647 0,822483 0,65046 6,7067 1,47981 22,80755 2,64848 54,17189 4,154318 106,1286
Pal 2 0,2601 1,164618 1,02665 9,5163 2,32325 32,2029 4,13914 76,49398 6,47757 148,6936
Pal 3 0,4054 1,465969 1,62616 12,008 3,68272 40,84935 6,57080 96,1846 10,29256 189,5559
50 cm Pal 1 0,2509 1,252642 0,99066 10,214 2,25375 34,7359 4,03364 82,50379 6,327026 161,6339
Pal 2 0,3962 1,773714 1,56359 14,493 3,53831 49,04502 6,30391 116,5003 9,865339 226,4603
Pal 3 0,6175 2,232671 2,47665 18,289 5,60879 62,21355 10,0073 146,4891 15,67557 288,6937
Table 4.8 gives information about the thrust produced and the power consumed by placing 3 propellers
vertically and moving them in the same direction and at different speeds. In this table, the distance
change between propeller centers is given in the left column, and thrust, power and speed numbers are
given in the right row. In the propeller order, blade 1 is the lowest propeller, and blades 2 and blade 3
are listed upwards, respectively. In the system placed vertically and as the distance between diameters
is minimized, the differences in the thrust values produced gradually increase. In table 4.9, blade 1 is
designed to rotate clockwise, blade 2 is designed to rotate counterclockwise, and blade 3 is designed to
rotate clockwise. As in Table 4.9, when the propellers are rotated in opposite directions, the efficiency
loss is significantly reduced. When the distance between the diameters of the propellers reached 50 cm,
the thrust power produced by a single propeller was produced. However, when the distance between
diameters is too short, up to 5 cm, efficiency loss occurs.
Table 4.9. Vertically placed cv-ccv-cv 3 pal Aerodynamic Parameters
2000 4000 6000 8000 10000
Distance thrust power thrust power thrust power thrust power thrust power
between
diameters
5cm Pal 1 0,24034 1,11606 0,93445 87,78 2 28,8252 3,753 70,7265 5,974 139,251
Pal 2 0,296835 1,33551 1,25196 11,286 2,73 38,0568 5,0787 92,07 7,956 180,084
Pal 3 0,342418 1,267794 1,356 10,1574 3,037 35,1052 5,474 8,23608 8,633 161,238
10cm Pal 1 0,264374 1,227666 1,027895 9,6558 2,2 31,70772 4,1283 77,79915 6,5714 153,1761
Pal 2 0,326518 1,469061 1,377156 12,4146 3,003 41,86248 5,58657 10,1277 8,7516 198,0924
Pal 3 0,37666 1,394573 1,4916 11,17314 3,3407 38,61572 6,0214 90,59688 9,4963 177,3618
Page 942
SEMIT 2023
20cm Pal 1 0,30403 1,411816 1,182079 11,10417 2,53 36,46388 4,747545 89,46902 7,55711 176,1525
Pal 2 0,375496 1,68942 1,583729 14,27679 3,45345 48,14185 6,424556 116,4686 10,06434 227,8063
Pal 3 0,433159 1,603759 1,71534 12,84911 3,841805 44,40808 6,92461 104,1864 10,92075 203,9661
50cm Pal 1 0,358756 1,665943 1,394854 13,10292 2,9854 43,02738 5,602103 105,5734 8,91739 207,86
Pal 2 0,443085 1,993516 1,868801 16,84661 4,075071 56,80739 7,580975 137,4329 11,87592 268,8114
Pal 3 0,511128 1,892436 2,024101 15,16195 4,53333 52,40153 8,17104 122,94 12,88648 240,68
Result
In this study, numerical analyzes of the existence of more than one thrust propeller in one arm of
quadcopter aircraft in terms of consumed power and thrust efficiency were made. The quadcopter arms
were examined according to the rotation of the propeller pairs in the horizontal and vertical directions,
in the same direction and in the opposite direction. However, the propeller pairs were evaluated
comparatively as data according to a single propeller. In the horizontal direction, the efficiency loss of
the propeller pairs is almost negligible compared to the single propeller, it is seen that the efficiency is
better as the distance between the diameters increases. It was observed that when the propeller pairs are
placed vertically, the efficiency loss is high when they are rotated in the same direction, but the
efficiency loss decreases as the distance between the centers increases. When the propeller pairs are
placed vertically and run in opposite directions, it has been observed that there is almost no efficiency
loss, even if the distance between their centers is very small.
Referance
[1] Cimoli, E., Marcer, M., Vandecrux, B., Bøggild, C. E., Williams, G., & Simonsen, S. B. (2017).
Application of low-cost UASs and digital photogrammetry for high-resolution snow depth
mapping in the Arctic. Remote Sensing, 9(11), 1144.
[2] De Michele, C., Avanzi, F., Passoni, D., Barzaghi, R., Pinto, L., Dosso, P., ... & Della Vedova, G.
(2015). Microscale variability of snow depth using UAS technology. The Cryosphere
Discussions, 9(1), 1047-1075.
[3] Bühler, Y., Adams, M. S., Stoffel, A., & Boesch, R. (2017). Photogrammetric reconstruction of
homogenous snow surfaces in alpine terrain applying near-infrared UAS imagery. International
Journal of Remote Sensing, 38(8-10), 3135-3158.
[4] Jones, C., Ryan, J., Holt, T., & Hubbard, A. (2018). Structural glaciology of isunguata sermia, West
Greenland. Journal of Maps, 14(2), 517-527.
[5] Brun, F., Wagnon, P., Berthier, E., Shea, J. M., Immerzeel, W. W., Kraaijenbrink, P. D., ... &
Arnaud, Y. (2018). Ice cliff contribution to the tongue-wide ablation of Changri Nup Glacier,
Nepal, central Himalaya. The Cryosphere, 12(11), 3439-3457.
[6] Burger, C. (2007). Propeller performance analysis and multidisciplinary optimization using a
genetic algorithm. Auburn University.
[7] Pauliny, J. O. Z. E. F., and P. Imris. "The overview of propellers in general aviation." Brno
University of Technology (2012).
[8] Weick, F. E. (1930). Aircraft propeller design. McGraw-Hill Book Company, Incorporated.
[9] W. Rankine. “On the Mechanical Principles of the Action of Propellers”. In: Transactions of the
Institute of Naval Architects 6, (1865), pp. 13–39.
Page 943
SEMIT 2023
[10] R. Froude. “On the Part Played in Propulsion by Difference in Fluid Pressure”. In: Transactions of
the Institute of Naval Architects 30 (1889), pp. 390–405.
[11] E. Bratt. Aircraft Propellers Design and Analysis, 2nd ed. Flygföretag HB, 1991.
[12] E. P. Hartman and D. Biermann. “The Aerodynamic Characteristics of Full-Scale Propellers
Having 2, 3, And 4 Blades of Clark Y and R. A. F. 6 Airfoil Sections”. In: National Advisory
Committee for Aeronautics, Report No. 640 (1938).
[13] D. Biermann and E. P. Hartman. “The Aerodynamic Characteristics of Six Full- Scale Propellers
Having Different Airfoil Sections”. In: National Advisory Committee for Aeronautics, Report
No. 650 (1939), pp. 145–178.
[14] Worobel, R., and Mayo, M. G., “Advance General Aviation Propeller Study,” NASA CR-
114289, April 1971.
[15] “Generalized Method of Propeller Performance Estimation,” Hamilton Standard, Preliminary
Design Group, Rept. PDB-6101A, Windsor Locks, CT, June 1963.
[16] McCormick, B. W., Jr, Aerodynamics of V/STOL Flight, Academic Press, New York, 1967, p.
93.
[17] Abbott, I. H., andVon-Doenhoff, A. E., Theory of Wing Sections, Dover, New York, 1959, pp.
449–668.
[18] McLemore, C. H., and Cannon, M. D., “Aerodynamic Investigation of a Four-Blade Propeller
Operating Through an Angle of Attack Range from 0 to 180 Degrees,” NACA TN-3228, June
1954.
[19] Evans, J. A., and Liner, G., “A Wind Tunnel Investigation of the Aerodynamic Characteristics of
a Full-Scale Sweptback Propeller and Two Related Straight Propellers,” NACA RM-L50J05,
Jan. 1951.
[20] Hassanalian, M., & Abdelkefi, A. (2017). Classifications, applications, and design challenges of
drones: A review. Progress in Aerospace Sciences, 91, 99-131.
[21] Hassanalian, M., Salazar, R., & Abdelkefi, A. (2019). Conceptual design and optimization of a tilt-
rotor micro air vehicle. Chinese Journal of Aeronautics, 32(2), 369-381.
[22] Fagiano, L. (2017). Systems of tethered multicopters: modeling and control design. IFAC-
PapersOnLine, 50(1), 4610-4615.
[23] Rajendran, P., & Jayaprakash, A. (2023). Numerical performance analysis of a twin blade drone
rotor propeller. Materials Today: Proceedings, 80, 492-498.
[24] Afthon, M. F., & Moelyadi, M. A. (2021). A Study in Aerodynamic Optimization of UAV
Helicopter Rotor-Blades Planform in Vertical Motion. AVIA, 2(1).
[25] Yoon, S., Lee, H. C., & Pulliam, T. H. (2016). Computational analysis of multi-rotor flows. In 54th
AIAA aerospace sciences meeting (p. 0812).
[26] Zhu, H., Nie, H., Zhang, L., Wei, X., & Zhang, M. (2020). Design and assessment of octocopter
drones with improved aerodynamic efficiency and performance. Aerospace Science and
Technology, 106, 106206. Cfd örnekleri
Page 944
SEMIT 2023
A Study on the Physical Properties of Elaborated Graphene

Oudjertli Salah1,*
1
Research Center In Industrial Technologies. (CRTI)
BP 64, Roade of Dely Brahim, Cheraga 16014 Algiers – Algeria
*Corresponding author: salah.oudjertli@gmail.com
Abstract
Graphene holds promise, for nanotechnology applications in information technology and computing. It
offers prospects for the development of electronic devices such as transistors, sensors, batteries and
supercapacitors. Additionally its spintronics property makes it a potential candidate for quantum
computing. In this regard, we have successfully synthesized graphene using the cost method of reducing
graphene oxide (rGO). Our research focuses on exploring the properties of the graphene encompassing
electron mobility calculations Hall Effect analysis, electrical conductivity assessment, mechanical
strength evaluation, cohesion measurements and flexibility. These experimental findings play a role in
understanding the characteristics of the synthesized graphene and gauging its potential, for electronic
and optical applications.
1. Introduction
Graphene [1] is an substance that has captured a great deal of attention in the realm of materials science.
It's a form of carbon that exists in two dimensions, with an structure and extraordinary characteristics.
Graphene consists solely of carbon atoms (C) that are bonded to three carbon atoms through covalent
bonds forming a hexagonal pattern. This hexagonal arrangement repeats throughout the layer of
graphene resulting in a smooth surface at the atomic level. The remarkable aspect of graphene is its thin
atomic thickness measuring only one carbon atom thick. In essence it can be thought of as a 2D material
with no thickness when compared to its length and width. This unique two dimensional structure grants
graphene properties, including electrical conductivity, impressive mechanical strength and impressive
optical transparency.
As discussed earlier thanks to its properties graphene [2] has opened up opportunities for revolutionary
applications [3, 4] across various fields. This two dimensional carbon based material presents qualities
such as conductivity, remarkable mechanical strength and excellent optical transparency. These
characteristics make graphene an adaptable substance, with potential to reshape industries and fuel
advancements. Graphene, a material being explored as a substitute, for silicon in circuits holds promise
in the field of electronics. It boasts high electron mobility, fast transfer speeds and exceptional
conductivity. This potential application of graphene could revolutionize devices by making them faster
more efficient and more compact. Imagine the possibilities for next generation smartphones, computers
and other advanced electronics. But that's not all – graphene has also captured attention in the energy
sector. Its use in cells could significantly enhance the conversion efficiency of light into electricity
potentially transforming the energy industry as we know it. Furthermore graphene has proven valuable
in supercapacitors due to its ability to store amounts of energy and charge/discharge rapidly. This
breakthrough could revolutionize energy storage systems and electric vehicles. The versatility of
graphene doesn't stop there; it is also making waves in materials. By incorporating graphene into
substances like polymers, ceramics and metals their mechanical strength, thermal properties and
electrical characteristics can be greatly improved. Industries such as aerospace, automotive
manufacturing, construction – any field requiring sturdy materials – have taken notice. There's more to
graphenes repertoire; it can be used for creating coatings and thin films that provide barriers against
corrosion and scratches. In summary; Graphene holds potential not in electronics but also in energy
applications such as solar cells and supercapacitors. It is transforming industries with its reinforcement
capabilities for materials while offering coatings, with anti corrosion properties. These coatings can be
Page 945
SEMIT 2023
used on surfaces ranging from construction materials to devices, which enhances their durability and
performance. Graphene also finds applications, in sensors that detect gases, biomolecules and
pollutants. Moreover it holds potential in the field, for drug delivery biosensing and tissue engineering.
The research discussed here primarily focuses on producing graphene using a process called reduced
oxide (rGO) [5]. It involves evaluating electron mobility through the Hall Effect, conductivity,
mechanical strength, cohesion and flexibility.
2. Procedure, for synthesizing graphene

The technique commonly used to produce high quality graphene is the reduced oxide (rGO) method [6,
7]. This method involves steps. I will provide you with a detailed explanation. To start take 0.5 g of
oxide (GO). Mix it with 50 ml of distilled water. Use a device to create a suspension, by sonication for
one hour. After that add 1.5 g of sodium nitrate (NaNO3) to the GO suspension. Stir the mixture for
half an hour. Now comes the careful part; slowly add 25 ml of sulfuric acid (H2SO4) to the GO NaNO3
mixture while stirring continuously.
You might notice that the mixture becomes hot and thick during this process. Continue stirring for
another half an hour so that a sulfonic ester forms on the surface of GO due to chemical reactions. Once
thats done gradually add 1 g of phosphoric acid (H3PO4) to the GO NaNO3 H2SO4 mixture while
ensuring a stir. Keep stirring for another half an hour so that phosphoric bonds form between oxygen
molecules on the surface of GO due, to these reactions. After that we add a 6 ml of hydrazine hydrate
to the mixture of GO NaNO3 H2SO4 H3PO4. We stir it for about 30 minutes. This stirring process
helps reduce the GO and transform it into graphene. Once this is done we transfer the resulting solution,
to an oven that has been preheated to 95°C. We keep it there for a period of 24 hours in order to dry out
the graphene that has been formed. Once the drying process is complete we take out the graphene from
the oven. Proceed to wash it with distilled water and acetone. This ensures that any impurities or
residues from the manufacturing process are completely eliminated. Finally we dry the washed
graphene by placing it in a set, at 60°C for 6 hours. This entire process allows us to obtain our high
quality graphene through reduction of graphene oxides (GO).

3.1 Calculation of electron mobility and Hall Effect in graphene
It is of importance, in the field of research to calculate the electron mobility and Hall Effect of graphene
produced through the chemically reduced graphene oxide (rGO) method. Here are detailed explanations
as to why these calculations hold significance:
- Understanding Electrical Transport Properties: The electron mobility serves as a parameter that helps
us understand how efficiently charge carriers (electrons) can move through a material when an electric
field is applied. By calculating the electron mobility of rGO researchers gain insights into the materials
transport properties. This information is particularly important for applications in electronics and
optoelectronics.
- Assessing Material Quality: The reduced oxide (rGO) method can introduce imperfections such as
defects, impurities and structural disorders in the graphene structure. These imperfections have an
impact on the properties of graphene including its electron mobility. Calculating the electron mobility
allows researchers to evaluate and determine the quality and purity of rGO materials. Lower values of
electron mobility may suggest the presence of defects or impurities whereas higher values indicate a
quality and cleaner sample of graphene.
- Understanding Scattering Mechanisms for Carriers: The calculation of electron mobility provides
insights, into scattering mechanisms that affect electron transportation in rGO materials.
-Scattering processes, like phonon scattering, impurity scattering and edge scattering can have an
impact on the mobility of electrons. Researchers can gain an understanding of the factors that limit
Page 946
SEMIT 2023
electron transport in rGO by quantifying electron mobility.

- Improving Device Performance: Graphene based electronic and optoelectronic devices rely on the
movement of charge carriers. The electron mobility of rGO directly affects the performance of these
devices. Accurately calculating electron mobility allows researchers to optimize device designs,
electrode configurations and material processing techniques to enhance device performance. This
optimization enables control and manipulation of charge carriers leading to device efficiency and
functionality.
- Hall Effect Measurements: The Hall Effect refers to a phenomenon where applying a field to a current
carrying conductor induces a voltage perpendicular to both the current and magnetic field. By
conducting Hall Effect measurements on rGO researchers can determine the type (electrons or holes)
concentration, sign and magnitude of charge carriers as valuable information about charge carrier
density and the nature of charge transport, in rGO.
Understanding the properties of graphene produced through reducing graphene oxide (rGO) is essential.
Calculating electron mobility and conducting Hall effect measurements helps evaluate material quality
identify scattering mechanisms, optimize device performance and determine charge carrier
characteristics. These calculations and measurements play a role, in advancing graphene-based
technologies and facilitating the development of performance electronic and optoelectronic devices.
In this context, we have experimentally calculated the electron mobility and Hall Effect of our graphene
[8] produced by the chemically reduced graphene oxide (rGO) method, to begin the experiment we start
by cleaning the sample using ethanol. Next, we clean the Hall bar device with ethanol. Then etch it to
remove any surface contaminants. Once cleaned we carefully mount the Hall bar device onto a substrate
using clean room technology.
After preparing the setup, we connect the Hall bar device to both the DC power supply and oscilloscope.
The positive terminal of the power supply is connected to the Hall bar device while the ground terminal
of the oscilloscope is also connected to it.
To ensure safety and avoid damaging the sample we apply a current to the Hall bar device. This current
is carefully set at a level.
In order to perform measurements accurately a magnetic field is applied in a direction, to both the Hall
bar device.
Finally we proceed by measuring what is known as Hall voltage, across the Hall bar. This voltage is
measured in a direction that runs perpendicular to both flow and magnetic field orientation.
To determine the electron mobility and Hall coefficient we can use the formulas:
μ = RH / ne (1)
R = V Hall / I (2)
The electron mobility μ is calculated by dividing the Hall coefficient, RH, by the carrier concentration,
n multiplied by the charge e. The Hall voltage is represented as V Hall and the current as I.
Upon analyzing our Reduced Graphene Oxide (RGO) sample that underwent reduction at 2500 K. we
found that the electron mobility is approximately 75 cm2/V•s. Additionally our calculations indicate a
Hall voltage for this sample. This suggests that electrons are the majority carriers and yields a Hall
coefficient of ~10−9 cm3/C. These results demonstrate the exceptional quality of graphene produced
through our method.
When conducting measurements for the Hall voltage using an oscilloscope it is crucial to ensure its
sensitivity due, to its magnitude. Furthermore to avoid inducing current in the sample [9] it is
recommended to apply any magnetic fields. In addition, it is advisable to measure the Hall voltage, at
magnetic field values to confirm linearity.
3.2 Calculation of electrical conductivity
Page 947
SEMIT 2023
The ability of graphene to carry current is measured by its electrical conductivity. Graphene, known for
its electrical conductivity is widely used in electronic applications, like transistors and solar cells.
Calculating the electrical conductivity of our graphene produced by the chemically reduced graphene
oxide (rGO) method is crucial for several reasons:
Firstly, it directly reflects the quality of the material. Imperfections such as defects and impurities can
lower the conductivity. Researchers can assess the materials quality. Identify areas for enhancement by
measuring rGOs conductivity.
Secondly, different reduction methods can influence rGOs electrical conductivity. These methods
produce rGO with varying properties, including conductivity. By measuring, the electrical conductivity
of rGO obtained through techniques researchers can optimize reduction processes to yield high quality
materials.
Lastly understanding rGOs electrical conductivity is essential for designing electronic devices. Armed
with knowledge, about the materials conductivity designers can optimize device designs to achieve
performance. Here are some specific instances that highlight the significance of calculating the
conductivity of rGO:
- Graphene Field Effect Transistors (FETs); FETs are devices utilized for amplifying and switching
signals. The speed and performance of FETs greatly rely on the conductivity of graphene.
- Graphene Solar Cells; Solar cells convert sunlight into energy. The efficiency of cells is determined
by the electrical conductivity of graphene.
- Sensors; Sensors are capable of detecting substances, like gases, chemicals and biomolecules. The
sensitivity and accuracy of sensors depend on the conductivity of graphene.
In general, evaluating the quality of reduced oxide (rGO) optimizing reduction processes and designing
high performance electronic devices all necessitate calculating the electrical conductivity of rGO.
In this context, we have experimentally calculated the electrical conductivity of our graphene. To
calculate the electrical conductivity of our sample of graphene produced by the chemically reduced
graphene oxide (rGO) method, we used specific materials. These materials include the sample itself a
four-point probe, a DC power supply and a digital multimeter. Our process began by cleaning the
sample and its substrate, with isopropyl alcohol. Next, we connected the four-point probe to the sample.
Set the DC power supply to a low voltage. Using the multimeter we measured the current flowing
through our sample. The electrical conductivity was then calculated using the equation;
σ=I/V (3)
It's worth mentioning that we took measures to ensure that our graphene sample had contact resistance
by maximizing its size. Additionally we made sure that there were no impurities or defects in our
sample.
The obtained value for the conductivity of our sample was 350 S/m indicating excellent quality. It's
important to note that in applications such, as transistors, sensors and composites the electrical
conductivity of graphene plays a role.
3. 3 Calculation of mechanical strength
The strength of graphene produced using the reduced oxide (rGO) method refers to its ability to
withstand deformation or breakage when subjected to external forces. Graphene is well known for its
strength, which makes it highly suitable, for a wide range of applications including composite materials,
electronic devices and nanostructures. Assessing the strength of rGO graphene serves purposes:
- It helps evaluate the quality of rGO graphene. The presence of imperfections like defects and
impurities can undermine its strength. By measuring this parameter researchers can assess the quality
of rGO graphene. Identify areas that need improvement.
- It enables prediction of the performance of rGO graphene in applications. Mechanical strength plays
a role in determining a materials capacity to bear loads withstand shocks and absorb energy.
Page 948
SEMIT 2023
Understanding the strength allows researchers to design performance composite materials, electronic
devices and nanostructures incorporating rGO graphene.
- It facilitates studying the properties of rGO graphene under conditions. Factors such as temperature
variations, pressure changes and application of fields can impact the materials strength.
Overall accurately determining the strength is essential, for optimizing aspects related to rGO graphenes
quality assessment, application design and understanding its behavior under diverse conditions. By
examining the attributes of rGO graphene, in scenarios scientists can gain valuable insights into its
mechanical properties and explore new potential uses for this material. Here are a few specific instances
that highlight the significance of assessing the strength of rGO graphene:
- Enhancing materials: Incorporating rGO graphene into polymers and metals can bolster their strength
and hardness. The mechanical strength of rGO graphene plays a role in determining how sturdy and
resilient these composite materials become.
- Powering devices: Employing rGO graphene in the production of transistors and solar cells holds
promise. The mechanical strength of rGO graphene is crucial in ensuring the reliability and longevity
of these devices.
- Building nanostructures: Crafting nanotubes and nanosheets with rGO graphene opens up possibilities.
The mechanical strength of rGO graphene directly impacts the stability and mechanical characteristics
of these nanostructures.
In general, calculating the strength of reduced graphene oxide (rGO) provides invaluable insights into
material quality predicts its performance, in specific applications and contributes to our understanding
of its physical properties across diverse conditions. This information can be utilized to create materials
manufacture electronic devices that are dependable and long lasting and explore new possibilities for
nanostructures.
In this context, we have experimentally calculated the mechanical strength of our graphene synthesized
by the chemical reduction method of graphene oxide (rGO). To measure the mechanical strength of our
graphene produced by the chemical reduction method of graphene oxide (rGO), we need specific
equipment, like a tensile testing machine, rGO samples, tweezers, a blade, a balance and solvents such
as water or methanol. We also require a dispersing agent like polyvinylpyrrolidone (PVP) and an oven.
Here is how we proceed:
1. We start with high quality rGO film obtained through the chemical reduction of oxide using
hydrazine.
2. After obtaining the rGO film we shape it into specimens with a sectional area of 1 mm². The cross
sectional area is crucial as it determines the force needed to cause rupture.
3. It is worth noting that rGO is a material comprising interconnected networks with oxygenated groups.
These oxygenated groups can impact the strength of rGO. Therefore, to ensure results we take measures
to minimize the presence of oxygenated groups during rGO preparation. One effective technique is
using chemical reduction methods, like hydrazine reduction.
4. Additionally we control the size of the rGO flakes since smaller flakes tend to exhibit strength.
Additionally to minimize the presence of oxygenated groups we utilized hydrazine at a temperature of
80°C, for a duration of 30 minutes since it has the ability to reduce rGO. Subsequently in order to
regulate the size of rGO particles we dispersed the rGO in a solvent like methanol along with a
dispersing agent known as polyvinylpyrrolidone (PVP). The resulting mixture was then subjected to
sonication to break down the rGO particles into flakes. Finally centrifugation was carried out to separate
the rGO flakes from the solution.
When it comes to mounting the rGO specimens for testing purposes they are securely placed within a
testing machine in such a way that controlled tensile forces can be applied. Typically two supports on
either end are used to hold and evenly distribute the force on the specimen. The application of force is
Page 949
SEMIT 2023
done at a strain rate which plays a role in how the specimen deforms and eventually fractures. In our
case this strain rate is set at 0.001 mm/s.
During data collection both force and deformation measurements are recorded throughout the test. A
force sensor is utilized for measuring force while deformation is measured using a strain sensor. These
measurements provide insights into how both force and deformation change, during testing.
The collected data is then used to determine the strength. In order to calculate the strength of rGO we
divide the breaking force by the specimen’s sectional area. The breaking force refers to the force that
can be applied before the specimen breaks. The cross sectional area represents the surface on which the
force is exerted. Based on our calculations we found that the mechanical strength of rGO is 30
megapascals indicating quality, for our sample.
3.4 Calculation of cohesion
The strength of the chemical bonds, between carbon atoms in graphene determines its cohesion.
Graphene is widely recognized for its strength and durability. Calculating the cohesion of graphene
produced through the reduced oxide (rGO) method holds significant importance for several reasons:
- It enables the assessment of rGO quality. Imperfections like defects and impurities can lower the
cohesion of rGO graphene. By measuring its cohesion, researchers can evaluate the materials quality.
Identify areas that need improvement.
- It aids in predicting the performance of rGO graphene. Cohesion plays a role in determining the
strength, hardness and toughness of rGO graphene. By understanding, its cohesion researchers can
predict how it will perform mechanically and tailor it for applications.
- It facilitates studying the properties of rGO graphene under conditions. Factors such as temperature,
pressure and electric fields can influence the cohesion of rGO graphene. By examining its cohesion
under conditions, researchers gain insights into its properties and explore novel applications, for this
remarkable material.
The significance of assessing the cohesion of rGO graphene can be illustrated through examples:
- Composite materials: rGO graphene plays a role, in strengthening materials like polymers and metals.
The cohesion of rGO graphene determines the robustness and hardness of these reinforced composites.
- Electronic devices: rGO graphene finds application in manufacturing devices like transistors and solar
cells. The cohesion of rGO graphene is vital for ensuring the reliability and durability of devices.
- Nanostructures: By utilizing rGO graphene nanostructures such as nanotubes and nanosheets can be
created. The cohesion of rGO graphene is pivotal in determining the stability and mechanical properties
of these nanostructures.
In general, calculating the cohesion of reduced oxide (rGO) produced by the method offers valuable
insights into material quality enables prediction of mechanical performance and facilitates studying
physical properties under various conditions. This information proves instrumental in designing
performance materials fabricating dependable electronic devices and exploring new applications, for
nanostructures.
In this context, we have experimentally calculated the cohesion in our graphene synthesized by the
chemical reduction method of graphene oxide (rGO). To determine the cohesion of our graphene
produced by the chemical reduction method of graphene oxide (rGO), we need equipment such, as a
balance, heating plate, beaker, rGO solution, chemical reduction solution and UV visible
spectrophotometer. For the reagents used in this experiment we utilized both rGO solution and chemical
reduction solution.
Here's the procedure:
1. We start by dispersing the rGO solution in a solvent.
2. Next we introduce the chemical reduction solution to the rGO solution.
Page 950
SEMIT 2023
3. The combined solution is then heated on a heating plate at 100 °C for 30 minutes.
4. Afterward we let the solution cool down to room temperature.
5. To remove any particles we filter the solution.
6. Finally at a wavelength of 270 nm we measure the absorbance of the filtered solution.
The cohesion is calculated using this equation:
σ = (A2-A1) / 2R (4)
In this equation:
σ represents cohesion measured in J/m²
A2 denotes absorbance after reduction
A1 represents absorbance before reduction
R indicates the radius of particles, in meters.
The cohesion of our graphene, which is synthesized using a chemical reduction method called oxide
(rGO) is measured to be 0.8 J/m². It's worth noting that cohesion refers to the force that holds carbon
atoms together within a sheet. We determine cohesion by measuring the force required to break the
sheet. In our process we calculate cohesion by assessing the absorbance of the rGO solution after
reduction. Through reduction we restore sp2 covalent bonds, between carbon atoms resulting in an
increase in solution absorbance. The obtained value, for our samples cohesion was 0.8 J/m² indicating
quality.
3.5 Calculation of flexibility
Calculating the flexibility of reduced oxide (rGO) holds significant importance as it helps in
understanding the materials mechanical properties and its ability to bend without fracturing. Here are a
few key reasons highlighting the importance of calculating rGO flexibility:
- Understanding properties: Flexibility plays a role in determining how well rGO can bend without
breaking. By calculating flexibility, we can gather information, about rGOs resistance to deformation
its capability to conform to curved surfaces and its ability to withstand stress while maintaining its
structural integrity. This understanding is vital for assessing the practicality of rGO in applications
where flexibility is required, such as devices or sensors.
- Designing devices: Accurate calculation of rGO flexibility is crucial during the design phase of
devices. Flexible materials are frequently utilized in creating devices, sensors or coatings that need to
adapt to curved surfaces or endure deformations without compromising their functionality. By
calculating rGO flexibility we can evaluate its compatibility with device requirements. Select the most
suitable material that guarantees both adequate flexibility and long lasting performance.
- Comparative analysis with materials: Calculating rGOs flexibility enables us to compare it with
materials, like monolayer graphene, polymers or metals.
It is crucial to compare and evaluate the advantages and limitations of rGO in terms of flexibility. For
instance, rGO might exhibit flexibility than monolayer graphene, which can be advantageous, for
applications that require deformation. By understanding how rGO compares to materials, one can
choose the suitable material for a specific application. To optimize the properties of rGO it is important
to calculate its flexibility. This calculation helps in refining chemical reduction methods and
experimental conditions to achieve desired properties. By adjusting chemical reduction, parameters or
implementing treatments it becomes possible to enhance the flexibility of rGO. This optimization plays
a role in ensuring mechanical performance for specific applications.
It should be noted that accurate calculation of rGOs flexibility requires techniques like bending or
tensile tests. These methods allow measurement of deformation and determination of flexibility.
Page 951
SEMIT 2023
Calculating rGOs flexibility is essential for characterizing its properties designing devices comparing
with other materials and optimizing its use in various applications where flexibility is crucial.
In this context, we conducted calculations, on the flexibility of graphene produced through the reduced
graphene oxide (rGO) method. To assess the flexibility of our graphene we utilized materials and
followed the steps outlined below:
Here is the procedure:
a. Prepare a solution of oxide (rGO) by dissolving graphene oxide (GO) in an organic solvent, like
acetone or ethanol.
b. Apply a drop of the rGO solution onto a substrate such as glass.
c. Let the solution evaporate, resulting in the formation of an rGO film.
d. Cut the rGO film into a shaped specimen for flexural testing.
-Fixation of the rGO specimen in the testing device:
a. Securely fasten the rGO specimen within a testing device to ensure it undergoes bending force.
b. Ensure that the fixation is sturdy enough to prevent any movement during testing.
-Application of force:
a. Gradually apply a force to the specimen at a rate of displacement.
b. Record any deformation exhibited by the specimen at each stage of testing.
-Measurement of deformation:
a. Determine deformation by comparing changes, in specimen curvature with its curvature.
Calculation of flexibility:
The flexibility of our rGO sample is determined by calculating the slope of the deformation force curve
using the equation:
f = Δy / F (5)
Where f represents flexibility in meters, per newton (m/N), Δy denotes the change in specimen curvature
in meters (m). F represents the applied force in newtons (N). Based on our calculations we found that
the flexibility value is 0.03 m/N indicating flexibility.
To ensure results it is crucial to calibrate the testing device. Using a device can lead to measurements
so conducting multiple tests for reliability is highly recommended. It's worth mentioning that this
experimental protocol can be used specifically to measure the flexibility of rGO synthesized using the
reduced graphene oxide method. These measurements help characterize the properties of rGO and allow
comparison between batches.
In addition to the described protocol, there is a method that utilizes Raman spectroscopy. This approach
measures changes, in Raman frequency as stress is applied to the specimen and flexibility is then
calculated from the slope of the frequency stress curve.
This new experimental procedure is less invasive compared to the one mentioned earlier because it
doesn't involve cutting the rGO sample. However there might be a decrease, in precision due, to
measurement errors. When comparing these two procedures the flexural testing device approach offers
precision but requires cutting the sample making it more invasive. The choice of which procedure to
use depends on the objectives of the experiment. If you need precision it is recommended to use the
flexural testing device approach. On the hand if you prefer an invasive method you can opt for the
Raman spectroscopy approach.
Page 952
SEMIT 2023
4. Conclusion
In summary using reduction on oxide (rGO) is a crucial technique for producing high quality graphene
with exceptional electronic properties. This makes it highly desirable for nanotechnology applications
in fields such, as information technology and computing.
Graphene, as a material possesses characteristics such as high electron mobility, surface area and
mechanical strength. These properties make it highly suitable for devices like transistors and sensors.
Understanding the properties of reduced graphene oxide (rGO) is crucial in optimizing its performance
and studying key factors such as electron mobility, Hall Effect, electrical conductivity, mechanical
strength, cohesion and flexibility. This knowledge can pave the way for advancements in fields like
information technology and computing where there is a growing demand for performance components.
In summary the reduced graphene oxide (rGO) method and its associated calculations play a role in
developing graphene based technologies for nanotechnology applications, in information technology
and computing.
References
[1] Rümmeli MH, Rocha CG, Ortmann F., Ibrahim I., Sevincli H., Börrnert F., Kunstmann J.,
Bachmatiuk A., Pötschke M., Shiraishi M. Graphene : Reconstitution. Adv. Mater. 2011 ; 23 : 4471-
4490. doi : 10.1002/adma.201101855. [ PubMed ] [ CrossRef ] [ Google Scholar ]
[2] AK Geim et KS Novoselov, La montée du graphène, dans Nanoscience and Technology: a
Collection of Reviews from Nature Journals , World Scientific, 2010, pp. 11–19 .
[3] CK Chua et M. Pumera, Réduction chimique de l'oxyde de graphène : un point de vue de la chimie
synthétique, Chem. Soc. Rév. , 2014, 43 (1), 291–312
[4] N. Karikalan, R. Karthik, S.M. Chen, C. Karuppiah, A. Elangovan Sonochemical synthesis of
sulfur doped reduced graphene oxide supported CuS nanoparticles for the non-enzymatic glucose
sensor applications, Sci. Rep., 7 (1) (2017), pp. 1-10
[5] H. Saleem, M. Haneef, H.Y. Abbasi, Synthesis route of reduced graphene oxide via thermal
reduction of chemically exfoliated graphene oxide, Mater. Chem. Phys., 204 (2018), pp. 1-7
[6] Cai, W., Zhu, Y., Li, X., Piner, RD, Ruoff, RS: Large-area few-layer graphene/graphite films as
transparent thin conducting electrodes. Appl. Phys. Lett. 95 , 123115 (2009). doi :
10.1063/1.3220807
[7] Li, X., Zhu, Y., Cai, W., Borysiak, M., Han, B., Chen, D., Piner, RD, Colombo, L., Ruoff, RS:
Transfer of large-area graphene films for high-performance transparent conductive electrodes. Nano
Lett. 9 , 4359 (2009). doi: 10.1021/nl902623y
[8] Luis E. F. Foa Torres, S. Roche et J-C. Charlier. Introduction to graphene-based nanomaterials:
From Electronic Structure to Quantum Transport. Cambridge University Press, 2 edition, 2020.
[9] Adam S., Hwang E. H., Galitski V. et Sarma S. D. A self-consistent theory for graphene transport.
Proceedings of the National Academy of Sciences, 104:18392-18397, 2007.
Page 953

SEMIT2023 Full - Conference - Book - 07 12 2023 .

Uploaded by

Copyright:

Available Formats

SEMIT2023 Full - Conference - Book - 07 12 2023 .

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SEMIT2023 Full - Conference - Book - 07 12 2023 .

Uploaded by

Copyright:

Available Formats

Table of Contents

Publisher and Indexing .............................................................................................................. 2

Indexing: SCIE – JCR (IF:1.7)

Operations Research Perspectives

Indexing: Scopus- JCR (IF: 2.5)

Special Issue Title:

International Journal of Supply

Journal of Turkish Operations

Indexing: EBSCO - TRDizin

Journal of Optimization and

International University of The University of Texas at Dallas, Prague University of

Kent Business School, UK University of Sfax, Tunisia University of Sfax, Tunisia

Manufacturing Engineering Abou Bekr Belkaid Tlemcen Suleyman Demirel University,

The conference team would like to thank to Laboratoire de Recherche

Prof. Dr. Hasan Okuyucu

RefConf organization holds International Conference on

The International Conference on Science, Engineering Management and Information

Figure 1. Geographical Diversity of Conference Committees

Figure 2. SEMIT 2023 Compared to SEMITs 2022

Other Countries’ Participation

Figure 3. Coutnries’ Participation in SEMIT 2023

Prof. Mete Gundogan

Ankara Yildirim Beyazit University Prof. Dr. Ali Cengiz KÖSEOĞLU

Prof. Hasan Okuyucu

Dr. Babek D. Erdebilli,

Conference Coordinators: Leyla Chehrghani

Dr. Zohreh Molamohamadi

Dr. Roya Soltani

Steering Committee Members Prof. A. Mirzazadeh

Assoc. Prof. Dr. Babek Erdebilli

Prof. Janny M.Y. Leung

Prof. Gerhard Wilhelm Weber

Prof. Kok Lay Teo

Prof. Ruben Ruiz Garcia

Prof. Sırma Zeynep Alparslan Gök

Dr. Yavuz Selim Ozdemir

Dr. Ibrahim Yilmaz

Dr. Abdullah Yildizbasi

Editorial Committee Members: Prof. Gerhard Wilhelm Weber

Dr. Arpan Kumar Kar

Dr. Erfan Babaee Tirkolaee

Prof. Elif Kılıç Delice

Dr. Zohreh Molamohamadi

Dr. Mustapha Oudani

Dr. Eloisa Macedo

Dr. Tatiana Tchemisova

Dr. Dinh Tran Ngoc Huy

Scientific Committee Members: Prof. Kathryn Stecke

Prof. Josef Jablonsky

Prof. Bernardo Almada-Lobo

Prof. Taicir Moalla Loukil

Prof. Stefan Pickl

Prof. Leopoldo Eduardo Cárdenas-Barrón

Prof. Jie Sun

Prof. Ergun Eraslan

Prof. Mehmet Kabak

Prof. Mete Gundogan

Dr. Sankar Kumar Roy

Prof. Sadia Samar Ali