978 3 031 29382 5

Lecture Notes in Management and Industrial
Engineering
Series Editor
Adolfo López-Paredes, Department of Economics and Business Administration,
TEP223 Operations & Sustainability/INSISOC, University of Malaga, Spain
This book series provides a means for the dissemination of current theoretical and
applied research in the areas of Industrial Engineering and Engineering Management.
The latest methodological and computational advances that can be widely applied by
both researchers and practitioners to solve new and classical problems in industries
and organizations contribute to a growing source of publications written for and by
our readership.
The aim of this book series is to facilitate the dissemination of current research
in the following topics:
• Strategy and Entrepreneurship
• Operations Research, Modelling and Simulation
• Logistics, Production and Information Systems
• Quality Management
• Product Management
• Sustainability and Ecoefficiency
• Industrial Marketing and Consumer Behavior
• Knowledge and Project Management
• Risk Management
• Service Systems
• Healthcare Management
• Human Factors and Ergonomics
• Emergencies and Disaster Management
• Education
Luis R. Izquierdo · José Ignacio Santos ·
Juan José Lavios · Virginia Ahedo
Editors
Industry 4.0: The Power

of Data
Selected Papers from the 15th International
Conference on Industrial Engineering
and Industrial Management
Editors
Luis R. Izquierdo José Ignacio Santos
Department of Management Engineering Department of Management Engineering
University of Burgos University of Burgos
Burgos, Spain Burgos, Spain
Juan José Lavios Virginia Ahedo

Department of Management Engineering Department of Management Engineering
University of Burgos University of Burgos
Burgos, Spain Burgos, Spain
ISSN 2198-0772 ISSN 2198-0780 (electronic)

Lecture Notes in Management and Industrial Engineering
ISBN 978-3-031-29381-8 ISBN 978-3-031-29382-5 (eBook)
https://doi.org/10.1007/978-3-031-29382-5
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2023
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
Management and Industrial Engineering constitutes a broad and diverse discipline

aimed at fostering not only firms’ economic success but also employees’ satisfaction
and the achievement of high-quality standards.
The unprecedented growth in the size and complexity of corporations over the last
decades—and, thus, of their management and operation—has led to the evolution of
the discipline toward a greater emphasis on mathematical and computer modeling.
This, together with the extraordinary increase in data availability and processing
capacities—the “Big Data Revolution”—has resulted into computers, mathematical
models, and artificial intelligence becoming an integral part of Management and
Industrial Engineering.
Along these lines, data and analytics are becoming one of the most critical factors
in determining competitive differentiation. Firms are undertaking a host of initiatives
to promote data-driven decision-making with the goal of increasing revenue, opera-
tional efficiency, and other core business outcomes. Nonetheless, challenges abound
throughout the whole data management and analysis lifecycle. The most common
difficulty is the lack of integration of data management platforms with analytics,
business intelligence, and data science platforms.
Due to the evolving nature of both Management and Industrial Engineering and the
Big Data Revolution, we are now immersed in what is known as the fourth industrial
revolution or Industry 4.0 and looking forward to Industry 5.0. The term Industry
4.0 has been widely used in the scientific literature to designate the trend toward
automation and data exchange in manufacturing processes and technologies, which
includes, but is not limited to: Industrial Internet of Things (IoT), cloud computing,
cognitive computing, and artificial intelligence. As for Industry 5.0, it complements
the existing Industry 4.0 paradigm by underscoring research and innovation as drivers
for a transition to a sustainable, human-centered, and resilient industry.
In an attempt to integrate all of the above, Industry 4.0: The Power of Data
was the motto of the “15th International Conference on Industrial Engineering
and Industrial Management (ICIEIM)—XXV Congreso de Ingeniería de Organi-
zación (CIO 2021)”, which was held online on July 8 and 9, 2021, and was hosted
by Universidad de Burgos. This conference was promoted by European Academy
v
vi Preface
for Industrial Management (AIM) and Asociación para el Desarrollo de la Inge-

niería de Organización (ADINGOR). The virtual space created for the conference
brought together more than 150 scholars and practitioners from several countries,
who actively discussed information on the most recent and relevant research, theo-
ries, practices, and educational approaches in Industrial Engineering and Operations
Management.
The abstracts of the 88 papers presented at the conference were published online
under CC BY 4.0 (DOI: 10.5281/zenodo.5109479). After the conference, a selection
of papers was made based on their reviews (a minimum of two reviews for each paper)
and the feedback provided by the chairperson of the session at which the paper was
presented. The authors of the selected papers were invited to submit a full paper to
be published in this book in the following months. Once submitted, another review
process was conducted for these full papers and, in the end, 38 full papers were
selected to be published in this book.
Thus, this book compiles extended and improved versions of selected papers
presented at the conference. In this way, it constitutes a representative compendium
of the state of the art and future trends in Management and Industrial Engineering.
The contributions have been structured into nine parts:
1. Education in Organizational Engineering
2. Management Information Systems and Knowledge Management
3. Operations Research, Modelling and Simulation
4. Product Design, Industrial Marketing and Consumer Behaviour
5. Production Planning and Control
6. Project and Process Management
7. Strategy, Innovation, Networks and Entrepreneurship
8. Supply Chain Management and Logistics
9. Sustainability, Eco-efficiency and Quality Management.
The editors would like to thank all contributors of this book, including the authors,
reviewers, and Springer’s production team. In addition, the editors deeply appreciate
the efforts and interest of all CIO2021 attendees, sponsors, and contributors. Thanks
to their generous dedication of time and expertise, the conference was held to a very
high standard, building on the experience of previous editions of the ICIEIM/CIO
conferences.
Burgos, Spain Luis R. Izquierdo

Juan José Lavios
José Ignacio Santos
Virginia Ahedo
Contents
Part I Education in Organizational Engineering

1 PROTOCOL—Assessing the Transversal Competence
of Teamwork in Bachelor’s and Master’s Degree by means
of the Competency-based Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Amable Juarez-Tarraga, Cristina Santandreu-Mascarell,
Pilar I. Vidal-Carreras, Julio J. Garcia-Sabater,
Juan A. Marin-Garcia, and M. Vicenta Fuster-Estruch
2 A Managerial Approach to Industry 4.0 Training . . . . . . . . . . . . . . . . 11
J. I. Igartua, J. Retegi, and J. A. Eguren
3 NetLogo Teaching Tool to Illustrate the Cooling Process
in Simulated Annealing Using the Metropolis Model . . . . . . . . . . . . . . 19
José Ignacio Santos, María Pereda, Virginia Ahedo,
and José Manuel Galán
4 Fostering Youth Entrepreneurship in STEM Students
for Industry 4.0 Era . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
L. Isasi-Sánchez, G. Castilla-Alcalá, F. A. Rivera-Riquelme,
and A. Durán-Heras
Part II Management Information Systems and Knowledge Management

5 Distributed Ledger Technology in Industry
4.0: An Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
S. Fernández-Vázquez, R. Rosillo, P. Priore, and J. Puente
6 A Bibliometric Analysis of the Time-Driven Activity-Based
Costing System. The Power of Cost Accounting
in Organizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Patxi Ruiz-de-Arbulo-López, Jesús Rodríguez-Martín,
Jordi Fortuny-Santos, and Beñat Landeta-Manzano
vii
viii Contents
7 Using Data Mining to Analyze Occupational Accidents

in the Construction and Manufacturing Sector . . . . . . . . . . . . . . . . . . . 57
Clodoaldo Polo Barrera, María Martínez Rojas,
and Juan Carlos Rubio Romero
8 Concept for Deployment Design of Machine Learning Models
in Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Henrik Heymann and Andrés Boza
Part III Operations Research, Modelling and Simulation

9 An MILP Model for the Lot-Sizing/Scheduling
of Automotive Plastic Components with Raw Materials
and Packaging Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
E. Guzmán, B. Andres, and R. Poler
10 Design of a Simulation Environment for Training or Testing
Algorithms to Solve the Workshop Sequencing Problem . . . . . . . . . . 93
Efraín Pérez-Cubero and Raúl Poler
11 Advanced Methods and Models of Optimization and Data
Visualization for the Management, Monitoring, and Control
of Operations in Companies Working in Collaborative
Manufacturing Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
12 Annualized Hours, Multiskilling, and Overtime on Annual
Staffing Problem: A Two-Stage Stochastic Approach . . . . . . . . . . . . . 109
Andrés Felipe Porto, Amaia Lusa, César Augusto Henao,
and Roberto Porto Solano
13 Conceptual Framework for Optimization Models in Industry
4.0 Context: Application to Production Planning . . . . . . . . . . . . . . . . . 119
Ana Esteso, Andrés Boza, M. M. E. Alemany,
and Pedro Gomez-Gasquet
14 Artificial Intelligence Techniques Applied to the Flowshop
and Jobshop Problems. A Review of Recent Literature . . . . . . . . . . . 129
Pedro Gomez-Gasquet, Alejandro Torres, Ana Esteso,
and Maria Angeles Rodriguez
15 Design and Implementation of an Experimentation Service
of the Production Scheduling Algorithms . . . . . . . . . . . . . . . . . . . . . . . . 137
María Ángeles Rodríguez, Pedro Gomez-Gasquet,
Llanos Cuenca, and M. M. E. Alemany
Contents ix
Part IV Product Design, Industrial Marketing and Consumer Behaviour

16 Machine Learning in Online Advertising Research:
A Systematic Mapping Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
María Cueto González, José Parreño Fernández,
David de la Fuente García, and Alberto Gómez Gómez
Part V Production Planning and Control

17 Redefinition of the Layout and the Impact on the Reduction
of Wastes: A Case Study in a Metalworking Industry . . . . . . . . . . . . . 163
Bruna Fernandes, Daniel Botelho, Francisco Fernandes,
Inês Aquino, João Ferreira, José Pinto, Maria Fevereiro,
Maria Machado, Nuno Rafael, and Rui M. Lima
18 Digital Twin Enabling Intelligent Scheduling in ZDM
Environments: an Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Julio C. Serrano-Ruiz, Josefa Mula, and Raúl Poler
19 Overview of Lean Production Under Uncertainty . . . . . . . . . . . . . . . . 183
Tania Rojas, Josefa Mula, and Raquel Sanchis
20 Defining Production Planning Problems in Additive
Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
J. de Antón, D. Poza, A. López-Paredes, and F. Villafáñez
21 Digital Twin for a Zero-defect Operations Planning in Supply
Chain 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
22 A Maturity Model for Industry 4.0 Manufacturing Execution
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Miguel Á Mateo-Casalí, Francisco Fraile, Andrés Boza,
and Raul Poler
23 Model Experimentation Environment for Production Planning . . . . 225
Andrés Boza, Pedro Gomez-Gasquet, David Pérez-Perales,
and Faustino Alarcón
Part VI Project and Process Management

24 BIM Implementation in Construction Project Management . . . . . . . 235
F. Acebes, R. Testa, J. Alonso, and D. Curto
Part VII Strategy, Innovation, Networks and Entrepreneurship

25 Airspace Operations Research Supported by EU Funds
and Industry 4.0 Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
J. A. Calvo-Fresno, J. Morcillo-Bellido, and B. Rodrigo-Moya
x Contents
26 Dealing with the Externalities of the Sharing Economy: Effect

of Airbnb in Long-term Rental Prices in the City of Madrid . . . . . . . 263
R. Marque, G. Morales-Alonso, Y. M. Núñez, and A. Hidalgo
27 Cognitive Ergonomics Perspective to Boost Human-centered
Innovations in Industry 4.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Juan Antonio Torrecilla-García, María Carmen Pardo-Ferreira,
and Juan Carlos Rubio-Romero
28 Business Model Patterns: A Systematic Literature Review . . . . . . . . 281
D. Ibarra, A. M. Valenciano, and J. I. Igartua
Part VIII Supply Chain Management and Logistics

29 The Potential of Industry 4.0 in Lean Supply
Chain Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
John Reyes, Josefa Mula, and Manuel Díaz-Madroñero
30 Enablers and Barriers to Industry 4.0 Implementation . . . . . . . . . . . . 303
Blanca Guerrero, Josefa Mula, and Guillermina Tormo
31 Blockchain Impact on Supply Chain Performance . . . . . . . . . . . . . . . . 317
Jesús Morcillo-Bellido, Lucía Romero Fernández-Cuartero,
and Jesús Morcillo-García
32 Proposal of a Methodology and Associated Techniques
for the Design and Management of the Global
Supply Chain Operations Strategy According to
a Circular Economy Criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Darwin Aldas-Salazar, Manuel Díaz-Madroñero, and Josefa Mula
33 Industry 4.0 Practices Applied in Pharma Sector Supply Chain . . . . 335
Jesús Morcillo-Bellido and Ramón Merino-Fuentes
34 A Conceptual Framework of a Blockchain Application
in a Manufacturing Supply Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Erick Ponce, Josefa Mula, and David Peidro
35 Mathematical Programming Model for Collaborative
Replenishment Between Competitive Supply Chains
in the Footwear Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Mario J. Seni and David Peidro
36 Moving Toward the Physical Internet: A Model that Moves
Toward Sustainability Against a Necessary Backdrop
of Industrial Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Carlos Alonso de Armiño, Roberto Alcalde Delgado,
Luis Santiago García Pineda, and Manuel Manzanedo
Contents xi
Part IX Sustainability, Eco-efficiency and Quality Management

37 Machine Learning Approaches to Predict the Use of Share
Bicycles According to Weather Conditions . . . . . . . . . . . . . . . . . . . . . . . 373
Alejandro Escudero-Santana, Andrea Beltrante,
Elena Barbadilla-Martín, and María Rodríguez-Palero
38 Green Aspects on Value Stream Mapping . . . . . . . . . . . . . . . . . . . . . . . . 383
Estefania Pilaloa-Morales and Pilar I. Vidal-Carreras
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
Part I
Education in Organizational Engineering
Chapter 1
PROTOCOL—Assessing the Transversal
Competence of Teamwork in Bachelor’s
and Master’s Degree by means
of the Competency-based Interview
Amable Juarez-Tarraga, Cristina Santandreu-Mascarell,

Pilar I. Vidal-Carreras, Julio J. Garcia-Sabater, Juan A. Marin-Garcia,
and M. Vicenta Fuster-Estruch
Abstract Teamwork competence is one of the soft skills. This paper proposes
and develops the protocol of a best teaching practice for the evaluation of the
transversal competence of teamwork, an important soft skill in the future of univer-
sity students, using a methodological tool used in the professional field of human
resources management: the competency-based interview. The proposal of experience
detailed in the protocol, includes different steps: (1) a prior analysis by the students
of the structure and contents of the competency-based interview, based on material
specifically developed for the practice; (2) the conduct of the interview, in which
each student carries out a double role, as interviewer and interviewee; (3) the use
of a rubric that the interviewer uses as support for the evaluation of the teamwork
competence; (4) and a self-evaluation questionnaire to be filled in by the students
involved.
A. Juarez-Tarraga · C. Santandreu-Mascarell · M. V. Fuster-Estruch

Dpto. Organización de Empresas, Universitat Politècnica de València, Valencia, Spain
e-mail: amjua@omp.upv.es
C. Santandreu-Mascarell
e-mail: crisanma@omp.upv.es
M. V. Fuster-Estruch
e-mail: vfuster@doe.upv.es
P. I. Vidal-Carreras (B) · J. J. Garcia-Sabater · J. A. Marin-Garcia
Dpto. Organización de Empresas, Grupo ROGLE, Universitat Politècnica de València, Valencia,
Spain
e-mail: pivicar@omp.upv.es
J. J. Garcia-Sabater
e-mail: jugarsa@omp.upv.es
J. A. Marin-Garcia
e-mail: jamarin@omp.upv.es
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 3

L. R. Izquierdo et al. (eds.), Industry 4.0: The Power of Data,
Lecture Notes in Management and Industrial Engineering,
https://doi.org/10.1007/978-3-031-29382-5_1
4 A. Juarez-Tarraga et al.
Keywords Teamwork · Competence · Competence-based assessment ·

Competency-based interview
1.1 Background
1.1.1 Detected Need and Proposal for Action
There is an inherent difficulty in assessing transversal competences (TCs) [1–4]. This

has been seen in the launching of the TC project at the university where the experience
was carried out. The practical application of these transversal competences is not
simple, presents multiple difficulties, and requires a good planning, which must
cover different stages, from the material delimitation of the competence itself to the
development of concrete tools for its evaluation.
In addition, this difficulty is also transferred to the professional sphere. It is a fact
assumed by experts in pedagogy, by academics and by human resources professionals
who carry out performance assessment processes, personnel selection processes, or
talent management processes in organizations [5, 6].
From the wide range of transversal competences that are promoted and strength-
ened in the academic field, we have focused on the teamwork competence, which is
one of the most demanded in the professional field [7–10].
Evidence-based HR practice is far from being known and practiced by HR practi-
tioners. The development and application of this good practice contribute to the chal-
lenge of its dissemination[11, 12]. We consider it necessary and urgent to reverse this
situation, and this is not possible without the appropriate training of students, who
will occupy positions of responsibility in the company in the future (either working
in the human resources department of an organization, or in other departments with
direct involvement in functions that affect human resources, such as the operations
department, or the management of industrial plants).
In this respect, it should be noted that in the original discipline of the evidence-
based practice movement (medicine), the advantages of this initiative have been
confirmed, and it has also become clear that the best strategy for its dissemination is
the training of future professionals in university classrooms [13, 14]. This is because
it develops habits and a professional culture that favors the deployment of this kind
of practices.
The aim of the experience is to develop a protocol to assess the level of devel-
opment achieved by the students’ transversal competence of teamwork, while they
develop an action-research project [15] related to the assessment of evidence-based
competencies using the competency-based interview as a tool.
1 PROTOCOL—Assessing the Transversal Competence of Teamwork … 5
1.1.2 Context Description
According to Aguilar Botija [16], “teamwork implies creating and developing a

climate of mutual trust among the components that allows working in a responsible
and cooperative way. The most appropriate term to describe this situation is sharing:
knowledge, commitment, and responsibility. It implies the sharing of tasks and roles
and the respect for the rules and regulations established by and for the group”. There-
fore, TC “Teamwork” can be defined as working and leading teams effectively to
achieve common goals, contributing to their personal and professional development
[16–18].
Teamwork is an essential and increasingly demanded competence in industry [19–
21]. In the organizational sphere, teamwork prevails over individual work when the
task or activity to be carried out is so complex that it is difficult for one person to
master all its problems [22]. Different studies indicate that in the field of engineering,
the coordination of groups of people, both from the organization itself and from
suppliers and clients, is part of the daily activity to provide products and services for
which engineers are ultimately responsible [23, 24].
However, for the assessment of whether a competence is acquired or not, and
even at what level, there are different proposals in academia [18, 25, 26]. Although
there is already a wide range of standardized instruments for competence assessment
by means of questionnaires in the professional field, [27–30], this instrumentation
is not directly transferable to the university environment for two reasons [31]: (a)
the substantive models on which the questionnaires are based respond to compe-
tency models developed in the organizational environment and do not necessarily
contemplate the competencies established in university models; and (b) the behav-
ioral evidence used as items involves samples of work behaviors that are not usually
representative of the behaviors developed by university students.
For all these reasons, the development of specific instruments is still required.
This should enable a standardized approximation to be made of the degree to which
university students display the different competences proposed.
1.2 Proposal for the Development of the Good Practice
It is proposed to develop the experience in the course Business Management. This

course is taught in the first year of the Master’s Degree in Industrial Engineering,
has a load of six credits, is taught to 350 students grouped into seven groups of an
average of 50 students and is taught in Spanish, Valencian, and English.
As a distinguishing feature of the subject, it has an important practical load, both
in the classroom and in the laboratory, which allows a much more intense teacher–
student contact, which is reinforced by the interest of the students who see themselves
at the point of jumping toward their professional activity.
Specifically, it has been considered appropriate to develop the experience in the

classroom practice of the said course, Business Management. For that, a classroom
practice has been designed in which students experience how a selection interview
is carried out, in which the transversal competence of teamwork is assessed, based
on the student’s description of situations, tasks, responsibilities ,and results from the
past where behaviors related to the competence to be assessed were manifested.
The stages of the development of the experience would be: (1) a prior analysis by
the students of the structure and contents of the competency-based interview, based
on material specifically developed for the practice; (2) the conduct of the interview, in
which each student carries out a double role, as interviewer and interviewee; (3) the
use of a rubric that the interviewer uses as support for the evaluation of the teamwork
competence; (4) and a self-evaluation questionnaire to be filled in by the students
involved.
The competency-based interview is a technique that aims to find out about a
candidate’s behavior and competencies based on situations experienced in the past.
This interview allows to assess the competences of a given person. This interview
consists of a formal conversation, which follows the methodology of a semistructured
interview, i.e., the interviewer asks a series of ordered questions that allow him/her to
indirectly assess the level of development of a given competence in the interviewee
and, consequently, to be able to understand and “predict” future behavior in the face
of certain work-related events.
This methodology is used to assess a group of competences or a specific compe-
tence, not by using direct questions, but by inquiring about how the interviewee has
performed in previously experienced situations. In other words, the interviewer must
obtain information about the candidate’s behavior by explaining experiences already
lived. This considers feelings, ideas that arise when facing a situation or carrying
out an activity, emotions, reactions, and, of course, the interviewee’s way of telling
stories. Because the behavior, the reactions, the way the candidate has behaved in past
experiences are strong indicators of how he/she will respond to similar experiences
in the future.
In order to develop the evaluation system, we have used the documentation
produced by Aguilar Botija [16]. The evaluation system includes the design of a
rubric for the students to develop the semistructured interview by competences, so
that the evaluation is obtained in a natural way during the execution of the practice.
1.3 Contribution/Relevance of the Developed Protocol
The relevance of this work is based on the principles that have governed the
development of the protocol:
• Raise awareness of the roles of managers in companies, focusing on talent
detection and retention.
• To highlight the importance of the atmosphere, we must create in a job interview.
• Detecting and valuing TC Teamwork in a colleague or classmate.

• Knowing and applying the competency-based interviewing method.
• Promote peer assessment.
• To introduce the competency-based interview.
• Define the basic dimensions of the competency “teamwork” to be analyzed in a
competency interview.
• Adapt the dimensions of the competency interview to a specific competency.
• To propose a reliable evidence-based method for assessing the competency
teamwork to Bachelor and Master students.
1.4 Action Plan/Protocol
1.4.1 Development
The main functions that would be assumed by the teachers in this best practice will
be:
• Designing and selecting awareness-raising dynamics to be used at the beginning of
the course, with the aim of demonstrating to students the advantages of teamwork
and allowing them to reflect on what they have learnt.
• Develop learning activities (case studies, group dynamics, collaborative learning,
Aronson’s Puzzle and essay writing) to develop the teamwork competence, the
system and the moment of evaluation.
• Follow up both the process and the result, providing feedback as quickly as
possible.
• Apply questionnaires and rubrics for self-evaluation, coevaluation, and heteroe-
valuation of the learning activities carried out by the learners.
• Organize the oral presentation of the solutions to the different exercises and work
carried out in order to achieve the objectives set.
• Tutoring on an individual or group basis, both at the request of the students and
at the request of the teaching staff, some of which is optional and others optional.
1.4.2 Analysis of Information and Evaluation of the Results

of the Experience
We consider that the main benefits for the participants in this project (teachers and
students) are:
• The development of a culture of evidence-based decision-making.
• The initiation in a method widely used in the professional field, useful for
empowering students.
• Identify and highlight individual strengths and weaknesses (soft skills).
• Clarify your strategy for greater personal effectiveness.

• Improve self-knowledge and deepen mutual understanding.
• Improve interactions and generate more fluid interpersonal relationships.
• Increase cohesion and productivity of individuals and teams, integrating skills.
1.5 Limitations, Future Research
It is important to note at this point that the developed protocol has been implemented
for the first time in the academic year 2020–2021, and now, the analysis is doing.
The main limitations are focused on:
• It is the first approach to the use of competency-based interviewing to assess it,
and the experience was only piloted in 4 of the 7 subject groups.
• No evidence was collected of how the students viewed this pilot experience,
although verbal feedback is available.
These limitations will become improvements to be made in future courses:
• To complete the rubric with a basic document for teachers, which would serve
as a script for the implementation of the protocol, which would be uniform and
measurable.
• Complete the peer evaluation with an evaluation of the teacher who works with
them in a team, so that a subsequent comparison of the results obtained can be
made.
However, we consider the protocol of this teaching experience that should be
disseminated in conferences and teaching journals, as its design can help other
teachers to complete the evaluation system for their students’ teamwork competence.
Acknowledgements The work described in this paper has been supported by the project “La
gestión de competencias basada en evidencias: aprendiendo a trabajar en equipo por medio de ABP
compartido entre asignaturas de grado y Máster (PIME/19–20/173)” of the Universitat Politècnica
de València, Spain.
References
1. Anthony S, Garner B (2016) Teaching soft skills to business students: an analysis of multiple
pedagogical methods. Bus Prof Commun Q [Internet]. 79(3):360–370. https://doi.org/10.1177/
2329490616642247
2. Ingols C, Shapiro M (2013) Concrete steps for assessing the “soft skills” in an MBA program.
J Manag Educ [Internet]. 38(3):412–435. https://doi.org/10.1177/1052562913489029
3. Lajara Camilleri N, Rovira Cardete A, Bañón Gomis AJ, Fernández Durán L, Cortés Meseguer
L, Fernández Zamudio MÁ et al (2015) Experiencias en el desarrollo y evaluación de la CT6:
Trabajo en equipo y liderazgo en la UPV. In: Libro de Actas IN-RED 2015—Congreso Nacional
de Innovación Educativa y de Docencia en Red [Internet]. Editorial Universitat Politècnica de
València [cited 2021 Feb 20], pp 1012–27. Available from: http://ocs.editorial.upv.es/index.

php/INRED/INRED2015/paper/view/1631
4. Marin-Garcia JA, Juarez-Tarraga A, Santandreu-Mascarell C (2018) Kaizen philosophy: the
keys of the permanent suggestion systems analyzed from the workers’ perspective. TQM J
5. Jonsson A, Svingby G (2007) The use of scoring rubrics: reliability, validity and educational
consequences. Educational Res Rev. Elsevier 2:130–44
6. Sancho-Thomas P, Fuentes-Fernández R, Fernández-Manjón B (2009) Learning teamwork
skills in university programming courses. Comput Educ 53(2):517–531
7. de Sousa Jabbour ABL, Jabbour CJC, Foropon C, Filho MG (2018) When titans meet—can
industry 4.0 revolutionise the environmentally-sustainable manufacturing wave? The role of
critical success factors. Technol Forecast Soc Change 132:18–25.
8. Azmi AN, Kamin Y, Noordin MK, Nasir ANM (2018) Towards industrial revolution 4.0:
employers’ expectations on fresh engineering graduates. Int J Eng Technol [Internet].
7(4.28):267–72. Available from: https://www.researchgate.net/profile/Aini_Najwa_Azmi2/
publication/329356058_Towards_Industrial_Revolution_40_Employers’_Expectations_on_
Fresh_Engineering_Graduates/links/5c038ba292851c63cab3f924/Towards-Industrial-Revolu
tion-40-Employers-Expectations-on-F
9. Baena F, Guarin A, Mora J, Sauza J, Retat S (2017) Learning factory: the path to industry 4.0.
Procedia Manuf 9:73–80
10. Shamim S, Cang S, Yu H, Li Y (2016) Management approaches for Industry 4.0: a human
resource management perspective. In: 2016 IEEE congress on evolutionary computation (CEC)
[Internet], pp 5309–16. Available from: https://ieeexplore.ieee.org/abstract/document/7748365
11. Gubbins C, Harney B, van der Werff L, Rousseau DM (2018) Enhancing the trustworthiness
and credibility of human resource development: evidence-based management to the rescue?
Human Resour Develop Q
12. Marler JH, Boudreau JW (2017) An evidence-based review of HR analytics. Int J Hum Resour
Manag
13. Luckmann R (2001) Evidence-based medicine: how to practice and teach EBM, 2nd Edition:
By Sackett DL, Straus SE, Scott Richardson W, Rosenberg W, Haynes RB, Livingstone C
2000. J Intensive Care Med [Internet], 16(3):155–6. Available from: https://doi.org/10.1177/
088506660101600307
14. Tranfield D, Denyer D, Smart P (2003) Towards a methodology for developing evidence-
informed management knowledge by means of systematic review. Br J Manag [Internet].
14(3):207–222. https://doi.org/10.1111/1467-8551.00375
15. Marin-Garcia, J.A.; Alfalla-Luque R. Teaching experiences based on action research: a guide to
publishing in scientific journals. WPOM - Work Pap Oper Manag [Internet]. 2021;12(No 1):42–
50. Available from: https://polipapers.upv.es/index.php/WPOM/article/view/7243/13813
16. Aguilar Botija A (2016) Competencia transversal trabajo en equipo y liderazgo
[Internet]. https://www.upv.es/contenidos/COMPTRAN/info/954872normalc.html. Available
from: ttps://www.upv.es/contenidos/COMPTRAN/info/954872normalc.html
17. Ballenato G (2005) Trabajo en equipo. Dinámica y participación en los grupos. Editor Piramide
18. Marin-Garcia JA, Villaescusa MM, Bonavia T (2019) Protocol: how to measure teamwork and
networking competencies. WPOM-Working Pap Oper Manag [Internet]. 10(2):55. Available
from: https://polipapers.upv.es/index.php/WPOM/article/view/12369
19. Korhonen-Yrjänheikki K, Tukiainen T, Takala M (2007) New challenging approaches to
engineering education: enhancing university–industry co-operation. Eur J Eng Educ
20. Davis S, Gervin D, White G, Williams A, Taylor A, McGriff E (2013) Bridging the gap between
research, evaluation, and evidence-based practice. J Soc Work Educ
21. Innobasque—Agencia Vasca de la Innovación. Tecnologías y competencias profesionales 4.0—
Análisis de la demanda empresarial [Internet]. [cited 2021 Feb 20]. Available from: https://
www.innobasque.eus/microsite/politicas_de_innovacion/publicaciones/publicacion-511/
22. Salas E, Sims DE, Shawn Burke C (2005) Is there A “big five” in teamwork? Small Group
Research
23. Stevens R, Johri A, O’connor K (2015) Professional engineering work. In: Cambridge handbook
of engineering education research
24. Trevelyan J, Tilli S (2007) Published research on engineering work. J Prof Issues Eng Educ
Pract
25. Solanes Puchol Á, Núñez Núñez R, Rodríguez Marín J (2008) Elaboración de un cuestionario
para la evalúación de competencias genéricas en estudiantes universitarios. Apunt Psicol
26. Bonavia T, Molina JG, Puchol A (2015) Structural validity of a questionnaire to measure
effective behaviors in work teams. An Psicol/Ann Psychol
27. Instituto de ingenieria del conocimiento (2021) eValue. Sistema online de evaluación de
competencias [Internet]. [cited 2021 Feb 14]. Available from: http://web.teaediciones.com/
CompeTEA.aspx
28. Arribas, D. (TEA ediciones), Pereña J (TEA ediciones) (2009) CompeTEA. Evaluación de
Competencias [Internet]. [cited 2021 Feb 14]. Available from: http://web.teaediciones.com/
CompeTEA.aspx
29. Blömeke S, Zlatkin-Troitschanskaia O, Kuhn C, Fege J (2013) Modeling and measuring
competencies in higher education: tasks and challenges. https://doi.org/10.1007/978-94-6091-
867-4
30. eGovlab_Stockholm University-Sweden, Fundació_Eurecat-Spain, DMC_Metrix-Ireland,
Universidad_Alcalá-Spain, Everis-Spain, Fondazione_Politecnico_di_Milano-Italy, et al.
skillsmatch.eu [Internet]. 2018 [cited 2021 Feb 21]. Available from: https://skillsmatch.eu/
31. Aguado D, González A, Antúnez M, de Dios T (2017) Evaluación de Competencias Transver-
sales en Universitarios. Propiedades Psicométricas Iniciales del Cuestionario de Competencias
Transversales. REICE Rev. Iberoam. sobre Calidad, Efic. y Cambio en Educ.
Chapter 2
A Managerial Approach to Industry 4.0
Training
J. I. Igartua , J. Retegi , and J. A. Eguren
Abstract Training Industry 4.0 is one of the challenges companies face in the context
of the fourth industrial revolution. However, existing training focuses mainly on
technology and does not pay much attention to the management implications of
implementing Industry 4.0 in organizations. The training approach presented in this
article and developed through a university-industry cooperation scheme is based
on a management perspective supported by an Industry 4.0 Advanced Management
Framework, a training case based on the IMPULS Industry 4.0 Maturity Model, and
a challenge-oriented technology training. The feedback received from the learners
reflects that the management approach developed is valuable for the training of senior
and middle management, underlining the importance of a goal-oriented strategic
approach when implementing Industry 4.0 technologies. The need for more person-
alized training (more focused on the business problems of the participants) and the
need to cope with “remote teaching” are the two training challenges to be addressed
in the future.
Keywords Industry 4.0 · Training · Advanced management · Maturity model
2.1 Introduction
The Industry 4.0 phenomenon is at the center of the agenda of companies and govern-
ments in Europe and around the world [1], where different agents are seeking to
position their businesses and industries within this new paradigm.
This digital transformation of organizations affects all people, departments, and
functions within companies [2]. All activities along the value chain will be impacted
[3], leading to changes in the way internal activities are carried out within the orga-
nization, as well as the functioning of the value chain beyond the activities of the
companies that implement these technologies.
J. I. Igartua (B) · J. Retegi · J. A. Eguren

Mondragon University, Loramendi 4, 20500 Mondragón, Spain
e-mail: jigartua@mondragon.edu

https://doi.org/10.1007/978-3-031-29382-5_2
12 J. I. Igartua et al.
The ability of company managers to assimilate these opportunities and changes

requires them to understand that the management of these organizations represents
a new challenge for people and the organization as a whole [4]. Companies need to
think strategically about their core resources, leadership, and knowledge portfolio to
take advantage of the Industry 4.0 paradigm [5].
In these circumstances, training programs aimed at coaching and raising the
awareness of top and middle managers play a key role in their upskilling and the
development of an advanced management approach toward the implementation of
Industry 4.0. Most of the existing training, as will be discussed in the next section,
focuses on the technological challenges of Industry 4.0 rather than on the manage-
ment challenges behind this industrial paradigm, which motivates the authors to work
on this training stream.
This article proposes a training approach in Industry 4.0 oriented to top and middle
managers. In particular, the article explains the structure and elements of the training
scheme, as well as the results of the training experiences developed, and the resulting
conclusions and challenges.
2.2 Industry 4.0 Training
As various studies have shown [6], the qualification of people and their competence
development will be key in the near future in order to respond to the challenges posed
by the Industry 4.0 paradigm.
Some of the existing training approaches have focused on the development of
immersive training actions of a technological nature through the use of technologies
such as virtual reality [7]. Other experiences, on the contrary, have been based on
the development of skills (technical, transformational, and social), through schemes
such as “learning factories” [6] or “teaching factories” [8].
Other approaches have focused on the training implications for specific functions
within the organization. For example, some training has focused on training needs
and capacity building actions in relation to operations management [9]. Finally, other
training approaches focus on technological skills, technical skills, and personal skills
[10].
In terms of training processes, some authors [11] detail their formative approach by
listing its elements (i.e., introduction, information gathering, learning, and training,
practice orientation, and testing and evaluation), as well as the timeframe assigned
to each part.
However, all these training approaches, while seeking the support of top manage-
ment (stressing the need to take into account the management impact of Industry
4.0 technologies and the need to manage different key aspects when implementing
Industry 4.0 in companies), do not directly address the management implications of
Industry 4.0 implementation.
Thus, in response to this identified shortcoming, in the following sections, we will
set out the training outline and experience developed in this ongoing training project.
2 A Managerial Approach to Industry 4.0 Training 13
2.3 Proposed Approach for Industry 4.0 Training
The approach to Industry 4.0 training described in this paper is based on the principle
of University-Business Cooperation [12].
In this context, the Association of Industrial Engineers of Bizkaia (Colegio de
Ingenieros Industriales de Bizkaia), EUSKALIT (Basque Foundation for Advanced
Management), Mondragon Unibertsitatea and Sisteplant have launched a training
program in Industry 4.0 aimed at executives and middle managers, which responds
to the shortfall detected in relation to the management implications of Industry 4.0
implementation.
As indicated by some authors [13], the support and leadership of managers is
key to the long-term development and implementation of Industry 4.0. Their beliefs,
approaches, and actions on Industry 4.0 shape the organization’s vision and strategy,
which is key to guiding the approach to Industry 4.0 implementation. Thus, the objec-
tive of this training action was focused on training top and middle management in the
implementation of Industry 4.0, through a practical scheme, based on an advanced
management approach, which favors the maximum use of technological advances,
integrating people, processes, and technology. The training scheme developed for
this objective focuses on five blocks (Fig. 2.1).
Thus, the “The Industry 4.0 Advanced Management Framework” block highlights
the importance of addressing the management implications of Industry 4.0 imple-
mentation. This reference framework [14], developed by EUSKALIT in collaboration
with Mondragon Unibertsitatea, focuses on four elements:
1. Industry 4.0 as Management Support: Industry 4.0 as a support for achieving
more advanced management in organizations.
2. Advanced Management for Industry 4.0: Development of advanced management
for the successful implementation of Industry 4.0.
Fig. 2.1 Industry 4.0 training approach

3. Innovation through Industry 4.0: Development of innovations through Industry

4.0.
4. Project Management for Industry 4.0: Guidelines for good Industry 4.0 project
management.
The second block, “Industry 4.0 Maturity Model”, emphasizes the need for compa-
nies and their managers to assess the starting situation of the company in relation
to Industry 4.0. The most commonly used models for this purpose are the maturity
models [15], which mainly focus on six dimensions (technology, people, strategy,
leadership, process, and innovation).
The third of the blocks, “Industry 4.0 Roadmap”, highlights the importance of
companies defining their own path toward Industry 4.0. The idea is that each organiza-
tion, depending on its competencies, motivations, capabilities, intentions, objectives,
priorities, and budgets, defines the roadmap to follow [16].
The fourth block, “Beacons”, underlines the importance of benchmarking existing
good practices at the global level [17], so that companies can model and incorporate
existing “best practices”.
Finally, the fifth block refers to “Industry 4.0 technologies”, focusing on the
understanding of Industry 4.0 technologies, as well as the potential of their application
and the opportunities they generate in industry [18].
2.4 Training Process and Results
This section presents the training process implemented under the Industry 4.0
Training Approach described in the previous section, as well as the results of this
process. Thus, the training implemented was based on two modules, which act in an
integrated manner (Fig. 2.2). One focused on the management of Industry 4.0, and
the other on the technological field of Industry 4.0.
Fig. 2.2 Industry 4.0 training process

The process proposed for the “management” module is based on three training
blocks (two face-to-face and one self-study). The first of the training blocks (face-
to-face training) responds to the objective of making managers and middle managers
aware of the importance of a strategic approach when implementing Industry 4.0, as
well as the discussion of existing best practices, both at a global level (Beacons) and
at a more regional level (business cases). This same block includes training on the
Industry 4.0 Advanced Management Framework.
The second of the training blocks (self-study) of this same module encourages
managers and middle managers to carry out a self-assessment of their company’s
situation based on the Industry 4.0 Advanced Management Framework, as well as
to study the case study based on which they will work in the third block.
The third block starts with a comparison of the online self-assessments to identify
key common patterns and establish a group discussion. From there, and with the aim
of enabling participants to define an Industry 4.0 plan, they work in teams on the
case study, first carrying out an assessment following the IMPULS maturity model
[19], and then identifying areas of opportunity, on which to develop a road mapping
activity [20]. The objective of this final activity is to encourage managers to establish
an Industry 4.0 implementation plan based on an ad hoc designed didactic case, in
coherence with the improvement areas identified through the maturity model and
aligned with the strategic development areas (opportunities) selected by the work
teams.
In addition, the “technology” module is structured on the basis of a first block in
which attendees carry out a self-assessment in relation to Industry 4.0 technologies
and the degree of implementation in their company, followed by a second block
of classroom training in which Industry 4.0 technologies are described, and their
potential is analyzed. In addition, in this same training module, attendees identify
and group together different objectives and challenges related to their companies and
then work in groups on different technological solutions.
As a result of this training process, managers and middle managers receive the
following training and managerial tools:
• The Industry 4.0 Advanced Management Framework.
• A self-assessment questionnaire based on the Advanced Management Model for
Industry 4.0.
• Industry 4.0 business application cases.
• A practical application of the IMPULS Maturity Model, with statement and
answer. In addition to the IMPULS model itself.
• An Industry 4.0 road mapping tool and its practical application.
• Specific training on Industry 4.0 technologies.
• An Industry 4.0 technology assessment questionnaire.
• A dynamic process for the development of Industry 4.0 technology proposals
based on challenges.
The analysis of the results of the training experiences is based on a question-
naire addressed to the learners (53 people), which evaluated six aspects: program,
training material, teaching staff, teaching methodology, response to their needs, and
usefulness for their professional development. In addition, the analysis also collected
information on priority training areas.
After three editions and two improvement cycles, the evaluations of the attendees
have a very positive impact on aspects such as:
• Program.
• Teaching staff (mastery of the subject and delivery skills).
• Usefulness for professional development.
• Training material.
Finally, it should be noted that the areas for improvement proposed focus on the
participants’ needs to solve their practical business cases (response to their needs).
This result could be explained by the analysis of the collected information on priority
training areas. Specifically, most of the learners (55%) focused on topics related to
leadership 4.0 and people management, an aspect that the middle and senior manage-
ment learners considered an important challenge in the development of Industry
4.0 projects. Other areas of training interest identified by the learners were: digital
manufacturing [21] (50%), machine learning systems [22] (35%), and transformation
process to a smart industry [23] (30%).
2.5 Conclusions
In this article, we have presented a training approach developed to train top and middle
managers in Industry 4.0. The training approach developed is based on a management
perspective supported by the Industry 4.0 Advanced Management Framework, as
well as a training case based on the IMPULS Industry 4.0 Maturity Model, and a
challenge-oriented technology training.
The feedback received from those trained reflect that the managerial approach
developed is valuable when training top and middle-level managers, underlining the
importance of a strategic goal-oriented approach when implementing Industry 4.0
technologies.
The conclusions obtained through three editions and two improvement cycles
reaffirm the need for this training approach and the training method used. It is neces-
sary to focus on the management aspects related to the implementation of Industry
4.0 technologies, or any other enabling technologies. Obtaining sustainable compet-
itive advantages from Industry 4.0 will only be possible if companies, and more
specifically their managers and middle management, approach the implementation
of this paradigm from a management perspective. This training scheme focuses on
this aspect.
This training experience is not without limitations that could be explored in future
developments. It would be of interest for future research to deepen on the study of the
training of senior and middle management in personal capabilities for digital transfor-
mation. This will help create the foundations for the development of organizational
capabilities for digital transformation [24].
In terms of lessons learned, there are two training challenges that need to be
addressed. On the one hand, the need expressed by the participants for more person-
alized training (more focused on their company’s problems), and on the other hand,
the need to deal with remote training. The first of these aspects will require the
development of customized training models seeking a balance between training and
consultancy activities; and the second challenge requires structuring training on a
modular basis, while making use of the tools and means of “remote training”.
Finally, it is worth highlighting the benefits of the University-Business Cooper-
ation scheme and the complementarity of the entities participating in the initiative.
The positive assessment of the experience by all the organizations involved, and
their commitment to the transformation of their territory and companies reinforces
the existing cooperation and the opportunities to develop new collaborations, which
will undoubtedly help to respond to the business and social challenges of the future.
Acknowledgements We would like to thank our project partners, the Association of Industrial
Engineers of Bizkaia (Colegio de Ingenieros Industriales de Bizkaia), EUSKALIT and Sisteplant
for their drive and commitment to this training approach to promote the management of Industry
4.0.
Compliance with Ethical Standards

In this article, the results of the learners’ experience were obtained by means of a questionnaire.
No personal data was collected in the questionnaire, and therefore, the answers are not traceable
(anonymous answers). The participating learners were informed of the process and the data manage-
ment policy, and voluntarily agreed to fill in the anonymous questionnaire. The authors declare that
they have no conflict of interest, nor do they work for, consult, own shares in, or receive funding
from any company or organization that may benefit from this article and have disclosed no relevant
relationships beyond their academic appointment. The Research Ethics Committee of Mondragon
Unibertsitatea (Ref. IEB-20221107) approved the entire procedure used in the research process.
References
1. Tay SI, Lee TC, Hamid NZA, Ahmad ANA (2018) An overview of industry 4.0: definition,
components, and government initiatives. J Adv Res Dyn Control Syst 10(14):1379–1387
2. Fettig K, Gacic T, Koskal A, Kuhn A, Stuber F (2018) Impact of Industry 4.0 on organizational
Structures. In: 2018 IEEE international conference on engineering, technology and innovation,
ICE/ITMC 2018–proceedings. ISBN 9781538614693
3. Fatorachian H, Kazemi H (2021) Impact of Industry 4.0 on supply chain performance. Prod
Plan Control. 32(1):63–81. https://doi.org/10.1080/09537287.2020.1712487
4. Sony M, Naik SS (2019) Ten lessons for managers while implementing industry 4.0. IEEE Eng
Manage Rev 47(2):45–52. https://doi.org/10.1109/EMR.2019.2913930
5. Agrawal A, Schaefer S, Funke T (2018) Incorporating industry 4.0 in corporate strategy.
ISBN 9781522534693
6. Schallock B, Rybski C, Jochem R, Kohl H (2018) Learning factory for industry 4.0 to provide
future skills beyond technical training. In: Procedia manufacturing, pp 27–32
7. Roldán JJ, Crespo E, Martín-Barrio A, Peña-Tapia E, Barrientos A (2019) A training system
for Industry 4.0 operators in complex assemblies based on virtual reality and process mining.
Robot Comput Integr Manuf 59:305–316. https://doi.org/10.1016/j.rcim.2019.05.004
8. Mourtzis D (2018) Development of skills and competences in manufacturing towards education

4.0: a teaching factory approach. ISBN 9783319666969
9. Koleva N, Andreev O (2018) Aspects of training in the field of operations management
with respect to industry 4.0. In: International conference on high technology for sustainable
development, HiTech 2018—proceedings. ISBN 9781538670392
10. Horrillo-Tello J, Triadó-Aymerich J (2018) Training gaps in engineering degrees for industry
4.0 in Spain. a proposal for actions | Carencias formativas de los grados de ingeniería para la
industria 4.0 en españa. Una propuesta de actuaciones. Dyna (Spain) 93(4):365–369. https://
doi.org/10.6036/8604
11. Yang S, Hamann K, Haefner B, Wu C, Lanza G (2018) A method for improving production
management training by integrating an industry 4.0 innovation center in China. In: Procedia
manufacturing, 213–218
12. Hewitt-Dundas N (2013) The role of proximity in university-business cooperation for
innovation. J Technol Transfer 38(2):93–115. https://doi.org/10.1007/s10961-011-9229-4
13. Prause M (2019) Challenges of industry 4.0 technology adoption for SMEs: the case of Japan.
Sustainability (Switzerland). 11(20). https://doi.org/10.3390/su11205807
14. EUSKALIT (2020) Marco para la orientación hacia la digitalización y la industria 4.0 desde
la perspectiva de la gestión avanzada
15. Hizam-Hanafiah M, Soomro MA, Abdullah NL (2020) Industry 4.0 readiness models: a system-
atic literature review of model dimensions. Information (Switzerland) 11(7):1–13. https://doi.
org/10.3390/info11070364
16. Ghobakhloo M (2018) The future of manufacturing industry: a strategic roadmap toward
Industry 4.0. J Manuf Technol Manage 2(6). https://doi.org/10.1108/JMTM-02-2018-0057
17. World Economic Forum (2019) Fourth industrial revolution beacons of technology and
innovation in manufacturing, January
18. Frank AG, Dalenogare LS, Ayala NF (2019) Industry 4.0 technologies: implementation patterns
in manufacturing companies. Int J Prod Econ 210:15–26. https://doi.org/10.1016/j.ijpe.2019.
01.004
19. Hamidi SR, Aziz AA, Shuhidan SM, Aziz AA, Mokhsin M (2018) SMEs maturity model
assessment of IR4.0 digital transformation. ISBN 9789811086113
20. Phaal R (2004) Technology roadmapping—a planning framework for evolution and revolution.
Technol Forecast Soc Chang. https://doi.org/10.1016/S0040-1625(03)00072-6
21. Cohen Y, Faccio M, Pilati F, Yao X (2019) Design and management of digital manufacturing
and assembly systems in the Industry 4.0 era. Int J Adv Manuf Technol 105(9):3565–3577.
https://doi.org/10.1007/s00170-019-04595-0
22. Diez-Olivan A, Del Ser J, Galar D, Sierra B (2019) Data fusion and machine learning for indus-
trial prognosis: trends and perspectives towards Industry 4.0. Inf Fusion 50:92–111. https://
doi.org/10.1016/j.inffus.2018.10.005
23. Ghobakhloo M (2018) The future of manufacturing industry: a strategic roadmap toward
Industry 4.0. J Manuf Technol Manag 29(6):910–936. https://doi.org/10.1108/JMTM-02-2018-
0057
24. González-Varona JM, López-Paredes A, Poza D, Acebes F (2021) Building and development
of an organizational competence for digital transformation in SMEs. J Indus Eng Manage.
https://doi.org/10.3926/jiem.3279
Chapter 3
NetLogo Teaching Tool to Illustrate
the Cooling Process in Simulated
Annealing Using the Metropolis Model
José Ignacio Santos , María Pereda , Virginia Ahedo ,

and José Manuel Galán
Abstract Simulated annealing is one of the most popular metaheuristic optimization

techniques used in engineering and management to solve combinatorial problems.
The algorithm is inspired by the thermodynamic process that occurs in the annealing
treatment in metallurgy. Although it is simple to implement, its general operating
mechanism and the rationale behind the search strategy are not always that intuitive.
In this work, we present a teaching tool implemented in NetLogo that illustrates the
metaphor of both processes and the effect of annealing cooling schedules on the
quality of the solutions obtained.
Keywords Simulated annealing · Optimization · Teaching resource · NetLogo ·

Cooling schedules
J. I. Santos (B) · V. Ahedo · J. M. Galán

Departamento de Ingeniería de Organización, Escuela Politécnica Superior, Universidad de
Burgos, Ed. A1, Avda. Cantabria S/N 09006, Burgos, Spain
e-mail: jisantos@ubu.es
V. Ahedo
e-mail: vahedo@ubu.es
J. M. Galán
e-mail: jmgalan@ubu.es
M. Pereda
Grupo de Investigación Ingeniería de Organización Y Logística (IOL), Departamento de
Ingeniería de Organización, Universidad Politécnica de Madrid, Administración de Empresas, y
Estadística. Escuela Técnica Superior de Ingenieros Industriales, C/ José Gutiérrez Abascal, 2,
28006 Madrid, Spain
e-mail: maria.pereda@upm.es
Unidad Mixta Interdisciplinar de Comportamiento y Complejidad Social (UMICCS), 28911
Leganés, Madrid, Spain
Grupo Interdisciplinar de Sistemas Complejos, Madrid, Spain

https://doi.org/10.1007/978-3-031-29382-5_3
20 J. I. Santos et al.
3.1 Introduction
Optimization is essential in any field of engineering and management. During the first
years of university education, it is common to teach the fundamentals of optimization
of continuous and derivable functions, continuous optimization with constraints, and
linear programming. Later, within the framework of integer programming, combi-
natorial optimization problems—characterized by discrete decision variables and a
finite search space—are usually introduced.
Remarkably, these problems are of paramount importance in management engi-
neering due to the large number of real problems they encompass (resource allocation,
ordering, portfolio selection, etc.) [1, 2]. In many of these cases, when the problem
reaches a certain size, enumerative search methods are often not suitable for finding
an optimum, as the size of the solution space is too large.
In such cases, it is common to resort to a family of approximate optimization
techniques (approximate algorithms) known globally as metaheuristic techniques.
These algorithms provide acceptable—although not necessarily optimal—solutions
in a reasonable computational time, thus, satisfactorily solving—in practical terms—
a multitude of problems in science, management, and engineering.
There exists a great variety of metaheuristics, which differ from each other (i) in
the way they combine strategies for exploring the solution space and (ii) in how they
exploit the information obtained to intensify their search in promising areas. Meta-
heuristics can be classified according to different taxonomies [3, 4], the most common
classification being into population-based and trajectory-based techniques. Gener-
alist optimization and operations research manuals typically include a chapter on the
two most popular and representative techniques of each of these two approaches:
genetic algorithms (population-based) and simulated annealing(trajectory-based).
In this paper, we present a teaching resource designed to help understand the
analogy between simulated annealing and the thermodynamic process on which it is
based: the annealing treatment performed in materials science.
Annealing is a heat treatment used to soften a metallic material so as to restore its
crystalline structure and eliminate internal stresses that may have arisen as a result of
a previous treatment and/or process. More specifically, annealing consists of heating
the material to a high temperature (above the recrystallization temperature), keeping
the metal at that temperature for some time, and then allowing the process to cool
down slowly. During the annealing process, atoms migrate through the crystalline
lattice reducing the number of dislocations in the material and hence increasing its
ductility and workability.
Inspired by this process, in 1983 Kirkpatrick, Gelatt and Vecchi [5] published
an article in the journal Science, in which they stated that “There is a deep and
useful connection between statistical mechanics (the behavior of systems with many
degrees of freedom in thermal equilibrium at a finite temperature) and multivariate
or combinatorial optimization (finding the minimum of a given function depending
on many parameters).”
3 NetLogo Teaching Tool to Illustrate the Cooling Process in Simulated … 21
In their proposed algorithm, trajectory search was used to look for solutions;
more precisely, solutions were selected after exploration of their neighboring envi-
ronment—i.e., of the solutions in their nearby solution space. Recall that such a
process, if naively implemented, is similar to a local search process, thus having a
strong tendency to get stuck at local minima in the search space. For this reason,
the simulated annealing algorithm incorporates a control mechanism specifically
conceived to allow escaping from such local optima: allowing to move to worse
solutions along the search path.
The basis of this control is governed by the Metropolis model [6], which describes
the cooling process by simulating the energy changes of a particle system as a
function of temperature. Specifically, a simplification of the Boltzmann probability
distribution known as the Metropolis acceptance criterion is used:
p[δ E] = e( )
−δ E
kT (3.1)
Equation (3.1) simulates a thermodynamic system from a sequence of states at a

given temperature. Each new sequence is obtained by randomly changing the energy
level of a single atom. If the movement of the atom is toward lower energy solutions,
the energy change is always accepted. On the contrary, if the new configuration
involves an increase in energy, it will be accepted with a probability that is given
by the negative exponential in Eq. (3.1). Recall that in (3.1), the change of energy
configuration depends on T (multiplied by a constant k), and on the energy jump
δ E that needs to be undertaken. Specifically, the higher the temperature, the greater
the probability of accepting the change, with smaller jumps being more likely than
larger ones to higher energy positions. The process is repeated indefinitely.
In the analogy between this thermodynamic model and the optimization process,
each configuration corresponds to a possible solution of the problem to be optimized,
the change of the atom to another energy level is an analogy to a movement to a
neighboring solution, and the energy of the system is an overall measure of the
quality of the solution obtained, of which a global minimum is ideally reached.
Under this framework, the fundamental state of the system—the minimum energy
equilibrium—is considered the global optimum solution of the optimization problem.
In contrast, a local optimum would be equivalent to a metastable state. A rapid cooling
process (temper) in the thermodynamic system can be understood as a local search
process in the optimization approach, in which the chances of reaching a metastable
state are very high. In both cases of the analogy, the role of temperature is very similar.
While in the thermodynamic process, the temperature corresponds to the physical
magnitude, in the system to be optimized, it is a variable, also called temperature,
which acts as a control parameter to balance the exploration-intensification process
of the search, exploring in the initial phases and intensifying afterward. Finally, to
complete the metaphor, what we identify as careful annealing in the thermodynamic
system corresponds in the optimization process to a simulated annealing correctly
parameterized (see [4] for more details).
The simulated annealing algorithm implements the exact Metropolis acceptance

criterion as its control mechanism for the search process. The general idea is to “heat
up” the process, starting at a high temperature—which relaxes the acceptance crite-
rion and thus intensifies the exploration process—and then lowering the temperature
in the search process, hence tightening the acceptance criterion toward worse solu-
tions and orienting the algorithm toward solution-improving movements. Ideally,
when the system has cooled down, the solution (atomic configuration in the analogy)
should be close to the global optimum (minimum energy state).
Although the implementation of this algorithm is straightforward, understanding
its general mechanism of operation and why it works is not always intuitive. Hence,
in this paper, we present a teaching tool specifically designed to facilitate (i) the
understanding of the analogy with the thermodynamic process, (ii) the effect of the
cooling rate, and (iii) to better capture the intuition of its general operation mechanism
and the influence of the different parameters. The Metropolis algorithm has been
developed in NetLogo [7]. Our work differs from the annealing implementation of the
NetLogo library [8] in its teaching and outreach vocation (we illustrate the Metropolis
model explicitly and visually), as well as in the variety of cooling mechanisms
implemented and in the possibility of studying the effect of their parametrization to
reach global minima.
3.2 Annealing Schedule Configuration
In the simulated annealing algorithm, different aspects that determine the complete
temperature decrease process need to be specified. In particular, it is necessary to
define: (i) an initial temperature, (ii) a final temperature (which establishes the stop
criterium of the algorithm), (iii) the time that the algorithm remains at a fixed tempera-
ture (the level-length L), and (iv) the sequence of temperatures from the initial temper-
ature until the algorithm has finished. Taken together, this whole process is called the
annealing schedule. The correct determination of this schedule has a fundamental
impact on the performance of the algorithm: an excessively slow cooling scheme can
significantly increase the computation time required to obtain a solution, while very
fast processes can lead to bad solutions.
There are two temperature cooling strategies: static and adaptive. Static programs
are characterized by an a-priori-determined cooling rate that is independent of the
search process, while adaptive programs adjust the descent rate according to the
search process itself.
The classical cooling mechanisms in simulated annealing are:
• Linear decrease, where the temperature decreases in each iteration according to
a constant c:
Tk+1 = Tk − c (3.2)
• Exponential or geometric decay, where the temperature is decreased in successive

iterations by multiplying by a positive number α smaller than one. Typically, α is
in the range (0.8,0.99):
Tk+1 = α · Tk (3.3)
• Logarithmic cooling schedule or Boltzmann annealing [9], an extraordinarily slow

descent mechanism that decreases temperature according to (3.4) and ensures
convergence to the global optimum given a sufficiently high T0 :
T0
Tk+1 = (3.4)
ln(1 + k)
• Cauchy cooling schedule [10] or fast annealing, in which the temperature

decreases hyperbolically in accordance with:
T0
Tk+1 = (3.5)
1+k
3.3 Description of the Teaching Tool
The tool can be downloaded from https://www.comses.net/codebases/6088b061-

d836-4f98-9945-0601aafe0570/releases/1.0.0/ in two different versions (A and B),
which differ in the type of dynamic visualization offered (see Fig. 3.1). In version A,
the atoms in each energetic level are grouped together to facilitate the visualization
of the number of particles per level (without the need to look at the histogram that
includes this information in both cases). As for version B, in it each of the atoms in the
model maintains its position on the abscissa axis to facilitate individual traceability.
The total energy of the system is given by the total sum of the energy of the
individual atoms (the energy level at which they are located), and the objective of
the annealing cooling program is to reach the global energy minimum, at which all
atoms are at the minimum energy level.
Irrespective of the individual atom energy display panel, the application interface
is divided into five sections: NetLogo’s general control panel (1), the run configura-
tion panel (2), the simulation control panel (3), the individual atom energy display
panel (4), and the general information panel (5) (see Fig. 3.2).
3.3.1 The Main Control Panel in NetLogo
This section consists of three tabs: run—where the necessary interface for the execu-
tion of the simulation is found; information—where general information about the
Fig. 3.1 Depending on the version selected, the dynamic representation corresponds either to the
energy level of each atom (left figure), where the position of each atom on the abscisse is kept fixed
along the simulation; or to the result of grouping the atoms together by their energy level (right
figure), to facilitate the visual counting of the number of atoms per energy level
Fig. 3.2 Application interface via desktop NetLogo app. The boxes and numbers in circles are not
part of the interface and have been included in this figure for explanatory purposes, specifically, to
structure and explain the different types of controls
program is offered; and code—where it is possible to view the program’s source

code.
Within the run tab, the most important control is the speed slider, which allows
to adapt the execution speed to the computational characteristics of the machine
on which the program is run. Typically, it will be necessary to reduce the speed to
facilitate following the dynamics.
3.3.2 The Run Configuration Panel
This section is the core of the program and consists of (i) different sliders for the
adjustment of various parameters, (ii) a drop-down menu to choose the desired type
of temperature decrease, and (iii) a drop-down menu to select one of the predefined
experiments.
To configure the simulation, the first control parameter to tune is use-cooling-
method. If this binary control is On, the evolution of the simulation is determined by
the cooling mechanism configured in the rest of the controls. On the contrary, if it is
Off , the cooling process is performed manually and interactively by the user through
the control of the temperature slider. This latter slider, in mode On, reflects the current
temperature at which the descent is taking place and its value is dynamically updated
as the execution progresses.
The level-length parameter (sometimes denoted by L) represents the number of
energy-change attempts that each atom makes before the temperature decreases. In
simulated annealing, this value represents the trajectory length in the solution space
for each constant temperature.
In the drop-down menu, the type of cooling schedule can be selected from the
different options explained in Sect. 3.2 of this document. Depending on the cooling
program selected, different sliders allow to configure the parameters specific to each
mechanism. Thus, it is possible to change the value of c in Eq. (3.2) using linear-beta,
or the value of α in Eq. (3.3) using geometric-alpha.
Finally, three preconfigured experiments are presented:
• Experiment 1: very fast cooling schedule or quenching: this is an example of a
metastable solution where the algorithm converges excessively fast (7 iterations).
• Experiment 2: fast cooling and good solution quality (68 execution steps). The
algorithm descends at a relatively fast rate giving a solution that is close to the
optimum.
• Experiment 3: slower cooling and global optimum (209 iterations) (Fig. 3.3).
3.3.3 The Simulation Control Panel
This panel consists of three buttons: the setup button, which allows you to initialize
the simulation (you need to press it before you can run your simulation); and two
Fig. 3.3 Typical results of the default experiments are displayed in version A: experiment 1 (left),
experiment 2 (middle), and experiment 3 (right)
additional buttons: go-step, which allows step-by-step execution, and the go button,
which runs the simulation continuously.
3.3.4 The Individual Energy Display Panel for Each Atom
This panel is the only one that changes in the two versions of the program (A and
B). As previously stated, it allows to choose the type of traceability desired for
the individual atoms. Recall that the number of energy levels implemented is 24
(including level 0) and that it is not modifiable through the control interface.
3.3.5 The General Information Panel
This block of the interface shows the dynamic evolution of the different elements of
the simulation. The upper graph shows the evolution of the temperature as the simu-
lation progresses. The middle graph shows the total energy of the system, which
corresponds to the sum of the individual energy of each atom and is strongly deter-
mined by the temperature at which the system is. Finally, the bottom part shows the
frequency distribution of the energy level of each atom.
3.4 Conclusions
Interactive and visual tools facilitate understanding and memorization of concepts.

In this work, we present a simple teaching resource that serves to illustrate in
a quick, straightforward and visual way the analogy between the thermodynamic
annealing process and the combinatorial optimization process that occurs in simu-
lated annealing using the Metropolis model. This analogy is not only relevant to
understand the origin of this optimization method but also to interactively under-
stand the effect of the different cooling schedules, key to properly parameterize the
algorithm.
The tool is openly available in two versions (A and B), which differ in the type of
visualization offered. As regards the programming language used, it was developed
in NetLogo (version 6.2.), which is high-level, and, hence, allows to follow the code
quite easily. The software can be run through the desktop NetLogo app—which
requires installation—or through NetLogo Web—which allows the code to be run in
a browser without any additional installation—thus facilitating its distribution and
use by students.
Acknowledgements The authors gratefully acknowledge financial support from the Ministry
of Science, Innovation and Universities (RED2018-102518-T and PGC2018-098186-B-I00),
the Spanish Research Agency (PID2020-118906GB-I00/AEI/10.13039/501100011033) and la
Fundación la Caixa (2020/00062/001).
References
1. Bautista-Valhondo J (2020) Metaheurísticas en Ingeniería. Dextra

2. Zäpfel G, Braune R, Bögl M (2010) Metaheuristic Search Concepts. Springer Berlin
Heidelberg. https://doi.org/10.1007/978-3-642-11343-7
3. Sörensen K, Glover FW (2013) Metaheuristics. In: Encyclopedia of operations research and
management science. Springer US, Boston, MA, pp 960–970. https://doi.org/10.1007/978-1-
4419-1153-7_1167.
4. Talbi EG (2009) Metaheuristics: from design to implementation. Wiley, Hoboken, NJ
5. Kirkpatrick S, Gelatt CD, Vecchi MP (1983) Optimization by simulated annealing. Science
80(220):671–680. https://doi.org/10.1126/science.220.4598.671
6. Metropolis N, Rosenbluth AW, Rosenbluth MN, Teller AH, Teller E (1953) Equation of state
calculations by fast computing machines. J Chem Phys 21:1087–1092. https://doi.org/10.1063/
1.1699114
7. Wilensky U (1999) NetLogo. Center for connected learning and computer-based modeling,
Northwestern University, Evanston, IL
8. Stonedahl F, Wilensky U (2009) NetLogo simulated annealing model. Center for connected
learning and computer-based modeling, Northwestern University, Evanston, IL
9. Geman S, Geman D (1984) Stochastic relaxation, gibbs distributions, and the Bayesian restora-
tion of images. IEEE Trans Pattern Anal Mach Intell PAMI-6, 721–741. https://doi.org/10.1109/
TPAMI.1984.4767596
10. Szu H, Hartley R (1987) Fast simulated annealing. Phys Lett A 122:157–162. https://doi.org/
10.1016/0375-9601(87)90796-1
Chapter 4
Fostering Youth Entrepreneurship
in STEM Students for Industry 4.0 Era
L. Isasi-Sánchez , G. Castilla-Alcalá , F. A. Rivera-Riquelme ,

and A. Durán-Heras
Abstract The present work is aimed to share the results of one of the initiatives
that Universidad Carlos III de Madrid (UC3M) has launched among those oriented
to secondary and high school students, more precisely within Tecnocamp activities.
It is more than obvious that entrepreneurship offers a feasible and successful way
to economic growth and personal fulfillment. In the same way, during the last two
decades it has clearly highlighted by most of the researchers that one of the most
important things to do in order to develop the entrepreneurial spirit into the citizens
is to academically train the citizens from the very early stages at the school and also
to develop their business-oriented and entrepreneurial skills.
Keywords Entrepreneurship education (EE) · Youth entrepreneurship (YE) ·

University · Industry 4.0 · STEM · Entrepreneurship intention
4.1 Introduction
Entrepreneurship Education (EE), and more specifically Youth Entrepreneurship

(YE), have notably gained importance among the academic researchers [1], and
also within economic, academic, and social worldwide institutions: [2–4], as one of
the strategic axes to develop the entrepreneurship spirit among the citizens.
As shown by some good research works, performed worldwide in different coun-
tries and even continents, like [5–9], a clear relationship exists between the early
business and entrepreneurship training, and the probability of those young students
becoming entrepreneurs in the future, normally improving their economic situation
and, consequently, the local economic development.
In the recent years, universities have been increasing their orientation to secondary
and high school students, to ease the transition from schools to colleges. However,
L. Isasi-Sánchez (B) · G. Castilla-Alcalá · F. A. Rivera-Riquelme · A. Durán-Heras

Escuela Politécnica Superior. Área de Ingeniería de Organización. Avenida de La Universidad nº
30, Universidad Carlos III de Madrid, 28911 Leganés (Madrid), Spain
e-mail: lisasi@ing.uc3m.es

https://doi.org/10.1007/978-3-031-29382-5_4
30 L. Isasi-Sánchez et al.
most of these activities have been focused on presenting themselves to the students,
and informing and steering them to better align their future desires and labor wisdom
to the adequate majors.
Nevertheless, in the last years some universities (among which the UC3M is), have
been thinking about the possibility to contribute, for the good of general society, to the
general education of the young students, and more precisely to their entrepreneurship
spirit, especially when STEM backgrounded talent is so important for Industry 4.0
consolidation and evolution. But once the final decision to collaborate is taken, the
real challenge is to decide what is the best way to collaborate.
It is absolutely true that entrepreneurship education has been, and indeed is, a
controversial topic [10], since a final consensus about what is the best way to start
developing the entrepreneurial skills and capabilities from the early academic stages
is far from being achieved. Nevertheless, and as it is shown at [3], it seems to be clear
that, no matter what is the best way to approach the capacitation activities, a good
entrepreneurial education clearly leads to economic growth and job creation.
The majority of the researches that have been related to entrepreneurship educa-
tion, agree that it should be approached from a global perspective [5]. From primary
education students [11], until graduates [12], but it is also imperative to enhance
the entrepreneurial orientation of all those workers and professionals that have
just completed the obligatory education [8, 13, 14]. This is especially important
in depressed areas [9, 15, 16] or those with high unemployment rates [17, 18].
The European Commission, [4], launched in 2016 a specific framework, inside the
“Growth” strategy, to foster entrepreneurship education as one of the key strategic
axes for mid- to long-term economic development of Europe.
There is also a certain consensus about the fact that the final skills and compe-
tencies of those that finally become businesspeople through entrepreneurship, are
wide, not so easy to detect and even vary with the economic cycle, the environ-
ment, or the activity sector [19, 20]. A very good analysis about the differences in
entrepreneurship situation and perception between two of the top countries of our
current society, like EEUU and Japan, can be found in [21], clearly showing that not
only economical aspects are important but also the cultural conditioning factors are
extremely important.
Specifically oriented to young students, there are really good initiatives focused
on developing the entrepreneurial spirit and detecting business potential among
secondary high school students. One of the most interesting ones, as it has been
celebrated already for some years with very good results, is The Diamond Chal-
lenge [22], organized by the University of Delaware. It is a global competition,
really well-organized, but it means a really hard task for the students that participate.
Consequently, for those schools and students that participate it is a really inter-
esting activity, but it is not something to be entered massively, since it requires high
commitment from the students and their teachers, and consequently it is extremely
high consuming.
It would then be really interesting to think about much fewer demanding activities,
for the students and their teachers, but that could reach that important objective of
creating among the students the seed of the entrepreneurial spirit.
4 Fostering Youth Entrepreneurship in STEM Students for Industry 4.0 Era 31
Taking all this into consideration, the authors of the present work, all of them with
broad college teaching experience, but also with vast experience in management
positions and mentoring, decided to perform the pilot program which results are
presented. Once it was decided that it was worth trying, from the university, to
contribute with youth entrepreneurial education and information, and after some
meetings with the team that, inside the UC3M is in charge of all those activities
oriented to secondary education (high school) students, it was finally decided to set
up a “hands-on” workshop, of just three hours duration, integrated into “Tecnocamp”
activity [23].
4.2 Objectives
The main objectives that were established for the present work could be summarized
as follows:
• The main aspect to be achieved was to test whether entrepreneurship intention
could be boosted on STEM high school students, through enjoyable, interesting,
and entertaining formative actions, carried out from the university environment.
• It was also intended to study the relationship between some non-cognitive char-
acteristics, evaluated through indirect questions, and the entrepreneurship spirit
and intention.
• Integrating the activity into the main program was also a must, since most of the
attendants would have never registered for a specific entrepreneurial seminar, and
it was important to detect whether it could be interesting for them, even if they
would have never thought about this possibility.
• In Spain, Science, Technology, Engineering and Mathematics (STEM students in
general, and engineering ones in particular, are normally not expected to become
entrepreneurs. However, it seems to be clear that STEM people are really perfor-
mant when becoming businesspeople, mainly when the knowledge of new tech-
nologies, programming skills, electronics, IT, etc., are so important for Industry
4.0 evolution.
• The activity was clearly clustered into entrepreneurship education, not into
entrepreneurship training. The main objective was to give an overview of all
the main aspects of entrepreneurship, and not to master the students into any
particular technique.
Fig. 4.1 Brief outline of the contents of the activity
4.3 Methodology
4.3.1 General Description of the Activity
Tecnocamp program [23] is a full week activity, mainly oriented to STEM students,
carried out in UC3M premises, which main objective is to entertain them, while
developing interesting engineering activities, together with more leisure oriented
ones like gymkhanas and sports.
Within this framework, a three-hour entrepreneurship education activity was
programmed, splitting the attendants into working groups of five people. Each group
had to think about a business idea and, based on that idea, they had to apply all
the concepts that the trainer described, along with the main phases through which a
typical company must complete, and using the basic tools that were described (see
Fig. 4.1).
4.3.2 Research Framework
Integrated into the activity that has been described and taking into consideration
the main objectives of the performed study, two surveys were conducted to all the
students, the first one just at the beginning of the activity, and the second one at the
end of it (see Fig. 4.2).
The activity and related surveys were conducted with two groups, in different
weeks, totaling 100 students (51 students on the first group, and 49 students on the
second group), with the characteristics that are detailed in Table 4.1.
Fig. 4.2 a Preliminary questionnaire: closeness to business world through family and acquain-
tances, main interests, and non-cognitive skills evaluation. b Final questionnaire: activity evaluation
and feedback, and detection if future entrepreneurship interest had been generated
Table 4.1 Main characteristics of the students

Family member with business
No Yes Total (%)
Group A 18 33 51
Degree 9th 1 1 2.0
10th 5 12 17 33.3
11th 11 9 20 39.2
12th 1 12 13 25.5
Group B 21 28 49
Degree 10th 10 12 22 44.9
11th 4 6 10 20.4
12th 7 10 17 34.7
Totals 39 61 100
4.4 Results
The main results that have been obtained from the collected data of the performed
activity are attached in Table 4.2. This table summarizes the general results and shows
the results that are obtained when considering two of the main aspects of the study:
• Different valuations depending on the main non-cognitive skills of the students.
Group I includes all the students declaring that one of their main objectives is to
have a good performance in all the school related homework; Group II is formed
for all those that have as a key priority to enjoy life, Group III includes all those
who are really thinking about how to “make money” in the future, and inside
Table 4.2 Summarized results of both. The first column (“ALL”) shows the average values of all
the students, for each of the questions. Detailed values corresponding to the main aspects are shown
in the next four columns. Values above the corresponding average are shown in italics
Main objectives Relative with
business
All I II III IV Yes No
Do you like to take risks 3,49 3,33 3,55 3,61 3,36 3,55 3,49
in your life? [Surely not
(1), May be (3), Surely
yes (5)]
How much time do you 4,39 4,53 4,36 4,33 4,57 4,37 4,41
think should be
dedicated to a business?
[Little (1), Normal (3),
A lot (5)]
Have you ever 2,86 2,53 3,03 3,14 2,64 3,30 2,68
considered becoming an
entrepreneur? Never (1),
Ever (3), Many times
(5)]
Do you like finances? 3,02 2,73 3,03 3,29 2,21 3,05 2,84
Nothing (1), Something
(3), A lot (5)]
Do you like to delegate? 3,14 3,60 3,28 3,43 4,43 3,77 3,16
Nothing (1), Something
(3), A lot (5)]
Do you like team 3,69 3,07 3,73 3,73 3,36 3,55 3,78
working? Nothing (1),
Something (3), A lot
(5)]
How do you think you 3,10 3,07 3,09 3,06 3,79 2,97 3,05
tolerate failures? Very
bad (1), Normal (3),
Very good (5)]
What is more important 3,88 3,80 3,88 3,90 3,86 3,88 3,86
for you? Price (1),
Quality (5)]
Imagine that you are the 3,65 3,67 3,49 3,37 3,57 3,73 3,11
owner of a good and
profitable business, and
an investor wants to buy
it from you for a lot of
money: Do you think
you would sell it?
Surely yes (1), I don’t
know (3), Surely not
(5)]
(continued)
Table 4.2 (continued)

Main objectives Relative with
business
All I II III IV Yes No
In general, what is your 3,76 3,73 3,65 3,75 3,46 3,76 3,64
opinion about
businessmen? Very bad
(1), Average (3), Very
good (5)]
And about 4,16 4,27 4,16 4,27 4,08 4,14 4,19
entrepreneurs?
Did you like the 4,02 3,93 3,99 4,10 3,79 4,10 3,89
workshop [Nothing (1),
Much (5)?
After having attended 3,61 3,73 3,52 3,61 3,29 3,60 3,46
the workshop, do you
plan to become an
entrepreneur much less
(1), the same (3), much
more (5) than before?
After having attended 3,73 3,67 3,87 3,86 4,00 3,92 3,73
the workshop, would
you recommend it to
your colleagues?
[Surely not (1), May be
(3), Surely yes (5)]
Do you think it is 4,04 4,20 4,09 4,10 4,07 4,20 3,86
convenient to do
something Similar in
ESO or in High School,
at the institute? [Surely
not (1), May be (3),
Surely yes (5)]
Group IV are all those who are really dedicated to programming applications, and
related technical tasks. As it can be seen in some research works, [12, 16, 19, 24],
these behavioral tendencies are absolutely related to some of the most important
skills for an entrepreneur, like the self-awareness, risk tolerance, resilience, or
resistance to frustration.
• The effect of having or not a close relative with business.
As shown in the table, some general conclusions can easily be obtained. The
students definitely liked the activity (4,04 average evaluation), even if it “com-
peted” with other activities that were initially supposed to be much more pleasant
for them. The “declared” interest to become an entrepreneur in the future clearly
moved from the value 2,86 to the astonishing 3,61. It will not be realistic to think
that with this short activity, a so important change into the students’ mind had been
achieved, but what is evident is that this kind of short, team-oriented, and “hands-on”
activities could enhance the entrepreneurial spirit, and contribute to what is called
entrepreneurial education. It is especially interesting to note, from the different results
that are obtained from those that have relatives with businesses, and those that have
not, that just having some closeness to business concepts, clearly shows a greater
propensity to create a business in the future.
Today, some of the largest and most important companies worldwide, like
Alphabet, Microsoft, and Tesla are good examples of the fact that an engineering
background is interesting for managers and officers, so it is surprising that both before
and after the workshop, those students close to “pure” engineering orientation, are
those less likely to entrepreneur.
4.5 Conclusions and Future Lines of Research
When this project was thought and created, it was initially planned to repeat the
described activity over several years. Unfortunately, the situation caused by COVID-
19 pandemic has made it impossible to maintain it in the last two years. Anyhow, the
authors have finally decided to share the main results since, although some aspects
should be confirmed with a greater sample of students, some interesting conclusions
have already been obtained. The main ones are:
• It is definitely possible and interesting to contribute, from the university and their
different colleges to the entrepreneurship education of high school students.
• Short and “hands-on” activities like the one that is described in the present work
clearly contribute to foster the entrepreneurship spirit, in one of the earlier stages
of senior education.
• Most of the aspects that have been highlighted by all the cited works and
researchers, with regard to non-cognitive skills, and to the students’ environment
(family, background, etc.) have been found to have clear correlations with their
entrepreneurship intention.
From the authors’ point of view, the present work should be completed in the
future, integrating into a global philosophy youth entrepreneurship, entrepreneurship
education, high education, and entrepreneurship training aspects. Thus, future lines
are:
• Work closely with the high school teachers to establish a good framework to
increase the knowledge of the students about business in general.
• Collaborate with the rest of the stakeholders to really create an “end-to-end”
process to improve all the entrepreneurship “ecosystem”.
References
1. Bezerra ÉD, Borges C, Andreassi T (2017) Universities, local partnerships and the promotion of
youth entrepreneurship. Int Rev Educ 63:703–724. https://doi.org/10.1007/s11159-017-9665-y
2. Kshetri N (2018) Global Entrepreneurship
3. Lackéus M (2015) Entrepreneurship in education. What, why, when, how
4. Commission E, Entrepreneurship Education Available online: https://ec.europa.eu/growth/
smes/supporting-entrepreneurship/education_en
5. Brüne N, Lutz E (2020) The effect of entrepreneurship education in schools on entrepreneurial
outcomes: a systematic review. Manag Rev Q 70:275–305. https://doi.org/10.1007/s11301-
019-00168-3
6. Dzomonda O, Fatoki O (2019) The role of institutions of higher learning towards youth
entrepreneurship development in South Africa. Acad Entrep J 25:1–11
7. Sowole OE, Hogue ME, Adeyeye OP (2018) Entrepreneurship: psychological factors influ-
encing youth’s desire for self-sustenance in Mpumalanga. Acad Entrep J 24:1–16
8. Beeka BH, Rimmington M, Esson J, Brixiová Z, Ncube M, Bicaba Z, Rtischev D, Geldhof
GJ, Porter T, Weiner MB et al (2015) Entrepreneurial development for U.S. minority homeless
and unstably housed youth: a qualitative inquiry on value, barriers, and impact on health. Acad
Entrep J 24:705–718. https://doi.org/10.1007/s11187-016-9798-6
9. Williams M, Hovorka AJ (2013) Contextualizing youth entrepreneurship: the case of
Botswana’s young farmers fund. J Dev Entrep 18:1350022. https://doi.org/10.1142/S10849
46713500222
10. Oosterbeek H, van Praag M, Ijsselstein A (2010) The impact of entrepreneurship education on
entrepreneurship skills and motivation. Eur Econ Rev 54:442–454. https://doi.org/10.1016/j.
euroecorev.2009.08.002
11. Barba-Sánchez V, Atienza-Sahuquillo C (2016) The development of entrepreneurship at school:
the Spanish experience. Educ Train 58:783–796. https://doi.org/10.1108/ET-01-2016-0021
12. Obschonka M, Hakkarainen K, Lonka K, Salmela-Aro K (2017) Entrepreneurship as a twenty-
first century skill: entrepreneurial alertness and intention in the transition to adulthood. Small
Bus Econ 48:487–501. https://doi.org/10.1007/s11187-016-9798-6
13. Beeka BH, Rimmington M (2011) Entrepreneurship as a career option for African youths. J
Dev Entrep 16:145–164. https://doi.org/10.1142/S1084946711001707
14. Cueto B, Mayor M, Suárez P (2017) Evaluation of the Spanish flat rate for young self-employed
workers. Small Bus Econ 49:937–951. https://doi.org/10.1007/s11187-017-9853-y
15. Olaniran SO, Mncube DW (2018) Barriers to effective youth entrepreneurship and vocational
education. Acad Entrep J 24:1–10
16. Brixiová Z, Ncube M, Bicaba Z (2015) Skills and youth entrepreneurship in Africa: analysis
with evidence from Swaziland. World Dev 67:11–26. https://doi.org/10.1016/j.worlddev.2014.
09.027
17. Papagiannis GD (2018) Entrepreneurship education programs: the contribution of courses,
seminars and competitions to entrepreneurial activity decision and to entrepreneurial spirit and
mindset of young people in Greece. J Entrep Educ 21:1–21
18. Vega LES, González-Morales O, García LF (2016) Entrepreneurship and adolescents. J New
Approach Educ Res 5:123–129. https://doi.org/10.7821/naer.2016.7.165
19. Geldhof GJ, Porter T, Weiner MB, Malin H, Bronk KC, Agans JP, Mueller M, Damon W,
Lerner RM (2014) Fostering youth entrepreneurship: preliminary findings from the young
entrepreneurs study. J Res Adolesc 24:431–446. https://doi.org/10.1111/jora.12086
20. Luis-Rico M-I, Escolar-Llamazares M-C, de la Torre-Cruz T, Herrero A, Jimenez A, Arranz
Val P, Palmero-Camara C, Jimenez-Eguizabal A (2020) The association of parental interest in
entrepreneurship with the entrepreneurial interest of Spanish youth. Int J Environ Res Public
Health 17:4744. https://doi.org/10.3390/ijerph17134744
21. Rtischev D (2017) A strategic behavior analysis of why ventures are risky for young people
in Japan but not in Silicon Valley. J Jpn Int Econ 44:78–89. https://doi.org/10.1016/j.jjie.2017.
03.003
22. Delaware U, The diamond challenge. Available online: https://diamondchallenge.org/

23. uc3m UC3M Tecnocamp Activities for Secondary Schools Available online: https://www.
uc3m.es/secondary/disclosure-science/tecnocamp
24. Yu CWM (2013) Capacity building to advance entrepreneurship education: lessons from the
teen entrepreneurship competition in Hong Kong. Educ Train 55:705–718. https://doi.org/10.
1108/ET-01-2013-0001
Part II
Management Information Systems
and Knowledge Management
Chapter 5
Distributed Ledger Technology
in Industry 4.0: An Implementation
S. Fernández-Vázquez, R. Rosillo, P. Priore, and J. Puente
Abstract Blockchain, also known as distributed ledger technology, is a type of

transformational technology that is currently regarded as one of the most important
instruments of the well-known Industry 4.0. Blockchain’s many properties, including
smart contracts, decentralization, transparency, traceability, data immutability, and
data protection, combined with a consensus structure, make it appropriate for appli-
cation in today’s fast-paced global businesses. As a result, businesses should eval-
uate and compare the value of traditional supply chains with new Blockchain-based
systems that add characteristics like transparency to the picture. The purpose of
this article is to use a literature review to demonstrate the benefits of Blockchain in
supply chain management, highlighting key features such as sustainability, decen-
tralization, data immutability, and the usage of smart contracts. It also seeks to offer
experts with positive consequences so that suitable measures may be taken to deploy
this technology.
Keywords Blockchain · Smart contracts · Industry 4.0
5.1 Introduction
At the Hannover Fair in 2011, the term Industry 4.0 was first introduced, referring
to how advances in technology would profoundly change the organization of global
S. Fernández-Vázquez (B) · R. Rosillo · P. Priore · J. Puente

Business Management Department, University of Oviedo, Oviedo, Spain
e-mail: fernandezsimon@uniovi.es
R. Rosillo
e-mail: rosillo@uniovi.es
P. Priore
e-mail: priore@uniovi.es
J. Puente
e-mail: jpuente@uniovi.es

https://doi.org/10.1007/978-3-031-29382-5_5
42 S. Fernández-Vázquez et al.
value chains. Although Industry 4.0 has generally been accepted by the community,
other concepts have also developed in this time period which refer to the use of digital
technologies in production [19].
Lately, our society has seen the advent of an innovative wave of transforma-
tive technology distributed through multiple industries called Industry 4.0 [6]. The
industrial sector was responsible for the term known as Industry 4.0. Nevertheless,
many industries have undergone an increase in their production by using disruptive
technologies [21]. This has led, for instance, to an increase in the use of these new
technologies in sectors such as banking or telecommunications [5].
Currently, these service companies either use or test these innovations to modify
the way they conduct business. The spectrum of the Industry 4.0 revolution involves
a broad variety of innovations such as cloud computing, the Internet of Things (IoT),
artificial intelligence (AI) or Blockchain [7]. In today’s world, incorrect and corrupted
data can lead to inaccurate choices and become a major challenge to connected,
dynamic development processes. The present manufacturing management typically
depends on a centralized network, with limited data traceability and fragile to failure
in processes [14].
In evolving conditions, the benefit of these technologies remains in their ability
to learn through AI, their highly secured processes and their capability to predict.
Customer knowledge and data can be combined through the use of cloud computing
[13].
Through the use of Blockchain’s groundbreaking technological framework that
has recently revolutionized the industry in device protection and performance, secu-
rity issues can be solved [2]. An open and shared framework for rendering transactions
in both enterprise and industry fields is provided by the Blockchain as a basis for
distributed ledgers. Blockchain’s innate features increase trust through clearness and
traceability of transactions [1]. The final aim is to enable machines such as computers
to develop and interpret concepts such as those of the human mind [18].
Industry 4.0 is a shift from a centralized planned production to a dynamic and
decentralized production in order to improve the quality of goods, tailor-made
processes, and the flexibility of systems [24]. In order to make collaboration choices,
a centrally controlled platform cannot prevent data privacy from other users, as it is
essential to know one another’s capacities and conditions. Manufacturing companies
also have to resolve the low robustness of centralized systems from a single key node,
leading to unreliable networking and data service [20].
5.2 Blockchain
The Blockchain is a distributed public database that can be configured for data sharing
and storage. Commonly defined as a distributed ledger, it consists of a chain of blocks
and is built around a peer-to-peer (P2P) or shared network [23].
It is composed, among others, by consensus protocols, methods of cryptography,
as well as smart contracts. It comprises modified blocks of data that are decentralized.
5 Distributed Ledger Technology in Industry 4.0: An Implementation 43
Fig. 5.1 Main

characteristics of Blockchain
Nodes agree to data
technology Consensus verification
Business terms
Shared encoded in the
Contract contract record
Data exchanged
Shared amongst the
Ledger network
Guarantees security,
verification and
Cryptography authentication of
transactions
A timestamp along with a connection to a previous block is included in each block of

data. In order to trace each transaction in the database back to the source, a Blockchain
includes full historical records. Blockchain is a modern secure and publicly available
computer model [3]. Its main characteristics can be seen in Fig. 5.1.
Its applications are typically built based on the technology offered by mainstream
Blockchain networks, such as Ethereum, EOS, Cardano, Hyperledger Fabric, or
Stellar. It is worth pointing out that different networks use different consensus algo-
rithms. This is the way in which the users in a specific Blockchain reach an agreement.
Below are some of the most important consensus algorithms and their principles [ 9]:
• Proof of Work (PoW): Consensus through mining by adding directly blocks to
the Blockchain.
• Proof of Stake (PoS): The higher the stake the nodes have the more chances they
will have in being accountants.
• Delegated Proof of Stake (DPoS): Nodes vote by the stake they hold.
• Notaries: Certifies that, for a particular transaction, no other transactions have
already been signed that consume all of the input states of the proposed transaction.
• Orderer: Through transaction ordering, alongside other orderer nodes forms
an ordering service.
• NeoScrypt: A PoW mining algorithm that has to be mined with graphics cards.
• Tangle: Miners do not validate transactions. Network participants jointly go
through the validation process.
• Stellar: Nodes go through rounds of federated voting.
Table 5.1 Record of some of

Networks Consensus algorithm
the main Blockchain
networks and their consensus Bitcoin PoW
algorithms Bitcore Timetravel 10
Cardano Ouroboros
Corda Notaries
EOS DPoS
Ethereum PoW
Fabric Orderer
Feathercoin NeoScrypt
IOTA Tangle
Qtum PoS
Stellar Stellar
Tezos PoS
Wanchain PoS
A list of the main networks and their consensus algorithms can be seen in Table
5.1.
5.3 Smart Contract Implementations
Smart contracts convey an independent, autonomous system that is encoded to carry

out a series of transactions without the intervention of a human being. In order for
this contract to execute the transactions, a code needs to be embedded in the smart
contract to perform a specific action if a series of requirements are met [11].
The incorporation of smart contracts arrived in IoT, mainly due to the security
of its application and the reduction of, for example, financial risk. This inclusion
allows transactions to be paid automatically or fully dedicated payment schemes. For
instance, in industries such as agriculture, farmers can use more efficient systems
in which the payments made to farmers are made in a different scheme than the
traditional fixed rate systems [15]. Smart contracts are nowadays included in many
Blockchain implementations. Some of those networks in which the deployment of
smart contracts are allowed include Ethereum or Hyperledger Fabric.
The integration of Blockchain and IoT into today’s payment system in the trans-
portation industry means having autonomous scenarios which are traceable and
secure. In the case of car rental, for instance, through the use of an application
in the user’s smartphone, a secure and transparent payment system can be deployed
through the use of smart contracts. Another sector in which smart contracts can be
introduced is fuel payment. In traditional mechanisms, the interaction between the
user’s credit card and the petrol pump takes place. In contrast, when using smart
contracts, there is no need for a central authority. In this case, the vehicle, which is
running a decentralized application (dApp) on the Blockchain, sends its cryptocur-

rency to the smart contracts. The gas station communicates with the Blockchain
explicitly to assess if the car has charged and tracks how much gas has been bought
[8].
Another area related to Internet of Things (IoT) in which this technology can
excel is micropayments. Traditional payment systems are not the best method for
a great deal of huge micropayments. The reasons being are their high transactional
costs and their limited capacity. Moreover, our credit card information when making
micropayments is shared between other devices [22]. In order to implement smart
contract systems in micropayments, current issues with more traditional methods
must be identified. Examples of these issues are:
• High transaction fees.
• High transaction timeframes.
• A distribution system with lack of transparency.
The use of Blockchain could help tackle these issues, as some of Blockchain’s
characteristics are [17]:
• Low processing fees: Through the use of Blockchain, third-party fees are avoided.
Payments are done in tokens and other users receive these tokens. The fees for
transactions in cryptocurrencies are extremely low in comparison to traditional
methods.
• Instant payment: When sending money, for instance, the transaction is completed
within a few seconds. This contrasts with the hours or days that it might take to
send money from one country to another (even more when it is done in different
currencies).
• Transparent distribution: Smart contracts hold in place the transaction and the
release of currency is automatic.
When using Blockchain technology in IoT, the system shifts toward a greater
control in trade processes without human interference. The devices could be authen-
ticated to ensure the security of the data transmitted and to deter unauthorized users.
Blockchain could improve the IoT by offering immutability to apps, redundancy,
openness, traceability, and durability of operations [12].
Industry 4.0 includes the smooth convergence of processes through all elements.
The production operations handled in the distributed shared ledger should be
organized and reconciled between the separate nodes on the Blockchain [4]. The
middleware is essential to the incorporation of Blockchain services in order to
provide stability, traceability, and decentralized manufacturing implementations
between participating nodes. It is important to define the interface framework
of Blockchain-driven manufacturers, to design the operating principles of the
Blockchain manufacturing partnership specifically, and to create adaptive Blockchain
logical structures for the manufacturing services [16].
5.4 Conclusion
The emerging use and implementation of Blockchain in Industry 4.0 are at a prelim-
inary phase, as this is a field that has a lot to explore. Some methods demonstrate
that most of the techniques are tailored for particular systems in which they aim to
simplify horizontal integration. Consequently, strategies for a vertical transformation
of manufacturing do require practice to go forward. Advances in sectors such as car
rental, fuel, or micropayments are just some examples of real applications in today’s
world that have been implemented. Nevertheless, there is still a long way to go before
these and other Blockchains are used worldwide on a daily basis.
References
1. Abeyratne S, Monfared R (2016) Blockchain ready manufacturing supply chain using

distributed ledger. Int J Res Eng Technol
2. Ahram T, Sargolzaei, Arman S, Saman D, Jeff A, Ben (2017) Blockchain technology
innovations, 137–141. https://doi.org/10.1109/TEMSCON.2017.7998367
3. Al-Jaroodi J, Mohamed N (2019) Blockchain in industries: a survey. IEEE Access, pp 1–1.
https://doi.org/10.1109/ACCESS.2019.2903554
4. Angrish A, Craver B, Hasan M, Starly B (2018) A Case study for blockchain in manufacturing:
“FabRec”: a prototype for peer-to-peer network of manufacturing nodes. Proc Manuf 26. https://
doi.org/10.1016/j.promfg.2018.07.154
5. Büchi G, Cugno M, Castagnoli R (2020) Smart factory performance and Industry 4.0. Technol
Forecast Soc Change. https://doi.org/10.1016/j.techfore.2019.119790
6. Chang SE, Chen Y-C, Lu M-F (2019) Supply chain re-engineering using Blockchain tech-
nology: a case of smart contract based tracking process. Technol Forecast Soc Chang 144:1–11.
https://doi.org/10.1016/j.techfore.2019.03.015
7. Chang V, Wang Y, Wills G (2020) Research investigations on the use or non-use of hearing
aids in the smart cities. Technol Forecast Soc Change. https://doi.org/10.1016/j.techfore.2018.
03.002
8. Ferreira C, Rabelo R, Sá Silva J, Cavalcanti C (2020) Blockchain for machine to machine inter-
action in industry 4.0. In book: blockchain technology for industry 4.0, secure, decentralized,
distributed and trusted industry environment, Springer, 99–116. https://doi.org/10.1007/978-
981-15-1137-0_5
9. Fu X, Wang H, Shi P (2020) A survey of blockchain consensus algorithms: mechanism, design
and applications. Sci China Inf Sci 64(2). https://doi.org/10.1007/s11432-019-2790-1
10. Garg P, Gupta B, Chauhan AK, Sivarajah U, Gupta S, Modgil S (2020) Measuring the perceived
benefits of implementing blockchain technology in the banking sector. Technol Forecast Soc
Change, 120407. https://doi.org/10.1016/j.techfore.2020.120407
11. Guadamuz A (2019) All watched over by machines of loving grace: a critical look at smart
contracts. Comput Law Secur Rev 35:105338. https://doi.org/10.1016/j.clsr.2019.105338
12. Hassan M, Rehmani M, Chen J (2019) Privacy preservation in blockchain based IoT systems:
integration issues, prospects, challenges, and future research directions. Future Gener Comput
Syst 97. https://doi.org/10.1016/j.future.2019.02.060
13. Larson D, Chang V (2016) A review and future direction of agile, business intelligence,
analytics and data science. Int J Inf Manage 36:700–710. https://doi.org/10.1016/j.ijinfomgt.
2016.04.013
14. Leng J, Ye S, Zhou M, Zhao J, Liu Q, Guo W, Cao W, Fu L (2021) Blockchain-secured smart
manufacturing in industry 4.0: a survey. IEEE Trans Sys Man Cybernetics: Syst 51. https://doi.
org/10.1109/TSMC.2020.3040789
15. Lim CH, Lim S, How BS, Ng WPQ, Ngan SL, Leong WD, Lam HL (2021) A review of industry
4.0 revolution potential in a sustainable and renewable palm oil industry: HAZOP approach.
Renew Sustain Energy Rev 135:110223. https://doi.org/10.1016/j.rser.2020.110223
16. Mohamed N, Al-Jaroodi J (2019) Applying blockchain in industry 4.0 applications. 0852–
0858. In: Proceedings IEEE 9th annual computing and communication workshop conference,
Las Vegas, NV, USA. https://doi.org/10.1109/CCWC.2019.8666558
17. Mushtaq A, Haq I (2019) Implications of blockchain in industry 4.0, 1–5. https://doi.org/10.
1109/CEET1.2019.8711819
18. Qu Y, Pokhrel SR, Garg S, Gao L, Xiang Y (2020) A blockchained federated learning framework
for cognitive computing in industry 4.0 networks. IEEE Trans Indus Inf, 1–1. https://doi.org/
10.1109/tii.2020.3007817
19. Schwab K (2017) The fourth industrial revolution. Crown Publishing Group, New York
20. Shen W (2002) Distributed manufacturing scheduling using intelligent agents. Intell Syst IEEE
17:88–94. https://doi.org/10.1109/5254.988492
21. Skilton M, Hovsepian F (2017) The 4th industrial revolution: responding to the impact of
artificial intelligence on business. Springer International Publishing, AG, Cham
22. Stjepandic J, Wognum N, Verhagen W (2015) Concurrent engineering in the 21st century.
Foundations, developments and challenges. Springer
23. Wang M, Wu Y, Chen B, Evans M (2020) Blockchain and supply chain management: a new
paradigm for supply chain integration and collaboration. Oper Supply Chain Manage: Int J,
111–122. https://doi.org/10.31387/oscm0440290
24. Zarreh A, Wan H, Lee Y, Saygin C, Al Janahi R (2019) Risk assessment for cyber security of
manufacturing systems: a game theory approach. https://doi.org/10.31224/osf.io/mb5t9
Chapter 6
A Bibliometric Analysis
of the Time-Driven Activity-Based
Costing System. The Power of Cost
Accounting in Organizations
Patxi Ruiz-de-Arbulo-López , Jesús Rodríguez-Martín ,

Jordi Fortuny-Santos , and Beñat Landeta-Manzano
Abstract The aim of this paper is to explore and evaluate, using bibliometric anal-
ysis, the papers published up to January 2021 on the time-driven activity-based
costing (TDABC) system. In recent decades, companies adopting innovations in
production management (e.g., lean manufacturing) have had to look for new methods
of cost control. For this purpose, the activity-based costing system was developed,
among others, with some implementation drawbacks. In 2004, Robert Kaplan, who
developed the ABC costing system, developed the TDABC method. The conclusions
of the paper show how the TDABC system has been mainly analyzed in healthcare
organizations and little in industrial organizations.
Keywords Time-driven activity-based costing · Bibliometric analysis · Network

analysis
6.1 Introduction
The last four decades have been characterized by changes in the business envi-
ronment, such as increased competition and growing customer demands in terms
P. Ruiz-de-Arbulo-López (B) · J. Rodríguez-Martín · B. Landeta-Manzano

University of the Basque Country (UPV/EHU), Bilbao, Spain
e-mail: patxi.ruizdearbulo@ehu.eus
J. Rodríguez-Martín
e-mail: jesus.rodriguez@ehu.eus
B. Landeta-Manzano
e-mail: benat.landeta@ehu.eus
J. Fortuny-Santos
Universitat Politecnica de Catalunya, Manresa, Spain
e-mail: jordi.fortuny@upc.edu

https://doi.org/10.1007/978-3-031-29382-5_6
50 P. Ruiz-de-Arbulo-López et al.
Table 6.1 Stages of the TDABC

Heading level
1. Identifies the activities that are carried out with the same means to constitute the “resource
groups”
2. Estimate the resources consumed by each “resource group”
3. Estimates the normal capacity of each resource group in terms of working hours
4. Calculate the unit costs of the drivers (the most common driver is the working minute) for
each resource group by dividing the cost of the resources consumed by the normal capacity
5. Determine for each task the time required based on its characteristics
6. To value each task, multiply the unit cost of the resources by the time required to perform it
of quality, price, customization and delivery times. In response to these changes,

innovations in production systems, such as lean manufacturing, have been introduced.
On the other hand, in the 1980s, it was observed that traditional cost accounting
evaluated unfavorably the innovations introduced by new approaches to production
management [13]. In response to the distrust of information derived from traditional
systems [5, 6], Robin Cooper and Robert Kaplan developed the activity-based costing
(ABC) method. The ABC system was designed to resolve the allocation of indirect
costs—increasingly important in companies—which was done in an almost arbitrary
way.
The ABC system has not been widely accepted [12]. The implementation of
an ABC system is a time-consuming process, as the development of interviews and
surveys necessary to understand the activities carried out in the company is very time-
consuming. In addition, companies are unsure of how to allocate costs to activities,
because it is often based on subjective calculations of the percentage of time spent
on each activity by each manager.
In summary, both academics and practitioners point out that the ABC model is
not accurate enough to capture the complexity of a company’s real operations.
To overcome the drawbacks of the ABC costing system, Robert Kaplan and Steven
Anderson developed the time-based ABC system (TDABC) in 2004 [3, 4] (Table 6.1).
6.2 Methodology
A systematic literature review was conducted to explore the current status of TDABC.
In order to minimize bias in the selection of the papers included in this study, a
systematic methodology was carried out. In contrast to the type of literature review
involved in any research, a systematic review can be defined as the review of a subject
matter using systematic methods to identify, select, and critically appraise relevant
research [10].
Articles were obtained from Web of Science and the Scopus database. These
sources ensure a selection of articles in high impact factor journals and refereed
manuscripts in reputable conference proceedings. Since WoS and Scopus are two
6 A Bibliometric Analysis of the Time-Driven Activity-Based Costing … 51
complementary databases, but in this case Scopus has more records than WoS, the
Scopus database has been chosen.
All searches were limited to the following conditions:
1. Type of document: Journal articles (mostly academic journals, but some are
practitioner journals), conference proceedings and book chapters.
2. Language: No language was set a priori, but the keywords used were in English.
3. Year: From 2004 to January 2021.
The search terms used to retrieve the articles from the databases were: “time-
driven activity-based costing” or “TDABC”. These words were entered either in
the title, in the keywords or in the abstract of the databases’ search engines. 178
files in Scopus matched these search criteria. The next step was to read the abstract
of each of the 178 papers and determine whether they matched our research topic.
A researcher and an assistant reviewed the abstracts in the first instance and then
a second researcher repeated the review, and some documents were excluded due
to various inconsistencies found, leaving 147 documents (120 journal articles, 16
conference papers, 8 book chapters, 2 notes, and 1 review).
6.3 Results and Discussion
6.3.1 General Trends in the Literature
Although it can be considered a young field of research, the last ten years have seen
a remarkable increase in the number of published articles (Fig. 6.1). The first three
publications were made in 2008. The rate of publications has gradually increased,
reaching 22 papers in 2016 and 19 in 2017. In terms of citations, an upward trend is
observed year after year. The most cited papers are Keel et al. [8] with 87 citations;
Laviana et al. [9] with 84 citations; Everaert et al. [2], cited 74 times; McLaughlin
et al. [11], with 47 citations; Kaplan, [7], cited 43 times. It is worth noting that of these
7 papers, 5 of them are from the hospital sector, one of them is from the distribution
sector, and finally, the last one is a doctrinal article on TDABC.
Table 6.2 shows the most productive authors. Siguenza-Guzman stands out with
eight publications. He is followed by Kaplan, creator of the ABC system and TDABC,
with seven publications. Feely and Guzman are in third and fourth place, both with
five publications.
The journals that publish papers on TDABC come from different fields of knowl-
edge. The three most relevant are medicine, business, management and accounting,
and engineering. In fourth place is computer science (Fig. 6.2).
30 300
26
25 253 250
22 22
20 200
19
Nº Publications
181
Nº Citations
15 150
15 150
10
9 108
10 7 100
7
65
5 3 30 50
3 4 42
1 12 14 18
7
1
0 1 1 0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Year
Publications Citations
Fig. 6.1 Papers and citations per year
Table 6.2 Most active organizations

Author Documents Citations Total link strength
Siguenza-Guzman 8 37 24
Kaplan 7 158 37
Feeley 5 84 36
Guzman 5 33 45
Balakrishnan 4 50 9
Burke 4 186 42
Polanczyk 4 13 15
Thaker 4 38 47
Agrawal 3 2 6
Bouami 3 28 4
Fig. 6.2 Disciplines of the journals or proceedings

6.3.2 Academic Performance: Country, Organizations,

and Authors
In terms of academic performance, the most productive countries, organizations,

and authors in the field of study were analyzed. The results are shown graphically in
Figs. 6.3, 6.4, and 6.5. The three figures were generated by the VOSviewer software
[14], following the procedure described by Calzado-Barbero et al. [1]. The most
productive were the USA (65 publications), followed by Brazil (10), China (9), and
Belgium (9) (Fig. 6.3). It should be noted that the USA has produced consistently
over the last ten years; however, in 2011 China (ranked 4) started publishing the
results of its TDABC research.
Fig. 6.3 Publication evolution by top countries per year
Fig. 6.4 Networking between countries

Fig. 6.5 Inter-university collaborations
To identify the main collaborative networks between countries, a network analysis

was carried out. The size of the node indicates the number of collaborations and, as
shown in Fig. 6.3, the most collaborating countries are in the core of the network and
are the USA and Sweden. We also observe other countries such as China, Belgium,
Brazil, and Portugal establishing intense network collaborative networks. At the
continental level, collaborations are mainly between North America, Europe, and
Asia.
In terms of organizations, the most productive ones between 2008 and 2021 were
Harvard Medical School (USA) and Harvard Business School (USA), both with 11
publications, followed by the MD Anderson Cancer Center of the University of Texas
(USA) and the University of Cuenca (Ecuador) with eight publications. However, as
can be seen in Table 6.3, the publications of the David Geffen School of Medicine at
UCLA (California) have an average number of citations that places them in second
position.
It is worth mentioning that a large proportion of the universities are medical
schools. This is because the TDABC costing system has been widely applied in the
medical field.
Table 6.3 Most active organizations

Publications Organization Average number of citations per
publication
11 Harvard Medical School 47,73
11 Harvard Business School 16,36
8 University of Texas MD Anderson 11,50
Cancer Center
8 University of Cuenca 4,63
7 Brigham and Women’s Hospital 6,29
5 David Geffen School of Medicine at 40,80
UCLA
4 Universidade Federal do Rio Grande 3,25
do Sul
4 Karolinska Institutet 23,50
4 KU Leuven 8,00
4 University of California, Los Angeles 39,25
4 Massachusetts General Hospital 14,25
4 Universiteit Gent 32,50
Fig. 6.6 Coword analysis
The following network of organizations (Fig. 6.5) reflects collaborations between

universities. As this is still a recent field, the collaborations are not as extensive
as shown in Fig. 6.5. Harvard University institution declared the world’s largest
producer of scientific articles on this topic.
Finally, a coword analysis was conducted to trace the interactions between themes,
the strength of these associations, and even research trends. Figure 6.6 shows
seventeen keywords in four different clusters.
6.4 Conclusions
This work contributes to the development of research on TDABC. The bibliometric

study has revealed which researchers work on CBABA and in which institutions or
organizations they carry out their work. The number of articles on this topic has
increased in recent years, which shows, on the one hand, that there is a real interest
in this issue, but this growth has not been very large and its practical application
in organizations has mainly focused on the case of hospitals. Some cases can be
mentioned in the automotive and retail sectors.
The TDABC system originated in the USA, specifically at Harvard University,
and most of the articles come from the USA, specifically from Harvard University.
However, as this is an underresearched field of knowledge, it is still difficult to
identify the main authors of the research stream. It is possible that the fact that it has
not had a great expansion or degree of implementation in organizations is due, as in

the case of the ABC costing system, to the complexity of its application.
References
1. Calzado-Barbero M, Fernández-Portillo A (2019) Educación emprendedora en la universidad.

J Manage Bus Educ 2(2):127–159
2. Everaert P, Bruggeman W, Sarens G, Anderson SR, Levant Y (2008) Cost modeling in logistics
using time-driven ABC. Experiences from a wholesaler. Inter J Phys Distrib Logistics Manage
38(3):172–191.
3. Kaplan RS, Anderson S (2007) Time-driven activity-based costing. A simpler and more
powerful path to higher profits. Harvard Business School Press, Boston
4. Kaplan RS, Anderson S (2004) Time-driven activity-based costing. Harv Bus Rev 82(11):131–
138
5. Kaplan RS (1983) Measuring manufacturing performance: a new challenge for managerial
accounting research. Account Rev 58(4):686–705
6. Kaplan RS (1984) Yesterday’s accounting undermines production. Harv Bus Rev 62:95–101
7. Kaplan RS (2014) Improving value with TDABC. Health Finance Manage 68:76–83
8. Keel G, Savage C, Rafiq M, Mazzocato P (2017) Time-driven activity-based costing in health
care: A systematic review of the literature. Health Policy 121(7):755–763
9. Laviana AA, Ilg AM, Veruttipong D (2016) Utilizing time-driven activity-based costing to
understand the short- and long-term costs of treating localized, low-risk prostate cancer. Cancer
122:447–455
10. Martín JLR, Tobías A, Seonane T (2006) Revisiones sistemáticas en ciencias de la vida,
Fundación para la Investigación Sanitaria en Castilla-La Mancha (FISCAM, Toledo
11. McLaughlin N, Burke MA, Setlur NP (2014) Time-driven activity-based costing: a driver for
provider engagement in costing activities and redesign initiatives. Neurosurg Focus 37(5):E3
12. Rigby DK (2003) Management tools. Bain and Company Publishing, Boston
13. Ruiz de Arbulo, P., Fortuny, J.: Innovation in cost management: from ABC to TDABC.
Dirección y Organización 43, 16–26 (2011)
14. Van Eck NJ, Waltman L (2010) Software survey: VOSviewer, a computer program for
bibliometric mapping. Scientometrics 84:523–538
Chapter 7
Using Data Mining to Analyze
Occupational Accidents
in the Construction and Manufacturing
Sector
Clodoaldo Polo Barrera, María Martínez Rojas ,

and Juan Carlos Rubio Romero
Abstract In this paper, we focus on two of the sectors with the highest accident
rates, the construction sector and the manufacturing sector. The aim of this paper is
to analyze the common causes and the variables with the greatest influence on the
occurrence of occupational accidents in Spain in these two sectors. This analysis
will allow us to find both similarities and differences that may be of interest in
order to take more effective action to prevent accidents in the future. To address the
aforementioned objective, a database provided by the Ministry of Labour and Social
Economy has been used, which contains all accidents registered in the ministry’s
Delt@ system from 2009 to 2018. After exploring the database, several variables
have been analyzed using the decision tree and clustering data mining technique.
Keywords Accidents · Occupational health and safety · Information systems ·

Construction sector · Manufacturing sector · KNIME platform
7.1 Introduction
An “accident at work” is defined as a discrete occurrence in the course of work

resulting in physical or mental harm. The term “in the course of work” means “while
performing an occupational activity or during working time” [1]. Accidents at work
are a problem that affects all work sectors as they represent a high human and
economic cost for both companies and society, although not all of them have the
same severity [2]. Every year in Spain, more than half a million accidents at work
are common, and the figures do not seem to improve over the years. This indicates
that there is a rigidity surrounding the prevention of occupational risks, probably
C. Polo Barrera · M. Martínez Rojas (B) · J. C. Rubio Romero

Dpto. de Economía y Administración de Empresas, Escuela de Ingenierías Industriales,
Universidad de Málaga, C/ Dr. Ortiz Ramos S/N, 29071 Málaga, Spain
e-mail: mmrojas@uma.es

https://doi.org/10.1007/978-3-031-29382-5_7
58 C. Polo Barrera et al.
due, in part, to the fact that accident rate studies are not evolving in the way they
analyze the variables that affect them [3].
In theory, the safety and health of workers are guaranteed by law. This implies
the obligation of companies and competent governmental bodies to ensure compli-
ance with this guarantee, ensuring the integrity of workers in the exercise of their
profession [4, 5].
Among all the characteristics and circumstances surrounding accidents at work,
some, such as the sector, the size of the company, or the type of place where it occurs,
as well as the form of the accident and its severity, become more important factors
for its analysis.
In particular, the construction and manufacturing sectors have the highest inci-
dence rates of all occupational sectors [6, 7]. In Spain, construction is twice the
average of the other rates, followed by the manufacturing industry. Moreover, it is in
these two sectors that accidents tend to be most serious.
Therefore, the purpose of this study is based on the application of data mining
and data analysis techniques to compare the two sectors mentioned to better under-
stand the situations in which accidents occur. These techniques make it possible to
contemplate a large number of variables related to each accident and to find relational
patterns between them [8].
The conclusions of this study may be of great interest to decision-makers in order
to prevent accidents in a more effective and specific way by considering the effects
of variables that may go unnoticed in traditional studies.
7.2 Methodology
This study proposes the use of data mining techniques to obtain relevant information
on occupational accidents that occurred in Spain between 2009 and 2018 and that
has been provided by the Ministry of Labour and Social Economy. These accidents
are provided to the Ministry through the Delt@ work report. The database has a total
of 58 variables that are classified into several blocks: personal data of the worker,
company data, accident site data, accident data, etc.
The dataset is of a size that implies a considerable computational demand for
its analysis due to the number of records. The first step to obtain the data for the
two sectors that are the object of this work is to filter the entire dataset based on
the variables referring to: classification of business activities (CNAE) and national
classification of occupations (CNO). Special care must be taken in this filter as these
two classifications underwent a coding change in 2009 and 2011, respectively.
In Table 7.1, the annual population of affiliates is shown for all sectors together
and for the construction and industry sectors separately. The number of accidents
and incidence rates corresponds to all sectors together.
In order to work with such a large dataset, the data mining platform KNIME
[9] has been used. It is open-source software designed to facilitate the extraction of
knowledge from databases with a visual working environment that is very intuitive.
7 Using Data Mining to Analyze Occupational Accidents … 59
Table 7.1 Affiliates and accidents in the sectors

Year No. of affiliated Industrial sector Construction No. of accidents Incidence rate
workers affiliates sector affiliates
2009 18.181.742,70 2.513.829,25 1.913.269,30 696.577 38,31
2010 17.546.011,05 2.308.471,63 1.577.475,94 645.964 36,81
2011 17.361.838,50 2.241.291,00 1.425.258,70 581.150 33,47
2012 16.958.267,14 2.157.154,66 1.203.003,61 471.223 27,78
2013 16.179.438,04 2.028.194,54 1.010.287,18 468.030 28,92
2014 16.173.609,52 1.992.501,85 942.375,80 491.099 30,36
2015 16.575.312,25 2.015.573,15 974.358,95 529.248 31,92
2016 17.104.357,25 2.075.165,73 1.006.443,52 566.235 33,10
2017 17.674.174,52 2.137.038,28 1.053.521,76 515.082 29,14
2018 18.282.030,81 2.207.617,81 1.133.305,86 617.488 33,77
The platform has a large repository of nodes that function as black boxes with different
algorithms implemented. These nodes are included in the workspace forming the
workflow. This workflow is composed of nodes that perform various operations on
the data depending on the analysis needs.
The first step consists of loading the database obtained in.csv format to proceed
with its study. This is followed by a cleaning operation on the dataset for both sectors.
Next, a filtering operation is carried out in order to filter out the accidents corre-
sponding to the construction and manufacturing industry sectors. The database is then
reduced from almost 6 million (5,920,749) to a total of 1,744,252 cases belonging
to the aforementioned sectors, which are filtered again by the same type of node
to separate construction and industry, obtaining 704,681 cases and 1,039,571 cases,
respectively.
7.3 Individual Analysis of Selected Variables
As mentioned above, the dataset provided by the Ministry has a total of 58 variables.
Among all these variables, a set of variables has been selected by a panel of experts to
be the focus of the study. By analyzing the following variables, the aim is to answer
questions such as the 5 Ws (who, when, what, how, where) in order to generate
knowledge in this domain.
Relative proportion (%)

4
3
2
1
0
16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64
Age
Construction Manufacturing
Fig. 7.1 Relative comparison of the ages for each sector
7.3.1 Age
As can be seen in Fig. 7.1, although the results are similar in both sectors, there is
a difference between younger and older workers in each sector. Younger workers
are more accident prone in the manufacturing sector, while older workers are more
accident prone in the construction sector.
7.3.2 Temporality
From the graphs in Fig. 7.2, it can be seen that accidents are concentrated on certain
days and time slots. This information may be of interest to take into account these
“black spots” in order to act with more awareness.
7.3.3 Physical Activities and Type of Contact
The following variables provide us with information on what exactly the victim was
doing at the exact time of the accident and how he/she was injured.
More than half of all accidents in both sectors are caused by trivial activities,
which should not imply a major risk for the workers by themselves as can be seen
in Table 7.2.
The most common type of contact suffered by the workers, representing more
than the third part of all causes, is the same for both sectors as can be seen in Table
7.3. This cause of damage is a common consequence of various physical activities
presented before.
Fig. 7.2 Relative occurrences along the week a and the day b for each sector
Table 7.2 Relative occurrences of most common physical activities

Actions Relative occurrences (%)
Construction Manufacturing
Hand-hold, grasp, hold, put-on a horizontal plane 20.51 24.19
Walk, run, go up, go down, etc. 19.57 14.97
Working with non-motorized hand tools 14.56 12.27
Table 7.3 Relative

Type of contact Relative occurrences (%)
occurrences of most common
types of contact Construction Manufacturing
Physical overexertion 35.76 35.07
Hit on or against, as result 13.76 9.69
of a worker fall
Hitting or tripping over a 8.73 7.93
stationary object
Indefinite contract: Full time
Fixed term contract by work or service:

Full time
0 10 20 30 40 50
Manufacturing Construction
Fig. 7.3 Most accidented types of contracts for each sector
7.3.4 Type of Contract
As Fig. 7.3 shows, the ratio of the two more common types of contracts for each sector
is inversed. Fixed-term contracts are the most common contracts of the workers in
the most accidented sector.
7.4 Multivariable Analysis with Decision Tree Technique

and Clustering
The following analysis will evaluate the variables contained in the dataset to deter-
mine which ones are the most relevant to predict the accident severity. To do this,
two techniques are selected: the decision tree technique and clustering. The first one
is a supervised data mining method that can serve as an effective tool for multivariate
data analysis is used [10, 11]. On the other hand, clustering algorithms attempt to
relate cases to each other on the basis of their common characteristics in distinct
groups [12].
The decision tree creates a top-down branching structure, consisting of a root
node that splits into a series of branches [13]. This technique provides simplicity
and ease of interpretation of the results, allowing them to be evaluated from the
beginning to the end of the tree visually node by node. In addition, decision trees
are useful for our evaluated dataset, which contains both quantitative and qualitative
variables. Different decision tree modelling techniques were tested. The one with the
best success rate was the gain ratio technique without pruning for both sectors. With
this technique, a result with 88% accuracy was obtained.
Figure 7.4 shows a portion of the decision tree obtained, as an example, from
the evaluation of occupational accidents in the construction sector. Next, the results
obtained for both sectors are detailed.
Construction. As can be seen, there are three main variables selected by the
algorithm to predict when the accident will be fatal. The first one refers to deviation,
Fig. 7.4 Part of the obtained decision tree. (*Note: Mortal = fatal, Grave = serious, Muy grave =
very serious)
which is the backward movement of the action that caused the accident. In this sense,
79.2% of accidents caused by an unrecorded deviation will be fatal. The second
factor, age, appears to be even more important. When the age of the worker is over
38 years, the probability of the accident being fatal increases to 82.6%. Finally, the
third important variable was physical activity. If the worker was walking, running,
climbing, or descending, the probability of the accident being fatal increases to 88%.
Manufacturing. The results for the manufacturing sector are different. The mate-
rial agent associated with the accident becomes the most influential cause of fatal
accidents. The number of codings included in the prediction involves several mate-
rial agents, so it is not a very specific result. The last important factor in the predic-
tion of fatal accidents in the manufacturing sector is, as in construction, physical
activity. Again, whether the worker was walking, running, climbing, or descending,
the probability of the accident being fatal increases to 84% in the manufacturing
sector.
The second method used in this work, clustering, is detailed below. In order to
apply this method, due to the fact that the values of the variables differ greatly, it is
necessary to apply a normalization for all the data prior to their introduction to the
“K-means” node, so that their values are between 0 and 1. The number of centroids
to be established will be K = 3; observing the results obtained, the differentiated
groups have been obtained for this value. Once this is done, the flow is executed and
the result is obtained, visualized by means of a scatter plot thanks to the “scatter
plot” and “scatter matrix” node.
In order to evaluate the solutions obtained more quickly, the results are analyzed
using the “scatter matrix” node instead of checking scatter plots for each pair of vari-
ables compared to those presented above. In this way, clusters formed by comparing
several variables with each other at the same time can be checked. For the sake of
simplicity, the results for both sectors will be shown for part of the analysis (Fig. 7.5).
Fig. 7.5 Y: Date of the accident (V31)—X: Age of worker (V57) for the construction industry
Fig. 7.6 Y: Date of the accident (V31)—X: Age of worker (V57) for the manufacturing industry
Three distinct clusters are observed in this case. The yellow cluster is distributed
across all values for both variables; however, there are two distinct groups, the green
cluster, present in the ages corresponding to younger workers and the red cluster,
present only for older workers. This means that there is a distinction in accident
characteristics for workers whose dividing line is set at the standardized value 0.5
corresponding to approximately age 40. This boundary is also slightly shifted to the
right when looking only at serious accidents, indicating that these tend to occur to
older people. When the severity variable (V48) is evaluated against the other variables
presented for clustering, it does not give results like this in any of the cases. These
two variables influence each other.
As can be seen in Fig. 7.6, in the case of industry, there are not as differentiated
groups as in construction. In this sector, the three clusters established are almost in
the same proportion for all ages, whereas in construction, this was only the case for
the yellow cluster and there was a distinction of two distinct groups for younger and
older workers.
7.5 Conclusions
In general, the results obtained for both sectors are similar, although interesting differ-
ences were also found. Regarding age, greater differences were observed between the
workers injured according to their age in the assessments relating to the construction
sector, while in industry this variable was not so influential.
Regarding the temporality variables, the more relevant information has been
obtained by evaluating them in the short term. The evaluation of accident concen-
trations on the days of the week and the time periods of the day, rather than the
occurrences for the days of the month and the months of the year. With the informa-
tion obtained from the time periods and days of the week, apparently more valuable
information has been obtained in terms of prevention, making it possible to establish

in a clearer and more focused way possible preventive actions on the days of the week
and time periods that can be considered “black spots”. In this sense, both sectors have
presented fairly similar results, with the greatest difference being found in the results
obtained for night work, where the accident rate in construction far exceeds that of
the sector.
In relation to the most frequent and relevant causes of accidents for both sectors
assessed through multivariate analysis, it has been observed that human error is in
most cases the main cause of the accident. Work activities that may seem more trivial
tend to cause the highest number of accidents in both sectors.
From the techniques selected to analyze the dataset, decision trees appear to have
provided the most interesting and valid results. This may be due to the fact that the
dataset with which we have worked is mainly made up of categorical variables once
they are decoded; therefore, clustering techniques have not been as effective in this
study as they are designed to group numerical data in order to establish categories
and bring together those that present similarities between them.
References
1. Eurostat (2016) European statistics on accidents at work (ESAW)

2. Aznar M, Página C (2016) Trabajo final de máster ingeniería de los recursos naturales
3. Aguilera AA, López-Alonso M, Martínez-Rojas M, Martínez-Aires MD (2017) Review of the
state of knowledge of the BIM methodology applied to health and safety in construction. In:
Occupational safety and hygiene V. CRC Press, pp 459–464
4. Carrillo-Castrillo JA, Rubio-Romero JC, Onieva L (2013) Causation of severe and fatal
accidents in the manufacturing sector. Int J Occup Saf Ergon 19(3):423–434
5. López MAC, Ritzel DO, Fontaneda I, Alcantara OJG (2008) Construction industry accidents
in Spain. J Safety Res 39(5):497–507
6. INSST (2010) Informe anual de accidentes de trabajo en España
7. INSST (2019) Informe anual de accidentes de trabajo en España, pp 1–38
8. Martínez-Rojas M, Torrecilla-García JA, Rubio-Romero JC (2020) Prediction model of
construction accidents during the execution of structures using decision tree technique. In:
Occupational and environmental safety and health II, Springer International Publishing,
133–140
9. Build End to End Data Science (2020) Accessed 25 Feb 2021. [Online]. Available: www.
knime.com
10. Martínez-Rojas M, Soto-Hidalgo JM, Martínez-Aires MD, Rubio-Romero JC (2021) An anal-
ysis of occupational accidents involving national and international construction workers in
Spain using association rule technique. Int J Occup Saf Ergon 1–37
11. Sarkar S, Patel A, Madaan S, Maiti J (2016) Prediction of occupational accidents using decision
tree approach
12. Xu R, Wunsch D (2008) Clustering, vol 10, Wiley
13. Martínez-Rojas M, Soto-Hidalgo JM, Marín N, Vila MA (2018) Using classification techniques
for assigning work descriptions to task groups on the basis of construction vocabulary. Comput-
Aided Civil Infrastruct Eng 33(11):966–981
Chapter 8
Concept for Deployment Design
of Machine Learning Models
in Production
Henrik Heymann and Andrés Boza
Abstract The application of artificial intelligence (AI) and machine learning (ML)
in production environments offers huge potential for the manufacturing industry. In
order to create added value, ML models must be deployed into production which
means making models available in a specific environment where the results are
needed. As an initial task in deployment, called the deployment design, decision
owners need to define the desired ML system architecture. The goal of this paper is
to provide a structured methodology in form of a morphological box containing the
available options for the deployment design. Through the review of gray literature,
the five most relevant parameters are identified as prediction approach, consuming
application, model serving, learning method, and hosting solution. Possible values
for each parameter are introduced and necessary considerations for the selection
of an option are discussed. By means of a case study in the context of predictive
quality, which describes the use of a ML model to predict the product quality based
on production data, the developed concept is applied and validated.
Keywords Artificial intelligence · Machine learning · Deployment ·

Manufacturing · Predictive quality
8.1 Introduction
Artificial intelligence (AI) and machine learning (ML) are experiencing an increase
in relevance in the areas of research and development, economy, and education across
the globe [1]. Applied to the production industry, ML enables the optimization of
products and processes in a data-driven manner [2]. For manufacturing companies,
H. Heymann (B) · A. Boza

Centro de Investigación Gestión E Ingeniería de La Producción (CIGIP), Universitat Politècnica
de València, Valencia, Spain
e-mail: henhey@etsii.upv.es
A. Boza
e-mail: aboza@cigip.upv.es

https://doi.org/10.1007/978-3-031-29382-5_8
68 H. Heymann and A. Boza
which plan on making use of ML models in production, a true benefit is only generated
by making predictions available to the appropriate users in production and using these
predictions for decision-making and action. On a technical level, the architecture of
the overall ML system needs to be designed as a basis for deploying any model.
Within this paper, the focus is set on the technical design of systems, where ML
models are deployed. Organizational challenges during the deployment [3] are not
addressed at this point. The achieved results represent a part of the first author’s
master thesis covering the deployment comprehensively.
8.2 Objective and Method
This paper aims to provide a concept for designing the ML model deployment in
manufacturing environments. It shall serve decision owners as a guideline during
the strategical and high-level selection process of the most adequate ML system
architecture under consideration of the company’s specific needs and restrictions.
In dynamic fields of investigation such as software engineering and ML, the
academic literature only gives an incomplete view on the topic. According to Garousi
et al. [4], publications by practitioners on specialized and acknowledged online plat-
forms represent valuable sources of information. Through the review of so-called
gray literature, the crucial parameters for the design step of the deployment are
identified.
In order to structure the identified parameters including the possible values each
parameter can assume, a morphological box as introduced by Zwicky and Wilson
[5] in 1967 comes to application. This technique allows to break down complex
problems into attributes and subsequently create new, unseen solutions.
8.3 ML Deployment
In 2000, Chapman et al. [6] introduced CRISP-DM, a step-by-step data mining

guide. Till this day, it serves as the standard process for managing the life cycle of
ML projects. The methodology is composed of the steps business understanding, data
understanding, data preparation, modeling, evaluation, and deployment. Deployment
describes the integration of the ML model into an organization’s decision-making
processes. Applied to the production context, the deployment is understood as making
a ML model accessible for the end user of the application [7].
The design of the deployment builds on previous phases, e.g., on the selection of
an algorithm in the modeling phase. Shalev-Shwartz and Ben-David [8] distinguish
between online and batch learning. An online learning model is continuously updated
with each new data point, whereas in the case of batch learning, the model is updated
using a whole set of new data at once. Kervizic [9] identifies one-off, batch and online
training as ways to train models once deployed into production.
8 Concept for Deployment Design of Machine Learning Models … 69
Similar to training, predictions can be made by batch or in real time [9]. Batch
predictions, also called offline predictions, are executed at a particular point in time
and have a forecast character as they do not consider real-time input [10]. In contrast,
real-time (or online) predictions are calculated at the exact required moment. The
predictions are triggered either by a user request or by the arrival of new data [9, 11].
Displaying the predictions of a ML model requires the distinction between web
apps and native apps [12]. A web app is an application that is accessible via network by
any kind of connected device without being downloaded onto the device. And native
apps are developed and installed on a particular device and enable local computation.
Different approaches for model serving are proposed by relevant authors. One
common way for the deployment is to embed the model in the main application [9,
10, 13, 14]. Alternatively, a model can be deployed as a separate service. In this case,
the model is either served through a web service or in a streaming manner [9–11, 15].
Regarding the hosting of the ML system, different cloud service levels are distin-
guished [16]. On-premises solutions are managed completely within the organization
with no external cloud provider involved. When opting for a cloud option, a provider
can supply an instant computing infrastructure known as infrastructure-as-a-service
(IaaS), a complete development and deployment environment in the cloud called
platform-as-a-service (PaaS), or a ready-to-use software solution which is referred
to as software-as-a-service (SaaS).
8.4 Results
Organizing the findings of the review in a morphological box allows to compress and
structure visually the huge and disorganized variety of deployment options. In doing
so, the terminology is harmonized as different authors use different denominations
for similar principles.
Table 8.1 shows the identified parameters as well as the corresponding technical
question each parameter aims to find an answer for. The final morphological box
with all relevant parameters for the deployment design is depicted in Fig. 8.1. By
selecting one option for each parameter, the design requirements for ML system
architecture are determined. Subsequently, all parameters and the available solutions
are explained focusing on the applicability in the context of production.
Table 8.1 Parameters and

Parameter Technical question
corresponding technical
questions Prediction approach How are predictions made?
Consuming application How are predictions consumed?
Model serving How are models served?
Learning method How are models updated?
Hosting solution How is the ML system hosted?
Parameters Options
Prediction Approach By batch In real time
Consuming Application Web app Native app
Model Serving Embedded Separate
Learning Method Offline Online
Hosting Solution On-premises Cloud
Fig. 8.1 Morphological box for deployment design
8.4.1 Prediction Approach
The approach in which the ML system performs the predictions determines its design
as batch predictions configure a different ML system than real-time predictions.
Batch predictions allow a distribution of the computational load over time [10]. The
predictions are calculated from a batch of data at the moment that is considered
most appropriate with no real-time input possible [9]. However, real-time predic-
tions, which are triggered by the arrival of new data in the system, require a higher
capacity of the real-time computing system. In addition, system monitoring and
debugging are more complex [11]. The arrival speed of the data from the produc-
tion system, compared to the speed of processing them by the ML system in real
time, determines the viability of its implementation, because slow algorithms are not
able to make predictions in real time. A batch system generates less complexity and
needs less maintenance effort but requires periodic review to confirm the validity of
its predictions over time.
8.4.2 Consuming Application
The design of the ML system is also conditioned by the way in which the end user
interacts with said system. Native apps can behave similarly to web apps. However,
the respective configurations of the system architecture are different. Native apps
require installation on each device, allow heavy use of device hardware, and can run
without a network connection. These applications have the disadvantage that they
require a significant development effort and that they are limited to the computing
capabilities of the device on which they are installed. Web applications have the
advantage of being accessible from multiple devices on the browser, which amplifies
the number of places from which they can be accessed. The disadvantage of these
applications is that they cannot access the built-in features of the device as they
are developed for multiple platforms. Thus, the decision will be conditioned by the
specific production context to make greater or lesser use of each type of application
[12, 17].
8.4.3 Model Serving
ML systems for prediction-making require first the establishment of the prediction

model through learning and then making use of the said model in a productive
environment. Thus, the construction of the model and its use are included in a single
application when the ML model is embedded in the application that makes use of it
(consuming application) [13, 14]. However, when the model is built separately from
the consuming application, the connection of the two is necessary so that the model
can be used by the consuming application. This connection is made through shared
databases or through REST API request-response services [15, 18]. In other cases, the
model is implemented separately but delivers a data stream to which the consuming
application subscribes [19]. This can increase the waiting time. However, it is easier
to deploy additional ML models in production if the models are served separately
and the development and operation of the services are decoupled. The latency and
scalability of the service required are decisive in deciding on an integrated or separate
model serving [20]. Although the implementation of additional ML is easier with
non-embedded models, the transmission services to the consuming application can
be complex to configure and are influenced by the need for real-time predictions.
8.4.4 Learning Method
A fundamental element in the ML system is the training process. Thus, the learning
method also conditions the design of the system. Online learning allows collecting
new data to feed the model and thus improve the prediction in real time. Thereby, these
online learning models are updated as new data becomes available. However, they are
more complex to manage because they require constant monitoring. When learning
is done offline, the training process can be treated separately from the prediction
process, making them less complex ML systems. The selection of offline or online
learning should be considered in the early phases of the design of the ML system
[9, 21].
8.4.5 Hosting Solution
Lastly, the facility where the system is hosted also determines its design, and it is
necessary to decide about an on-premises solution or a cloud solution. On-premises
software is installed on the computers of the organization. Cloud solutions, on the

other hand, are hosted on external computer systems of a cloud provider company.
Choosing on-premises hosting requires properly sized own resources and trained
personnel for operation on the company’s servers and networks. The solution hosted
in the cloud implies less internal effort for the architecture of the system and, in
general, lower costs, but it generates dependency on the external provider, the need
to act according to the mechanisms of the external provider, and the uncertainty
regarding data privacy [16].
8.5 Case Study
In form of a case study, the methodology is applied in the context of predictive quality.
Deploying a ML model in order to predict the product quality in a production process
represents a common use case in the manufacturing industry, especially for high-tech
products with strict quality standards. Consultation with industry experts shows that
similar requirements toward the deployment can be found across companies. On the
basis of the generalized requirements from the experience in practice, the best fitting
option for each parameter of the morphological box is selected.
Prediction approach: In real time. Real-time capability is required as the predic-
tions for a produced item are requested as soon as the last process step is
finished.
Consuming application: Web app. Employees in the quality department require
the predictions on their devices used in quality control. Other departments are also
interested in the data and need to have access to the information. Therefore, installing
a native app on every device is not worthwhile.
Model serving: Separate. Serving the model separately from the web app allows
scalability and independence from the consumer. Moreover, wrapping the model in
a web service and delivering the results via REST API when requested represents an
adequate level of complexity for the given requirements.
Learning approach: Offline. In order to build a performing model for predictive
quality, the production process has to be in a mature stage with stable behavior.
Thus, updating the model with high frequency is not necessary and offline learning
algorithms come to application.
Hosting solution: On-premises. Due to the sensitivity of production data, an on-
premises solution is to be strived for. Regarding the representative circumstances
in this case study, sufficient resources are available to manage the system and its
complexity internally.
For validating purposes, the results are compared to existing architectures of
realized deployments and discussed with industry experts. Examples from practice
show that the identified ML system architecture is a common deployment pattern for
medium-sized manufacturing companies which do not have a high level of expertise
and maturity in ML operations as the deployment of complex ML systems do not
belong to their core competencies. The resulting architecture is not only common in
practice but also is described by Samiullah [19] as the best trade-off for many use
cases in terms of performance versus complexity.
8.6 Conclusions
In this paper, a concept for the deployment design of ML models in production is

presented which serves as a support system to define the most suitable system archi-
tecture for a given set of requirements. As the result of a gray literature review, rele-
vant decisions for the deployment concern the prediction approach, the consuming
application, the model serving, the learning method, and the hosting solution. For
each aspect, the available options are provided in form of a morphological box and
relevant considerations for the selection of an option are discussed. With the aid of
a case study from the field of predictive quality, the methodology is applied and
validated.
As future lines of research, it is to be investigated how further parameters, which
are currently not covered, can be included in the concept. Moreover, implications of
increasing the scope of the methodology to companies outside of the manufacturing
industry can be analyzed.
Acknowledgements This research has been funded by the Fondo Europeo de Desarrollo Regional
(FEDER)/Ministerio de Ciencia e Innovación (MCI)—Agencia Estatal de Investigación (AEI) of
Spain, in the framework of the project entitled “Integración de la Toma de Decisiones de los Niveles
Táctico-Operativo para la Mejora de la Eficiencia del Sistema de Productivo en Entornos Industria
4.0 (NIOTOME)” (Ref. RTI2018-102020-B-I00).
References
1. Perrault R, Shoham Y, Brynjolfsson E, Clark J, Etchemendy J, Grosz B, Lyons T, Manyika J

(2019) The AI index 2019 annual report, https://hai.stanford.edu/sites/default/files/ai_index_
2019_report.pdf
2. Schmitt RH, Kurzhals R, Ellerich M, Nilgen G, Schlegel P, Dietrich E, Krauß J, Latz A, Gregori
J, Miller N (2020) Predictive quality—data analytics in produzierenden Unternehmen. Internet
of Production—Turning Data into Value, Fraunhofer-Institut für Produktionstechnologie IPT
and Werkzeugmaschinenlabor WZL der RWTH Aachen, pp 226–253
3. Baier L, Jöhren F, Seebacher S (2019) Challenges in the deployment and operation of
machine learning in practice. In: Twenty-seventh European conference on information systems
(ECIS2019), Stockholm-Uppsala, Sweden
4. Garousi V, Felderer M, Mäntylä MV (2019) Guidelines for including grey literature and
conducting multivocal literature reviews in software engineering. Inf Softw Technol 106:101–
121
5. Zwicky F, Wilson AG (1967) New methods of thought and procedure. In: Contributions to the
symposium on methodologies. Pasadena. Springer, New York
6. Chapman P, Clinton J, Kerber R, Khabaza T, Reinartz T, Shearer CR, Wirth R (2000) CRISP-
DM 1.0—Step-by-step data mining guide. Copenhagen
7. Krauß J, Pacheco B, Zang H, Schmitt RH (2020) Automated machine learning for predictive
quality in production. Procedia CIRP 93:443–448
8. Shalev-Shwartz S, Ben-David S (2019) Understanding machine learning: from theory to
algorithms (12th printing). Cambridge University Press, 24
9. Kervizic J (2019) Overview of the different approaches to putting machine learning (ML)
models in production. https://medium.com/analytics-and-data/overview-of-the-different-app
roaches-to-putting-machinelearning-ml-models-in-production-c699b34abf86. Last accessed
06 Mar 2021
10. Akyildiz B (2020) How to serve models. https://bugra.github.io/posts/2020/5/25/how-to-serve-
model/. Last accessed 06 Mar 2021
11. Dawson R (2020) Towards data science—navigating ML deployment. https://www.seldon.io/
navigating-ml-deployment/. Last accessed 06 Mar 2021
12. Bignu A (2019) Web apps vs native apps: what is the best choice for a data
scientist? https://medium.datadriveninvestor.com/web-apps-vs-native-apps-what-is-the-best-
choice-for-a-data-scientist-3d31169d2335. Last accessed 06 Mar 2021
13. Grek T (2018) There are two very different ways to deploy ML models, here’s both. https://
towardsdatascience.com/there-are-two-very-different-ways-to-deploy-ml-models-heres-both-
ce2e97c7b9b1. Last accessed 06 Mar 2021
14. Sato N, Wider A, Windheuser C (2019) Delivery for machine learning. https://martinfowler.
com/articles/cd4ml.html. Last accessed 06 Mar 2021
15. Pinhasi A (2020) Deploying Machine learning models to production—inference service
architecture patterns. https://medium.com/data-for-ai/deploying-machine-learning-models-to-
production-inference-service-architecture-patterns-bc8051f70080. Last accessed 06 Mar 2021
16. Watts S, Raza M (2019) SaaS vs PaaS vs IaaS: what’s the difference and how
to choose. https://www.bmc.com/blogs/saas-vs-paas-vs-iaas-whats-the-difference-and-how-
to-choose/. Last accessed 06 Mar 2021
17. Konstantinidis F (2020) Why and how to run machine learning algorithms on edge
devices. https://www.therobotreport.com/why-and-how-to-run-machine-learning-algorithms-
on-edge-devices/. Last accessed 07 Mar 2022
18. Patruno L (2020) The ultimate guide to deploying machine learning models. https://mlinprodu
ction.com/deploying-machine-learning-models/. Last accessed 07 Mar 2022
19. Samiullah C (2019) How to deploy machine learning models. https://christophergs.com/mac
hine%20learning/2019/03/17/how-to-deploy-machine-learning-models/#monitoring. Last
accessed 06 Mar 2021
20. Google (2019) Minimizing predictive serving latency in machine learning. https://cloud.google.
com/architecture/minimizing-predictive-serving-latency-in-machine-learning. Last accessed
02 May 2022
21. Hunt X (2017) Online learning: machine learning’s secret for big data. https://blogs.sas.
com/content/subconsciousmusings/2017/10/17/online-learning-machine-learnings-secret-
big-data/. Last accessed 07 Mar 2022
Part III
Operations Research, Modelling
and Simulation
Chapter 9
An MILP Model
for the Lot-Sizing/Scheduling
of Automotive Plastic Components
with Raw Materials and Packaging
Availability
E. Guzmán , B. Andres , and R. Poler
Abstract This paper examines the lot-sizing /scheduling problem for plastic auto-
motive components manufacturing. The scenario in which the problem is tackled
refers to a second-tier supplier in the automotive supply chain. Here, the studied
second-tier supplier is characterized by transforming plastic granules in injection
machines using specific moulds that produce components or finished products. Each
mould can be set up on distinct machines to inject one same automobile component,
or even two different components or more in the same mould. The same mould is
assembled on different injection machines and can have distinct production rates
subject to the machine on which it is set up. Our research work puts forward a mixed
integer linear programming (MILP) model to minimize setup, the inventory of raw
materials and plastic components, stockout, backorder costs, and machine-mould
assignation costs. We demonstrate the usability of this model with randomly gener-
ated instances. The results of the experiments show that our MILP converges toward
optimal solutions in large instances by reaching efficient solutions in reference to
both quality and execution times. The novelty of this model lies in it considering
the arrival of materials as raw material for the injection of parts into moulds, the use
of raw materials and the availability of containers for packaging finished products.
Moulds can also be set up only during specific time periods in accordance with the
quantity of available labor during each time period.
Keywords Scheduling · Lot-sizing · Raw materials · Packaging · Automotive

industry · Mixed integer linear programming
E. Guzmán (B) · B. Andres · R. Poler

Research Centre on Production Management and Engineering (CIGIP), Universitat Politècnica de
València (UPV), Calle Alarcón, 03801 Alcoy, Spain
e-mail: eguzman@cigip.upv.es
B. Andres
e-mail: bandres@cigip.upv.es
R. Poler
e-mail: rpoler@cigip.upv.es

https://doi.org/10.1007/978-3-031-29382-5_9
78 E. Guzmán et al.
9.1 Introduction
Production and scheduling planning are central functions in manufacturing industries

whose relevance is increasingly important due to the complexity of the operations
required to manufacture final products from raw materials and the growing attention
paid to supply chain management. The production planning and scheduling problem
represent an important area of production planning and operations research [1].
Production decisions for a manufacturing environment are concerned about estab-
lishing the most efficient utilization of available resources to produce items, while
also meeting customer requirements. The lot-sizing /scheduling problem frequently
appears in manufacturing systems with complex configurations and finite capaci-
ties. In both practice and theory, lot-sizing/scheduling decisions are often made in
parallel at the production planning and scheduling levels. The objective at the plan-
ning level is to draw a production plan, i.e., determining the production quantities
(corresponding to the batch sizes processed in workshops) for each horizon period to
meet demands and to minimize different costs (production, maintenance, and setup
costs). These batches are sequenced in production assets at the scheduling level [2].
A substantial number of papers have dealt with lot-sizing /scheduling, the majority
of which are mathematical models for this problem, where the objective function
seeks to minimize production costs. Our study centers on modeling a real industrial
case to solve the lot-sizing/scheduling problem that is subject to internal/external
materials requirement planning (MRP) restrictions. The problem is linked with an
automotive plastic component producer that acts as a second-tier supplier in the
automotive supply chain. The herein studied second-tier supplier is characterized for
its specific moulds for producing components or finished products. This problem is
particularly characteristic of the automotive industry because:
(i) The aim of having to produce the plastic components of a specific car model
is to supply them during most of the model’s lifetime, e.g., five years.
(ii) The increasing costs of plastic raw materials caused by the pandemic crisis have
led second-tier suppliers to purchase larger amounts of plastic pellets (raw
materials), which always entails contemplating warehouse space limitations
and discount prices.
(iii) Specific reusable containers are purchased by the first-tier supplier to receive
the second-tier supplier’s components. As reusable containers are expensive,
there is only a limited number of them. The number of reusable containers
can be slightly adjusted to the agreed demand for the supply period, which
normally coincides with a car model’s lifetime. When reusable containers are
not available for second-tier suppliers, injected parts have to be stored in card-
board containers until reusable containers arrive. Then the components stored
in cardboard containers must be moved to the reusable ones, which incurs an
extra handling cost [3].
The case study is framed within the European project Zero-Defect Manufacturing
Platform (ZDMP) in the Preparation Stage: start-up optimization, in which tasks
9 An MILP Model for the Lot-Sizing/Scheduling of Automotive Plastic … 79
like the optimization of equipment, materials, energy and energy efficiency are
addressed [4].
This research work proposes a mixed integer linear programming (MILP) model
for the lot-sizing /scheduling problem to manufacture plastic automotive compo-
nents that contemplates the use, availability, and arrival of materials, including raw
materials, to inject parts into moulds, as well as containers for packing the finished
components to be delivered to the first-tier supplier. It aims to minimize setups, the
inventory of raw materials and plastic components, stockouts, backorder costs, and
machine-mould assignation costs.
This work is set out as so. Section 9.2 starts by reviewing the related litera-
ture. Section 9.3 describes the studied problem and the mathematical formulation.
Section 9.4 discusses the computational experiments and the results. Section 9.5
offers some concluding remarks and future research lines.
9.2 Literature Review
Substantial research has been conducted on various aspects of lot-sizing /scheduling

problems in distinct industries [5] like those presented by Almada-Lobo et al. [6], who
studied two linear mixed integer programming formulations for a multi-item capaci-
tated lot-sizing problem with sequence-dependent setup costs and times for the glass
container industry. de Armas and Laguna [7] developed an MILP formulation for
a capacitated lot-sizing/scheduling problem toward pipe insulation manufacturing,
which included multiple- and single-level items processed on parallel machines
according to a planning horizon.
The literature also describes several articles that have addressed injection
moulding lot-sizing/scheduling problems. They include Nagarur et al. [8], who
present a goal programming model for the injection moulding of PVC pipe fittings.
This model aimed to minimize total production costs, inventory and shortages.
Ghosh Dastidar and Nagi [9] address the production scheduling problem in an injec-
tion moulding facility that produces healthcare products. Their work presents an
MILP model that schedules parallel work centers with changeover costs, sequence-
dependent setup times and multiple capacitated resources in a single-stage case.
Martínez et al. [10] describe an MILP model that addresses the lot-sizing/scheduling
problem for a Brazilian moulded pulp packaging plant. With their model, they seek to
establish which moulding patterns can be utilized, for how long, and how they can be
sequenced. Ríos-Solís et al. [11] present an MILP model and a heuristic method based
on a mathematical programming method for a lot-sizing/scheduling problem. The
aim is to determine the maximum profit made with assembled products during many
periods. This model deals with plastic injection moulding as part of manufacturing,
pursues precise production assignment from parts to moulds and from moulds to
machines, seeks to maximize the total value of manufactured products, and deduces
maintenance costs. Mula et al. [12] propose an MILP model for solving the capaci-
tated lot-sizing problem with sequence-dependent setups and parallel machines for
injection moulding in the automotive industry. Moulding involves injecting two

different parts or products into the same mould. Both parts need the same sequence
order and available capacity at the same time.
Andres et al. [13] set up an MILP model for the production/lot-sizing /scheduling
problem on parallel flexible injection moulding machines with common setup oper-
ators. To produce automotive plastic components, the model allocates moulds to
machines during a given time period and calculates the number of components to
be manufactured. A sequence-dependent setup time is followed for this purpose.
The model also bears in mind the common setup operators who change moulds on
machines.
As far as we know, mathematical models do not contemplate the availability of
materials/packaging for the delivery of components from the second-tier to the first-
tier supplier. The herein proposed model extends that by Andres et al. [13] because
the proposed MILP model takes into account the arrival, use and availability of not
only the raw materials for injecting parts in moulds but also the packaging for the
finished components. Moreover, moulds can be changed only during time windows
and depend on the amount of labor available during each period. The proposed
model contemplates similar assumptions to those reported by Andres et al. [13],
which envisages that moulds can be set up on different injection machines, and
MILP output supplies mould-machine assignments.
9.3 Problem Description and Formulation
The proposed MILP the lot-sizing /scheduling of automotive plastic components

with common setup labor and limited raw and packaging materials availability to
transport components from the second- to the first-tier supplier is incorporated in a
source and make a scheme which is classified according to SCOR views [14]. The
Plan Source (S) deals with the calculation of the raw materials, items or components
to be supplied during each time period and on a specific planning horizon so that the
Plan Make (M) can be fulfilled with no backorder penalizations. For the Source and
Make Plans (SM), the production plan (M) is computed according to the production
requirements identified in the procurement plan (S) [15, 16].
The SM planning scheme is followed by the second-tier supplier to deliver auto-
motive plastic components to assemble them at the first-tier supplier and original
equipment manufacturers (OEMs). The SM plan is generated to identify the period
and quantity of: (i) the materials and components to be purchased from suppliers
(plan S); (ii) the components to be manufactured in the company to assemble and
produce the final product (plan M); see Fig. 9.1.
The firm under study has several moulds that are set up on different injection
machines to produce the range of plastic components to be delivered to the first-tier
supplier and finally to the various OEMs forming part of the distinct automotive
supply chains characterized by selling to several car brands. The MILP model under
discussion is based on the following assumptions:
Fig. 9.1 Outline of the source and make plans
• Plastic components are injected into moulds, which are assembled on parallel
flexible injection machines. Injection machines inject plastic granules which are
transformed into automotive semifinished products.
• The second-tier supplier has specific moulds for producing each automotive plastic
component. When two moulds are available to produce the same plastic compo-
nent, these moulds can come into play at different processing times because of
their technical characteristics.
• Each mould can produce one part, or two parts or more, in the same mould
• Each mould can be placed on distinct injection machines to manufacture the same
automotive component. However, the same mould set up on different machines
has several production rates depending on the machine it is assembled on.
• The company works three shifts per day five days a week and works overtime
shifts on day 6 of the week if production does not end during normal working
hours. On day 6, no setup operators are available.
• One of the company’s study requirements is that, after installing the mould on a
machine, the mould must remain at least 24 h to not saturate operators’ work and
involve too many setups because the installation time is estimated to go from 1
to 3 h, and it obviously has an associated setup cost. If a longer production time
is necessary, the mould is set up for the required time periods without incurring
installation costs.
• When the production time lasts longer than 24 h, the mould remains assembled
for the necessary time periods with no incurred installation costs.
• Backorders are highly penalized in the automotive sector because they work with
just-in-time (JIT) models.
• The mould can be changed only during specific time windows. Mould changes are
counted to not exceed setup operators’ capacity. Table 9.1 describes the indices,
parameters and variables of this problem.
Next the formulation of the MILP model proposed for the lot-sizing /scheduling
of automotive plastic components with available raw materials and packaging
takes place. The objective function minimizes the setup and labor costs, machine-
mould assignation, raw materials/packaging and plastic components inventory costs,
backorder costs, and costs for coverage stockouts.

Min z = cs j · S Ai l j t + scl i j l · S Ai l j t
i l j t i j t l

+ r oi j · r ci j · S A i l j t + ci k
i l j t k t

· I N V kt + ci m r · I N V m r t
r t

+ cst k · ST kt + cbk · B kt (9.1)
k t k t
Subject to:
Sequence constraints

Siljt · roij ≤ 1 ∀i, t (9.2)
j l

Siljt · roij ≤ a j ∀ j, t (9.3)
i l
Constraint (9.2) establishes that 1 or 0 moulds j are set up by setup operator l

to be produced during each time period t. Constraint (9.3) guarantees that the total
number of available moulds j can only be set up for production as a maximum by
setup operator l during each time period t.
Production and capacity constraints

X kt = pjk · roij · Siljk ∀k, t (9.4)
i j l
Table 9.1 Notation

Index
i Index of machines i ∈{1, …, I}
j Index of moulds j ∈{1, …, J}
k Index of parts k ∈{1, …, K}
l Index of setup operators l ∈{1, …, L}
r Index of materials (raw materials/packaging) r ∈{1, …, R}
t Index of time periods t ∈{1, …, T}
Model parameters
aj Total amount of moulds j available for production
cakr Use of material r required to produce each unit of part k
cbk Backorder cost of part k
cik Inventory cost of part k
cimk Inventory cost of materials r
covkt Stock coverage defined as number of time periods for the stock minimum
coverage of part k during time period t
csj Setup cost of preparing mould j
cst k Coverage stockout cost of part k
d kt Demand of part k during time period t
INVk0 Initial inventory of part k
INVr 0 Initial inventory of material r
INVMAXk Maximum inventory units for part k during time period t
INVMINk Minimum inventory units for part k during time period t
INVMAXmatr Maximum inventory units for material r during time period t
nct Number of mould changes permitted during time period t
pjk Number of parts k produced when mould j is set up
roij 1 if mould j can be set up on machine i
and 0 otherwise
rcij Assignation cost of mould j on machine i
rprt Quantity received of material r during each period t
slaijl Amount of setup operators l required to setup mould j on machine i
scl ijl Cost of setup operator l to setup mould j on machine i
slst Number of workers l available during each period t
Decision variables
Bkt Backorder of part k during time period t
INVkt Inventory level of part k at the end of time period t
SAiljt 1 if mould j is set up on machine i by setup operator l during time period t, and
is not set up on machine i during time period t-1;
0 if mould j is set up by setup operator l on machine i during time period t-1
(continued)

S iljt 1 if mould j is set up by setup operator l on machine i during time period t;
0 otherwise
STkt Coverage stockout of part k during time period t
Camrt Material (raw material, packaging) r consumed during period t
INVmrt Inventory of material r during period t
X kt Amount of part k to be produced during time period t

Camrt = cakr · X kt ∀k, r, t (9.5)
k
Constraint (9.4) determines the number of parts k to be manufactured during time

period t, and ensures that a specific mould j can be set up on machine i during time
period t while producing product k. Constraint (9.5) establishes the amount of raw
material and packaging r used during time period t.
Setup constraints
SAiljt = Siljt ∀i, l, j, t = 1 (9.6)
SAiljt ≥ Siljt − Siljt−1 ∀i, l, j, t > 1
SAiljt ≤ 1 ∀i, l, j, t (9.7)

SAiljt ≤ nct ∀l, t, (9.8)
i j
Constraint (9.6) records the first setup of mould j carried out by operator l on
machine i to identify the first time that mould j is set up during time period t on
machine i. Constraint (9.7) ensures that SAiljt takes binary values. Constraint (9.8)
limits the number of mould j changes allowed during time period t, which are set up
by operator l on machine i.
Labor constraint

SAiljt · slaijl ≤ slst ∀l, t (9.9)
i j
Constraint (9.9) limits the number of mould changes permitted during time period t
to the number of available workers l by bearing in mind the number of setup operators
l needed to set up mould j on machine i.
Inventory balance equations
INVkt = INVk0 + X kt − dkt + Bkt ∀k, t = 1 (9.10a)
INVkt = INVkt−1 + X kt − dkt + Bkt − Bkt−1 ∀k, t > 1 (9.10b)

INVmrt = INVr 0 + rprt − cakr · X kt ∀k, r, t = 1 (9.11a)
k

INVmrt = INVmkt−1 + rprt − cakr · X kt ∀k, r, t > 1 (9.11b)
k
Inventory balance Eqs. (9.10a) and (9.10b) limit the appropriate values for inven-
tories, the quantities to produce, and the backorders for each time period t = 1 and
t > 1, respectively. Constraints (9.11a) and (9.11b) ensure the uninterrupted supply
of raw materials and packaging r for time periods t = 1 and t > 1.
Stock coverage constraint
INVkt ≥ INVMINk ∀k, t (9.12)
INVkt ≤ INVMAXk ∀k, t (9.13)

INVmrt ≥ cakr · X kt ∀k, r, t (9.14)
k
INVmrt ≤ INVMAXr ∀r, t (9.15)
STkt ≥ covkt − INVkt ∀k, t (9.16)
Constraints (9.12) and (9.13) restrict the inventory levels for each part k during
time period t. Constraint (9.14) guarantees that the materials inventory corresponds
to the quantity of material that need to be produced during the same period by
considering a lead time of 0 and the batching technique is lot-for-lot. Constraint
(9.15) limits the inventory levels for raw materials and packaging r during time
period t. Constraint (9.16) is for the stock coverage of parts.
Bound and nature variables
SAiljt , Siljt ∈ {0, 1} ∀i, l, j, t (9.17)

X kt , INVkt , Bkt , STkt , ∈ Z ∀k, t (9.18)
Camrt , INVm rt ∈ Z ∀r, t (9.19)
Constraint (9.17) determines the binary nature of both variables’ setup S iljt and
setup amount SAiljt Constraints (9.18) and (9.19) determine the represented variables’
integer nature.
9.4 Computational Experiments
An MILP model for the lot-sizing /scheduling of automotive plastic components,

along with the availability of raw materials /packaging, was developed in Python 3.9.2
with Pyomo [17], employed as an extensible python-based open-source optimization
modeling language for linear programming, and with Gurobi 9.0. All the experiments
were run on a PC equipped with an Intel(R) Core (TM) i7- 1165G7 CPU @ 2.80 GHz,
16 GB of RAM with the Windows 10 Pro operating system.
9.4.1 Generating Datasets
This section presents the experimental results. The conducted model’s performance
is depicted by 13 test problems. Data values are generated to reflect real automotive
component industry data (see Table 9.2). The datasets needed for the experiments
were built as in Andres et al. [13]. Data values are defined as shown below:
The algorithm developed to build the synthetic datasets is found at http://hdl.han
dle.net/10251/172395
9.4.2 Computational Results
This section offers details of the case study of a second-tier supplier in an automotive
supply chain. The results derived from the run time and the objective function value
for solving problems are tabulated in Table 9.3. A simplified view of the solution is
seen in Fig. 9.2 to provide details of the problem that the second-tier supplier faces.
The size of datasets, including the number of machines (I), moulds (J), parts
(K), material (R), setup labor (L), and periods (T ), appears in the second column
of Table 9.3. In most resolved instances (small—S, medium—M, large—L), the
model’s computational performance (CPU time) is efficient for all instances. The
solution for large instances provides optimal solutions in computational times under
20 s. The results obtained in the objective function do not include the backorder cost.
Table 9.2 Value generation for the data parameters

Parameter Value Parameter Value
aj 1 INVMINk Random (10, 100)
cakr 1 INVMAXmatr 99,999
cbk 99,999 INVMINmatr Random (100, 150)
cik, cimr U (0.1, 1) nct Random (1, 2)
covkt Random (10, 100) pjk Random (20, 50)
csj Random (50, 100) roij Random (0, 1)
cstk 1 rcij Random (1, 2)
dkt Random (10, 100) if T = first of the five rpkt Random (0, 50)
periods of the week, otherwise 0 if T =
period 6 and T = period 7 of the week
INVk0 Random (10, 150) slailj 1
INVr 0 Random (1000, 1500) sclilj U (2.5, 8.5)
INVMAXk Random (10,000, 50,000) slsl nct
Table 9.3 MILP model results

Data-set Problem size Objective Lower Upper GAP CPU
I J K R L T bound bound (%) (sec)
S1 2 4 6 3 1 3 505.84 505.84 505.84 0.00 0.03

S2 6 8 2 3 2 7 6503.13 6503.13 6503.13 0.00 0.29
S3 8 10 30 3 2 7 8673.92 8673.92 8673.92 0.00 0.26
S4 10 12 40 3 2 7 13,841.98 13,841.98 13,841.98 0.00 0.41
M1 12 14 60 3 2 14 40,097.58 39,696.60 40,097.58 0.01 1.64
M2 14 16 80 3 4 14 50,359.57 49,855.97 50,359.57 0.01 1.51
M3 16 18 100 3 4 14 59,831.37 59,233.05 59,831.37 0.01 2.08
M4 18 20 110 3 4 14 68,995.23 68,305.28 68,995.23 0.01 5.67
L1 20 24 140 3 4 14 80,587.91 80,587.91 80,587.91 0.00 2.04
L2 22 26 160 3 4 14 132,509.80 131,184.70 132,509.80 0.01 14.09
L3 25 28 180 3 6 21 165,415.31 163,761.16 165,415.31 0.01 11.94
L4 30 35 200 3 6 21 182,957.97 181,128.39 182,957.97 0.01 18.52
L5 40 45 300 3 6 21 271,586.75 268,870.88 271,586.75 0.01 15.92
Small instance S1 comprises two machines and four moulds, six parts, one oper-
ator, and three periods. Figure 9.2 depicts how moulds can produce one part or more.
In this case, the data is generated synthetically, mould 1 produces parts 4, 5 and
6, mould 3 produces part 3, and mould 4 generates parts 1 and 2. The obtained
results appear in Tables 9.4, 9.5 and 9.6. With regard to the results of the sequence
of the moulds on the machines, Fig. 9.2 illustrates that the operator puts mould 1 on
machine 1 and manufactures for three periods, once mould 1 is placed the operator
Fig. 9.2 Representation of the realistic lot-sizing /scheduling model with raw materials and
packaging availability
puts mould 3 on machine 2, and mould 4 is put on the same machine in period 2 (see
Table 9.5). Table 9.6 describes the consumption and inventory, where r = 1 corre-
sponds to the raw material (plastic granules) and r = 2 and r = 3 to the packaging
of the automotive semifinished products.
Table 9.4 Numerical results of instance S1: backorders, inventories, stockout, lot-sizing
k t Bkt INV kt ST kt X kt
1 1 0 20 0 0
1 2 0 30 0 20
1 3 0 40 0 20
2 1 10 10 10 0
2 2 10 10 10 50
2 3 10 10 10 50
3 1 0 10 10 50
3 2 10 10 10 0
3 3 20 10 10 0
4 1 0 10 5 50
4 2 0 10 5 50
4 3 0 10 5 50
5 1 0 10 25 50
5 2 0 10 25 50
5 3 0 10 25 50
6 1 0 10 0 50
6 2 0 10 0 50
6 3 0 10 0 50
Table 9.5 Numerical results of instance S1: scheduling

i l j t Siljt SAiljt
1 1 1 1 1 1
1 1 1 2 1 0
1 1 1 3 1 0
1 1 2 1 0 0
1 1 2 2 0 0
1 1 2 3 0 0
1 1 3 1 0 0
1 1 3 2 0 0
1 1 3 3 0 0
1 1 4 1 0 0
1 1 4 2 0 0
1 1 4 3 0 0
2 1 1 1 0 0
2 1 1 2 0 0
2 1 1 3 0 0
2 1 2 1 0 0
2 1 2 2 0 0
2 1 2 3 0 0
2 1 3 1 1 1
2 1 3 2 0 0
2 1 3 3 0 0
2 1 4 1 0 0
2 1 4 2 1 1
2 1 4 3 1 0
Table 9.6 Numerical results

r t Camrt INVmrt
of instance S1: use and
inventory of materials 1 1 200 850
1 2 220 680
1 3 220 460
2 1 200 800
2 2 220 630
2 3 220 410
3 1 200 800
3 2 220 580
3 3 220 410
9.5 Conclusion
This research work develops an MILP model to integrate lot-sizing /scheduling deci-
sions about automotive plastic components with raw materials /packaging avail-
ability to minimize setup and labor costs, components and raw materials inven-
tory costs, backorder costs, machine-mould assignations, and penalization costs
for coverage stockouts. Both moulds and parts are employed as central indices for
planning/scheduling on parallel machines. This work also contemplates the mould
changes time window, the several setup times according to the number of workers
assigned to mould change and mould-machine assignments. It also includes the
arrival of materials, use of raw materials and availability of packaging containers.
This paper validates MILP performance and proves computationally efficient for
different instance types, including large datasets that replicate the amount of data
employed in real automotive industries. In future studies, the model’s assumptions
can be extended by adopting other practical conditions, such as constraints for trans-
porting finished products, waiting times for containers for packing finished products
to be delivered and limited space to store finished products.
Funding This work was supported by the Conselleria de Educación, Investigación, Cultura y
Deporte (Generalitat Valenciana) for hiring predoctoral research staff with Grant (ACIF/2018/170)
and European Social Funds with the Grant Operational Program of FSE 2014–2020, the Valencian
Community (Spain). The research leading to these results obtained funding from the European
Union H2020 Program with grant agreement No. 825631 “Zero-Defect Manufacturing Platform”
(ZDMP).
References
1. Mohammadi M, Esmaelian M, Atighehchian A (2020) Design of mathematical models for

the integration of purchase and production lot-sizing and scheduling problems under demand
uncertainty. Appl Math Model 84:1–18. https://doi.org/10.1016/j.apm.2020.03.021
2. Wolosewicz C, Dauzère-Pérès S, Aggoune R (2015) A Lagrangian heuristic for an integrated
lot-sizing and fixed scheduling problem. Eur J Oper Res 244(1):3–12. https://doi.org/10.1016/
j.ejor.2015.01.034
3. Guzman E, Andres B, Poler R (2021) A MILP model for reusable containers management in
automotive plastic components supply chain. In: 22nd IFIP WG 5.5 working conference on
virtual enterprises, PRO-VE 2021, p 8, [Online]. Available: https://hal-emse.ccsd.cnrs.fr/emse-
03338406
4. Campbell S (2019) D2.1: inception and vision document. [Online]. Available: https://portal.
effra.eu/result/show/3722
5. Copil K, Wörbelauer M, Meyr H, Tempelmeier H (2017) Simultaneous lotsizing and scheduling
problems: a classification and review of models. OR Spectr 39(1):1–64. https://doi.org/10.1007/
s00291-015-0429-4
6. Almada-Lobo B, Klabjan D, Carravilla MA, Oliveira JF (2007) Single machine multi-product
capacitated lot sizing with sequence-dependent setups. Int J Prod Res 45(20):4873–4894.
https://doi.org/10.1080/00207540601094465
7. de Armas J, Laguna M (2019) Parallel machine, capacitated lot-sizing and scheduling for the
pipe-insulation industry. Int J Prod Res. https://doi.org/10.1080/00207543.2019.1600763
8. Nagarur N, Vrat P, Duongsuwan W (1997) Production planning and scheduling for injection
moulding of pipe fittings: a case study. Int J Prod Econ 53(2):157–170. https://doi.org/10.1016/
S0925-5273(97)00109-6
9. Dastidar SG, Nagi R (2005) Scheduling injection molding operations with multiple resource
constraints and sequence dependent setup times and costs. Comput Oper Res 32(11):2987–
3005. https://doi.org/10.1016/j.cor.2004.04.012
10. Martínez KYP, Toso EAV, Morabito R (2016) Production planning in the molded pulp
packaging industry. Comput Ind Eng 98:554–566. https://doi.org/10.1016/j.cie.2016.05.024
11. Ríos-Solís Y, Ibarra-Rojas OJ, Cabo M, Possani E (2020) A heuristic based on mathematical
programming for a lot-sizing and scheduling problem in mold-injection production. Eur J Oper
Res 284(3):861–873. https://doi.org/10.1016/j.ejor.2020.01.016
12. Mula J, Díaz-Madroñero M, Andres B, Poler R, Sanchis R (2021) A capacitated lot-sizing model
with sequence-dependent setups, parallel machines and bi-part injection moulding. Appl Math
Model 100:805–820. https://doi.org/10.1016/j.apm.2021.07.028
13. Andres B, Guzman E, Poler R (2021) A novel MILP model for the production, lot sizing, and
scheduling of automotive plastic components on parallel flexible injection machines with setup
common operators. Complexity 2021:16. https://doi.org/10.1155/2021/6667516
14. Supply Chain Council SCC, Supply chain operations reference model SCOR version 11.0.
2012
15. Orbegozo A, Andres B, Mula J, Lauras M, Monteiro C, Malheiro M (2016) An overview
of optimization models for integrated replenishment and producction planning decisions. In:
Building bridges between researchers and practitioners. Book of Abstracts of the International
Joint Conference CIO-ICIEOM-IISE-AIM (IJC2016), p 68
16. Andres B, Sanchis R, Poler R, Saari L (2017) A proposal of standardised data model for cloud
manufacturing collaborative networks. In: Collaboration in a Data-Rich World, pp 77–85
17. Hart WE, Laird C, Watson JP, Woodruff DL (2012) Pyomo—optimization modeling in Python,
1st ed. Springer Publishing Company, Incorporated
Chapter 10
Design of a Simulation Environment
for Training or Testing Algorithms
to Solve the Workshop Sequencing
Problem
Abstract Simulation is a frequently used tool in engineering fields, in particular its

use as a testing mechanism has been one of its longest-lived functions. In the current
context, simulation has become an important component in the training of artificial
intelligence algorithms. This paper presents an approach to develop a simulation
environment for testing or training novel algorithms to solve the sequencing problem
in jobshops. This environment has the novelty of being as close as possible to the
reality of jobshops in terms of variability and dynamism. Specifically, it details the
methodology to be followed, the minimum components or features to be included and
the considerations to be considered to integrate it with a machine learning algorithm
for training and suggests a possible approach to test results for statistical validation.
Keywords Simulation · Jobshop · Sequencing.
10.1 Using Simulation to Test or Train JSSP Solutions
Simulation involves developing descriptive computer models of a system and exer-

cising those models to predict the operational performance of the underlying system
being modeled [8].
Based on this definition, various applications have been made in business and
industrial processes. For decades, it has been used as a tool to support decision
making in manufacturing systems [7].
Tunali [9] states that computer simulation is an extremely powerful tool for
studying the behavior of advanced manufacturing systems. Traditionally, discrete
event simulation has been mainly used to analyze and design manufacturing systems.
E. Pérez-Cubero (B)
Universidad de Costa Rica, Alajuela, Costa Rica
e-mail: efrain.perezcubero@ucr.ac.cr
R. Poler
Universitat Politècnica de València, Valencia, Spain

https://doi.org/10.1007/978-3-031-29382-5_10
94 E. Pérez-Cubero and R. Poler
In recent years, simulation has been widely used not only in the design, but also
in the planning, scheduling, and control of advanced manufacturing systems [9]. It is
in this sense where it is of interest to the authors to apply it as a tool to train machine
learning algorithms as a means of solving the jobshop scheduling problem (JSSP).
Recent research has shown that the evaluation or training of algorithms to solve
JSSP and its derivations has been done mainly through simulation, as shown in
[6]. These simulations facilitate the testing of such algorithms without risking the
performance of real processes, which makes it an excellent option for testing new
paradigms for solving NP-Hard problems in industry.
The simulation environments that have been used recently show a reduction of
the real conditions found in manufacturing environments, as shown by [2, 3, 10] and
[11]. It is for this reason that when contemplating the development of a simulation
environment for testing and training novel algorithms that provide a solution to JSSP,
an effort should be made to reflect as closely as possible the reality of the shop floor.
Variables such as: variation in orders, machine downtime due to breakdowns, operator
absenteeism, quality problems, delays in the delivery of raw materials, among others,
should be considered. These conditions would bring the algorithms closer to more
realistic scenarios and thus increase their effectiveness and efficiency.
Furthermore, the target functions used in the problem are of special interest, from
a recent literature review [6] it can be extracted that the makespan is the preferred
option of the authors, as can be seen in Fig. 10.1. However, there is no mention of
the reason why this is preferred over other options, it would be worth conducting
a survey in the industry to identify which is the preferred objective function at a
practical level. Makespan is understood as the maximum completion time of jobs
[3].
At this point, we can state that simulation is the most widely used option for
testing or training new algorithms, but also that simulations are usually based on
several assumptions that keep them far from the reality of the workshop.
10.2 Advantages of Creating a Simulation Environment

Versus Specific Software
When starting a simulation project of a workshop or a production plant, a decision

must be made regarding the tool in which the simulation will be carried out. There
are three possible options: (a) using a commercial simulator, (b) using a simulation
language, and (c) using a general-purpose language.
When making the decision, one must consider, as stated by [4], a series of special
modeling features that may or may not be included in a simulation tool. It should
also be considered that simulating a production plant design to evaluate its operation
is not the same as simulating a process to test or train a novel sequencing algorithm,
since the latter exercise requires the application of the scientific method and quite
possibly repeated adjustments to both the algorithm and the simulation environment.
10 Design of a Simulation Environment for Training or Testing Algorithms … 95
Fig. 10.1 Objective function in JSSP algorithms, drawn from data presented in [6]
Finally, it should be borne in mind that in order to carry out the tests or training,
the sequencing algorithm must be available in the test environment, which would
be very complicated in a commercial simulator because these are not designed to
add algorithms to those already available. The final decision is to use a general-
purpose language, such as Python, but using a specialized simulation library such as
Sympi, which makes it a hybrid scenario between a general-purpose language and
a simulation language. This justifies the creation of an environment in a general-
purpose programming language.
10.3 Simulation Environment
10.3.1 Design Methodology
As already mentioned, the main objective of the simulation environment to be devel-

oped is the testing or training of novel artificial intelligence algorithms that provide
a solution to the JSSP. Currently, one of the most frequently used software for the
implementation of artificial intelligence algorithms is Python, which has specialized
libraries, is open, and has a very dynamic global community in its use. It is there-
fore logical to design the simulation environment in the same language as this will
facilitate communication and integration.
Regarding the construction of the framework, [10] propose four components:
• Data preparation.
• Model generation.
• Model validation.
• Scenario simulation.
Wilson and Evans [12] propose a more detailed set of steps than those proposed by
[10]. These are:
1. Problem formulation (definition) and project planning.
2. Data collection and model formulation.
3. Model validation.
4. Coding and verification of a computer program.
5. Execution and analysis of pilot tests.
6. Validation.
7. Design of experiments.
8. Execution of production runs.
9. Output data analysis.
10. Documentation, presentation, and implementation of results.
Ciaburro [1] proposes the following steps.
1. Problem analysis.
2. Data collection.
3. Setting up the simulation model.
4. Simulation software selection.
5. Verification of the software solution.
6. Validation of the simulation model.
7. Simulation and analysis of results.
The authors would favor an adaptation of these methodologies for the development
of the simulation environment:
1. Parameterization of a generic jobshop
2. Collection and preparation data.
3. Simulation framework coding.
4. Functional and logical testing.
5. Running the simulated experiments or machine learning agent training.
6. Results analysis.
10.3.2 Components of the Simulation Environment
Once the methodology to be followed for the design of the simulation environ-
ment has been defined, the minimum components of the simulation environment
must be established. For the development of a simulation environment, the objects,
the attributes of these objects, and the functions must be identified. In a jobshop,
the relevant objects to simulate are machines, material handlers (depending on the
configuration of the jobshop), and jobs or parts.
The machine object owns the attributes: (1) duration (probability distribution) of
breakdowns; (2) frequency (probability distribution) of breakdowns; (3) time avail-
able per day; (4) maximum WIP it is able to store; (5) maximum queue size in front
of the machine and (6) quality rate (probability distribution). The machine object
must also have functions that tell it which the next job is to be processed. This func-
tion must be able to accept everything from conventional heuristics, such as FIFO,
to novel machine learning algorithms. The inclusion of failure rate and quality rate
through a probability distribution, as well as WIP and availability rate, serve to create
an environment that emulates the reality of shop floor processes.
Regarding the object piece or job, the minimum attributes to consider are route,
size, frequency (probability distribution) of arrival, frequency (probability distri-
bution) of cancelation, arrival date, and required delivery date. By including the
frequency of arrival and cancelation through probability distributions, it is possible
to emulate the reality of these processes in a good way, which brings the environment
closer to the reality of industrial processes.
Relevant attributes for material handlers (people or machines) would be: (1) the
frequency of pick-up and delivery, (2) the route, (3) the load capacity, (4) the time
available per day, and (5) the absenteeism of personnel or machine downtime. In the
case of people a probability distribution should be considered that models absen-
teeism from the process, in the case of machines a probability distribution should be
considered that describes the time lost. The last two attributes, being stochastic, seek
to emulate real industrial processes.
Another function that the simulation environment must have is that related to
certain key indicators of the process. As already mentioned, there are many indicators
that are optimized in the sequencing of a jobshop: makespan, total tardiness, shortest
delivery time, etc. These indicators must be estimated by the simulation environment
and displayed as part of the simulation outputs.
Finally, to ensure that the simulation environment runs realistic simulations, it
will be validated using real process data whereby it will be verified that the outputs
of the simulation environment do not differ statistically from those obtained in the
real process.
10.3.3 Integration with Machine Learning Agent
The purpose of the simulation environment includes the training of new algorithms,
so the integration of the environment with the algorithm is essential to achieve the
proposed objective. To guarantee this integration, it is proposed to use the same
programming language for both: Python. The variables or data that the simulation
environment can generate as outputs must also be taken into account, so that these
are the inputs used by the sequencing algorithm.
Using the same language would make it easier to adjust the code of one or the
other to improve communication between the two if necessary.
It is proposed that the simulation environment considers the artificial intelligence
agent as a function that it will call to decide which is the next job to be executed, as
it would do with a conventional heuristic.
10.4 Designing an Experiment to Test Solutions Through

Simulation
The other objective pursued with the development of the simulation environment is
to compare different algorithms, including heuristics, that provide solutions to the
JSSP. The aim of this is to identify if there are differences between the solutions
generated, and if there are differences, which would be the best one.
Since the use of dynamic variables is envisaged, the best possible approach would
be to consider the system as a black box. That is, keep all conditions with the same
parameters, including the probability distributions used, and study only the effect of
the algorithms on the objective functions. This configuration would allow using an
ANOVA as a tool to contrast the performance of the algorithms and complement it
with Tukey or Fisher tests to identify the best of the algorithms used.
If the data generated does not meet the assumptions required for an ANOVA, the
choice would be to increase the number of runs in order to use ANOVA or to perform
a Kruskal–Wallis test [5].
10.5 Conclusions and Lines of Research
The use of simulation environments as a mechanism for testing and training algo-
rithms to solve JSSP has been used in recent years. However, these environments are
often considerably abstracted from reality.
This condition opens room to question the effectiveness of such algorithms in real
processes. It also makes it possible to question research results that claim that a given
algorithm is better than others. This situation makes it necessary to design simula-
tion environments that retain more of the typical characteristics of manufacturing
processes.
It has been interesting to discover that in the literature consulted, the use of one
objective function over the others is not justified to any extent, i.e., there is no apparent
reason why using the makespan is better than total tardiness. This situation leads us
to consider, as a future line of research, to answer the question: what are the most
used objective functions in industry when optimizing the sequencing of a jobshop?
This is of relevance because algorithms could be generated that are not optimized
for what industry requires, which would make them inefficient and therefore unlikely
to be implemented at the industrial level.
The limitation of commercial simulation software to cope with the testing or
training of novel algorithms is another conclusion drawn from this paper, since as
discussed, it does not allow for easy interaction with external agents.
References
1. Ciaburro G, Hands-on simulation modeling with python: develop simulation models to get
accurate results and enhance decision-making processes. Packt Publishing. Kindle edition
2. Kim H, Lim DE, Lee S (2020) Deep learning-based dynamic scheduling for semiconductor
manufacturing with high uncertainty of automated material handling system capability. IEEE
Trans Semicond Manuf 33(1):13–22. https://doi.org/10.1109/TSM.2020.2965293
3. Kundakci N, Kulak O (2016) Hybrid genetic algorithms for minimizing makespan in dynamic
job shop scheduling problem. Comput Ind Eng 96:31–51. https://doi.org/10.1016/j.cie.2016.
03.011
4. Miller S, Pegden D (2000) Introduction to manufacturing simulation. Winter Simul Conf Proc
1:63–66. https://doi.org/10.1109/WSC.2000.899699
5. Montgomery DC (2017) Design and analysis of experiments. In WILEY (Ninth Edit)
6. Pérez-Cubero E, Poler R (2021) Application of machine learning algorithms to production order
scheduling in job shops: a review of recent literature. Direccion y Organizacion 72(72):82–94.
https://doi.org/10.37610/DYO.V0I72.588
7. Seleim A, Azab A, AlGeddawy T (2012) Simulation methods for changeable manufacturing.
Procedia CIRP 3(1):179–184. https://doi.org/10.1016/j.procir.2012.07.032
8. Smith JS (2003) Survey on the use of simulation for manufacturing system design and operation.
J Manuf Syst 22(2):157–171. https://doi.org/10.1016/S0278-6125(03)90013-6
9. Tunali S (2004) A simulation-based scheduling system for a textile plant. Int J Comput Appl
Technol 19(2):119–124. https://doi.org/10.1504/IJCAT.2004.003643
10. Wang Y, Liu H, Zheng W, Xia Y, Li Y, Chen P, Guo K, Xie H (2019) Multi-objective work-
flow scheduling with deep-Q-network-based multi-agent reinforcement learning. IEEE Access
7:39974–39982. https://doi.org/10.1109/ACCESS.2019.2902846
11. Wei Y, Pan L, Liu S, Wu L, Meng X (2018) DRL-Scheduling: an intelligent QoS-Aware job
scheduling framework for applications in clouds. IEEE Access 6:55112–55125. https://doi.
org/10.1109/ACCESS.2018.2872674
12. Wilson R, Evans GW (1992) Simulation of advance manufacturing systems
Chapter 11
Advanced Methods and Models
of Optimization and Data Visualization
for the Management, Monitoring,
and Control of Operations in Companies
Working in Collaborative Manufacturing
Environments
Abstract The main objective of this research is the generation of artificial intel-
ligence algorithms for the solution of the jobshop scheduling problem (JSSP) with
real-time adjustments to changing conditions. The advent of technologies such as the
Internet of Things (IoT) and data science has boosted the use of advanced machine
learning techniques to solve complex problems (NP-Hard) that until recently were
impractical using conventional techniques due to their computational times. The
ability to have production plants that respond to variations in manufacturing envi-
ronments has been a requirement of process managers for a long time, having to
deal with this problem most of the time through expert judgment, contemplating the
effects of variations in the overall performance of the system is a necessity in the
face of the high demands of customers and global markets. Therefore, the search for
mechanisms that provide solutions to this problem is a very promising line of applied
research.
Keywords Deep learning · Reinforced learning · Shop floor sequencing problem ·

Operations management · Control
E. Pérez-Cubero (B)
Universidad de Costa Rica, Alajuela, Costa Rica
e-mail: efrain.perezcubero@ucr.ac.cr
R. Poler
https://doi.org/10.1007/978-3-031-29382-5_11
11.1 Introduction
This research seeks to integrate tools and technologies of the Fourth Industrial Revo-
lution with manufacturing environments to improve decision-making regarding the
sequencing, control, and monitoring of the operation.
Specifically, based on a review of the state of the art, the project seeks to develop a
proposal of methods and tools for sequencing, control, and monitoring of operations
in collaborative manufacturing environments using machine learning for decision
support, specifically the application of deep learning (DL) and reinforcement learning
(RL) to jobshop scheduling problem (JSSP) [1].
As for the collaborative manufacturing environment, this can occur at different
levels, between different companies, within the same company at different levels of
the logistics network, and within the same factory between departments. The project
will develop algorithms at the latter level, but with the perspective of being applied
to the other levels in later stages of research, validation, and development.
Deep learning (DL) and reinforced learning (RL) are part of the current axes
of study and research in the planning, control, and monitoring of operations [1],
together with elements such as information technologies, artificial intelligence and
metaheuristics, robotics, and automation [2], mean that, among others, the role of
the human resource decision-maker is affected by these tools and their daily use.
11.1.1 Problem Statement and Motivation
In today’s manufacturing environments where there are production systems with

multiple plants located in different geographical locations interconnected by
supply/supply networks of materials, subassemblies, and products, not all of which
are owned by the same company, some of which are small- and medium-sized enter-
prises, variations in the production plan of each plant have an effect on the perfor-
mance of the whole system and can affect customers and partners. Variations in
the production schedule of individual plants influence the performance of the whole
system and can affect customers and partners.
These variations, generated by a variety of elements such as machine failures,
staff absenteeism, defects, reduced equipment speed among others; require decision-
making during the execution of processes in real time and, in many cases, with little
information on the impact of these decisions on the entire production system, which
affects the quality of the same and the performance of the entire system, this would
eventually involve affecting customers or making the operation more expensive.
This is how it has already been indicated that in today’s environment operations
scheduling must deal with a smart manufacturing system supported by novel and
emerging manufacturing technologies, such as cyberphysical systems (CPS), big
data, Internet of Things (IoT), artificial intelligence (AI), virtual twins, and social,
11 Advanced Methods and Models of Optimization and Data Visualization … 103
mobile, analytics, cloud (SMAC). Production scheduling research needs to shift its
focus toward modeling and optimizing intelligent distributed scheduling [2].
On the other hand, Erol and Sihn [3] indicate that process control, in manufacturing
companies, tracks the overall state of the process (order) and also tracks all subprocess
states. Therefore, progress control and production control can be achieved in high
detail. This control should be immersed within the whole collaborative manufacturing
system and in the cloud where all actors would be able to review changes and their
impacts on the value chain, also becoming a possible input as variables or constraints
for future production scheduling models.
This scheme can allow real-time visualization of the system status and alert on
changes as they occur, as well as support decision-making through optimization algo-
rithms or artificial intelligence for resource reallocation at each stage of the process.
This information management and process optimization would not be feasible with
traditional operations schemes due to the lack of connectivity and isolated informa-
tion that is normally handled, allowing at most a local optimization, which does not
guarantee reaching global optima.
This research project seeks to compile the existing literature research in the field
of sequencing, monitoring, and control of operations in collaborative manufacturing
environments and the optimization tools applied to these. Then, based on these find-
ings, a novel proposal of an algorithm for sequencing, monitoring, and control of
operations in collaborative manufacturing environments should be made. At this
point, it is expected to use data science tools for capturing and visualizing information
in real time, as well as machine learning to support decision-making.
11.1.2 Research Questions and Objectives of the Doctoral

Thesis
Starting from the importance of generating optimization algorithms and tools for
visualizing the status of processes in real time, which achieve the best use of manu-
facturing resources in the workshops, while allowing them to react to the variations
inherent to these systems.
The following research questions are posed to support the design of artificial
intelligence algorithms for optimization of sequencing, as well as visualization tools
for monitoring and control of operations and for the purpose of integration with the
objective of this research, a general research question is inferred:
GRQ. What would be the best solution to the problem of sequencing production
in real workshops in a collaborative environment?
The following specific research questions are derived from this general research
question:
RQ1. What algorithms have been used to solve the jobshop scheduling problem
and how successful have they been?
RQ 2. What artificial intelligence algorithms have been used to solve the workshop
scheduling problem and how successful have they been?
RQ 3. What kind of information do the workshops generate to feed artificial
intelligence algorithms capable of solving the workshop sequencing problem?
RQ 4. What are the target functions most frequently used by industry to optimize
a production plan?
RQ 5. What are the most appropriate objective functions for artificial intelligence
algorithms to solve the jobshop sequencing problem?
Could an enhanced learning agent be the best solution to the workshop sequencing
problem?
RQ 7. What role should people play in an artificial intelligence solution to the
jobshop sequencing problem?
RQ 8. What is the information that should be provided to people to develop their
role in an artificial intelligence solution to the jobshop sequencing problem?
In contrast to the above, the general research objective is defined:
GRO. To develop advanced solutions in methods and tools for the sequencing,
monitoring, and control of operations, as a support for decision-making in companies
working in collaborative manufacturing environments.
From which the following specific objectives can be derived:
SO1. Conduct a literature review on methods and conceptual frameworks for
planning, monitoring, and control of operations to know the current state of the art
and identify lines of work not addressed so far by other researchers to make a novel
contribution to this branch of engineering.
SO 2. Identify the most suitable artificial intelligence algorithms to solve the
workshop sequencing problem.
SO 3. Identify the data generated by the workshops that can feed artificial
intelligence algorithms capable of solving the workshop sequencing problem.
SO 4. Design an artificial intelligence algorithm to provide an efficient solution
to the JSSP problem in a novel way to improve and facilitate decision-making by
those in charge of production processes.
SO 5. To implement an artificial intelligence algorithm to provide an efficient
solution to the JSSP problem in a novel way to improve and facilitate decision-making
by those in charge of production processes.
SO 6. Validate an artificial intelligence algorithm that allows an efficient solution
to the JSSP problem in a simulated environment and in real environments.
SO 7. Define the role of people in the production scheduling process using an
artificial intelligence algorithm that solves the JSSP problem.
SO 8. Identify future lines of research in the application of artificial intelligence
to the problem of workshop sequencing.
11.2 Methodology
The following methodology has been proposed for the development of the research
project, which responds to the part-time dedication of the doctoral student, which is
why the proposed duration is 60 months.
Phase I: Review of the state of the art. This phase includes a literature review of
methods and tools for sequencing, monitoring, and control of operations using or
based on artificial intelligence, specifically for companies working in collaborative
manufacturing environments. The above with the aim of detecting spaces that allow
the generation of new knowledge or new applications of existing knowledge.
Phase II: Identification of available data in a workshop. The purpose of this phase
is to define the data available in a workshop that will function as input parameters to
an artificial intelligence algorithm capable of solving the sequencing problem. The
output data and performance indicators of the artificial intelligence algorithm must
also be defined.
Phase III: Construction of algorithms for operations planning. This phase proposes
the use of artificial intelligence to propose a solution to the sequencing problem of
the workshop, which can be applied to a collaborative manufacturing environment.
Phase IV: Validation of the proposed algorithms. Within this phase, validation is
proposed by means of simulations of the proposals in the previous phases, and it is
also expected that tests will be carried out in real environments.
Figure 11.1 graphically presents the relationship between the stages, objectives,
and research questions.
Fig. 11.1 Proposed research methodology

Fig. 11.2 Outline of the development of the thesis
11.3 Scheme
The development of the research project, with which the general objective and the
specific objectives will be achieved, will be developed through a process of four
phases, as shown in the previous section, these phases will cover the six chapters of
the thesis. These are divided into activities (A) and tasks (T), and a timeline for the
execution of these is also presented in Fig. 11.2.
11.4 Current State of Research
The current progress of the research is in phase 1, with an advance of 70%. The state
of the art is in the final stages of drafting for publication in a specialized scientific
journal.
11.5 Conclusions
The developments of recent years in the fields of computer science and robotics have
made possible the arrival of the so-called Fourth Industrial Revolution. This context
makes it necessary to search for advanced artificial intelligence algorithms that allow
solutions to complex problems such as JSSP and its derivatives, which in turn are
applicable to collaborative environments.
Specifically, the aim will be to develop algorithms to solve the workshop

sequencing problem.
References
1. Cunha B, Madureira AM, Fonseca B, Coelho D (2020) Deep reinforcement learning as a job
shop scheduling solver: a literature review. Adv Intell Syst Comput 923:350–359. https://doi.
org/10.1007/978-3-030-14347-3_34
2. Zhang J, Ding G, Zou Y, Qin S, Fu J (2019) Review of job shop scheduling research and its
new perspectives under Industry 4.0. J Intell Manuf 30(4):1809–1830. https://doi.org/10.1007/
s10845-017-1350-2
3. Erol S, Sihn W (2017) Intelligent production planning and control in the cloud—towards a
scalable software architecture. Proc CIRP 62:571–576. https://doi.org/10.1016/j.procir.2017.
01.003
Chapter 12
Annualized Hours, Multiskilling,
and Overtime on Annual Staffing
Problem: A Two-Stage Stochastic
Approach
Andrés Felipe Porto , Amaia Lusa , César Augusto Henao ,

and Roberto Porto Solano
Abstract This study evaluates the potential benefits of a labor flexibility strategy
that simultaneously incorporates: (i) annualized working hours, (ii) multiskilled staff,
and (iii) overtime; into the annual staffing problem for a retail store. A two-stage
stochastic optimization model is proposed to determine: How many staff is required in
each store department; what amount of weekly working time (ordinary and overtime
hours) is required per employee in an annual planning horizon; how many employees
will be multiskilled, and in which departments they will be trained. This formulation
considers uncertainty in the staff demand. Using real data from a Chilean retail outlet,
the obtained results allow us to design a flexible and cost-effective workforce at a
strategic level. That is, the proposed triple strategy reports the lowest total annual
cost (a saving of 69%) and also requires a smaller staff size for the store.
Keywords Labor flexibility · Annualized hours · Multiskilling · Overtime ·

Retail · Uncertainty
A. F. Porto · R. Porto Solano

Corporación Universitaria Americana, Barranquilla, Colombia
e-mail: rporto@coruniamericana.edu.co
A. F. Porto (B) · A. Lusa
Universitat Politècnica de Catalunya, Barcelona, Spain
e-mail: aporto@coruniamericana.edu.co; andres.felipe.porto@upc.edu
A. Lusa
e-mail: amaia.lusa@upc.edu
C. A. Henao
Universidad del Norte, Barranquilla, Colombia
e-mail: cahenao@uninorte.edu.co
https://doi.org/10.1007/978-3-031-29382-5_12
110 A. F. Porto et al.
12.1 Introduction
For industries, especially those in the service sector (e.g., call centers, hospitals,
transportation, retail), maintaining a high level of service for customers while main-
taining or improving their profits is not an easy task [1, 2]. In addition, personnel
planning is affected by external and unpredictable factors in their environment. As
a sample of these factors, the uncertainty in demand (e.g., unexpected increases or
decreases in the staff demand) and the unscheduled staff absenteeism stand out.
These unpredictable factors produce mismatches between the staff supply and the
demand for personnel, ultimately resulting in an increase in labor costs [3–6]. In
the personnel planning problems, these mismatches between supply and demand
are known as overstaffing and understaffing (i.e., staffing level higher or lower than
required, respectively) and have been studied in the operations management area, for
a long time [7, 8]. The personnel planning problems constantly prompt industries
and economic sectors to define new and better labor flexibility strategies [5, 9].
In an uncertain context such as the one described above, the retail industry stands
out for its constant and rapid growth worldwide, in addition to its intensive use of
personnel [10, 11]. Henao et al. [5] listed a set of restrictions that hinder proper
personnel planning: (i) contracts with fixed daily working hours, typically full-time;
(ii) legal restrictions, such as the regulated amount of weekly working hours, ordinary,
or overtime; (iii) institutional restrictions, regarding shifts and their duration; (iv)
personal employee’s preferences; and finally; (v) single-skilled personnel; that is,
employees can perform a single task type. Henao et al. [5] also mentioned that
these restrictions generate inflexible labor plans for facing a seasonal and uncertain
personnel demand.
Service industries have used traditional flexibility strategies for personnel plan-
ning, such as varying lengths and number of shifts, as well as overtime. In addition,
the literature on labor flexibility shows that there are typically four flexibility strate-
gies that are individually implemented, these are: (i) Flexible working time, which
allows to relax the shifts length and the number of weekly/annual working hours of
the employees (e.g., [12–14]). (ii) Multiskilled staff, employees trained to work on
multiple task types (e.g., [2, 3, 5, 11, 14–21]). (iii) Teamwork, groups of employees
who carry out the tasks together, and not individually (e.g., [22, 23]). (iv) Temporary
workers are those employees who are hired to work short periods of time in pressing
situations (e.g., [24, 25]). Finally, several authors have expressed that in the service
sector, very few efforts have been made to implement flexibility strategies jointly in
personnel planning [8, 26].
12 Annualized Hours, Multiskilling, and Overtime on Annual Staffing … 111
Considering uncertain demand, this work seeks to solve an annual staffing problem
through the joint use of three labor flexibility strategies: annualized hours, multi-
skilling, and overtime. Regarding the annualized hours’ literature, some papers have
not considered multiskilling into the personnel planning problem (e.g., [27–30]),
while some other papers have considered it (e.g., [31–35]). However, in this last group
of studies, multiskilling was considered as a parameter, which means that training
decisions are not made, but rather a set of pre-established skills was assumed in the
employees.
Regarding the combined use of annualized hours, multiskilling, and overtime, it
is possible to highlight studies that simultaneously considered these three flexibility
strategies (e.g., [31, 33, 35]). However, as for our knowledge, none of them consid-
ered multiskilling as a decision variable, and very few considered uncertainty in the
personnel demand (i.e., [35]). Furthermore, these studies were limited to tactical
(e.g., shift scheduling problems) and operational (e.g., tasks assignment problems)
decisions, but these did not consider strategic decisions (e.g., staffing, training), in the
personnel planning problems. Finally, this paper is an extension of Porto et al. [36]
and fills said gap in the literature, since we solve a staffing problem under uncertain
demand and considering three labor flexibility strategies. In addition, we consider
multiskilling as a decision variable. Thus, this proposal addresses the following
strategic decisions in the personnel planning: (i) how many employees must be
hired; (ii) how many of them should be multiskilled and in which departments;
(iii) how many weekly working hours (ordinary and overtime) should be assigned
per employee in a one-year planning horizon.
12.3 Problem Statement
Oriented to the service sector, and particularly for a retail store under uncertain
demand, the problem consists in planning a flexible workforce, which minimizes
the expected costs of over/understaffing for a one-year planning horizon. Thus, we
propose a triple strategy of labor flexibility that considers: (i) annualized hours’
schemes, which allow contracting an employee for a fixed number of working hours
per year, and then, these hours can be assigned irregularly over the weeks and months;
(ii) multiskilled staff; and (iii) overtime. In short, the solution to the problem seeks
to answer the three strategic decisions described in the previous section.
The formulation of the proposed problem considers the following assumptions:
(1) The personnel demand is aggregated weekly by the department. (2) Understaffing
cost is included in the cost function and corresponds to the expected cost of lost sales.
(3) Overstaffing cost is also considered, which quantifies the incurred opportunity
cost for having idle personnel. (4) The cost of training is also included in the cost
function. It is also assumed that costs of under-and-overstaffing and costs of training
are the same by the department. (5) There is no absenteeism of personnel. (6) The
workforce is homogeneous; that is, the individual productivity of the employees is
the same, even if they are multiskilled. (7) Multiskilled employees can only work in
a total of two departments. (8) It is assumed that all employees are hired with the
same number of hours per year. (9) Constraints on the maximum number of weekly
working hours, both ordinary and overtime, are based on the legal context of each
country.
12.4 Methodology
We propose a two-stage stochastic optimization (TSSO) model to explicitly incor-

porate the uncertain demand on the annual staffing problem. To achieve this, we rely
on the sample average approximation approach (SAA) [37].
This section presents the sets, parameters, and variables associated with the
proposed TSSO model. Note that, a TSSO model considers first-stage and second-
stage variables. The first-stage variables do not explicitly depend on the demand
scenarios, while the second-stage variables do depend on the realizations of the
uncertain demand. Below, the mathematical notation of the problem is presented.
Sets:
S Weeks in the annual planning horizon, indexed by s ∈ {1, 2, 3, …, 52}

L Store departments, indexed by l ∈ {1, 2, 3, 4, 5}
W Set of skill sets, indexed by w. Includes sets with a unique skill (i.e., single-skilled) and sets
that have two skills (i.e., multiskilled); w ∈ {1, 2, 3, …, 25}
K Demand scenarios, indexed by k ∈ {1, 2, …, 10}
Parameters:
rlsk Weekly demand hours in department l, week s, and demand scenario k,

∀l ∈ L , s ∈ S, k ∈ K
u Understaffed cost per hour in any department
b Overstaffed cost per hour in any department
a Cost per ordinary annualized hour
e Cost per overtime hour
m Training cost
Variables:
First-stage variables
Zw Number of employees required with skill set w, w ∈ W
Second-stage variables
(continued)
(continued)
Pwlsk Number of weekly ordinary working hours assigned to employees with skill set w, in
department l, week s, and demand scenario k, ∀w ∈ W, l ∈ L , s ∈ S, k ∈ K
Swlsk Number of weekly overtime working hours assigned to employees with skill set w, in
department l, week s, and demand scenario k, ∀w ∈ W, l ∈ L , s ∈ S, k ∈ K
K lsk Number of hours that cannot be covered (understaffing) in department l, week s, and
demand scenario k, ∀l ∈ L , s ∈ S, k ∈ K
L lsk Number of hours in excess (overstaffing) in department l, week s, and demand
scenario k, ∀l ∈ L , s ∈ S, k ∈ K
The objective function of the TSSO model is formulated as follows:

Min m Zw
w∈W

1
+ (u K lsk + bL lsk ) (12.1)
| K | l∈L s∈S k∈K

+ (a Pwlsk + eSwlsk )
w∈W l∈L s∈S k∈K
In the objective function, the first-stage variables are Z w , which define the hiring
and training decisions of employees. The second-stage variables K lsk , L lsk , Pwlsk ,
and Swlsk are operational adjustments, which are taken once the random realizations
of the demand in each department are known. The objective function (1) minimizes
the following annual costs: (a) training of employees, (b) under-and-overstaffing,
and (c) salary associated with ordinary and overtime working hours. The constraints
associated with this model, and not shown here due to limited space, are described
below: (i) It is guaranteed that the employees must work the total number of annual
ordinary working hours stipulated in their contracts. (ii) A maximum and minimum
of weekly ordinary hours of work are limited according to labor regulations. (iii) A
maximum of weekly overtime hours is also limited according to labor regulations.
(iv) It is ensured that in a moving horizon of q consecutive working weeks (e.g., 12),
employees cannot exceed an average number of ordinary weekly working hours (e.g.,
35 h per week). Throughout the year, this last constraint seeks to assign balanced
weekly work schedules for each employee.
12.5 Test Scenarios and Results
Using data real from a Chilean retail [10], this section shows the proposed test
scenarios and the obtained results. The TSSO model was written in AMPL, and it
was solved by using ILOG CPLEX 12.4.0.1 software. Test instances were executed
on a laptop with INTEL® Core ™ i7—8550U processor, 2.0 GHz, 8 GB of RAM

memory, and with 64 bits of Windows.
12.5.1 Test Scenarios
We consider a retail store with five departments, and it was also assumed a minimal
training cost. This last assumption allows us to obtain an upper bound for the multi-
skilling requirements. In addition, we use historical data to obtain 10 scenarios for
the weekly demand in each department, such that |K | = 10.
Finally, we proposed four test scenarios to measure the benefits of our triple labor
flexibility strategy. Note that, Scenario 1 represents an individual labor flexibility
strategy, Scenarios 2 and 3 represent a double strategy, and Scenario 4 represents our
proposed triple strategy.
1. Annualized hours’ scheme with single-skilled staff.
2. Annualized hours’ scheme with multiskilled staff.
3. Annualized hours’ scheme with overtime.
4. Annualized hours’ scheme with multiskilled staff and overtime.
12.5.2 Preliminary Results
Below, for each scenario, Tables 12.1 and 12.2 present the preliminary results asso-
ciated with the incurred costs (Table 12.1) and the required staff levels (Table 12.2)
in the solution of the annual staffing problem with uncertain demand.
Table 12.1 Costs associated for each scenario. Note that, savings percentages in Scenarios 2, 3,
and 4 are calculated in relation to Scenario 1, which represents the most basic and inflexible scen
ario
Costs per scenario
1 2 3 4
Annual cost ($) 587,720 573,423 216,073 184,422
Overstaffing ($) 244,335 364,920 31,485 6840
Understaffing ($) 156,695 7805 7345 261
Ordinary salary ($) 186,690 200,692 158,687 154,019
Overtime salary ($) – – 18,557 23,296
Training ($) – 7 – 6
Saving (%) – 2 63 69
Table 12.2 Staff levels for the solution of the annual staffing problem
Staff size per scenario
1 2 3 4
Hired staff 40 43 34 33
Single-skilled staff 40 36 34 27
Multiskilled staff – 7 – 6
12.6 Conclusions
In this article, an annual staffing problem is solved considering uncertain demand.

Mathematical formulation simultaneously incorporated three labor flexibility strate-
gies: annualized working hours, multiskilled staff, and overtime. The objective was
to minimize the expected costs of over/understaffing. The methodology was applied
a case study associated with a retail store in Santiago (Chile). The following strategic
questions were addressed: (i) how many staff is required in the store; (ii) number of
weekly working hours, ordinary and overtime that must work each employee for a
one-year planning horizon; and (iii) how many employees should be multiskilled,
and in which departments they should be trained.
Preliminary results showed a reduction in the costs of understaffing and over-
staffing, as well as a reduction in the total annual salary cost (see Table 12.1). From
Table 12.1, it was observed that Scenario 1 was the most expensive. This result
is intuitive since this scenario considered a single labor flexibility strategy (i.e.,
an annualized hours’ scheme). Regarding Scenario 2, which includes multiskilled
employees and an annualized hours’ scheme, it was observed a minimal decrease in
the total annual cost. Regarding Scenario 3, which includes overtime and an annual-
ized hours’ scheme, the total annual cost decreased by 63% compared to Scenario 1.
This last result showed the complementarity between overtime and annualized hours’
scheme. Regarding Scenario 4, which includes overtime, multiskilled employees, and
an annualized hours’ scheme, savings of 69% were achieved in relation to Scenario
1. This showed that the proposed triple strategy (Scenario 4) was superior to an indi-
vidual strategy (Scenario 1) and the double strategies (Scenarios 2 and 3). Finally,
our formulation can be extended to other industries in the service sector, being useful
for decision-makers regarding their personnel planning problems.
References
1. Muñoz R, Muñoz JC, Ferrer JC, González VI, Henao CA (2021) When should shelf stocking
be done at night? A workforce management optimization approach for retailers. Submitted to
Computers & Industrial Engineering
2. Mac-Vicar M, Ferrer JC, Muñoz JC, Henao CA (2017) Real-time recovering strategies on
personnel scheduling in the retail industry. Comput Ind Eng 113:589–601
3. Henao CA (2015) Diseño de una fuerza laboral polifuncional para el sector servicios: caso
aplicado a la industria del retail. Ph.D thesis, Pontificia Universidad Católica de Chile in
Santiago de Chile
4. Kabak Ö, Ülengin F, Aktaş E, Önsel Ş, Topcu YI (2008) Efficient shift scheduling in the retail
sector through two-stage optimization. Eur J Oper Res 184(1):76–90
5. Henao CA, Muñoz JC, Ferrer JC (2015) The impact of multi-skilling on personnel scheduling
in the service sector: a retail industry case. J Oper Res Soc 66(12):1949–1959
6. Álvarez E, Ferrer JC, Muñoz JC, Henao CA (2020) Efficient shift scheduling with multiple
breaks for full-time employees: a retail industry case. Comput Ind Eng 150:106884
7. Ernst AT, Jiang H, Krishnamoorthy M, Owens B, Sier D (2004) An annotated bibliography of
personnel scheduling and rostering. Ann Oper Res 127(1–4):21–144
8. Qin R, Nembhard DA, Barnes II WL (2015) Workforce flexibility in operations management.
Surv Oper Res Manage Sci 20(1):19–33
9. Porto AF, Henao CA, López-Ospina H, González ER (2019) Hybrid flexibility strategy on
personnel scheduling: retail case study. Comput Ind Eng 133:220–230
10. Porto AF, Henao CA, López-Ospina H, González ER, González VI (2020) Dataset for solving
a hybrid flexibility strategy on personnel scheduling problem in the retail industry. Data Brief
32:106066
11. Henao CA, Muñoz JC, Ferrer JC (2019) Multiskilled workforce management by utilizing
closed chains under uncertain demand: a retail industry case. Comput Ind Eng 127:74–88
12. Lusa A, Pastor R, Corominas A (2008) Determining the most appropriate set of weekly working
hours for planning annualised working time. Int J Prod Econ 111(2):697–706
13. Lusa A, Pastor R (2011) Planning working time accounts under demand uncertainty. Comput
Oper Res 38(2):517–524
14. Porto AF, Henao CA, Lusa A, Polo Mejía O, Porto Solano R (2021) Solving a staffing problem
with annualized hours, multiskilling with 2-chaining, and overtime: a retail industry case.
Submitted to Computers and Industrial Engineering
15. Henao CA, Batista A, Porto AF, González VI (2021) Multiskilled personnel assignment
problem under uncertain demand: a benchmarking analysis. Sub Ann Oper Res
16. Henao CA, Ferrer JC, Muñoz JC, Vera J (2016) Multiskilling with closed chains in a service
industry: a robust optimization approach. Int J Prod Econ 179:166–178
17. Mercado YA, Henao CA, González VI (2021) A two-stage stochastic optimization model for
the retail multiskilled personnel scheduling problem: A k-chaining policy with k≥2. Math
Biosci Eng 18(5):1–25
18. Fontalvo Echavez O, Fuentes Quintero L, Henao CA, González VI (2021) Two-stage stochastic
optimization model for personnel days-off scheduling using closed-chained multiskilling struc-
tures. In: Rossit DA, Tohmé F, Mejía Delgadillo G (eds) Production research. ICPR-Americas
2020. Communications in computer and information science, vol 1407, 19–32. Springer, Cham.
https://doi.org/10.1007/978-3-030-76307-7_2
19. Abello MA, Ospina NM, De la Ossa JM, Henao CA, González VI (2021) Using the k-chaining
approach to solve a stochastic days-off-scheduling problem in a retail store. In: Rossit DA,
Tohmé F, Mejía Delgadillo G (eds) Production research. ICPR-Americas 2020. Communica-
tions in computer and information science, vol 1407, 156–170. Springer, Cham. https://doi.
org/10.1007/978-3-030-76307-7_12
20. Mercado YA, Henao CA (2021) Benefits of multiskilling in the retail industry: k-chaining
approach with uncertain demand. In: Rossit DA, Tohmé F, Mejía Delgadillo G (eds) Production
Research. ICPR-Americas 2020. Communications in Computer and Information Science, vol

1407, 126–141. Springer, Cham. https://doi.org/10.1007/978-3-030-76307-7_10
21. Vergara, S., Del Villar, J., Masson, J., Pérez, N., Henao, C.A., González, V.I. (2021). Impact
of labor productivity and multiskilling on staff management: A retail industry case. In: Rossit,
D.A., Tohmé, F., Mejía Delgadillo, G. (eds.) Production Research. ICPR-Americas 2020.
Communications in Computer and Information Science, vol 1408, 223–237. Springer, Cham.
https://doi.org/10.1007/978-3-030-76310-7_18
22. Powell SG (2000) Specialization, teamwork, and production efficiency. Int J Prod Econ
67(3):205–218
23. Wu Y, Ramachandran K, Krishnan V (2014) Managing cost salience and procrastination in
projects: compensation and team composition. Prod Oper Manag 23(8):1299–1311
24. Chambel MJ, Castanheira F (2012) Training of temporary workers and the social exchange
process. J Manage Psychol
25. Lyons E (2020) The impact of job training on temporary worker performance: field experimental
evidence from insurance sales agents. J Econ Manage Strat 29(1):122–146
26. Van Den Bergh J, Belien J, De Bruecker P, Demeulemeester E, De Boeck L (2013) Personnel
scheduling: a literature review. Eur J Oper Res 226(3):367–385
27. Hung R (1999) A multiple-shift workforce scheduling model under annualized hours. Naval
Res Logist (NRL) 46(6):726–736
28. Corominas A, Lusa A, Pastor R (2004) Planning annualised hours with a finite set of weekly
working hours and joint holidays. Ann Oper Res (1–4):217–233.17
29. Lusa A, Corominas A, Munoz N (2008) A multistage scenario optimisation procedure to plan
annualised working hours under demand uncertainty. Int J Prod Econ 113(2):957–968
30. Hertz A, Lahrichi N, Widmer M (2010) A flexible MILP model for multiple-shift workforce
planning under annualized hours. Eur J Oper Res 200(3):860–873
31. Azmat CS, Widmer M (2004) A case study of single shift planning and scheduling under
annualized hours: a simple three-step approach. Eur J Oper Res 153(1):148–175
32. Corominas A, Lusa A, Pastor R (2007) Planning annualised hours with a finite set of weekly
working hours and cross-trained workers. Eur J Oper Res 176(1):230–239
33. Corominas A, Pastor R (2010) Replanning working time under annualised working hours. Int
J Prod Res 48(5):1493–1515
34. Sureshkumar MR, Pillai VM (2013) An efficient method to reduce relative capacity shortage
using annualised hours planning. Int J Adv Manuf Technol 65(1–4):571–580
35. Hasan MG, Ali I, Hasan SS (2016) Annualized hours planning with fuzzy demand constraint.
In: ProbStat forum, vol 9, pp 50–56
36. Porto AF, Lusa A, Henao CA, Porto Solano R (2021) Annualized hours, multiskilling and
overtime on annual staffing problem: an uncertainty approach. In: XXV Congreso de Ingeniería
de Organización, p 123
37. Birge J, Louveaux F (2011) Introduction to stochastic programming. Springer, New York
Chapter 13
Conceptual Framework for Optimization
Models in Industry 4.0 Context:
Application to Production Planning
Ana Esteso , Andrés Boza , M. M. E. Alemany ,

and Pedro Gomez-Gasquet
Abstract Production planning has traditionally been supported by mathematical

programming. In recent years, Industry 4.0 and the use of its related techniques
have been gaining strength in the production planning field. To develop mathe-
matical programming models integrating Industry 4.0 techniques, it is necessary to
migrate from optimization software normally used in academia to optimization pack-
ages developed in high-level programming languages. This paper aims to propose
a conceptual framework to facilitate this migration by transcribing mathematical
programming models to the programming languages of the MPL commercial opti-
mization software and the optimization package Pyomo developed in Python. This
framework can be used by both, academics and practitioners, to translate already
implemented models in one software to the other, as well as to implement from
scratch a mathematical programming model in any of this software. The proposed
conceptual framework is validated through its use for the translation of a production
planning model implemented in MPL to Pyomo.
Keywords Optimization · Production planning · Industry 4.0 · MPL · Pyomo
A. Esteso (B) · A. Boza · M. M. E. Alemany · P. Gomez-Gasquet

Research Centre On Production Management and Engineering (CIGIP), Universitat Politècnica de
València, Camino de Vera S/N, 46022 Valencia, Spain
e-mail: aesteso@cigip.upv.es
A. Boza
e-mail: aboza@omp.upv.es
M. M. E. Alemany
e-mail: mareva@omp.upv.es
P. Gomez-Gasquet
e-mail: pgomez@cigip.upv.es
https://doi.org/10.1007/978-3-031-29382-5_13
120 A. Esteso et al.
13.1 Introduction
The main purpose of digital transformation is to redesign the organizational busi-

ness through the introduction of digital technologies and achievement benefits such as
productivity improvements, cost reductions, and innovations [1]. To production enter-
prises, Industry 4.0 (I4.0) has become one of the most commented industrial busi-
ness concepts in recent years. This new paradigm in industry promotes autonomous
decision-making, interoperability, agility, flexibility, efficiency, and cost reduction
among others [2].
According to [3], the four main characteristics of Industry 4.0 include: (a) vertical
integration of smart production systems; (b) horizontal integration through global
value chain networks; (c) through-engineering across the entire value chain; and (d)
acceleration of manufacturing. To achieve this, new technological trends emerge such
as big data and analytics, horizontal and vertical system integration, the Industrial
Internet of Things, or augmented reality.
Production planning systems must adapt to these new integrated smart produc-
tion systems. These new industrial environments include [4]: (a) Internet of Things:
objects with capabilities that allow them to communicate with one another and with
other devices and services; (b) Smart Data: a large amount of data that need to be
saved, processed and analyzed; (c) Advanced Processing Analytics: massive amounts
of detailed data can be combined and analyzed by predictive analytics, data mining,
simulation or statistics. Relevant advanced processing analytics-based technologies
are artificial intelligence, predictive analytics, or simulation models.
Thus, production planning systems must continue to reinforce their vertically
integrating role in organizations but making use of this new set of technologies. In
this context, this paper proposes a conceptual framework (CF) to facilitate the migra-
tion of mathematical programming models implemented in optimization software to
optimization packages developed in high-level programming languages that allow
their integration into the techniques mentioned above.
The rest of the paper is structured as follows. Section 2 reflects on the use of opti-
mization software in the field of production planning. Section 3 compares the MPL
software and Pyomo optimization package. In Sect. 4, a CF is proposed that allows
the transcription of mathematical programming models (MPM) between both opti-
mization software. Section 5 validates this framework by applying it to a production
planning model. Finally, Sect. 6 displays conclusions and future research lines.
13.2 Modeling of Production Planning Problems
Mathematical programming, and more specifically linear programming, has been

traditionally employed to support the production planning process [5]. To formulate
and solve this type of models, there are various optimization software on the market.
13 Conceptual Framework for Optimization Models in Industry 4.0 … 121
Fig. 13.1 Optimization software
Figure 13.1 shows some of the most used commercial and free optimization soft-
ware in the academic and professional environment. Most of this software is based
on the algebraic modeling language, whose advantage is the similarity between its
syntax and the notation of optimization problems. Algebraic modeling languages
have proven to be the most efficient method of formulating and maintaining opti-
mization models because they are easier to learn, quicker to formulate, and require
less programming [6].
It can be observed in literature how, in the academic sphere, commercial software
has been used to a greater extent. Commercial optimization software usually provides
free licenses to academics so they can use the software without limitations. However,
the academic nature of these licenses makes it difficult to transfer models developed in
the academy to the business sector, where paying for a license may not be profitable. In
addition, research projects tend to request both, the developed models, and their code,
to be openly published. To meet this requirement, it makes sense for the academy to
migrate to free software that allows the free dissemination of MPM and their codes.
On the other hand, tools based on artificial intelligence have begun to be developed
to support production planning. In the context of the project “Integration of Decision
Making of the Tactical-Operational Levels for the Improvement of the Efficiency of
the Productive System in Industry 4.0 Environments” in which this paper is devel-
oped, it is intended to build tools that combine MPM with heuristics and artificial
intelligence algorithms. For this, it is important to work with software that allows us
not only to solve optimization models but also to connect them with heuristics and
artificial intelligence algorithms. This will facilitate the later interoperability with
other systems and programming the required agility and flexibility achieving more
efficiency, and cost reduction that is in line with I4.0. In the case of free software
such as Pyomo and JuMP, they are optimization packages developed in Python and
Julia high-level programming languages, respectively, so the integration of MPM
with other techniques is facilitated.
From this point, the paper focuses on the commercial software MPL that has
been widely used both in teaching and research contexts, and the free optimization
Table 13.1 MPL and Pyomo main characteristics

MPL Pyomo
License Free academic license Open source
Code structure Grouped in sections Plain code
Model type Concrete or abstract model Concrete or abstract model
Opt. problems LP, MILP, QP, NP, MINLP… LP, MILP, QP, NP, MINLP…
Database connection Connection with Microsoft Through data portals: tab, csv,
Access, Excel, ODBC, Paradoc, json, yaml, xml, excel, dat, …
FoxPro, Dbase, SQL Server,
Oracle
Solver integration Commercial and open-source Commercial and open-source
solvers: CPLEX, GUROBI, solvers written in Python or in
GLPK, … compiled, low-level languages
Result prints Printed in .sol file in matrix Unattractive printing on console
format
Integration with other Not possible in the software Possible to integrate Pyomo with
techniques Python programming
package Pyomo for Python, since it is one of the most used programming languages
and allows a faster transfer of academic work to the industry given its open nature.
13.3 Comparison of MPL and Pyomo Optimization

Software
Table 13.1 collects the main characteristics of the MPL optimization software and
the optimization package Pyomo developed in Python.
13.4 Conceptual Framework for the MPM Transcription
In this section, a CF to facilitate the transcription of MPM to MPL and Pyomo

languages is proposed (Fig. 13.2). This CF, which is defined for the implementation
of abstract models, can be used as a reference tool to: a) Implement a mathemat-
ical programming model from scratch in the programming languages of the MPL
optimization software and Pyomo optimization package; b) transcribe a model previ-
ously implemented in one of the programming languages (MPL or Pyomo) to the
other one.
As seen in Fig. 13.2, the MPL optimization software groups the code into different
sections, so that each element of the MPM must be defined in the corresponding MPL
section. It is remarkable that each line of code is separated by a semicolon. On the
Fig. 13.2 Conceptual framework to transcribe MPM to MPL and Pyomo languages
other hand, MPL has the section MACROS where all the elements that make up
the objective function are calculated, and where different indicators can also be
calculated to facilitate the subsequent analysis of obtained solutions.
With respect to the language used in Pyomo, it is not organized into sections,
but rather different instructions are used for the different elements of an MPM. It
is remarkable that there is no instruction to declare set of indexes in Pyomo, so
binary parameters should be defined to contemplate such sets. Unlike MLP, there is
only one instruction to define decision variables in Pyomo. Therefore, it is necessary
to indicate which is the domain of the variable (continuous, integer, Boolean). In
addition, when working with non-negative variables, it is necessary to define its
lower bound (lb) to zero. In case, there is no upper bound for the variable (ub), and
the upper bound should be fixed to none. On the other hand, the expressions where
the different elements that make up the objective function are calculated as well as
the objective and constraints instructions are defined through a rule, which must also
be defined in the Pyomo code.
It is remarkable that the model developed in Pyomo can be part of a larger code
developed in Python, which can also contain other modules related to metaheuristics
or other techniques typical of Industry 4.0, thus facilitating its integration.
13.5 CF Application to a Production Planning Model
To validate the proposed CF, it is applied to transcribe a production planning model

implemented in MPL to Pyomo language. Table 13.2 shows the nomenclature used
Table 13.2 Nomenclature

Parameters
dit Expected demand for product i at period t
pci j Cost for producing one unit of product i at machine j
pri j Production rate of product i at machine j
oci Cost of outsourcing one unit of product i
hci Holding cost for one unit of product i
ca jt Capacity of machine j at period t
fp Production of the family of products in the planning horizon
Decision variables
X i jt Production of product i at machine j at period t
Oit Outsourced production of product i at period t
I nvit Inventory of product i at period t
by the MPM, where i refers to products, j to machines, and t to periods (in this case,
months).
The MPM is formulated as follows, where the objective is to minimize production,
outsourcing, and holding costs (1). Production and outsourcing of products over the
horizon should equal to the production required for the family of products (2). This
constraint connects the model with a hierarchically superior one. The time required
to produce products at each machine cannot exceed its capacity (3). Finally, the
expected demand should be met with produced, outsourced, or in stock products (4).
∑∑∑ ∑∑ ∑∑
MinZ = pci j ∗ X i jt + oci ∗ Oit + hci ∗ I nvit (13.1)
i j t i t i t
subject to:
∑∑∑ ∑∑
X i jt + Oit = f p (13.2)
i j t i t
∑ X i jt
≤ ca jt ∀ j, t (13.3)
i
pri j
∑
I nvit = I nvit−1 + X i jt + Oit − dit ∀i, t (13.4)
j
Figure 13.3 shows the presented model implemented in MPL (left side) and its
transcription to Pyomo language (right side). The relationship between the different
sections of the model in both languages is established to reveal their equivalences.
It should be noted that it is necessary to assign data to the defined indexes and
parameters so the model can be solved. As previously mentioned, both MPL and
MPL PYOMO
INDEXES CONF
TITLE from pyomo.environ import *
ProdPlanning; m=AbstractModel('ProdPlanning')
INDEX #INDEX
i; m.I=Set()
i; m.J=Set()
t; m.T=Set()
DATA #DATA
d[i,t]; m.d=Param(m.I, m.T)
PARAMETERS
pc[i,j]; m.pc=Param(m.I, m.J)

pr[i,j]; m.pr=Param(m.I, m.J)
oc[i]; m.oc=Param(m.I)
hc[i]; m.hc=Param(m.I)
ca[j,t]; m.ca=Param(m.J,m.T)
fp; m.fp=Param()
INTEGER VARIABLES #VARIABLES
X[i,j,t]; m.X=Var(m.I, m.J, m.T,
VARIABLES
domain=Integers,bounds=(0,None))
O[i,t]; m.O=Var(m.I, m.T, domain=Integers,
bounds=(0,None))
Inv[i,t]; m.Inv=Var(m.I, m.T, domain=Integers,
bounds=(0,None))
MACROS #EXPRESSIONS
ProdCost := def PC_r(m):
SUM(i,j,t:pc*X); return sum(m.pc[i,j] * m.X[i,j,t] for i
OBJECTIVE FUNCTION ELEMENTS
in m.I for j in m.J for t in m.T)

m.ProdCost = Expression(rule=PC_r)
OutCost := def OC_r(m):
SUM(i,t:oc*O); return sum(m.oc[i]*m.O[i,t] for i in m.I
for t in m.T)
m.OutCost = Expression(rule=OC_r)
InvCost := def IC_r(m):
SUM(i,t:hc*Inv); return sum(m.hc[i]*m.Inv[i,t] for i in
m.I for t in m.T)
m.InvCost = Expression(rule=IC_r)
TotCost := def TC_r(m):
ProdCost+OutCost+InvCost return m.ProdCost+m.OutCost+m.InvCost
m.TotCost = Expression(rule=TC_r)
MODEL #OBJECTIVE
OBJECTIVE
FUNCTION
Min TotCost; def obj_r(m):

return m.TotCost
m.obj = Objective(rule=obj_r, sense=minimize)
SUBJECT TO #CONSTRAINTS
C1: SUM(i,j,t:X) + def C1_r(m):
SUM(i,t:O) = fp; return sum(m.X[i,j,t] for i in m.I for j
in m.J for t in m.T)+ sum(m.O[i,t] for i
in m.I for t in m.T) == m.fp
m.C1 = Constraint(rule=C1_r)
C2[j,t]: SUM(i:X/pr) <= def C2_r(m,j,t):
CONSTRAINTS
ca; return sum(m.X[i,j,t]/m.pr[i,j] for i in

m.I) <= m.ca[t]
m.C2 = Constraint(m.J,m.T,rule=C2_r)
C3[i,t]: Inv = Inv[t:=t- def C3_r(m,i,t):
1] + SUM(j:X) + O - d; if t==1:
return I[i,t] == sum(m.X[i,j,t] for j in
m.J) + m.O[i,t] - m.d[i,t]
else:
return I[i,t] == I[i,t-1] + m.O[i,t] +
sum(m.X[i,j,t] for j in m.J) - m.d[i,t]
END m.C3 = Constraint(m.I,m.T,rule=C3_r)
Fig. 13.3 CF application to a production planning problem

Pyomo languages allow data to be read from different types of files and databases.
In addition, the instructions to be used for reading data differ depending on the type
of the datafile or database used.
As an example, Fig. 13.4 shows how to connect MPL with a database created in
Microsoft Access, as well as the reading of the index i and parameter dit . To do this,
a section called OPTIONS is first included between the TITLE and INDEX sections.
There, the type and name of the database used are defined. On the other hand, when
declaring indexes and parameters, the instruction DATABASE(“name1'' ,” name2'' )
is included, where name1 and name2 indicate the table and column from which the
data should be imported. Once the data reading is defined in MPL, the model can be
solved.
Figure 13.5 shows how to read data in Pyomo from .tab files as well as the reading
of the index i and parameter dit . A Data Portal is created once the whole abstract
model is defined. Later the data is loaded using the instructions shown in the Figure,
which differ depending on whether the data refers to indexes or parameters. After
reading the input data, the model should be instantiated and the instructions for its
resolution and printing of results must be included. In this example, the glpk solver
is called to solve the model.
Note that names assigned to the databases, files, and tables as well as other
elements colored in red in Figs. 13.2 and 13.3 are defined by the model designer
and do not necessarily have to be those shown in these examples.
OPTIONS
DatabaseType = ACCESS;
DatabaseACCESS = “Planning.mdb”;
INDEX
i:=DATABASE(“Index_i”,”i”);
DATA
d[i,t]:=DATABASE(“Par_d”,”d”);
Fig. 13.4 MPL connection to Microsoft Access database
dp = DataPortal()
dp.load(filename='Index_i.tab',set=m.I)
dp.load(filename='Par_d.tab',param=m.d,index=m.I,m.T)
inst = m.create_instance(dp)
opt = SolverFactory('glpk')
results = opt.solve(inst)
inst.solutions.store_to(results)
print(results)
Fig. 13.5 Pyomo connection to .tab files and resolution instructions

13.6 Conclusions
In this paper, mathematical programming has been identified as the main tool to
support production planning as well as the need to combine this type of models
with other resolution tools in the context of Industry 4.0 such as heuristics or arti-
ficial intelligence algorithms. This paper proposes a CF that allows academics and
practitioners to migrate from the commercial optimization software MPL to the free
optimization package Pyomo. This migration is supported by the incipient character-
istics of Industry 4.0, which not only require optimizing decision-making, but also
creating tools that allow continuous improvement of its results, taking into account
the possible variations in the environment.
In future lines of research, the connection between different optimization models
implemented in the Pyomo environment will be investigated, as well as their integra-
tion with heuristics to compose the well-known metaheuristics and their connection
with algorithms related to artificial intelligence.
Acknowledgements This research has been funded by the Fondo Europeo de Desarrollo Regional
(FEDER)/Ministerio de Ciencia e Innovación (MCI)—Agencia Estatal de Investigación (AEI) of
Spain, in the framework of the project entitled “Integración de la Toma de Decisiones de los Niveles
Táctico-Operativo para la Mejora de la Eficiencia del Sistema de Productivo en Entornos Industria
4.0 (NIOTOME)” (Ref. RTI2018-102020-B-I00).
References
1. Hess T, Matt C, Benlian A, Wiesböck F (2016) Options for formulating a digital transformation
strategy. MIS Q Exec 15:123–139
2. Pérez Perales D, Alarcón Valero F, Boza García A (2018) Industry 4.0: a classification scheme.
In: Closing the gap between practice and research in industrial engineering, pp 343–350
3. Gilchrist A (2016) Introducing Industry 4.0. In: Industry 4.0. Apress, Berkeley, CA, pp 195–215
4. Boza A, Alarcón F, Perez D, Gómez-Gasquet P (2019) Industry 4.0 from the supply chain
perspective. In: Technological developments in Industry 4.0 for business applications. pp 331–
351
5. Bagshaw KB (2019) A review of quantitative analysis (QA) in production planning decisions
using the linear programming model. Am J Oper Res 09:255–269
6. Maximal software—MPL. http://www.maximalsoftware.com/mpl/
Chapter 14
Artificial Intelligence Techniques Applied
to the Flowshop and Jobshop
Problems. A Review of Recent Literature
Pedro Gomez-Gasquet , Alejandro Torres , Ana Esteso ,

and Maria Angeles Rodriguez
Abstract This paper makes a review of the different contributions that the artificial
scheduling in flowshops and jobshops using approaches or techniques from the disci-
pline of artificial intelligence. High quality and growing production is maintained
for a promising line of research. The work carries out a literature review following a
contrasted methodology that has allowed us to analyze in detail 17 articles published
in 2020 and 2021 as a continuation of a previous review. The most outstanding conclu-
sions allow us to see a tendency to improve the use of deep learning techniques in
the field of sequencing, both to generate sequences directly and to complement parts
of other algorithms.
Keywords AI · Machine learning · Jobshop · Flowshop · Scheduling
14.1 Introduction
For almost 80 years the discipline of operations research (OR) has tackled produc-
tion scheduling problems, incorporating in each decade the contributions that have
been made from other areas (modeling, statistics, biology, etc.) with which the disci-
pline has been growing and consolidating her body. After a strong start, artificial
intelligence (AI) declined and it seemed not to be able to solve the returns that were
being tested. However, in recent years it has been able to solve these aspects, and
at this time, we could say that it seems to provide a different vision that is being
incorporated into many disciplines. How could it be otherwise, the OR is choosing
to incorporate approaches and techniques that come from AI in many of its papers.
P. Gomez-Gasquet (B) · A. Esteso · M. A. Rodriguez

Centro de Investigación de Gestión e Ingeniería de Producción (CIGIP), Universitat Politècnica de
Valencia, Cno/ de Vera S/N, 46022 Valencia, Spain
A. Torres
Departamento de Organización de Empresas, Universitat Politècnica de Valencia, Cno/ de Vera
S/N, 46022 Valencia, Spain
https://doi.org/10.1007/978-3-031-29382-5_14
130 P. Gomez-Gasquet et al.
This work is carried out review of the important volume of scientific progress
published in 2020 and early 2021 as a continuation of a previous one [1], in which the
contributions that from an OR perspective included techniques clearly coming from
AI were analyzed, where we already emphasized that most of the initial contributions
were focused on machine learning, although the trend in the last years it was to use
reinforced learning and neural networks. In this work, the problem of the flowshop
and jobshop in its different configurations is addressed [2].
14.2 Methodology
Once the problem has been identified, a methodology based on four steps is proposed
for its study [3]: (1) select the sources of information; (2) collect contributions; (3)
select the most appropriate contributions; (4) analyze the results. In this case, the
main source of information has been Scopus, since it includes the highest quality
contributors in this field and has a large number of entries. For the compilation, the
following keywords have been used: “artificial intelligence”, “machine learning”,
and “reinforcement learning”, on the one hand, in combination with “flowshop” and
“jobshop”, and on the other hand, in combination with “production scheduling”. In
the next step, papers older than ten years were discarded, as well as work not directly
related to sequencing in the context of our problem. Within this same step, duplicate
papers were discarded. Finally, and once the trend in the last decade was analyzed,
we focused only on the works published in 2020 and 2021, where we obtained a total
of 17 articles of interest.
14.3 Analysis of the Results
A first analysis of the results obtained can be seen in Fig. 14.1, which shows a
temporal evolution of the number of annual articles that have been found on the
subject in question, differentiated by keyword for a better appreciation. The graph
shows a clear increase in the number of publications over time, so that it can be
associated with a greater interest in this topic. It is also interesting to differentiate
the amounts published by type of workshop, which can be seen in the interior graph.
Regarding the contents, a greater interest in problems related to jobshops can be
observed and also that technologies related to neural networks are an almost common
factor in the different proposals. Another interesting aspect is the fields of learning
that are identified as “machine learning” and “reinforcement learning” or the most
used algorithms such as those already mentioned associated with neural networks or
those of Q-learning, k-means, etc. In short, a first analysis makes us optimistic about
the evolution of these techniques in the field of sequencing, but it also alerts us to a
certain bias in the methods used where there seem to be clear favorites.
14 Artificial Intelligence Techniques Applied to the Flowshop and Jobshop … 131
Fig. 14.1 Analysis of the temporal evolution of the papers
Table 14.1 Selected contributions for detailed analysis

Supervised Unsupervised Reinforcement learning
Q-learning NN
Flowshop Guo et al. [4], Pan et al. [7], Yang
Oztop et al. [5], et al. [8], Ren et al.
Zhao et al. [6] [9]
DNN Clustering Q-learning DQN
Jobshop Kramer et al. [10] Lara-Cardenas et al. Wang [11], Wang Altenmüller et al.
[14] et al. [12], Chen [17], Luo et al. [18],
et al. [13], Park et al. [19], Liu
Lara-Cardenas et al. et al. [20]
[14], Kardos et al.
[15],
Martines-Jumenez
et al. [16]
The last step of the proposed methodology consists of a detailed analysis of the
selected proposals. This will allow us to focus on the most relevant papers and with
the greatest contribution to the selected topic.
Table 14.1 shows the most detailed references in relation to the subject under
study. They will be discussed in detail below.
14.3.1 Analysis of the Contributions on the Flowshop

Problem
Contributions made during 2020, on the problem of the flowshop, are mainly focused
on the application of reinforced learning (RL). Papers [4–6] use the more traditional
approach based on a Q-learning algorithm in order to minimize makespan. In the

proposal of [4], a hybrid flowshop problem is addressed, the decisional system is
modeled as a Markov chain, where the state is denoted as a tuple (stage, workpiece),
the action is the machines in the next stage. The reward function is based on the
completion time of each job, and the greedy rule depends on the number of iteration,
going from more to less. In the contribution [5], a no-idle flowshop is addressed
and a general variable neighborhood search is used in such a way that a Q-learning
algorithm is executed in each iteration in the inner loop. This algorithm allows to
adjust the parameters of the shaking and construction procedures of the GVNS. In the
case of [6], a distributed no-idle flowshop is addressed using the “Cooperative Water
Wave” optimization algorithm. In its first phase, a system based on reinforcement
learning is used, which in order to balance the exploration and exploitation stages is
framed within a VNS heuristic. In the second phase, the algorithm tries to improve
the operations with sequences that use combined routes.
Within the scope of RL [7–9] introduce the use of neural networks, thus config-
uring what is known as deep reinforced learning. Authors [7] and [8] are using the
well-known critic-actor method. Through this proposal, the system tries to reach a
better result using two different roles. The “actor” makes decisions, and the “critic”
evaluates the value function and informs the actor of possible improvements. A TD
approach based on policy gradient is followed. Paper [7] addresses a permutation
flowshop problem in order to minimize the makespan. The authors propose the use of
the actor-critic approach generating two networks that adopt these roles. The actors
use an encoding–decoding model taking as input the processing times of the jobs. The
network of actors generates a solution in which the sequence of jobs to be processed
in the different resources is arranged. The balance is offered by the network of crit-
ical agents that, using job properties, evaluate the result of the value function. Paper
[8] faces a dynamic permutation flowshop where the objective is minimizing the
tardiness cost. The authors propose a neural network that is fed with the processing
times of the jobs and the states of the machines. The network calculates a value that
is used to choose a dispatch rule that allows choosing the next job to be scheduled.
Once scheduled, it obtains the new status of the workshop and the neural network is
updated, starting the process again.
Although the proposal [9] also deals with a flowshop with the objective function of
minimizing the makespan, this is divided into stages, establishing a different decision
for each stage. For its resolution, reinforced learning is proposed in which there are
nine possible states. The states always contain updated information on the status of
jobs on each machine. The authors propose the use of State and Action Network
(SA-NET) that allows to relate states and actions of the neural network (NN). When
facing a problem with 5 machines, the proposed network has 45 nodes (5 machines
× 9 states) in each layer, with a final network in which 5 actions are obtained, one
for each machine.
14.3.2 Analysis of the Contributions on the Jobshop Problem
In [10], authors focus on regression methods within supervised learning. The methods
learn to estimate the processing times of each job from a set of problem data. Although
sequences are not obtained, they help to assess possible solutions. After applying
various methods to a real example, it can be deduced that deep neural networks offer
the best results in terms of accuracy in predicting the optimal lead time.
In [11] and [12], basic Q-learning approach is proposed. In [11], it is proposed a
method based on dynamic greedy search as a means of providing more information to
subsequent phases, obtaining an approximate value of the operation time. Likewise,
the proposal deals with a modification of the value function (Q) using fuzzy logic
approximations in the calculation. In [12], Q-dual learning method with unknown
assembly times is proposed. The top level of Q-learning is in charge of finding the
optimal sequencing to minimize machine downtime and balance machine workload.
Once all the jobs are associated with the corresponding machines, the lower level
of Q-learning focuses on the global objectives to establish a scheduling policy that
allows minimizing the completion time of all the jobs.
Other contributions [13, 14] use Q-learning to improve a heuristic. In [13], the
authors try to minimize the completion time (makespan) of the jobs in a workshop.
A proprietary genetic learning algorithm is established that optimally adjusts the
key parameters used in genetic algorithms based on reinforced learning (Pc and
Pm, crossover, and mutation probabilities). In [14], the research focuses on opti-
mizing hyperheuristics in workshops, that is, the selection of the optimal heuristics
depending on the step of the problem in which they are. Unsupervised learning tries
to group the “items” that have common properties (it is known as clustering), and for
this, the k-means algorithm is used. Instead, reinforced learning focuses on learning
the characteristics of the problem, using Q-learning.
In [15] and [16], a multiagent reinforcement learning tool is proposed. In [15], the
goal is the reduction of lead times in a jobshop type production system. To implement
reinforced learning, a simulation model is defined, composed of a set of machines
(SP) with an infinite warehouse at the entrance of these and a series of intelligent
products (IP) whose processing and preparation times are known. In [16], the authors
propose a tool that makes it possible to obtain a very good first solution to the problem
using the Q-learning algorithm and even, after the first solution, to impose particular
conditions or additional restrictions to adapt the problem and once again obtain an
optimal solution to the sequencing problem is proposed.
Finally, [17–20] employ what is known as a Deep Q-Network (DQN) agent, which
is a well-defined algorithm for combining reinforced learning and neural networks.
In [17], the state space defined in this paper is quite rich and has 210 entries (current
machine status, job status, etc.) As a result of the process, the agent selects an output
among the 20 possible actions or idle action. As a reward, the agent receives a value
that depends on the time constraints, and that is also used in the reference heuristics.
These restrictions are valued according to their impact so that the greater the number
of violations, the less interesting the solution. For similar solutions, work in progress
is a second criterion used. In [18], authors address the problem of the dynamic
workshop with the aim of minimizing the maximum delay. The DQN is improved
with a double DQN and the “soft card weight update”. As data input to DQN, 7
(rules) with values between 0 and 1 are proposed to make the problem approachable
and that the DQN can be extrapolated directly to other production environments. In
[19], authors try to minimize the makespan in the sequencing problem, and a setup
change scheduling method is proposed using reinforcement learning (RL) in which
each agent determines the configuration decisions in a decentralized way and learns
a centralized policy sharing a neural network between agents to cope with changes in
the number of machines. In [20], the work does not take into account preparation times
or delivery dates and comparing the developed method with various alternatives such
as heuristic rules, Q-learning, and genetic algorithm. The authors propose a method
based on deep reinforcement learning in which uses an actor-critic network as a
learning algorithm to obtain the predictions of the results and the behavior of the
system. The system is shown to be flexible in the face of changes and can deal with
problems such as machine breakdowns and unexpected job changes.
14.4 Conclusions and Future Research
The work that has been presented has tried to analyze the contributions made to
date on the selected problem, focusing especially on those published in 2020. The
publications selected in the period under review have not significantly changed the
focus of interest in techniques that have been widely used for many years in the
field of sequencing, such as neural networks or reinforced learning, in particular
Q-learning algorithms. Although it may be somewhat circumstantial, we have seen
that the number of publications in such a short period is high compared to those
found in [1] in a much longer period. In any case, it can be concluded that the
contribution of AI to OR is gradually consolidating the focus on reinforced learning
and neural network techniques, but above all the use of Deep Q-Network is beginning
to appear strongly. Due to the upward trend and the quality of the results obtained,
everything seems to indicate that in the near future this is going to be a work area of
maximum interest. The development of the general concept of deep learning has led
to the adaptation and creation of methods and algorithms that experts are applying
to the field of sequencing, especially in more complex problems such as jobshop
scheduling.
After conducting this review and also considering the reality previously [1], the
authors of this paper believe that the current AI techniques mentioned are relatively
mature to be successfully applied to flowshop and jobshop problems. However, we
believe that there is still a lot of work to be done in terms of modeling sequencing
problems in the terms that AI needs. In other words, we believe that the AI methods
analyzed as applied are not achieving the desirable results in terms of quality achieved
in efficiency measures, nor in terms of computational speed, let alone simplicity.
However, a new way of formulating the same sequencing problem in other terms
would perhaps make it possible to move forward and correct the deficiencies we
have discussed using the same AI techniques.
Acknowledgements This research is being funded by project entitled NIOTOME (Ref. RTI2018-
102020-B-I00) (MCI/AEI/FEDER, UE). The author María Ángeles Rodríguez was supported by
the Generalitat Valenciana (Conselleria de Edu-cación, Investigación, Cultura y Deporte) under
Grant-Agreement ACIF/2019/021.
References
1. Gomez-Gasquet P, Boza-Garcia A, Navarro A (2020) Artificial intelligence for solving flow-

shop and jobshop scheduling problems: a literature review. In: 14th international confer-
ence on industrial engineering and industrial management, XXIV Congreso de Ingeniería de
Organización, Madrid, Spain, July 9–10, 2020 (publication in progress)
2. Brucker P (2007) Scheduling algorithms, 5 edn. Springer, Heidelberg. https://doi.org/10.1007/
978-3-540-69516-5
3. Medina-López C, Marín-García JA, Alfalla-Luque R (2010) Una propuesta metodológica para
la realización de Búsquedas sistemáticas de bibliografía. (A methodological proposal for the
systematic literature review). In: Working papers on operations management, vol 1, no 2, pp
13–30
4. Guo F, Li Y, Liu A, Liu Z (2020) A reinforcement learning method to scheduling problem of
steel production process. J Phys Conf Ser. https://doi.org/10.1088/1742-6596/1486/7/072035
5. Oztop H, Tasgetiren MF, Kandiller L, Pan QK (2020) A novel general variable neighbor-
hood search through Q-learning for No-Idle flowshop scheduling. In: 2020 IEEE congress on
evolutionary computation, CEC 2020—conference proceedings. https://doi.org/10.1109/CEC
48606.2020.9185556
6. Zhao F, Zhang L, Cao J, Tang J (2021) A cooperative water wave optimization algorithm with
reinforcement learning for the distributed assembly no-idle flowshop scheduling problem.
Comput Ind Eng. https://doi.org/10.1016/j.cie.2020.107082
7. Pan R, Dong X, Han S (2020) Solving permutation flowshop problem with deep rein-
forcement learning. In: Proceedings—2020 prognostics and health management conference,
PHM-Besancon 2020. https://doi.org/10.1109/PHM-Besancon49106.2020.00068
8. Yang S, Xu Z, Wang J (2021) Intelligent decision-making of scheduling for dynamic permuta-
tion flowshop via deep reinforcement learning. Sensors (Switzerland). https://doi.org/10.3390/
s21031019
9. Ren J, Ye C, Yang F (2021) Solving flow-shop scheduling problem with a reinforcement
learning algorithm that generalizes the value function with neural network. Alex Eng J. https://
doi.org/10.1016/j.aej.2021.01.030
10. Kramer KJ, Wagner C, Schmidt M (2020) Machine learning-supported planning of lead times
in job shop manufacturing. IFIP Adv Inf Commun Technol. https://doi.org/10.1007/978-3-030-
57993-7_41
11. Wang YF (2020) Adaptive job shop scheduling strategy based on weighted Q-learning
algorithm. J Int Manuf. https://doi.org/10.1007/s10845-018-1454-3
12. Wang H, Sarker BR, Li J, Li J (2020) Adaptive scheduling for assembly job shop with uncertain
assembly times based on dual Q-learning. Int J Prod Res. https://doi.org/10.1080/00207543.
2020.1794075
13. Chen R, Yang B, Li S, Wang S (2020) A self-learning genetic algorithm based on reinforcement
learning for flexible job-shop scheduling problem. Comput Ind Eng. https://doi.org/10.1016/j.
cie.2020.106778
14. Lara-Cardenas E, Silva-Galvez A, Ortiz-Bayliss JC, Amaya I, Cruz-Duarte JM, Terashima-

Marin H (2020) Exploring reward-based hyper-heuristics for the job-shop scheduling problem.
In: 2020 IEEE symposium series on computational intelligence, SSCI 2020. https://doi.org/10.
1109/SSCI47803.2020.9308131
15. Kardos C, Laflamme C, Gallina V, Sihn W (2021) Dynamic scheduling in a job-shop production
system with reinforcement learning. Proc CIRP. https://doi.org/10.1016/j.procir.2020.05.210
16. Martínez Jiménez Y, Coto Palacio J, Nowé A (2020) Multi-agent reinforcement learning tool
for job shop scheduling problems. Commun Comp Inf Sci. https://doi.org/10.1007/978-3-030-
41913-4_1
17. Altenmüller T, Stüker T, Waschneck B, Kuhnle A, Lanza G (2020) Reinforcement learning for
an intelligent and autonomous production control of complex job-shops under time constraints.
Prod Eng Res Devel. https://doi.org/10.1007/s11740-020-00967-8
18. Luo S (2020) Dynamic scheduling for flexible job shop with new job insertions by deep
reinforcement learning. Appl Soft Comput J. https://doi.org/10.1016/j.asoc.2020.106208
19. Park IB, Huh J, Kim J, Park J (2020) A reinforcement learning approach to robust scheduling of
semiconductor manufacturing facilities. IEEE Trans Autom Sci Eng. https://doi.org/10.1109/
TASE.2019.2956762
20. Liu CL, Chang CC, Tseng CJ (2020) Actor-critic deep reinforcement learning for solving job
shop scheduling problems. IEEE Access. https://doi.org/10.1109/ACCESS.2020.2987820
Chapter 15
Design and Implementation
of an Experimentation Service
of the Production Scheduling Algorithms
María Ángeles Rodríguez , Pedro Gomez-Gasquet , Llanos Cuenca ,

and M. M. E. Alemany
Abstract This work details, in its design and implementation phase, a service that
by means of a web application (http://niotome.cigip.upv.es/) allows any user to carry
out the execution of a design of experiments for the validation of an algorithm,
heuristic, or metaheuristic, for the production scheduling. The work justifies its need
and the interest it has for researchers. Likewise, the scope of the problems that can
be addressed is defined, and how to use it is shown.
Keywords Production scheduling · Design of experiments · Web service
15.1 Introduction
For almost 80 years, researchers have been analyzing problems related to production
scheduling. It is well known that a large part of the problems addressed can be clas-
sified as NP-Complete, which often leads researchers to propose resolution methods
that do not guarantee optimal solutions, such as heuristics or metaheuristics. Unlike
the methods that provide an optimum, the suboptimal methods should accompany
their formal proposals with a study that shows the validity of the results achieved
and their quality.
M. Á. Rodríguez · P. Gomez-Gasquet (B) · L. Cuenca · M. M. E. Alemany

València, Camino de Vera S/N, 46022 València, Spain
M. Á. Rodríguez
e-mail: marodsa4@cigip.upv.es
L. Cuenca
e-mail: llcuenca@cigip.upv.es
M. M. E. Alemany
e-mail: mareva@cigip.upv.es
https://doi.org/10.1007/978-3-031-29382-5_15
138 M. Á. Rodríguez et al.
The validation of a new algorithm should be accompanied by a design of exper-

iments (DoE) that will be supported with statistical data related to the measure or
measures that are used for its evaluation as is highlighted in [1, 2]. This experimental
study can have a double objective. On the other hand, make the parametric adjust-
ment of the proposed method, in a context where in most cases there are one or more
parameters have a range of candidate values and it is dificult to set the best one. On
the other hand, they would allow algorithms to be compared with each other, with
the idea of establishing a classification against a well-contextualized problem.
However, in many cases the experimentation is not carried out or is not carried
out in an adequate way due to the difficulty of creating a test environment and the
resources that may be necessary for it. In this work, the design and subsequent imple-
mentation of a service through a web application are proposed, which allows imple-
mentation of the execution of all the necessary experiments to obtain the required
results in a previously designed experiment. The service allows you to create experi-
ments for a wide group of production scheduling problems, apply them to algorithms
already available or to your own, and obtain the results when they are available
without waiting and without consuming your own resources. In the next sections,
we explained the different production scheduling problem supported to experimen-
tation, the production scheduling web application for experiments designed and
implemented, and finally conclusions of this research are presented.
15.2 The Production Scheduling Problem Supported
The theory of scheduling is characterized by a virtually unlimited number of problem

types [3–6]. In this work, we have tried to address a large number of problems type
and for this we have based ourselves on the following bases.
Suppose that m machines M j (j = 1, …, m) have to process n jobs J i (i = 1, …, n).
A schedule is for each job and allocation of one or more time intervals to one or more
machines. A job J i consists of a number ni of operations Oi1 , ..., Oi , ni . Associated
with operation Oij is a processing requirement pij . Furthermore, a release date r i , on
which the first operation of J i becomes available for processing may be specified.
Associated with each operation Oij is a set of machines µij ⊆ {M 1 , …, M m }. Oij may
be processed on any of the machines in µij . In this case, all machines are considered
unrelated. No preemption for jobs is allowed, and sequence-dependent setup times
are considered, such S ijm is the time required for setting up the job j after job I in
machine m.
A General Shop or Jobshop model [7] without machine repetition is considered,
assume that µij /= µi , j + 1 for j = 1, ..., ni – 1. Finally, there is an objective function
f i (t) which measures the cost of completing J i at time t. A due date di and a weight
wi may be used in defining f i . In general, all data pij , r i , d i , wi are assumed to be
integer.
Regarding the objective function, several measures have been considered. We
denote the completion time of job J i by C i , and the associated makespan as f max (C)
15 Design and Implementation of an Experimentation Service … 139
= max{f i (C i )|i = 1, ..., n}, the Lateness as max{f i (L i )/L i = C i − d i }, the Earliness
as max{f i (E i )/E i = max{0, d i − C i }}, and the Tardiness as max{f i (T i )/T i = max{0,
C i − d i }}.
Considering all these assumptions, users of this service can use a class written in
Python language. This allows initializing the problem, as first step in each execution
of the experimenter as:
• Type: Machine, flowshop, or jobshop problem. Allows adapted some calculation
methods to make them more efficient.
• And the following True/False parameter to define the problem context. Multi
(True if a set of machines µij ⊆ {M 1 , …, M m } where m > 1), Setup (True if setup
is considered), Ready (True if any machine is not available at the beginning),
Release (True if any job is not available at the beginning), Due date (True if due
date is considered for jobs), and Weight (True if weights are considered for jobs).
It only refers to inform about the type, the rest of the parameters are added False by
default. The second step in the use of this class would be the instantiation of a concrete
problem, which implies loading or generating the concrete data of an instance of the
previously initialized problem. The class allows generating data randomly if it is
indicated in parameters. The required data is the number of jobs and the number of
operations, the rest of the data is entered only if required. For example, due date is
only needed if the due date parameter has been set to True during initialization.
After instantiation, the class is in a position to offer a service for calculating a
schedule and for calculating any of the objective functions previously indicated to
the algorithm that requires it. In addition to the maximum value of C i , L i , S i, and
T i , the average and accumulated values are calculated, as well as with or without
weights associated with the jobs.
15.3 Production Scheduling Web Application

for Experiments
15.3.1 Methodology
The web application (http://niotome.cigip.upv.es/) has been designed and imple-

mented according to Web information system development methodology (WISDM)
[8].
This methodology is divided into five unordered stages: organization analysis,
information analysis, work design, technical design, and human–computer interac-
tion (HCI). Organization and information analysis have achieved the requirements
for describing the functionalities of the web application. Work design evaluates user
satisfaction. The technical design describes the data model and software employed
and HCI design and implements the web interfaces.
In this paper, we detail the functionalities obtained in phases organization and

information analysis; the web software got in stage technical design; and web inter-
faces made in phase HCI. The user satisfaction and data model will be developed in
the future.
15.3.2 Functionalities
As has been mentioned in the Introduction section, the web application is mainly
designed to create experiments for a wide group of production scheduling problems,
apply them to algorithms already available or to your own, and obtain the results.
This translates into basic functionalities that are obtained due to the user
requirements, which are described as follows:
• Initialize the problem (between a wide range of workshops).
• Instantiation of a concrete problem with data (randomly generated by machine or
uploaded by user).
• Calculation of a schedule of the instantiated problem.
• Calculation of any of the objective functions associated with schedule.
• Show results of calculation in Gantt diagram format and the summary of
experimentation.
These functionalities are implemented as methods in Python class explained in
the previous section and the web application called the methods through API Rest
designed.
15.3.3 Web Software
The web software employed is classified into five distinct areas: graphic design, web
programming, database software, web server, and production scheduling algorithms
which are shown in Table 15.1.
15.3.4 Web Interfaces
After describing the functionalities and the software to be used, the web interfaces
were designed. The web interfaces design approach focused on providing the user
with a friendly and intuitive way to run experiments. For that purpose, the interfaces
are intended to be self-explanatory, and it offers extra information through help pop-
up web dialogues, if the user requires more information to understand how to use
the interface.
Table 15.1 Web software used

Area Web software Main use
Graphic design CSS (bootstrap library) Customize the appearance of the
web in responsive mode
Web programming HTML and JavaScript (React Create interactive user
library, Redux library, Webpack interfaces in single page
library) applications (SPAs) method
Database software MongoDB Storage data in NoSQL database
system
Web server Nginx (HTTP server) and Linux Provide web application in
HTTP server, which is hosted in
open-source operating system
Production scheduling Python and API rest Implement class for production
algorithms scheduling algorithms in an
interpreted programming
language. The methods of class
are called from API rest
It is noted that all software used is open source.
The web application starts with the sign in/new user interface where the user logs
in or registers if he/she does not already have an account. This tool can be available
to any researcher, access is not limited. The main screens of the application are
presented below:
• Initial interface: This screen is shown in Fig. 15.1, and it covers the functionality
of initialize the problem. Once you are logged in, the user must choose the type
of problem (machine, flowshop, or jobshop problem) to start the experimenta-
tion. Also, some True/False parameters must be provided (if applicable), such as
multi (if a set of machines µij ⊆ {M 1 , …, M m } where m > 1), setup (if setup is
considered), ready (if any machine is not available at the beginning), release (if
any job is not available at the beginning), due date (if due date is considered for
jobs), and weight (if weights are considered for jobs). Afterward, the user can
choose between generating cases with random data or uploading files to upload
cases already created by the user.
• Generator interface: If the user chooses to generate data randomly to instantiate the
problem, the screen requests several data as can be seen in Fig. 15.2. This interface
(along with the interface of upload files) provides the functionality of instantiation
of a concrete problem with data (randomly generated by machine or uploaded
by user). The generation requires of following mandatory data: number of jobs,
number of operations, and time units. The rest of the data is optional, and only
fill if parameter checked in the initial interface. For example, if multiparameter
is checked, the user should offer the number of machines. Also, the summary of
the instantiation and the possibility to instance more than once is provided on this
screen.
Fig. 15.1 Initial interface
Fig. 15.2 Generator interface

Fig. 15.3 Results interface
• Results interface: After instantiation of the problem, the web application calculates
a schedule and any of the objective functions to the algorithm selected and the
number of execution repetitions previously indicated to the results interface. The
screen displayed in Fig. 15.3 offers the functionality of show results of calculation
in Gantt diagram format and also, the summary of experimentation (number of
instances, algorithm used, number of repetitions, calculation of objective function
and sequence). Moreover, the information can be downloaded in Microsoft Excel
format or printed. It is important to highlight that if it takes a long time for
the results to appear, a pop-up web dialogue will be displayed to indicate the
identification number of the results so that they can be displayed when they are
available in the home interface of the user.
15.4 Conclusions
This paper presents a tool for the execution of experiments in the field of produc-
tion scheduling. This support tool has been designed and implemented as a web
application. Furthermore, has been created for the execution of experiments not for
the design of experiments. With this application, it is possible to evaluate several
production scheduling algorithms and tune up the parameters of these algorithms.
This application web is in developing progress and there are still some details to be
completed. For this purpose, in future work we will evaluate the user satisfaction to
web application using and we will develop the data model associated with algorithms
production scheduling, as well as uploading the web application to the cloud via
Docker containers.
Acknowledgements This research is being funded by Ministerio de Ciencia e Innovación (MCI),

Agencia Estatal de Investigación (AEI), and Fondo Europeo de Desarrollo Regional (FEDER)
project entitled “Integración de la Toma de Decisiones de los Niveles Táctico-Operativo para la
Mejora de la Eficiencia del Sistema de Productivo en Entornos Industria 4.0 (NIOTOME)” (Ref.
RTI2018-102020-B-I00). The author María Ángeles Rodríguez was supported by the Generalitat
Valenciana (Conselleria de Educación, Investigación, Cultura y Deporte) under Grant-Agreement
ACIF/2019/021.
References
1. Blackstone JH, Phillips DT, Gary L. A state-of-the-art survey of dispatching rules for
manufacturing job shop operations. Int J Prod Res. https://doi.org/10.1080/00207548208947745
2. Ruiz R, Stützle T (2007) A simple and effective iterated greedy algorithm for the permutation
flowshop scheduling problem. Euro J Oper Res.https://doi.org/10.1016/j.ejor.2005.12.009
3. Allahverdi A, Gupta JND, Aldowaisan T (1999) A review of scheduling research involving setup
considerations. Omega. https://doi.org/10.1016/S0305-0483(98)00042-5
4. Blazewicz J, Ecker KH, Pesch E, Schmidt G, Weglarz J (1997) Scheduling computer and
manufacturing processes. J Oper Res Soc. https://doi.org/10.1057/palgrave.jors.2600793
5. Graham RL, Lawler EL, Lenstra JK, Kan AHGR (1979) Optimization and approximation in
deterministic sequencing and scheduling: a survey. Ann Discrete Math. https://doi.org/10.1016/
S0167-5060(08)70356-X
6. Pinedo ML (2016) Scheduling: theory, algorithms, and systems, 5th edn. https://doi.org/10.1007/
978-3-319-26580-3
7. Manne AS (1960) On the job-shop scheduling problem. Oper Res. https://doi.org/10.1287/opre.
8.2.219
8. Vidgen R (2002) Constructing a web information system development methodology. Inf Syst J
12(3):247–261.https://doi.org/10.1046/j.1365-2575.2002.00129
Part IV
Product Design, Industrial Marketing
and Consumer Behaviour
Chapter 16
Machine Learning in Online Advertising
Research: A Systematic Mapping Study
María Cueto González , José Parreño Fernández ,

David de la Fuente García , and Alberto Gómez Gómez
Abstract In order to consolidate a study framework on the academic production

about digital marketing and artificial intelligence, this paper aims to provide an
overview of the state of research on this a specific topic and to decide on the axes
where to dig by using a systematic mapping study (SMS) methodology. As extended
scope research areas both fields require to become less general to face a systematic
literature review. For this reason, this study introduces a previous phase in which
an initial systematic mapping study is performed combined with a subsequent text
analysis to obtain the most frequent bigrams in the literature and to narrow down
more specific and interconnected study areas. As a result, online advertising and
machine learning were identified as parameters to perform a final complete systematic
mapping study. The results of this paper allow a framework for all academic produc-
tion about online advertising and machine learning studied together, by providing a
review of this corpus, analyzing annual production rate, sources and cites received.
Keywords Online advertising · Digital marketing · Forecasting · Artificial

intelligence · Machine learning · Systematic mapping study
16.1 Introduction
Online advertising and web analytics have revolutionized the study of customer
behavior. They enable continuous analysis, forecasting, and the creation of user
definition ecosystems. One of many indicators of this online media revolution has
been that, in 2019, online advertising surpassed television historical leadership [1]
for the first time in twenty-six editions of InfoAdex Study of Advertising Investment
in Spain. This movement in the media by investment ranking in Spain remained in
2020, being, in addition, the medium of those so-called controlled media that least
suffered the fall in investment with a decrease of − 5.3% [2].
M. Cueto González (B) · J. Parreño Fernández · D. de la Fuente García · A. Gómez Gómez

Escuela Politécnica de Ingeniería, Campus de Viesques, Gijón, España
e-mail: uo194189@uniovi.es
https://doi.org/10.1007/978-3-031-29382-5_16
148 M. Cueto González et al.
In this context, the present research begins running a superficial initial systematic
mapping study (iSMS) that identifies “online advertising” and “machine learning”
(ML) as the most repeated fields in the joint study of their respective disciplines,
“digital/online marketing/advertising” and “artificial intelligence” (AI). Once this
context has been identified, an advanced systematic mapping study (aSMS) is then
carried out, which aims to generate a detailed review of the existing literature on the
specific study of machine learning in online advertising research.
16.2 Methodology
To have an overview of the state of research in a specific topic and to decide on the
axes where to dig, a literature study is requested, as Rachad and Idri [3] describe.
Thus, as it is quoted in these authors’ paper, “a SMS offers a superficial overview of a
particular topic by providing a count and classification of research works published
in this topic”. The following sections identify, describe, and conduct this type of
study on the proposed research area.
16.2.1 Research Questions
Research questions provide a comprehensive analysis and help to obtain relevant

information in the study area [4]. They are entered in Table 16.1.
16.2.2 Research Process
Following the methodology used by Noorbehbahani et al. [4] and defined by Petersen
et al. [5], the process to carry out an SMS consists of three phases: Identification of
Table 16.1 Research questions

ID Research question
RQ1 What is the context in which digital/online marketing and AI are related in academic
research. Identification of suitable search terms
RQ2 What is the proportion of academic production about online advertising in general
related to ML
RQ3 Which sources have published the most about ML and online advertising
RQ4 How the annual production of publications about ML and online advertising has
evolved
RQ5 What are the most cited publications to date
16 Machine Learning in Online Advertising Research: A Systematic … 149
search strings according to the topic, choice of valid and relevant databases, and
application of search strings to libraries.
In this work, terms related to a primary research area (PRA) have been identified
as parameters, together with terms related to a secondary research area (SRA). As
databases, WOS and Scopus have been chosen since these are the main databases
of bibliographic references and citations sources that provide a complete overview
of world research production. Once all parameters have been defined and databases
have been identified, search strings have been applied to the libraries in two different
systematic mapping studies, identified in Table 16.2 in different columns: initial
systematic mapping study (iSMS) and advanced systematic mapping study (aSMS).
Focusing on the initial systematic mapping study (iSMS), a wide variety of param-
eters related to the primary research area have been identified and applied to both
databases using topic match with parameters related to the secondary research area.
Findings of this first study have contributed to set a corpus of abstracts that had been
used to perform further text analysis on bigrams in R in order to contextualize the
most relevant terms in both primary and secondary research areas.
In a second stage, an advanced systematic mapping studio (aSMS) is performed.
The reason behind this is that it’s not a matter of looking up a specific segment of
the titles included in the iSMS, but instead to generate a new extraction which can
include all kinds of publications that contain as parameters the most relevant terms
identified through the text analysis by bigrams in R, whose results are shown in
Fig. 16.1.
16.2.3 Study Filtering
Raw results need to be filtered to reach a selection of the most relevant publications
in the investigated areas. To do this, a common set of inclusion criteria (IC) and
exclusion criteria (EC) have been identified for both SMS, which is detailed in Table
16.2.
Studies that meet all of the following criteria will be included: (IC1) Include one
or more parameters related to the primary research area based on topic match; (IC2)
Include the parameter related to the secondary research based on the topic too.
In contrast, studies that coincide with one or more of the following criteria will
be excluded: (EC1) Results after 2020, as 2021 is still in progress by the date of
the drafting of this research; (EC2) Results not published in English or Spanish;
(EC3) Duplicate results in WOS and Scopus. A manual exclusion criterion—by
title, abstract, or source quick review—is added at the end of the process (EC4) to
remove some document types, like patents, conference reviews, or non-related titles
to the topic.
150
Table 16.2 Research process and study filtering with results

iSMS aSMS
2.1 Search string definition Primary research área (PRS) (digital-marketing) (online-advertis*)
OR (digital-advertis*) OR (online-ads)
OR (digital-ads)
OR (online-marketing)
OR (online-advertis*)
OR (online-ads)
OR (e-marketing)
OR (e-advertising)
Secondary research área (SRA) AND (artificial-intelligence) AND (machine-learning)
2.2 Choice of databases WOS Scopus WOS Scopus
2.3 Application of search strings to libraries N° % N° % N° % N° %
(IC1) topic: PRS 9355 100 7467 100 3117 100 3007 100
(IC2) topic: SRA 264 2.8 230 3.1 123 3.9 172 5.7
3. Study filtering (EC1) year: > 2020 217 2.3 186 2.5 103 3.3 145 4.8
(EC2) lang.: EN&ES 204 2.2 180 2.4 103 3.3 140 4.7
Unification 384 2.3% 143 4.0%
(EC3) refine: duplicates 341 2.0% 169 2.8%
(EC3) refine: manual 336 2.0% 161 2.6%
M. Cueto González et al.
online advertising
digital marketing
machine learning
artificial intelligence
social media
e commerce
big data
deep learning
social networks
online marketing
neural network
decision making
marketing strategy
data mining
0 50 100
Fig. 16.1 Frequency graph of bigrams in R with abstracts of 336 titles extracted from the iSMS
16.3 Results
At this stage, answers corresponding to each of the research questions defined in the
methodology are detailed in order.
RQ1. What is the context in which digital/online marketing/advertising and
AI are related in academic research. Identification of suitable search terms. To
answer this question, an initial text analysis is carried out on the 336 titles extracted
from the iSMS. Studying this corpus will allow us to visualize the context in which
the two research areas are studied to identify the main suitable terms for a more
specific systematic mapping study, the aSMS.
Figure 16.1 shows the frequency graph of bigrams about all the abstracts of these
336 titles extracted from the iSMS. In this superficial analysis, “online advertising”
is shown as the most recurrent term of those that make up the primary study area, and
“machine learning” as the most recurrent term in the secondary study area. Followed
by the generic terms “digital marketing” and “artificial intelligence”, the difference
in frequency of these two terms with respect to the rest of bigrams provides more
concrete basis to start the main systematic mapping study in this research. The aSMS
will be performed focusing directly on knowing in depth all the studies in the field
of academic research related to online advertising and machine learning.
Likewise, the appearance of “deep learning” (DL) and “neural networks” (NN)
among the fifteen most repeated bigrams strengthens the predominance of machine
learning (ML) as a discipline studied in online/digital marketing/advertising area.
RQ2. What Is the Ratio of Academic Production about ML over Online Adver-
tising general production. Detailed results to answer this research question are
shown in Table 16.1, aSMS column, phases 2.3 and 3. Altogether, 161 titles are
obtained, which represent 2.6% of the total publications in the online advertising
study area. By database, before the unification process, 3.3% (103) of the WOS titles
about online advertising are related to ML, compared to 4.7% (140) of titles extracted
from Scopus.
RQ3. Which Sources Have Published the Most about ML and Online Adver-
tising. In this regard, the top five sources from the aSMS with the greatest number
of publications included have been identified. These have been divided depending
on whether they are WOS Journals (Table 16.3), Scopus Journals (Table 16.4), or
conference papers (Table 16.5).
With three titles registered, Electronic Commerce Research and Applications has
been the journal which publishes the most about the study area and Proceedings of
the World Wide Web Conference, with 15 titles registered, the conference series that
has dealt with the subject the most.
How the Annual Production of Publications about ML and Online Advertising
Has Evolved. Figure 16.2 shows the annual number of publications evolution based
on the aSMS final 161 results. The first study that includes a joint reference to
online advertising and machine learning dates from 2007 [5] being the fourth title
Table 16.3 Top 5 WOS journals with most titles published ordered by JCR 2020
Source JCR2020 JCR5YEAR Q N°
Foundations and trends in information retrieval 8.000 6.611 Q1 2
Expert systems with applications 6.954 6.789 Q1 2
Information processing and management 6.222 5.789 Q1 1
Electronic commerce research and applications 6.014 6.433 Q1/Q2 3
Applied intelligence 5.086 4.602 Q2 1
Table 16.4 Top five Scopus journals with most titles published ordered by SJR 2020
Source SJR2020 CiteScore 2020 N°
Foundations and trends in machine learning 4.292 37.8 1
Journal of intelligent and fuzzy systems 1.851 1.797 1
Neural networks 1.396 10.9 1
Soft computing 0.626 5.1 1
Information systems 0.547 7.3 1
Table 16.5 Top five conferences with most titles published ordered by the number of titles included
Source N°
Proceedings of the World Wide Web Conference 15
Proceedings of the ACM SGKDD Int. Conference on Knowledge Discovery and Data 13
Mining
International Conference on Machine Learning (ICML) 5
Proceedings of the ACM International Conference on Web Search and Data Mining 5
Proceedings of the Annual Int. ACM SIGIR Conference on Research Development in 4
Information Retrieval
with the highest number of citations in the aSMS. Motivated by the needs of search
advertising, the authors propose a methodology for building a practical robust query
classification in search engine traffic, primarily focused on rare queries.
In 2008, machine learning approaches were explored to improve query classifi-
cation accuracy in eBay contextual advertising [6], for targeting users based on their
past behavior maximizing ad network revenue and minimizing user annoyance [7]
and to promote a community of researchers interested in online advertising area and
yield future collaboration and exchanges [8].
2009 is the year in which the first publication appears in a scientific journal [9].
Its authors propose a methodology using a pseudo-relevance feedback technique for
building a robust query classification system that can identify thousands of query
classes, while dealing in real time with the query volume of a commercial Web
search engine.
From 2010 to 2016, both included, the annual production of academic research
in this field is held captive with between 6 and 9 publications per year. However, it
was precisely during this period when the four remaining titles with more than 100
citations to date (Table 16.6) were published. The first one was in 2011, a high-level
rigorous survey to provide a modern overview of online learning [10], which was
furthermore published in the second WOS source by JCR2020 identified in Table
16.3. Later, in 2013 a selection of case studies and topics drawn from experiments
in the setting of a deployed cost through rate (CTR) prediction system that includes
improvements in the context of traditional supervised learning based on an FTRL-
Proximal online learning algorithm and the use of per-coordinate learning rates [11]
was published. Then, two in 2014: A model which combines decision trees with
logistic regression to predict clicks on Facebook ads [12] and a standard Five Factor
Model personality questionnaire to exam how users’ behavior—captured by their
website choices and Facebook profile features—relates to their personality [13]. It
is also in 2014 when the article in the journal with the highest WOS impact factor
is published. A survey focused on discussing problems and solutions pertaining to
the information retrieval, machine learning, and statistics domain of computational
advertising (CA) was published on Foundations and Trends in Information Retrieval
[14], and it has been cited ten times to date.
A considerable increase in academic production in terms of online advertising
and machine learning occurred in 2017, with eighteen titles. Most cited ones, in
Table 16.6 Publications in the aSMS with more than ten cites to date ordered by ascendant number
of citations from left to right
Group Cites Bibliographic references N° %
1 + 100 [5, 10–13] 5 3
2 51–100 [15, 16, 34–37] 6 4
3 11–50 [6, 9, 17, 23, 24, 38–41, 43–49] 22 14
4 1–10 Not detailed 71 44
5 0 Not detailed 57 35
43
27
18 16
7 7 7 9 9 7
6
1 3 1
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Fig. 16.2 Annual number of publications evolution based on the aSMS final 161 results
descending order, are concerned with predicting user response in display advertising
with field-aware factoring machines (FFM) [15], customer acquisition via display
advertising using multiarmed bandit (MAB) methods [16], transforming low-quality
ads into positive quality ad predictions through exploiting tree-based set classifiers
[17], deep character-lever click-through rate (CTR) prediction for sponsored search
[18], optimizing feature selection in video-based recognition using max–min ant
system for the online video contextual advertising [19], and advert value calculation
in cost per thousand (CPM), cost-per-click (CPC) and cost per acquisition (CPA)
networks using a methodology based on deep learning [20].
In 2018, production of online advertising and machine learning decreased by
two units compared to the previous year. In this period, it is found one of the arti-
cles grouped in Table 16.6, 1–10 citations block, which delves into the theoretical
development of deep learning by introducing a learning approach of deep neural
networks to localized manifold learning [21]. During this year, it is also published a
paper in the fourth journal of WOS by impact factor: Electronic Commerce Research
and Applications, identified in Table 16.3. This paper presents a novel methodology
for optimizing the microtargeting technique in direct response display advertising
campaigns by using genetic algorithms as the basis optimization model and a machine
learning-based click-through rate (CTR) model [22].
An early beginning of the rise of publications about this field of study can be
inferred from the number of publications registered in 2019. Growing until the sum
of 27 titles, the two most cited in descending order are about a web-based service
for the automation of contextual advertising management in the Google AdWords
system [23] and deep learning techniques from different aspects of study [24].
Finally, 2020 registers a growth of more than 50% compared to the previous
year. In this block, four articles stand out for having been published in sources
with a high impact index, all of them identified in Tables 16.3 and 16.4 of this
study. In descending order of citations received, these papers study click-through
rate (CTR) prediction through a novel attentive deep interest-based network model
[25], user response prediction through operation-aware neural networks [26], display
advertising campaigns optimization using genetic algorithms [27], and web user
preferences identification and behavior clustering based on a new neural model which
learns different representations for different operations [28].
Springer included this same year one paper on this respect in Springer Proceedings
in Mathematics and Statistics [29]. In this paper, the authors describe how to use some
well-known machine learning tools to make groups of textual queries of similar
semantic meaning, to improve the performances of bidding algorithms for online
advertising.
To conclude, it is worth mentioning that in 2020 IEE Access is the source with the
largest number of publications on the study area. Four papers about the following
issues: A detailed analysis in user behavioral to provide the framework of how Enter-
prise Resource Planning systems track the targeted audience and show their content
[30]; a prediction method of peak time popularity based on Twitter hashtags [31];
and two using model-free reinforcement learning model applied to perform better
results in dynamic bidding strategy in display advertising [32, 33].
RQ4. What Are the Most Cited Publications to Date. This analysis has been
carried out by grouping all the 161 titles into blocks according to the total number
of citations that each of them has received to date.
Group 1 is made up of publications that have more than 100 citations. Specifically,
five, which represent 3% of the total number of titles, with 766, 274, 253, 151, and 105
citations, respectively. Group 2 corresponds to the titles that have 51–100 citations.
It includes six titles, 4% of the 161 ones extracted from the advanced systematic
mapping study (aSMS). In order of reference, these have 83, 74, 62, 54, 53, and 52
citations. Group 3 gathers a higher volume of titles, twenty-tree, which are in a range
of citations between 11 and 50.
Finally, groups 4 and 5 include titles with 1–10 or non-citations. Its titles will not
be detailed in this publication because together all of them add up to 128 titles, 79%
of all, and have too few or no citations.
16.4 Discussion
Observations related to the research questions proposed in this study show that,
although it currently accounts for 2.6% of the total academic production on online
advertising, the growth of publications related to machine learning in this discipline
has registered a notable take-off since 2017, especially in 2020, emerging as a notable
trend.
Regarding databases, Scopus has a greater number of documents about the
research areas, both in pSMS and aSMS. However, it is a rather non-significant
difference.
In relation to sources in which the aSMS research has been published, a consis-
tent dispersion is identified, with three being the maximum number of publications
included in the same journal.
Annual production of academic research in this field shows an exponential trend
line since 2017, stepping up from 2019 and remaining in 2020.
16.5 Conclusions and the Future of the Investigation
Once the corpus of publications has been identified, results have been observed and
research questions have been answered, it is time for conclusions and the future of
investigation.
Regarding conclusions, online advertising and machine learning were found
as the most recurrent terms of those that were included in the primary research
area (online/digital marketing/advertising) and the secondary (artificial intelligence).
Through a superficial initial systematic mapping study (iSMS), with the subsequent
analysis by bigrams of the abstracts of those 336 titles, it has been possible to face
an advanced systematic mapping study (aSMS) in a less general and more precise
way. Thus, this variation in the classic SMS methodology that introduces a previous
initial superficial SMS allows the study area to be narrowed, becoming more specific
and obtaining results that are more related to each other.
In terms of the future of the investigation, it has been performed an ultimate text
analysis by bigrams in R with all the abstracts corresponding to the 161 titles obtained
as a result of the complete research process in the aSMS. Frequencies of each bigram
are detailed in Fig. 16.3.
The frequency analysis shows bigrams that have more than 10 repetitions, with
33 being the maximum frequency of the same bigram. Search strings used in the
libraries (online advertising and machine learning) have been eliminated from the
results drawn in the graph when considering that their frequencies will not be relevant
when it comes to help revealing the future of research in its field.
deep learning
ctr prediction
sponsored search
rate ctr
experimental results
online advertisement
learning techniques
contextual advertising
computational advertising
social media
user behavior
search engine
reinforcement learning
information retrieval
web search
learning algorithms
high dimensional
ad requests
time bidding
online ads
e commerce
display advertising
bidding strategy
results show
neural network
logistic regression
ad networks
ad network
0 10 20 30
Fig. 16.3 Frequency graph of bigrams in R with abstracts of 161 titles extracted from the aSMS
As it can be noticed, in the primary research area—related to online adver-

tising—most repeated frequencies, in a descending order, point to cost through rate
(CTR) prediction, sponsored search, contextual advertising, computational adver-
tising, social media (advertising), search engine, web search, (real) time bidding,
online ads, e-commerce, display advertising, and ad networks. On the other side, in
the secondary research area—related to machine learning—most repeated frequen-
cies in descending order point to deep learning, reinforcement learning, information
retrieval, learning algorithms, and neural network.
In this framework, it would be of interest for future research to complete these
superficial contributions by studying the specific production on these identified
aspects in detail, and its relationship with this paper bibliographic notes. Likewise, it
would be of interest in the future to establish a specific classification of publications
aimed to associate each title to the moment of the user’s life cycle to which it refers.
References
1. Resumen Estudio InfoAdex de la inversión publicitaria en España 2019. http://www.infoad

ex.e. Last accessed 24 May 2021
2. Resumen Estudio InfoAdex de la inversión publicitaria en España 2020. http://www.infoad
ex.es. Last accessed 24 May 2021
3. Rachad T, Idri A (2020) Intelligent mobile applications: a systematic mapping study. Mob Inf
Syst. https://doi.org/10.1155/2020/6715363
4. Noorbehbahani F, Salehi F, Jafar Zadeh R (2019) A systematic mapping study on gamification
applied to e-marketing. J Res Interact Mark 13(3):392–410. https://doi.org/10.1108/JRIM-08-
2018-0103
5. Broder AZ, Fontoura M, Gabrilovich E, Joshi A, Josifovski V, Zhang T (2007) Robust classi-
fication of rare queries using web knowledge. In: Proceedings of the 30th annual international
ACM SIGIR conference on research and development in information retrieval, SIGIR’07,
231–238. https://doi.org/10.1145/1277741.1277783
6. Wu X, Bolivar A (2008) Keyword extraction for contextual advertisement. In: Proceeding of
the 17th international conference on world wide web 2008, WWW’08, 1195–1196. https://doi.
org/10.1145/1367497.1367723
7. Jaworska J, Sydow M (2008) Behavioural targeting in on-line advertising: an empirical study.
Lecture notes in computer science (including subseries lecture notes in artificial intelligence
and lecture notes in bioinformatics), 5175 LNCS, 62–76. https://doi.org/10.1007/978-3-540-
85481-4_7
8. Dominowska E (2008) First workshop on targeting and ranking for online advertising. In:
Proceeding of the 17th international conference on world wide web 2008, WWW’08, 1269–
1270. https://doi.org/10.1145/1367497.1367762
9. Gabrilovich E, Broder A, Fontoura M, Joshi A, Josifovski V, Riedel L, Zhang T (2009) Clas-
sifying search queries using the web as a source of knowledge. ACM Trans Web 3(2). https://
doi.org/10.1145/1513876.1513877
10. Shalev-Shwartz S (2011) Online learning and online convex optimization. Found Trends Mach
Learn 4(2):107–194. https://doi.org/10.1561/2200000018
11. Mcmahan HB, Holt G, Sculley D, Young M, Ebner D, Grady J, Nie L, Phillips T, Davydov E,
Golovin D, Chikkerur S, Liu D, Wattenberg M, Hrafnkelsson AM, Boulos T, Kubica J (2013)
Ad click prediction: a view from the trenches
12. He X, Pan J, Jin O, Xu T, Liu B, Xu T, Shi Y, Atallah A, Herbrich R, Bowers S, Candela JQ

(2014) Practical lessons from predicting clicks on ads at Facebook. Proc ACM SIGKDD Int
Conf Knowl Disc Data Min. https://doi.org/10.1145/2648584.2648589
13. Kosinski M, Bachrach Y, Kohli P, Stillwell D, Graepel T (2014) Manifestations of user person-
ality in website choice and behaviour on online social networks. Mach Learn 95(3):357–380.
https://doi.org/10.1007/S10994-013-5415-Y
14. Dave K, Varma V (2014) Computational advertising: techniques for targeting relevant ads.
Found Trends Inf Retrieval 8(4–5):263–418. https://doi.org/10.1561/1500000045
15. Juan Y, Lefortier D, Chapelle O (2017) Field-aware factorization machines in a real-world
online advertising system. In: 26th international world wide web conference 2017, WWW
2017 companion, 680–688. https://doi.org/10.1145/3041021.3054185
16. Schwartz EM, Bradlow ET, Fader PS (2017) Customer acquisition via display advertising using
multi-armed bandit experiments. Market Sci 36(4):500–522. https://doi.org/10.1287/MKSC.
2016.1023
17. Tolomei G, Silvestri F, Haines A, Lalmas M (2017) Interpretable predictions of tree-based
ensembles via actionable feature tweaking. In: Proceedings of the ACM SIGKDD international
conference on knowledge discovery and data mining, Part F129685, 465–474.https://doi.org/
10.1145/3097983.3098039
18. Edizel B, Mantrach A, Bai X (2017) Deep character-level click-through rate prediction for
sponsored search. In: SIGIR 2017—proceedings of the 40th international ACM SIGIR confer-
ence on research and development in information retrieval, 305–314. https://doi.org/10.1145/
3077136.3080811
19. Bao LN, Le DN, Nguyen GN, Bhateja V, Satapathy SC (2017) Optimizing feature selection in
video-based recognition using Max-Min ant system for the online video contextual advertise-
ment user-oriented system. J Comput Sci 21:361–370. https://doi.org/10.1016/J.JOCS.2016.
10.016
20. Miralles-Pechuán L, Rosso D, Jiménez F, García JM (2017) A methodology based on deep
learning for advert value calculation in CPM, CPC and CPA networks. Soft Comput 21(3):651–
665. https://doi.org/10.1007/S00500-016-2468-4
21. Chui CK, Lin SB, Zhou DX (2018) Construction of neural networks for realization of localized
deep learning. Front Appl Math Stat 4:14. https://doi.org/10.3389/FAMS.2018.00014/BIBTEX
22. Miralles-Pechuán L, Ponce H, Martínez-Villaseñor L (2018) A novel methodology for opti-
mizing display advertising campaigns using genetic algorithms. Electron Commer Res Appl
27:39–51. https://doi.org/10.1016/J.ELERAP.2017.11.004
23. Chyrun L, Kowalska-Styczen A, Burov Y, Berko A, Vasevych A, Pelekh I, Ryshkovets Y (2019)
Heterogeneous data with agreed content aggregation system development. CEUR Works Proc
2386:35–54
24. Mu R, Zeng X (2019) A review of deep learning research. KSII Trans Internet Inf Syst
13(4):1738–1764. https://doi.org/10.3837/TIIS.2019.04.001
25. Li H, Duan H, Zheng Y, Wang Q, Wang Y (2020) A CTR prediction model based on user
interest via attention mechanism. Appl Intell 50(4):1192–1203. https://doi.org/10.1007/S10
489-019-01571-9
26. Yang Y, Xu B, Shen S, Shen F, Zhao J (2020) Operation-aware neural networks for user response
prediction. Neural Netw 121:161–168. https://doi.org/10.1016/J.NEUNET.2019.09.020
27. Miralles-Pechuán L, Ponce H, Martínez-Villaseñor L (2020) A 2020 perspective on a novel
methodology for optimizing display advertising campaigns using genetic algorithms. Electron
Commer Res Appl 40:100953. https://doi.org/10.1016/J.ELERAP.2020.100953
28. Tang X, Li S, Yu W (2020) Web user preferences and behavior clustering based on BP neural
network. J Intell Fuzzy Syst 38(2):1189–1196. https://doi.org/10.3233/JIFS-179480
29. Fodra P, Pasquet E, Goutorbe B, Mohr G, Cornec M (2020) Semantic keywords clustering to
optimize text ads campaigns. Proc Math Stat 339:203–212. https://doi.org/10.1007/978-3-030-
57306-5_19
30. Alojail M, Bhatia S (2020) A novel technique for behavioral analytics using ensemble learning
algorithms in e-commerce. IEEE Access 8:150072–150080. https://doi.org/10.1109/ACCESS.
2020.3016419
31. Yu H, Hu Y, Shi P (2020) A prediction method of peak time popularity based on twitter hashtags.
IEEE Access 8:61453–61461. https://doi.org/10.1109/ACCESS.2020.2983583
32. Liu M, Jiaxing L, Hu Z, Liu J, Nie X (2020) A dynamic bidding strategy based on model-free
reinforcement learning in display advertising. IEEE Access 8:213587–213601. https://doi.org/
10.1109/ACCESS.2020.3037940
33. Shih WY, Lu YS, Tsai HP, Huang JL (2020) An expected win rate-based real time bidding
strategy for branding campaign by the model-free reinforcement learning model. IEEE Access
8:151952–151967. https://doi.org/10.1109/ACCESS.2020.3016824
34. Hajian S, Bonchi F, Castillo C (2016) Algorithmic bias: from discrimination discovery to
fairness-aware data mining. In: Proceedings of the ACM SIGKDD international conference on
knowledge discovery and data mining, 13–17 Aug 2016, 2125–2126. https://doi.org/10.1145/
2939672.2945386
35. Li H, Xu J (2013) Semantic matching in search. Found Trends Inf Retr 7(5):343–469. https://
doi.org/10.1561/1500000035
36. Crussell J, Stevens R, Chen H (2014) MAdFraud: investigating ad fraud in android applications.
In: MobiSys 2014—proceedings of the 12th annual international conference on mobile systems,
applications, and services, 123–134. https://doi.org/10.1145/2594368.2594391
37. Eriksson B, Barford P, Sommers J, Nowak R (2010) A learning-based approach for IP
geolocation. In: Passive and active measurement, proceedings
38. Attenberg J, Provost F (2010) Why label when you can search? Alternatives to active learning
for applying human resources to build classification models under extreme class imbalance.
In: Proceedings of the ACM SIGKDD international conference on knowledge discovery and
data mining, 423–432.https://doi.org/10.1145/1835804.1835859
39. Fan TK, Chang CH (2010) Sentiment-oriented contextual advertising. Knowl Inf Syst
23(3):321–344. https://doi.org/10.1007/S10115-009-0222-2
40. Li C, Lu Y, Mei Q, Wang D, Pandey S (2015) Click-through prediction for advertising in
twitter timeline. In: Proceedings of the ACM SIGKDD international conference on knowledge
discovery and data mining, 2015 Aug, 1959–1968. https://doi.org/10.1145/2783258.2788582
41. Agarwal D, Long B, Traupman J, Xin D, Zhang L (2014) LASER: a scalable response prediction
platform for online advertising. In: WSDM 2014—proceedings of the 7th ACM international
conference on web search and data mining, 173–182. https://doi.org/10.1145/2556195.255
6252
42. Dalessandro B, Chen D, Raeder T, Perlich C, Han Williams M, Provost F (2014) Scalable
hands-free transfer learning for online advertising. In: Proceedings of the ACM SIGKDD
international conference on knowledge discovery and data mining, 1573–1582.https://doi.org/
10.1145/2623330.2623349
43. Kabbur S, Han EH, Karypis G (2010) Content-based methods for predicting web-site demo-
graphic attributes. In: Proceedings—IEEE international conference on data mining, ICDM,
863–868.https://doi.org/10.1109/ICDM.2010.97
44. Ta AP (2015) Factorization machines with follow-the-regularized-leader for CTR prediction in
display advertising. In: Proceedings—2015 IEEE international conference on big data, IEEE
big data 2015, 2889–2891. https://doi.org/10.1109/BIGDATA.2015.7364112
45. Agarwal D, Gurevich M (2012) Fast top-K retrieval for model based recommendation. In:
WSDM 2012—proceedings of the 5th ACM international conference on web search and data
mining, 483–492. https://doi.org/10.1145/2124295.2124354
46. Lang KJ, Moseley B, Vassilvitskii S (2012) Handling forecast errors while bidding for display
advertising. In: WWW’12—proceedings of the 21st annual conference on world wide web,
371–380. https://doi.org/10.1145/2187836.2187887
47. Attenberg J, Provost F (2011) Online active inference and learning. In: Proceedings of the ACM
SIGKDD international conference on knowledge discovery and data mining, 186–194.https://
doi.org/10.1145/2020408.2020443
48. Li SM, Mahdian M, McAfee RP (2010) Value of learning in sponsored search auctions.
Lecture notes in computer science (including subseries lecture notes in artificial intelligence
and lecture notes in bioinformatics), 6484 LNCS, 294–305. https://doi.org/10.1007/978-3-642-
17572-5_24
49. Shatnawi M, Mohamed N (2012) Statistical techniques for online personalized advertising: a
survey. In: Proceedings of the ACM symposium on applied computing, 680–687.https://doi.
org/10.1145/2245276.2245406
Part V
Production Planning and Control
Chapter 17
Redefinition of the Layout
and the Impact on the Reduction
of Wastes: A Case Study
in a Metalworking Industry
Bruna Fernandes, Daniel Botelho, Francisco Fernandes, Inês Aquino,

João Ferreira, José Pinto, Maria Fevereiro, Maria Machado, Nuno Rafael,
and Rui M. Lima
Abstract This study reports a work carried out in a production process of a metal-
lurgical company in the automotive industry. The main objective is to demonstrate
that through the reorganization of the layout is possible to significantly reduce the
associated wastes. This is an aspect frequently disregarded by companies due to the
difficulties associated with the movement of heavy machines, which is a situation
identified in the case under study. The dimensions of the space, the machines, and
the work areas, as well as the respective restrictions, are factors to be considered
in the construction of a logical and organized layout. Initially, the group faced an
inadequate dispersion of the machines on the factory floor, causing high distances
traveled by the operators in the transportation of the materials. High numbers of inter-
sections resulted in defects as parts were often exchanged in different processing
phases. Thus, the idea was to bring together all workstations. However, since the
deburring/washing workstation is shared with other productive processes the change
could only take place in the workstations that precede this. With the changed layout,
such intersections were reduced, as well as the distances traveled by the operators
and, consequently, the associated costs. Wastes such as defects and transportation
have been reduced with the changes made, which allows the flow of material to be
continuous and effective. Hence, with the reduction of 606.51 m covered by workers
per day, the company obtains a sales increase of 830.76 e per month.
Keywords Layout redesign · Waste reduction · Lean manufacturing
B. Fernandes · D. Botelho · F. Fernandes · I. Aquino · J. Ferreira · J. Pinto · M. Fevereiro ·

M. Machado · N. Rafael · R. M. Lima (B)
Algoritmi Centre, Department of Production and Systems, School of Engineering, University of
Minho, Guimarães, Portugal
e-mail: rml@dps.uminho.pt
https://doi.org/10.1007/978-3-031-29382-5_17
164 B. Fernandes et al.
17.1 Introduction
Currently, for a company to stand out in the market, it must guarantee a good concept
of efficiency. To ensure the implementation of that concept, it is necessary to identify
and reduce the wastes in the production systems.
The “Toyota Production System” (TPS) as described by Monden [1] and Ohno [2],
allows distinguishing seven production wastes. This TPS approach was disseminated
as “Lean Manufacturing”, from the work of Womack, Jones, and Roos and the work of
Krafcik [3, 4]. The TPS’s main objective is to reduce costs and increase productivity,
by reducing all waste inherent in the production system [1].
Waste is “any human activity that consumes resources, but that does not create
value for the product” [5]; that is, it is any activity that does not bring advantages,
contributing to an increase in costs, time, and customer dissatisfaction. The seven
major types of waste are [2]: overproduction, waiting, transport, movements, stock,
overprocessing, and defects.
The adoption of “Lean” thinking presupposes organizational changes that make
production systems more efficient and more responsive to customer requests [6].
According to Maia et al. [7], it means “doing more with less”, where less means less
space, less transport, fewer stocks and, most importantly, less human effort and less
need to use natural resources.
For an organizational-level improvement of the shop floor, optimization of the
production layout may be the most effective solution. This optimization is, neces-
sarily, related to production efficiency, combining the following factors: reduction
of movements and material transportation, reduction of additional costs, improving
the quality of the products, among others [8].
According to Cury [9], a new layout must consider the placement of machines,
raw materials, and semifinished products in strategic places in order to fill, in the
best possible way, the available space. These decisions should be made considering
the best possible way for the operator to perform his/her function and to guarantee
job satisfaction and quality. Peinedo and Reis [10] reinforce the importance of the
layout, stating that it is “the most visible and exposed part of any organization”.
When there is a need to form a new company or reformulate an existing one, it is
mandatory to have a detailed study of the production to create a layout that meets
the imposed needs.
Given this context, the present article aims to demonstrate that the reduction of
production wastes may be accomplished through the reorganization of the layout of
a production cell.
This work starts with the methodology and an analysis of the current situation of
the company is made. Next, a layout improvement proposal is shown. It is followed
by a discussion of results where a comparative analysis between the current layout
of the company with the proposed one is presented.
17 Redefinition of the Layout and the Impact on the Reduction of Wastes: … 165
17.2 Methodology
The company is divided into three major production areas: lathing, stamping, and
forging, where production plans vary on orders. Lathing consists of five production
processes, stamping by six and forging by eight. The production system in the three
areas is similar, being characterized by a workshop production that is typically used
when there are large varieties of products being produced in small quantities. The
workshops are divided into work centers consisting of a machine and an operator
responsible not only for carrying out the process but also for transport to the next
work center.
However, the problem in focus is related to a production cell that performs different
manufacturing operations for a car part, the connecting rod. The company has a
production system integrated by nine operations spread over eight work centers.
Although, two operations, “Deburring” and “Wash”, are performed in the same work
center. A number will be assigned to each workstation that will be used across the
article: Press (1), Milling (2), Deburring (3), Drill (4), Variomatic (5), Top part debur-
ring (6), Ream + clean (7), wash (8) and, finally, the quality wall (9). The raw material
enters the production cell and, after sequentially going through all the workstations,
becomes the final product. The operator, in addition to being responsible for its oper-
ation, is also responsible for transporting the batch of 600 pieces for subsequent
activity.
The data was collected while visiting the shop floor of the company through direct
observation and measurement of traveled distances, the velocity of the workers while
transporting the materials, measurements of the space required and of each machine
involved in the productive process. With these distances, a Spaghetti diagram was
drawn up with the objective of mapping and visually demonstrating the route that
the collaborators take to produce the piece. To make the mapping of the flows more
understandable, different colors can be used to draw the spaghetti by distinguishing
the resources that move in the system or the time bands in which they move [11].
According to the wastes identified, a new layout design was developed. The main
steps for a layout redesign are, according to Kovács [12], the following:
Step 1: Define the objectives of the design. Generally, the main objectives are to
minimize the total distance of goods flow and the material handling cost.
Step 2: Define the main activities of the process. Requirements relating to the
activities (e.g., workstations), human resources and material flow must be
specified.
Step 3: Determine the space requirements for all objects and material flow paths.
Step 4: Create alternatives for facility layouts. Variations of layouts must be formed.
Step 5: Select the most effective layout.
Step 6: Implement the best layout. The best plan must be selected and implemented.
17.3 Analysis of the Current Situation
It is a company that competes in different business areas such as the automotive

industry. In addition, the company´s great advantage is its ten production processes
that allow the company to become highly competitive. It is the only one with these
conditions that approve the production of its parts to exactly fulfill its functions and
with the maximum level of efficiency.
For better clarification of the movements made by the operators, as well as the
routes made by them in the transport of the product, a diagram of Spaghetti was
generated [13] (Fig. 17.1).
Currently, the machines do not have a logical distribution and are located far
from each other. Derived from these conditions, workers are forced to travel long
distances to move the lots, which leads to high intersections of both operators and
parts, consequently, making the workspace confusing.
In addition to the conditions presented above, the existence of machines that are
not part of the production system under study and that are placed in the space available
to produce the connecting rod increases entropy.
In order to map the production process, the elaboration of a sequence diagram
made it possible to determine the percentages of activities that add and do not
add value to each job [14]. All tasks related to each workstation described in
Sect. 17.2 were detailed in a flow process chart. A summary of this flow process
chart, represented in Table 17.1, shows that the main waste identified is transport.
Table 17.2 shows the distances currently performed by operators in the transporta-
tion processes related to the production of the connecting rod.
Following the calculation of these distances, the need to recognize the cost asso-
ciated with transporting the product between the production cell’s workstations is
instigated. Thus, in addition to recording distances, the number of movements per
day was calculated, dividing the daily production rate by the number of pieces per
lot (Table 17.3).
Workstations that are

part of the productive
system
Operator 1
Operator 2
Operator 3
Operator 4
Operator 5
Operator 6
Operator 7
Operator 8
Fig. 17.1 The current layout of the production cell

Table 17.1 Flow process chart summary

Workstation
1 2 3 4 5 6 7 8 9
Operation ◯ 1 1 1 2 1 2 2 1 0
Transport ⇨ 1 1 1 1 1 0 1 1 0
Control  1 1 0 0 0 0 0 0 0
Wait 0 0 0 0 0 0 0 0 0
Motion 3 6 2 5 4 1 2 2 2
Stock ∇ 1 1 0 1 1 1 2 0 1
Total 7 10 4 9 7 4 7 4 5
VA (%) 14 10 25 22 14 50 29 25 0
NVA (%) 86 90 75 78 86 50 71 75 100
Distance (m) 28.5 44.1 36.5 13.8 7.5 0 38.1 24.5 –
Time (s) 45.8 92.7 55.2 33.3 19.7 7.4 57.3 38.4 4.7
Table 17.2 Traveled

Production sequence Distances (m)
distances per day in the
current layout Press – milling 342.50
Milling – deburring 485.25
Deburring – drilling 547.56
Drilling – variomatic 138.28
Variomatic – top part deburring 76.44
Top part deburring – ream + clean 0.00
Ream + clean – wash 360.88
Wash – quality wall 367.80
Total 2318.71
There are several methods for layout design or redefinition. One of the most used
is the Craft method (Computerized Relative Allocation of Facilities Technique) [15],
which helps to improve the layout of installations. Its main objective is to reduce the
total transport cost, obtained by Eq. (17.1) of the material movement cost (MMC).
MMC corresponds to the sum of the multiplication of the material flows ( f i j ) by the
cost of the associated transport (ci j ) and by the distance between workstations (di j )
[16].
∑
n−1 ∑
n
MMC = ci j × f i j × di j (17.1)
j=1 i= j+1
Table 17.3 Daily

Workstations Production rate Pieces/lot Movements/day
movements
daily
Press 7176 600 12
Milling 6325 11
Deburring 9200 15
Drill 5934 10
Variomatic 6010 10
Top part 7420 12
deburring
Ream + 5991 10
clean
Wash 9200 15
Quality wall 5849 10
For the application of the Craft method, it was necessary to define a transport cost
per meter. This cost could be calculated considering several inputs, namely: energy
costs of the equipment used, maintenance costs, and human work costs. In this case,
only the human working costs were used, which is a cautious estimation. Thus, it
was defined in this case as 0.0016 e/m.
Through the utilization of the Craft formula, the MMC is e 3.72/day with an
associated distance of 2318.71 m. The objective is to reduce these values and,
consequently, the associated wastes without compromising the production flow.
17.4 Layout Improvement
In order to reduce the time spent on transportation and the number of intersections,
a new layout proposal was developed. As the objective is to reduce the distances
traveled by the operators in the workplace and, consequently, the MMC, the criterion
to be used will be the distance covered by the components. At the same time, it is
a scenario closer to a production cell layout than what is currently implemented on
the factory floor.
Since the main limitation is the location of the Deburring/Washing workstation,
the idea is to bring together all workstations that precede deburring or washing, to
reduce the distance covered and, at the same time, locate these workstations at the
extremities of the production cell to reduce the number of intersections.
Thus, the proposed layout was conceived as represented in Fig. 17.2. The
route taken by the operators in the production process was mapped, to clarify the
movements made as well as the routes taken by them.
With the help of the Spaghetti diagram, Table 17.4 shows the distances that would
be made by the operators in the proposed layout. In order to make a comparison
between the two scenarios, the MMC was calculated in relation to the proposed
Workstations that are part

of the productive system
Operator 1
Operator 2
Operator 3
Operator 4
Operator 5
Operator 6
Operator 7
Operator 8
Fig. 17.2 Proposed layout of the production cell
Table 17.4 Traveled

Production sequence Distances (m)
distances per day in the
proposed layout Press – milling 163.2
Milling – deburring 264.0
Deburring – drilling 541.2
Drilling – variomatic 40.00
Variomatic – top part deburring 60.0
Top part deburring – ream + clean 0.0
Ream + clean – wash 276.0
Wash – quality wall 367.8
Total 1712.2
layout. Since the number of movements and the cost of transport does not change in
the two scenarios, only the factor of the distances covered will vary the value of the
MMC to 2.72 e/day.
17.5 Discussion of Results
There is a reduction in the distances covered per day of 606.51 m (Table 17.5), which
is equivalent to a reduction of approximately 26% when comparing the current layout
to the proposed one. This distance reduction allows the saving of 10 min. During
this time the company can produce approximately 43 connecting rods which equal
a sales increase of 39.56 e every day, which equals 830.76 e every month.
Comparing the MMC of the current layout with the proposed one, it appears that
there is a decrease in this cost of 1 e/day.
In addition to the decrease in the distance traveled in the proposed layout and,
consequently, material movement costs, it is also possible to verify a decrease in the
Table 17.5 Comparison of the distances traveled between the two layouts
Production sequence Current layout distances (m) Improved layout Reduction (m)
distances (m)
Press – milling 342.50 163.20 179.30
Milling – deburring 485.25 264.00 221.25
Deburring – drilling 547.56 541.20 6.36
Drilling – variomatic 138.28 40.00 98.28
Variomatic – top part 76.44 60.00 16.44
deburring
Top part deburring 0.00 0.00 0.00
– ream + clean
Ream + clean – wash 360.88 276.00 84.88
Wash – quality wall 367.80 367.80 0.00
Total 2318.71 1712.20 606.51
number of intersections. The culmination of these two factors leads to a continuous

production flow and prevents the mixing of parts at different stages of processing.
Thus, wastes such as transportation and defects can be reduced, which meets the
main objective: by redefining the layout there is a waste reduction.
17.6 Conclusion
Continuous improvement and increasing the efficiency of production processes must

be a constant concern of the company. Several factors influence this efficiency, namely
the layout of the manufacturing space.
In this article, the group compared the distances traveled by operators when trans-
porting parts, as well as the costs associated with the company’s current layout and
the improvement proposal. Analyzing the obtained results, it was concluded that the
arrangement of machines on the shop floor in an organized and logical manner has
an impact on reducing distances and, consequently, on costs. In addition, there was a
reduction in the number of intersections, which allows a continuous production flow
and avoids the mixing of parts at different processing stages. Furthermore, the time
saved with the reduction of the distance traveled allows the company to improve the
daily production up to 43 connection rods, which means a sales increase of 39.56 e
every day. This is, 830.76 e every month.
The presented proposal could be more advantageous if it did not have the limitation
imposed by the company regarding the deburring/washing workstation location. As it
is a workstation shared by other production processes, its location cannot be changed.
Acknowledgements This work has been supported by FCT—Fundação para a Ciência e

Tecnologia within the R&D Units Project Scope: UIDB/00319/2020.
References
1. Monden Y (1998) Toyota production system: an integrated approach to just-in-time. Engi-

neering and Management Press
2. Ohno T (1988) Toyota production system: beyond large scale production. Productivity Press,
New York
3. Krafcik JF (1988) Triumph of the lean production system. Slogn Manage Rev:41–52
4. Womack JP, Jones DT, Roos D (1990) The machine that changed the world: the story of lean
production—Toyota’s secret weapon in the global car wars that is now revolutionazing world
industry, vol 30. Free Press, New York
5. Womack JP, Jones DT (1996) Lean thinking. Siman & Schuster, New York, USA
6. Rao HA (1999) A genetic algorithms-based approach for design of manufacturing systems: an
industrial application. Int J Prod Res:557–580
7. Maia LC, Alves AC, Leão CP (2014) Implementar o modelo de produção Lean na ITV para
promover sistemas eco-eficientes. Nova Têxtil:18–25
8. Olivério JL (1985) Projeto de fábrica: produtos processos e instalações industriais. São Paulo
9. Cury A (2007) Organização e métodos: uma visão holística, prespectiva comportamental e
abordagem contingencial. São Paulo
10. Peinedo J, Reis AG (2007) Administração da produção: operações industriais e de serviços.
Curitiba
11. Sendrska K, Mares A, Václav S (2017) Spaghetti diagram application for workers’ movement
analysis. UPB Sci Bull Ser D 79:139–150
12. Kovács G (2019) Layout design for efficiency improvement and cost reduction. Bull Acad Pol
Sci 67(3):547–555
13. Institute for Innovation and Improvement (2018) Quality and service improvement tools:
spaghetti diagram. NHS Improvement
14. Nanthasamroeng N (2012) Systematic layout planning for germinated brown rice mill under
GMP and ISO22000:2005 requirements. IOSR J Eng 2(10):35–40
15. Hari N, Rajyalakshmi G, Screenivasulu A (2014) A typical manufacturing plant layout design
using CRAFT algorithm. Procedia Eng:1808–1814
16. Gonçalves RGDA (2020) Reconfiguração de um sistema produtivo e melhoria de processos
aplicando Lean thinking numa carpintaria. Universidade do Minho
Chapter 18
Digital Twin Enabling Intelligent
Scheduling in ZDM Environments:
an Overview
Abstract As at any decision level in operations planning and control (OPC), oper-
ational decisions are influenced by the technological advances underpinning the
Industry 4.0 (I4.0) paradigm. In this increasingly digitized environment, scheduling
problems have to cope with stochastic demand, dynamic task allocation flow, routing
flexibility, or task rescheduling. The ability to virtually replicate the scheduling
process in an I4.0 environment enables its optimization, simulation, prediction, and
automatic analysis in real time. These features are necessary in manufacturing envi-
ronments with a zero-defect manufacturing strategy (ZDM) because these factors
allow scheduling problems to be adapted to this strategy’s requirements. Therefore,
a scheduling problem in a ZDM environment driven by a digital twin (DT) will
favor better production system performance. With this approach, the present article
provides an overview of the scientific literature for this combined set of concepts. It
presents the academic and research implications of the present research, discusses
its results and limitations, and indicates where future research into this theme is to
be directed.
Keywords Industry 4.0 · Scheduling · Digital twin · Zero-defect manufacturing
J. C. Serrano-Ruiz (B) · J. Mula · R. Poler

Centro de Investigación en Gestión e Ingeniería de la Producción (CIGIP), Universitat Politècnica
de València, 03801 Alcoy, Spain
e-mail: jserrano@cigip.upv.es
J. Mula
e-mail: fmula@cigip.upv.es
R. Poler
https://doi.org/10.1007/978-3-031-29382-5_18
174 J. C. Serrano-Ruiz et al.
18.1 Introduction
Production scheduling is a fundamental process that manufacturing companies face

for production to be efficient and effective [1]. The relevance of scheduling can
be extended to any manufacturing environment, regardless of its typology, char-
acteristics, and circumstances. Within an Industry 4.0 (I4.0) context, scheduling
can be addressed by the smart manufacturing (SM) system through the digital and
emerging manufacturing technologies that characterize it, such as cyberphysical
systems (CPS), big data, Internet of Things (IoT), artificial intelligence (AI), DT
and social, mobile, analytics, cloud (SMAC) technologies [2]. Of these, the DT
technology has been revealed in recent years as one of the most synergic tools for
scheduling. Something similar has taken place with the ZDM strategy, which is
clearly interrelated to the scheduling issue.
The main purposes of this paper are to: (i) provide an overview of the most
relevant contributions of the academic and research community in the field of DT-
driven scheduling in ZDM environments; (ii) identify the main perspectives raised
by authors in the field; (iii) identify the main knowledge gaps in the current state
of the art; (iv) establish the first implications; and (v) determine additional research
directions in which advances in the defined field would significantly contribute to
academic knowledge, to point which might be the next steps of this research that
currently is in its initial stage.
The remaining of this paper is organized as follows. Section 18.2 introduces
the review methodology. Section 18.3 presents an overview of the related litera-
ture. Section 18.4 summarizes the outcomes by providing a taxonomy. Section 18.5
exposes the main implications resulting from the overview. Finally, Sect. 18.6
provides conclusions and outlines future research in this domain.
18.2 Review Methodology
Scheduling, whose origin as a research objective dates back to the start of the twen-
tieth century, is a popular and mature topic in the scientific community, especially
since the 1970s from which time interest in it has gradually increased to the present
day. Today tens of thousands of publications consider it every year. On the contrary,
the DT as a research objective has a noticeably shorter history, to such an extent that
a search in Scopus for “digital twin” in the thematic areas of engineering, computer
science, science decision, business, management and accounting, and multidisci-
plinary, from its origin to 2017, gives only 152 publications. However, for the 2018–
2020 triennium, 2570 publications are registered, with 95% of all publications in
its history. The ZDM strategy, unlike previous ones, has a long-standing history
with publications since the 1970s. However, interest in its research dwindled until
2018, when suddenly publications increased by 190% compared to the previous year.
This interest remains today and coincides with the new approach that both the I4.0
18 Digital Twin Enabling Intelligent Scheduling in ZDM Environments: … 175
paradigm and its potential synergy represent for ZDM. For all these reasons, the most
representative time period to explore the approach herein sought in believed to be
that starting from 2018.
The relevant bibliography for conducting this study was compiled from articles
and conference papers obtained from the Scopus, Web of Science, Science Direct,
ProQuest, and IEEE Xplore databases by considering a time window from 2018
to the present day, and a search strategy based on the Keywords: “Industry 4.0”,
“scheduling”, “digital twin”, and “zero-defect”
18.3 Overview of DT-Based Smart Scheduling in ZDM
In the reviewed literature, the following groups of conceptual frameworks were

identified: (i) DT enabling smart scheduling in a ZDM environment (Table 18.1),
which contemplates the three concepts considered in Table 18.1 ; (ii) DT enabling
scheduling (Table 18.2), where the implementation of a DT into the scheduling
process allows the replication of the process and its virtual simulation, analysis,
prediction, and optimization, but in an environment without ZDM characteristics;
(iii) scheduling in ZDM environments (Table 18.3), where the scheduling process
takes place in a ZDM environment, but without the assistance of a DT; (iv) other
frameworks (Table 18.4), which approach to the DT-based smart scheduling in ZDM
schemes, but from different perspectives.
Table 18.1 DT-based scheduling in a ZDM environment

References Role of each concept in the model
Lindström et al. [3] Scheduling: One of the seven main areas of the posed ZDM strategy is
production rescheduling
DT: It is mentioned several times in the paper, but without going into
great detail
ZDM: ZDM plays a leading role. A cost function associated with it is
proposed to reflect the condition, quality, and safety aspects of a
production process
Dreyfus et al. [4] Scheduling: An automatic scheduling algorithm uses information about
quality and maintenance to optimize when to do maintenance, tune
machines, and schedule production orders
DT: It is only mentioned as the conclusion and a final objective to be
addressed by future research
ZDM: The ZDM goal is managed by a tuning assistant. If a problem
cannot be automatically repaired by it, production can quickly be
interrupted, and the technician is informed
Table 18.2 DT enabling scheduling

Hu et al. [5] Scheduling: An issue solved using Petri nets and deep-Q learning with the aid
of a DT
DT: A tool to offer a high-fidelity simulation and a visualization environment
in which the DRL scheduling agent interacts with
Liu et al. [6] Scheduling: An issue solved by introducing the supernetwork technology into
the DT jobshop
DT: A tool to perform real-time monitoring and face both the uncertainty of
the process route and the existence of external and internal dynamic
disturbance factors from multiple sources
Xia et al. [7] Scheduling: One of the main tasks of the digital engine model
DT: A tool termed digital engine that works as a dynamic scheduling agent of a
robotic manufacturing cell based on machine learning, and represents,
simulates, and predicts
Zhang et al. [8] Scheduling: An issue to optimize in the jobshop by a DT-enhanced dynamic
methodology
DT: A tool to enhance jobshop planning by predicting machinery availability,
detecting production disturbances, and evaluating performance to perform
dynamic scheduling
Fang et al. [9] Scheduling: An issue to cope with reducing scheduling deviation and
improving accuracy and the robustness of the jobshop scheduling scheme
DT: A tool to arrange jobshop scheduling and provide sufficient data support
for production
18.4 Taxonomy
A summary taxonomy based on 11 different aspects is provided to contribute in this

way to a better general understanding of the main characteristics of each reviewed
article and the degree of progress in this topic (Table 18.5).
18.5 Discussion
Real-time production and flexibility are two SM objectives. Given the large number
of operations associated with an SM environment, its complex cooperative relation,
its strong continuity characteristic and the rapid changes in this context, the failure
of a certain part often affects the entire production system’s operation [9]. Therefore,
scheduling failure in an SM environment can lead the production system as a whole
to fail. Consequently, it is necessary to identify the causes that lie behind scheduling
failures, such as unexpected events, information asymmetries or abnormal distur-
bances in the actual scheduling process, which deviate SM execution, and affect
both its efficiency and quality [9]. All these causes can also disrupt normal manu-
facturing system operation by posing risks, incurring additional costs and reducing
Table 18.3 Scheduling in ZDM environments

Paprocka et al. [10] Scheduling: A non-detailed production task to be improved with a
strategy headed to zero machine defects, zero-defect, and zero
accidents at work
ZDM: The effect of predictive scheduling and the application of Total
Productive Maintenance
Psarommatis et al. [11] Scheduling: An issue solved by using the Tabu search algorithm
(TSA) for both scheduling and buffer size optimization in a ZDM
strategy
ZDM: A strategy implemented by a procedure run to reschedule
production, provided that new actions are needed, as the main means
to reduce faults and defects
Psarommatis et al. [12] Scheduling: An issue studied from the perspective of the rescheduling
caused by unexpected events occurring, such as receiving new orders,
faulty products, and product or predicting machinery defects in a
ZDM strategy context
ZDM: A strategy that models manufacturers’ response time by
rescheduling production in view of unexpected events, such as new
orders, faulty products, or machine defects
Psarommatis et al. [13] Scheduling: An issue solved by using the Tabu search algorithm from
the rescheduling perspective in a ZDM strategy
ZDM: A strategy implemented by applying mitigation actions to
counteract problems whenever they arise, actions that must be
implemented in manufacturing planning rescheduling as often as
necessary, but by looking for high-quality initial solutions from the
beginning to reduce rescheduling frequency
Psarommatis et al. [14] Scheduling: An issue dynamically solved in a ZDM strategy with an
intelligent decision support system (DSS), which is not detailed
ZDM: A strategy implemented with two tools, a smart DSS and
dynamic scheduling, to reschedule production when required as the
main means to reduce faults and defects
system efficiency [12]. Here DT technology can assist in the scheduling process by
generating virtual replicas of assets or processes with which to visualize, model,
simulate, and analyze "what-if" scenarios, predict, generate alternative management
scenarios, learn or optimize, among other possible actions [20, 21], to address this
troublesome situation and to reschedule as optimally and as soon as possible.
The main challenge in virtual scheduling replication when using the DT focuses
on providing its functions in SM with intelligence and overcoming the limitations
of traditional methods. Apart from agility, flexibility, speed, and predictability, the
main SM objectives also include quality. Nowadays, manufacturers increasingly pay
special attention to quality improvement using the ZDM paradigm, which can be
enhanced by implementing I4.0 [11]. As far as we know, there is no complete ZDM
solution for manufacturers in scheduling decision terms based on today’s state of
the art. As real-time events can disrupt production, scheduling tools must be able
to maintain production at a certain level of efficiency [14], and ZDM can provide
Table 18.4 Other frameworks

References Specific features
Hu et al. [15] Approach: DT enabling scheduling and quality control in the
semiconductors equipment industry
Scheduling: The advantages of using DT for scheduling are explained
from a general perspective
DT: A bridge between the real world and information that lays the
basis for product quality tracing and continuous improvement
searching
Quality control: One of the physical processes considered in the
model
Li et al. [16] Approach: DTs enabling enhanced production planning optimization
Scheduling: A manufacturing task to optimize by using
manufacturing task (MT) semantic modeling and manufacturing
resource (MR) dynamic recommendation (MT&MR method)
DT: A tool to support production planning simulation and optimization
Wang and Wu [17] Approach: DT enabling both scheduling and quality control
Scheduling: An advanced scheduling management enabled with the
DT-driven production management system (DTPMS) is described
DT: The DTPMS is a tool to monitor shop-floor situations in such a
way that shop-floor staff can centrally follow every procedure and
record every step
Quality control: Product quality is considered in the model from a
general perspective
Bilberg and Malik [18] Approach: DTs enabling the dynamic workload balancing of cell
assembly tasks based on the skills of both humans and robots
Scheduling: A non-detailed issue solved in such a way that tasks
performed by human and robots along the assembly line are optimally
balanced
DT: A tool that contributes to: (i) rapid skills-based workload
balancing between human and robots for a product variety; (ii)
dynamic workload balancing during operations to account for human
factors; (iii) optimizing robot trajectories; (iv) generating robot control
programs
Zupan et al. [19] Scheduling: An issue to solve and optimize in a jobshop environment
by a multistart local search heuristic “remove and reinsert” algorithm
(RaRA)
DT: A tool to build “what-if” jobshop problem scenarios and to
simulate manufacturing
Quality: The quality measure of scheduling is considered and
determined from the time used by the algorithm for calculations and
the calculated makespan
it with both detection and correct reaction [11]. This implies zero-defect-oriented
scheduling having to be able to trigger the rescheduling process and facilitate it in
real time whenever required and, as previously commented, the DT is a useful tool
for facing this particular need.
Table 18.5 Summary of the overview
References Research OPC Scope in Product Shop Modeling Resolution DT purpose DT enabling ZDM tactics ZDM
approach issue supply typology typology approach approach technologies regarding integration
chain defects
Lindström ECS PS S/M/D – – – – OP/S BD/S/CC/SMs/ML DE/RE IL
et al. [3] DM/IoT/PHM PD/PV
Dreyfus D PS M – – – – – – DE/RE IL
et al. [4] PD/PV
Hu E PS M MISL FJS AI DES OP/S SMs/ML – –
et al. [5] DQN O/A
Liu E PS M MIML - AI PSN OP/S BD/SMs/DM – –
et al. [6] A/MPI
Xia E PS M SIML FS S/AI DQL OP/S/PA/MPI BD/SMs – –
et al. [7] ML/IoT
Zhang E PS M MIML JS A GA OP/S BD/S/SMs – –
et al. [8] P/A
Fang E PS M MIML JS A NS OP/S BD/S/CC – –
et al. [9] GA-II SMs/IoT
Paprocka E – M – – – – – – – –
et al. [10]
Psarommatis E PS M SIML FJS – TSA – – DE/RE/PV IL
et al. [11]
Psarommatis E PS M MISL FJS A EMA – – DE/RE IL
18 Digital Twin Enabling Intelligent Scheduling in ZDM Environments: …
et al. [12] PR/PV

Psarommatis E PS M MIML FJS H TSA – – – IL
et al. [13]
(continued)
179
180
References Research OPC Scope in Product Shop Modeling Resolution DT purpose DT enabling ZDM tactics ZDM
approach issue supply typology typology approach approach technologies regarding integration
chain defects
Psarommatis E PS M MIML - – – – – DE/RE IL
et al. [14] PR/PV
Hu E PS MD MIML - – – S/P BD/S/SMs/ML DE/RE/PR IL
et al. [15] PHM/IoT
Li E AAL S/M/D MIML - AI IGSO OP/S/PA/MPI BD/S/SMs – –
et al. [16] DNN ML/DM
Wang and Wu E PS M SIML FS – – OP/S S/SMs/DM DE/RE/PR IL
[17] P/A
Bilberg and E PS M MIML – – – OP/S/A S/SMs – –
Malik [18]
Zupan et al. E PS M MIML JS H RaRA S/A SMs PV OL
[19]
Note Research approach: C conceptual, D descriptive, E empirical, ECS exploratory cross-sectional, EL exploratory longitudinal; OPC issue: AP aggregate planning, MPS
master planning scheduling, CRP capacity resource planning, MRP material requirement planning, PS production scheduling, DP distribution planning, PC operation
control, AAL at all levels; Scope in supply chains: S sourcing, M manufacturing, D distribution; Product typology: SISL single item and single level, SIML single item
and multilevel, MISL multi-item and single level, MIML multi-item and multilevel; Shop typology: SM single machine, PM parallel machine, FS flowshop, FFS flexible
flowshop, JS jobshop, FJS flexible jobshop, OS open shop, CPA complex product assembly; Modeling approach: C conceptual, A analytical, H heuristics, S simulation,
AI artificial intelligence; Resolution approach: DES discrete event simulation algorithm, DQN deep-Q network, PSN processing supernetwork, DQL deep-Q learning,
GA genetic algorithm, NSGA-II fast non-dominated sorting genetic algorithm, RaRA remove and reinsert algorithm, TSA Tabu search algorithm, EMA events management
algorithm, IGSO improved glowworm swarm optimization algorithm, DNN deep neural network; DT purpose: OP optimization, S simulation, P prediction, A analysis,
MPI multiphysics integration; DT key enabling technology: BD big data, S sensoring, CC cloud computing, SMs simulation methods, ML machine learning, DM data
mining,IoT Internet of Things, AR augmented reality, AM additive manufacturing, PHM prognostic & health management; ZDM/Quality tactic regarding defects: DE
detection, RE repair, PR prediction, PV prevention; ZDM/Quality integration into manufacturing systems: IL in-line, OL off-line, LAB laboratory; and, finally, the symbol
“–” indicates that there is no contribution in this regard
J. C. Serrano-Ruiz et al.
18.6 Conclusions
This paper reviewed the concepts, terms, and conceptual frameworks toward DT-
driven scheduling in ZDM environments. It is worth noting that the literature on this
subject is still scarce, and the few works found have not worked on all three concepts
(scheduling, DT, ZDM) and provided very little detail. However, several authors
have begun to increasingly look in-depth at less researched aspects in the last few
years and provide contributions that evidence the growing interest in the subject.
Among the incomplete approaches, but close to our main research objective, the
most frequent was DT-driven scheduling, with fewer contributions than scheduling
in ZDM environments. Of all the contributions, the addressed problem is restricted
mostly to the specific manufacturing domain, which does not contemplate sourcing
and distribution processes. The most studied manufacturing configuration is the
jobshop, but the flowshop and some other configurations like the open shop or the
assembly shop for complex products are relegated to a small number of studies. Tradi-
tional analytical or heuristic modeling approaches coexist with increasingly present
AI approaches, which are gradually overcoming some of the limitations of more
traditional proposals. The DT is generally conceived as a simulator and optimizer,
while its other possible roles are less frequently raised and studied. No contributions
were identified in which the DT plays a prescriptive or decision-making role.
Hence, future research lines were identified based on: (i) the implications on
the sourcing and distribution domains; (ii) the applicability of the studied frame-
work to shop-floor configurations other than the jobshop; (iii) the use of AI; (iv)
exploring other roles for the DT, such as prescribers or decision-makers; (v) the
need to contribute to the state of the art of conceptual, descriptive, and empirical
research for the scheduling driven by the DT in ZDM environments.
Acknowledgements The research leading to these results received funding from the European
Union H2020 Program under grant agreement No. 825631 “Zero-Defect Manufacturing Platform
(ZDMP)” and under grant agreement No. 958205 “Industrial Data Services for Quality Control in
Smart Manufacturing (i4Q)” and from the Spanish Ministry of Science, Innovation and Univer-
sities under grant agreement RTI2018-101344-B-I00 “Optimisation of zero-defects production
technologies enabling supply chains 4.0 (CADS4.0)”.
References
1. Negri E, Ardakani HD, Cattaneo L, Singh J, MacChi M, Lee J, Barari A (2019) A digital
twin-based scheduling framework including equipment health index and genetic algorithms.
IFAC-PapersOnLine 52(10):43–48
2. Zhang J, Ding G, Zou Y, Qin S, Fu J (2019) Review of job shop scheduling research and its
new perspectives under industry 4.0. J Intell Manuf 30(4):1809–1830
3. Lindström J, Kyösti P, Birk W, Lejon E (2020) An initial model for zero defect manufacturing.
Appl Sci MDPI AG 10
4. Dreyfus P-A, Kyritsis D (2018) A framework based on predictive maintenance, zero-defect

manufacturing and scheduling under uncertainty tools, to optimize production capacities of
high-end quality products. IFIP Adv Inf Commun Technol 536:296–303
5. Hu L, Liu Z, Hu W, Wang Y, Tan J, Wu F (2020a) Petri-net-based dynamic scheduling of flexible
manufacturing system via deep reinforcement learning with graph convolutional network. J
Manuf Syst 55:1–14
6. Liu Z, Chen W, Zhang C, Yang C, Cheng Q (2020) Intelligent scheduling of a feature-process-
machine tool supernetwork based on digital twin workshop. J Manuf Syst
7. Xia K, Sacco C, Kirkpatrick M, Saidy C, Nguyen L, Kircaliali A, Harik R (2020) A digital
twin to train deep reinforcement learning agent for smart manufacturing plants: environment,
interfaces and intelligence. J Manuf Syst 58:210–230
8. Zhang M, Tao F, Nee A (2020) Digital twin enhanced dynamic job-shop scheduling. J Manuf
Syst
9. Fang Y, Peng C, Lou P, Zhou Z, Hu J, Yan J (2019) Digital-twin-based job shop scheduling
toward smart manufacturing. IEEE Trans Industr Inform 15(8821409):6425–6435
10. Paprocka I, Kempa WM, Kalinowski K, Grabowik C (2014) A production scheduling model
with maintenance. Adv Mater Res 1036:885–890
11. Psarommatis F, Boujemaoui A, Kiritsis D, Lalic B, Marjanovic U, Majstorovic V, von Cieminski
G, Romero D (2020a) A computational method for identifying the optimum buffer size in the era
of zero-defect manufacturing. In: IFIP advances in information and communication technology,
vol 592. IFIP Springer, pp 443–450
12. Psarommatis F, Gharaei A, Kiritsis D (2020b) Identification of the critical reaction times
for re-scheduling flexible job shops for different types of unexpected events. Procedia CIRP
93:903–908
13. Psarommatis F, Vuichard M, Kiritsis D (2020) Improved heuristics algorithms for re-scheduling
flexible job shops in the era of zero-defect manufacturing. Procedia Manuf 51:1485–1490
14. Psarommatis F, Kiritsis D, Kiritsis D, Moon I, Park J, von Cieminski G, Lee GM (2018) A
scheduling tool for achieving zero defect manufacturing (ZDM): a conceptual framework. IFIP
Adv Inf Commun Technol 536:271–278
15. Hu C, Gao W, Xu C (2020b) Study on the application of digital twin technology in complex
electronic equipment. Lecture Notes Electr Eng 589:123–137
16. Li X, Wang L, Zhu C, Liu Z (2020) Framework for manufacturing-tasks semantic modelling
and manufacturing-resource recommendation for digital twin shop-floor. J Manuf Syst
17. Wang Y, Wu Z (2020) Digital twin-based production scheduling system for heavy truck frame
shop. Proc Inst Mech Eng Part C J Mech Eng Sci
18. Bilberg A, Malik AA (2019) Digital twin driven human-robot collaborative assembly. CIRP
Ann 68:499–502
19. Zupan H, Žerovnik J, Herakovič N (2018) Local search with discrete event simulation for the
job shop scheduling problem. Stud Comput Intell 762:371–380
20. Jones D, Snider C, Nassehi A, Yon J, Hicks B (2020) Characterising the digital twin: a
systematic literature review. CIRP J Manuf Sci Technol 2020(29):36–52
21. Kritzinger W, Karner M, Traar G, Henjes J, Sihn W (2018) Digital twin in manufacturing: a
categorical literature review and classification. IFAC-PapersOnLine 51:1016–1022
Chapter 19
Overview of Lean Production Under
Uncertainty
Tania Rojas , Josefa Mula , and Raquel Sanchis
Abstract This article presents a literature review on the application of lean manu-
facturing (LM) techniques under a context of uncertainty. Forty articles have been
identified, reviewed, and classified according to the following criteria: keywords,
application context, modeling approach, LM techniques/tools, type of LM waste, type
of uncertainty, and software tool. This classification emphasizes the types of uncer-
tainty inherent in lean production planning processes and the modeling approaches
for optimization. The selection of the articles has been based on those scientific
journals containing a higher representation of papers within this context. The main
findings of this literature review point to the use of three main lean manufacturing
tools; the most used modeling approach, which is interpretive structural modeling;
and the main uncertainty studied, which is demand. From the results of this study, it
was found that research and experimentation in LM applications under an uncertainty
context only represent 10% of the selected articles, making it an underresearched
topic that requires future research efforts.
Keywords Lean manufacturing · Uncertainty · Production planning ·

Optimization
T. Rojas · J. Mula · R. Sanchis (B)

València, c/Alarcón, 1, 03801 Alcoy, Alicante, Spain
e-mail: rsanchis@cigip.upv.es
T. Rojas
e-mail: taropar@doctor.upv.es
J. Mula
T. Rojas
Department of Industrial Engineering, Universidad Politécnica Salesiana, Chambers 227, 090114
Guayaquil, Ecuador
https://doi.org/10.1007/978-3-031-29382-5_19
184 T. Rojas et al.
19.1 Introduction
The technical complexities of production systems require the development of

complete industrial sectors and efficient interaction between them and/or with the
international economy [1]. In this context, lean manufacturing (LM) practices are
oriented toward a waste reduction approach [2–5]. However, although LM is applied
in numerous industries, it presents problems when it is applied under a context of
uncertainty. Uncertainty is the difference between the amounts of information needed
to perform a task and the information that is possessed [6]. The main types of uncer-
tainty in production processes are: (i) system uncertainties, which are those inherent
to the production process itself; and (ii) environmental uncertainties, which arise
beyond the production processes, such as demand and supply uncertainties [7].
The objective of this article is to review the scientific research literature about
proposals and applications of LM techniques in a production planning context under
uncertainty. The following sections of the article are structured as follows: Sect. 2
describes the review methodology. Section 3 presents the current findings and status
in the context of LM under uncertainty. Finally, Sect. 4 provides conclusions and
indications for future research.
The review methodology was based on the search for articles using the Scopus
database and their corresponding selection according to whether they contribute
to fulfill the objective of this review. Thus, the search was carried out using the
following keywords: “lean” and “mathematical”, “lean” and “programming”, “lean”
and “variables”, “lean” and “modeling”, “lean” and “uncertainty”, “lean” and “opti-
mization/sation”. A total of 522 documents were obtained, including scientific arti-
cles and conference proceedings. From this first result, 146 articles were identified
and selected from those related to production planning in the context of lean manufac-
turing under uncertainty. The articles were then grouped by journal and the journals
containing at least two papers were selected. Time window covers from 2003 to
2020; however, it is worth mentioning that one article, dating from 1996, is included
with 13 citations. The articles reviewed have been published in 15 different scientific
journals (Table 19.1).
19.3 Overview of the Lean Manufacturing
The articles were classified according to the following criteria: keywords, appli-
cation context, modeling approach, LM techniques/tools, type of LM waste, type
19 Overview of Lean Production Under Uncertainty 185
Table 19.1 Journals of the reviewed articles

Journals References % of total
International Journal of Lean Six Sigma 6 4.11
Advanced Materials Research 4 2.74
International Journal of Production Research 3 2.05
Benchmarking 3 2.05
International Journal of Advanced Manufacturing Technology 3 2.05
Procedia Manufacturing 3 2.05
IFIP Advances in Information and Communication Technology 2 1.37
540
International Journal of Computer Integrated Manufacturing 2 1.37
International Journal of Productivity and Quality Management 2 1.37
Journal of Advanced Manufacturing Technology 2 1.37
Journal of Manufacturing Systems 2 1.37
34
Journal of Manufacturing Technology Management 2 1.37
Journal of Modeling in Management 2 1.37
TQM Journal 2 1.37
ZWF Zeitschrift fuer Wirtschaftlichen Fabrikbetrieb 2 1.37
Total 40 100
of uncertainty and software tool. Moreover, Table 19.2 shows the scientific arti-
cles according to the following categories per criteria: keywords: L/Mo (lean and
modeling), LU (lean uncertainty), LO (lean optimization); application context: I
(industry), S (services), A (agri-food), C (construction); modeling approach: DOE
(design of experiments), SEM (structural equation modeling), ISM (interpretive
structural modeling), DES (discrete events simulation), IPA (intuitive and prag-
matic approach), SD (system dynamics), MH (metaheuristics), FL (fuzzy logic),
HS (hybrid simulation), AM (analytical model), WF (workflow), CM (conceptual
model), CoM (cost model), MCDM (multiple criteria decision-making method), MO
(multi-objective model); LM techniques/tools: VSM (value stream mapping), SMED
(single-minute exchange of die), JIT (just in time); PY (poka-yoke), KB (kanban),
A (andon), TPM (total productive maintenance), 5S, H (heijunka), G (gemba), JK
(jidoka), CEL (celular manufacturing), K (kaizen); TQM (total quality management),
SS (six sigma); type of LM waste: T (transportation), M (motion), ST (standby), INV
(inventory), OP (overproduction), R (rework), ER (process and waiting time errors);
software tool: SPSS, ARENA, AMOS (analysis of moment of structures), MATLAB
(matrix laboratory), Promodel, Excel Solver, Simul8, Minitab, N/A (not applicable);
type of uncertainty: SU (system uncertainty), EU (environment uncertainty).
Table 19.2 Overview of the LM
186
References Keywords Application context Modeling approach LM techniques/tools Type of LM waste Type of uncertainty Software tool
[2] LU S MCDM-DES VSM ES SU/EU Arena
[3] LU I SD – ES EU –
[4] LO I DOE – ER-ES SU
[5] L/Mo I ISM TPM ER SU
[8] L/Mo I IPA KB-A INV-ES SU –
[9] L/Mo I DES JIT INV SU Promodel
[10] L/Mo I WF-DES VSM INV SU Simul8
[11] LU I AM SMED INV-T-ES EU SPSS-Excel
Solver
[12] LU S N/A LPS N/A SU –
[13] LO I N/A VSM INV EU –
[14] LO I N/A – INV EU –
[15] L/Mo I SEM JK ER SU –
[16] L/Mo I FL JIT R SU –
[17] L/Mo I CoM PY ER SU –
[18] LO A DOE SS ER SU Minitab-Excel
Solver
[19] LO I MH – – SU
[20] L/Mo - SD SMED ES SU
[21] LO S DES VSM ES SU Arena
[22] L/Mo I ISM JIT – SU –
[23] L/Mo I ISM – – SU
(continued)
T. Rojas et al.
References Keywords Application context Modeling approach LM techniques/tools Type of LM waste Type of uncertainty Software tool
[24] L/Mo I CM JIT INV SU SPSS
[25] L/Mo I ISM VSM SP SU
[26] L/Mo C DOE VSM SP EU Minitab
[27] L/Mo I SEM TQM – SU AMOS
[28] LO I MO VSM ER-ES SU
[29] L/Mo – ISM – – SU
[30] L/Mo I ISM – – SU
[31] L/Mo I SEM SS ER EU AMOS
[32] L/Mo – ISM – – EU
[33] L/Mo I DES JIT M SU Arena
[34] L/Mo I CM CEL INV SU
19 Overview of Lean Production Under Uncertainty
[35] L/Mo I N/A JIT INV-ES SU

[36] L/Mo I SEM TPM – SU AMOS
[37] L/Mo I ISM TQM ER SU SPSS
[38] L/Mo S CM JIT – SU
[39] L/Mo - HS VSM – SU
[40] L/Mo I-S ISM – – SU
[41] LO S HS CEL ER EU MATLAB
[42] LO S ISM TQM ER-ES SU
[43] L/Mo S ISM SS ER-ES-M SU
187
188 T. Rojas et al.
19.4 Results
According to the literature review, it was found that 67.5% of the articles reviewed
correspond to the keywords L/Mo, 22.5% to LO, and 10% to LU. It should be noted
that the largest number of applications are related to the industrial context. Some
of the industrial applications belong to sectors such as automotive and robotics.
Whereas the service applications are referred to health and logistics, among others.
Regarding the modeling approach, ISM is one of the most used ones, specifically,
in 11 of the 40 articles reviewed, followed by those classified with HS. Regarding
LM techniques, VSM is one of the most used methodologies in the reviewed articles
being useful to identify the activities that add or not value to the processes, followed
by the JIT production planning approach.
The most frequently addressed type of waste is process errors and waiting times.
The type of uncertainty EU is addressed by 9 articles of which demand is mentioned
as the primary uncertain aspect in the context of such investigations. Finally, the most
frequently used software are Arena, MATLAB, and AMOS, respectively.
19.5 Conclusions
This paper has presented an overview of the scientific literature oriented to the
application of LM techniques in the context of production planning under uncertainty.
From the literature review carried out, it has been found that the largest number of
applications is related to the industrial field (sectors such as automotive or robotics
are studied in the different papers) and to a lesser extent to services field (highlight the
applications in the health and logistics sector). The most commonly used modeling
approach is ISM, although some of the research does not mention it directly. Addi-
tionally, SEM, DOE, MO, and MH modeling approaches are also widely used. The
most frequently used LM techniques are VSM followed by JIT, being the method-
ologies related to the initial assessment of problems in the industry. The type of LM
waste identified with the highest representation refers to errors in the processes. It is
important to highlight that there is little research on the application of LM techniques
under uncertainty. Among the articles reviewed, 77.5% addresses, in a general way,
the SU related to the internal processes of organizations and 22.5% refer to the EU
and, mainly, related to the demand. In this sense, the MCDM modeling approach
contemplates both SU and EU showing significant results of takt time optimiza-
tion for production planning [2]. The software tools used include Arena, SPSS, and
AMOS.
Finally, there is a great opportunity to deepen the relationship of the proposed clas-
sification criteria in the future research within the LU context. In this sense, novel
mechanisms to break down uncertainty into lean manageable chunks are welcome.
Additionally, more optimization and simulation optimization models to LM produc-
tion planning systems under uncertainty are required, fuzzy set could be a useful
theory to integrate uncertainty in optimization and simulation LM models, as it

previously was done for material requirement planning (MRP) and supply chain
planning systems under uncertainty [7, 44, 45].
Acknowledgements The research leading to these results received funding from the Grant
RTI2018-101344-B-I00 funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A way
of making Europe”.
References
1. Miranda J, Toirac L (2010) Indicadores de productividad para la industria Dominicana. Ciencia

y Sociedad, 235–290
2. Lu JC, Yang T, Wang CY (2011) A lean pull system design analysed by value stream mapping
and multiple criteria decision-making method under demand uncertainty. Int J Comput Integr
Manuf 24(3):211–228
3. Deif AM (2012) Dynamic analysis of a lean cell under uncertainty. Int J Prod Res 50(4):1127–
1139
4. Garzaniti NAC (2018) Optimization of multi-part 3D printing build strategies for lean product
and process development. IFIP Adv Inf Commun Technol 540:488–497
5. Mohammad IS, Oduoza CF (2019) Interactions of lean enablers in manufacturing SMEs using
interpretive structural modelling approach—a case study of KRI. Procedia Manuf 38:900–907
6. Galbaith J (1973) Designing complex organizations. Addison Wesley, Reading Mass, MA
7. Mula J, Peidro D, Díaz-Madroñero M, Vicens E (2010) Mathematical programming models
for supply chain production and transport planning. Eur J Oper Res 204(3):377–390
8. Burcher P, Dupernex S, Relph G (1996) The road to lean repetitive batch manufacturing:
modelling planning system performance. Int J Oper Prod Manage 16(2):210–220
9. Green FB, Barbee J, Cox S, Rowlett C (2006) Green Belt Six Sigma at a small company. Int J
Six Sigma Competitive Adv:179–189
10. Ajaefobi JO, Weston RH (2007) Enterprise modelling in support of the application of lean
manufacturing in SMEs. Adv Mat Res 18:359–364
11. McWilliams DL, Tetteh EG (2009) Managing lean DRC systems with demand uncertainty: an
analytical approach. Int J Adv Manuf Technol 45(9–10):1017–1032
12. Qiu X (2011) Uncertainty in project management based on lean construction implementation.
Adv Mater Res 328–330:194–198
13. Abele EMA (2012) Optimization of value streams—with lean methods and material flow
simulation to operational excellence|[Optimierung von wertströmen - Mit lean-methoden
und materialflusssimulation zu operativer exzellenz]. ZWF Zeitschrift fuer Wirtschaftlichen
Fabrikbetrieb 107(4):212–216
14. Xiong F, Yao J, Li JM (2012) Improve manufacturing execution and cycle time through database
LEAN optimization. Adv Mater Res 479–481:194–197
15. Vinodh S, Joy D (2012) Structural equation modelling of lean manufacturing practices. Int J
Prod Res 50(6):1598–1607
16. Anvari A, Zulkifli N, Yusuff RM (2013) A dynamic modeling to measure lean performance
within lean attributes. Int J Adv Manuf Technol 66(5–8):663–677
17. Wasim A, Shehab E, Abdalla H, Al-Ashaab A, Sulowski R, Alam R (2013) An innovative cost
modelling system to support lean product and process development. Int J Adv Manuf Technol
65(1–4):165–181
18. Besseris G (2014) Multi-factorial lean six sigma product optimization for quality, leanness and
safety: A case study in food product improvement. Int J Lean Six Sigma 5(3):253–278
190 T. Rojas et al.
19. S˛ekala A, Gwiazda A, Monica Z, Banaś W (2014) Optimization of the lean production process
using the virtual manufacturing cell. Adv Mater Res 1036:858–863
20. Dombrowski U (2014) Modeling the lean production system|[Auswirkungsmodell
Ganzheitliches Produktionssystem]. ZWF Zeitschrift fuer Wirtschaftlichen Fabrikbetrieb
109(12):904–908
21. Yang T, Kuo Y, Su CT, Hou CL (2015) Lean production system design for fishing net
manufacturing using lean principles and simulation optimization. J Manuf Syst 34(1):66–73
22. Dubey R, Singh T (2015) Understanding complex relationship among JIT, lean behaviour, TQM
and their antecedents using interpretive structural modelling and fuzzy MICMAC análisis. TQM
J 27(1):42–62
23. Vasanthakumar C, Vinodh S, Ramesh K (2016) Application of interpretive structural modelling
for analysis of factors influencing lean remanufacturing practices. Int J Prod Res 54(24):7439–
7452
24. McLeod AA, Stephens MP, McWilliams DL (2016) Empirical modeling of lean adoption in
small to medium size manufacturers. J Adv Manuf Syst 15(4):173–188
25. Sharma V, Dixit AR, Qadri MA (2016) Modeling lean implementation for manufacturing
sector. J Model Manage 11(2):405–426
26. Nagi M, Chen FF, Wan HD (2017) Throughput rate improvement in a multiproduct assembly
line using lean and simulation modeling and analysis. Procedia Manuf 11:593–601
27. Khalili A, Ismail MY, Karim ANM (2017) Integration of lean manufacturing and quality
management system through structural equation modelling. Int J Prod Qual Manage 20(4):534–
556
28. Uriarte AG, Ng AH, Moris MU (2018) Supporting the lean journey with simulation and
optimization in the context of Industry 4.0. Procedia Manuf 25:586–593
29. Chaple AP, Narkhede BE, Akarte MM, Raut R (2018) Interpretive framework for analyzing
lean implementation using ISM and IRP modeling. Benchmarking 25(9):3406–3442
30. Ghobakhloo M, Fathi M, Fontes DBMM, Ching NT (2018) Modeling lean manufacturing
success. J Model Manage 13(4):908–931
31. Ramadas T, Satish KP (2018) Identification and modeling of process barriers: implementing
lean manufacturing in small-and medium-size enterprises. Int J Lean Six Sigma
32. Chaple AP, Narkhede BE, Akarte MM, Raut R (2018) Modeling the lean barriers for successful
lean implementation: TISM approach. Int J Lean Six Sigma
33. Mahmood WW, Yusup MZ, Salleh MR, Muhamad MR, Ab Rahman MN (2018) Simulation
modelling for lean production: case study. J Adv Manuf Technol 12(1):223–232
34. Abdullah R, Hashim H, Rahman MNA, Salleh MR (2018) Development of enterprise human
system modelling framework in support of cellular modelling framework in support of cellular
manufacturing lean operation. J Adv Manuf Technol 12(1):235–246
35. Camuffo A, Gerli F (2018) Modeling management behaviors in lean production environments.
Int J Oper Prod Manage 38(2):403–423
36. Khalili A, Ismail MY, Karim ANM, Daud MRC (2018) Soft total quality management and
lean manufacturing initiatives: model development through structural equation modelling. Int
J Product Qual Manage 28(1):1–30
37. Singh M, Kumar P, Rathi R (2018) Modelling the barriers of lean six sigma for indian micro-
small medium enterprises: an ISM and MICMAC approach. TQM J 31(5):673–695
38. Baliga R, Raut R, Kamble S (2019) The effect of motivators, supply, and lean management on
sustainable supply chain management practices and performance: systematic literature review
and modeling. Benchmarking 27(1):347–381
39. Oleghe O, Salonitis K (2019) Hybrid simulation modelling of the human-production process
interface in lean manufacturing systems. Int J Lean Six Sigma 10(2):665–690
40. Sindhwani R, Mittal VK, Singh PL, Aggarwal A, Gautam N (2020) Modelling and analysis
of barriers affecting the implementation of lean green agile manufacturing system (LGAMS).
Benchmarking 26(2):498–529
41. Kant R, Pattanaik LN, Pandey V (2020) Sequential optimisation of reconfigurable machine
cell feeders and production sequence during lean assembly. Int J Comput Integr Manuf 33(1–
2):62–78
42. Vaishnavi V, Suresh M (2020) Modelling of readiness factors for the implementation of lean
six sigma in healthcare organizations. Int J Lean Six Sigma 11(4):597–633
43. Ajmera P, Jain V (2020) A fuzzy interpretive structural modeling approach for evaluating
the factors affecting lean implementation in Indian healthcare industry. Int J Lean Six Sigma
11(2):376–397
44. Peidro D, Mula J, Poler R, Verdegay JL (2009) Fuzzy optimization for supply chain planning
under supply, demand and process undertainties. Fuzzy Sets Syst 160(18):2640–2657
45. Campuzano F, Mula J, Peidro D (2010) Fuzzy estimations and system dynamics for improving
supply chains. Fuzzy Sets Syst 161(11):1530–1542
Chapter 20
Defining Production Planning Problems
in Additive Manufacturing
J. de Antón , D. Poza , A. López-Paredes , and F. Villafáñez
Abstract Additive manufacturing (AM) introduces a set of technology-specific

constraints that increase the complexity of production planning. As research in the
production planning of AM facilities is gaining attention, a literature review revealed
a lack of uniformity in the design of efficient approaches. At this stage, it is crucial
to bring clarity to the identification of all the problems to solve while focusing on
the processing sequence. For this reason, this paper presents a first approach for
production planning in AM in conjunction with an unambiguous definition of their
problems and subproblems.
Keywords Production planning · Additive manufacturing · 3D printing ·

Nesting · Scheduling
20.1 Introduction
Contrary to traditional manufacturing, additive manufacturing (AM) offers desirable

features for customized production such as flexibility and easy integration within the
cloud manufacturing framework [1, 2]. Moreover, AM brings a great opportunity
for developing sustainable production systems, mainly by decentralizing the supply
chain due to its energy-saving potential [3–5]. Whereas topics on production planning
in traditional manufacturing have been extensively studied, issues regarding opera-
tions management (OM) in AM-based production systems have not been completely
addressed. This is especially noticeable with respect to planning and scheduling
problems [6].
J. de Antón (B) · D. Poza · F. Villafáñez

Department of Business Organisation and Market Research, INSISOC—University of Valladolid,
Valladolid, Spain
e-mail: juan.anton@uva.es
A. López-Paredes
Department of Economics and Business Management, Universidad de Málaga, Málaga, Spain
https://doi.org/10.1007/978-3-031-29382-5_20
194 J. de Antón et al.
As AM continues its journey to maturity as a production technique, the focus

has been set on reaching mass production of customized products [7]. From the
OM perspective, this is not a trivial problem. Reaching mass production entails the
manufacture of large quantities of heterogeneous parts [8] which involves a variety
of subproblems such as manufacturing multiple parts in the same batch, scheduling
one machine to produce several batches, or scheduling multiple machines to produce
several batches in parallel and so forth [9]. These problems are not new to production
planning. However, when it comes to AM, production planning triggers a set of
additional problems associated with this technology such as part orientation, part
location within the manufacturing surface, or part grouping [10, 11]. These additional
problems appear in the AM literature under the umbrella term “nesting”. According
to Araújo et al. [12], nesting refers to the problem of packing as many parts as
possible in a given build volume. Frequently, nesting is accompanied by the term
“scheduling” in works addressing AM planning issues. In the AM field, scheduling
is associated with the problem of properly distributing the batches resulting from the
nesting to one or several AM machines [13]. These two concepts must be understood
as complementary and interrelated problems in the production planning process for
AM.
Nesting and scheduling problems have already been addressed in AM production
contexts following different approaches. First, the research focus was on solving the
nesting problem alone [14]. Later, scheduling concerns started to be incorporated into
nesting solutions [15]. More recently, works addressing an integrated formulation of
nesting and scheduling have been presented [16, 17]. We claim that, when merging
both nesting and scheduling problems, it is common to fall into vague definitions
of the subproblems included in each concept. This may lead to imprecisions that
lead to a myriad of approaches to solving only some portions of the whole planning
problem. Furthermore, since nesting in AM already integrates scheduling concerns,
the optimization of each subproblem individually does not guarantee an optimized
global solution.
In this context, the aim of this paper is to propose a scheme that includes a
clear definition of the specific problems faced by production planning in AM. These
problems will be presented sequentially, starting from the reception of part orders
from customers and ending with the production orders programmed for each machine.
This first approach aims to provide researchers with a reference to find an explicit
definition of the different subproblems that arise in production planning in AM.
The research literature on production problems in AM has evolved significantly over

time. In the beginning, the nesting problem was defined for AM production contexts
and the first models seeking its optimization were developed. Later, the nesting
problem was extended by including scheduling questions. More recently, nesting
20 Defining Production Planning Problems in Additive Manufacturing 195
and scheduling have been addressed together, which results in more complex and
comprehensive models.
Research about nesting in AM started in the 1990s with the adventure of stere-
olithography (SLA). Wodziak et al. [18] studied the problem of packing parts into
the vat volume of an SLA machine seeking to efficiently occupy the available space.
However, it was not until the late 2000s when it started to receive greater atten-
tion [19]. In the middle of the 2010s, the nesting problem began to incorporate
scheduling concerns. The problem was then trying to assign objects to AM machines
and schedule their production to minimize time and cost [7]. In the last six years,
research topics regarding production planning and scheduling in AM have experi-
enced exponential growth and this trend is expected to continue at the same pace
[20]. Throughout this period, some of these works individually deal with nesting
and scheduling, and other works address nesting and scheduling problems following
an integrated approach. Several models have been proposed to describe some of the
production problems in AM. Also, several algorithmic solutions have been devel-
oped to solve these problems [21, 22]. However, very few reviews and taxonomies
have been proposed to categorize these complex problems.
The first reviews and taxonomies were proposed to tackle nesting problems exclu-
sively. Zhang et al. [19] reviewed the previous works on nesting and classified them
according to seven parameters. They considered issues related to placement, orienta-
tion, and rotation problems. Besides, they developed an interesting classification for
nesting problems based on production context in which they drew an analogy with
classical problems from operations research (OR). Shortly after, Araújo et al. [12]
proposed a taxonomy for nesting problems based on four criteria: dimensionality of
the problem, optimization criteria, build volume, and attributes of parts. Also, they
reviewed and classified under this taxonomy the existing works about nesting. In the
following year, Li et al. [21] reviewed the literature on production planning in AM
already including works about both nesting and scheduling. Nesting was defined as a
bin packing problem, while scheduling was regarded as a batch processing problem.
A yes/no classification according to the optimization objectives and the contempla-
tion of profit/cost and time concerns was presented. Under a similar framework,
Aloui & Hadj-Hamou [22] recently extended this review with new contributions
and also added data regarding both the solving approaches employed in each work
and the AM technologies to which they were applied. The only review systemati-
cally covering the approaches for nesting, scheduling, and the integration of both
was presented by Oh et al. [7]. They proposed a taxonomy based on a physical
hierarchy consisting of part, machine, and AM Machine levels to classify nesting
(NfAM), scheduling (SfAM), and nesting-scheduling (NSfAM) problems. They also
considered eight AM-specific supplementary criteria to refine the classification.
Although there already exist reviews on AM production planning, their proposals
still consider only parts of the whole range of production planning problems. At the
same time, these works fall short in providing researchers and practitioners with a
robust scheme for identifying the wide variety of subproblems enclosed in nesting
and scheduling. Thus, new contributions are necessary to ensure a straight path of
AM to maturity in OM. To bridge this gap, in the following section we propose
a sequential scheme for production planning which can be used as a reference for
approaching these AM problems.
20.3 Scheme for Production Planning in Additive

Manufacturing
In this section, the concepts of nesting and scheduling are reviewed, and the partic-
ularities introduced by the AM field to them are exposed. Next, a scheme for the
classification of production planning problems in AM is presented.
20.3.1 Nesting and Scheduling Concepts in Additive

Manufacturing
The nesting problem is a classical cutting and packing problem from the OR field.
It describes the problem in which a set of two-dimensional irregular objects has to
be laid out on a rectangular large object [24]. This definition has been adapted by
AM researchers to describe the problem of determining an optimized layout of parts
in a 3D printer. Moreover, it has been extended to include other issues that appear
when facing the AM nesting, such as part orientation, part location, and part rotation.
Anyhow, it is commonly agreed that the two main subproblems regarding nesting
in AM are the allocation of parts to batches in a printer and the placement of those
parts in the manufacturing surface of the printer.
Sometimes the words scheduling and production planning are used as one unique
concept, which may lead to misconceptions. Production planning refers to medium-
term decisions such as the assignment of production targets and transportation plan-
ning. On the contrary, scheduling refers to short-term planning at the production
level. Hence, scheduling is concerned about the daily or weekly assignment of
tasks to resources and the sequencing of tasks on each resource unit [25]. Although
the term scheduling was originally used to describe only the allocation of tasks to
resources over time, this term has evolved to not only include assignment issues but
also consider the sequencing and the timing of tasks [26]. In this work, we under-
stand scheduling as the integration of the following three subproblems: allocation,
sequencing, and timing. These subproblems are interrelated and can be addressed
integrally. On the other hand, we restrict the concept of production planning to
production issues, thus excluding the transportation side.
In AM, the allocation of parts to a machine for their manufacture is solved
throughout the nesting step, along with the placement problem. In turn, the deter-
mination of the proper placement of a part in the surface (or the volume in the
3D nesting) prompts two subproblems: finding the best orientation for the part and
finding the best location for the part in the surface. As we have introduced above, the
Fig. 20.1 Problem sequence for the production planning in AM
scheduling results from the integration of allocation, sequencing, and timing prob-
lems. In AM, the allocation is solved in the nesting step together with the placement
in many production contexts. Consequently, we redefine the scheduling subproblems
for the AM context as nesting, sequencing, and timing.
20.3.2 Definition of Production Planning Problems

in Additive Manufacturing
The scheme that we propose describes the sequence followed by the manufacturer
when he or she addresses the production planning of the 3D factory. It is assumed
that several printers with different speed and size features are available. Figure 20.1
shows this sequence in a two-stage planning procedure.
The procedure starts with the reception of several part orders from distributed
customers. Then, these parts, whether grouped or not, must be assigned to a suitable
AM machine (i.e., 3D printer). These steps comprise the first stage, which is the
Part/Machine assignment (Fig. 20.1, left).
Hence, the allocation and placement of parts in each machine are solved by
creating batches in the nesting step. Finally, the sequence for processing all those
batches and their production start-up times are determined. These activities belong
to the second stage: Machine scheduling (Fig. 20.1, right). In the second stage, we
consider two possible variants: the single-batch planning case and the multi-batch
planning case.
A thorough definition of each problem within the AM production sequence is
provided below. The main features of each problem are summarized in Table 20.1.
• Grouping: Parts are sorted and grouped based on one or more criteria, which may
be only dependent on the technology constraints or include process and service
considerations. Technology-dependent criteria recurrently used are volume, mate-
rial, accuracy, or surface quality. A process-related criterion used on occasion is
Table 20.1 Description of the production problems in AM and their possible variants
Problem Variant Input Output Objective
List of subsets of Group parts to improve
Grouping – Pool of parts
parts the planning process
Single part Pool of parts Identify the printers capa-

List of capable ma-
Matching ble of manufacturing each
chines
Multi-part List of subsets of parts part/group
List of capable ma- The machine as-

Choose the most suitable
Selection – chines for each signed for each
printer
part/group part/group
List of parts assigned List of parts in a Determine the parts for
Allocation –
to a printer batch each batch
Determine the locations

Layout distribution
Placement – List of parts in a batch and orientations of the
in a batch
parts in a batch
List of batches and

Multi- Batch sequence on Determine the sequence of
Sequencing their corresponding
batch each printer the batches in a printer
printers
Single The layout distribution Start time for pro-
batch of the batch ducing the batch
Determine the start time
Timing Start time for pro-
Multi- The sequence of for each batch in a printer
ducing each batch in
batch batches on printers
their printer
a similar height of parts, while the service-related main criterion is a similar due
date. The output of this problem is a list of subsets of parts.
• Matching: Parts are assigned a list of machines whose features enable them to
produce the parts. The matching might be done for each individual part or it might
find suitable printers for a group of parts previously sorted in the grouping step.
This variant is indicated in Table 20.1 as single part or multipart. In any case, the
output of the matching is a list of capable machines. In this step, the availability
of machines is not considered.
• Selection: Among the list of capable printers for each part or group of parts, one
must be selected. This choice should respond to some optimization objectives. A
common practice is to select the lowest-performance printer from the list, since
the fact that it appears in the list already guarantees its capability for the task.
The output of the selection phase is the machine assigned for the part/group
manufacturing. This selection step might be merged with the matching one so
that the list generation and final selection are solved integrally.
• Nesting: Once the machine selection for a part or group is determined, the next
is to create the batches in which the parts will be manufactured by the printer.
Each batch is characterized by two complementary pieces of information: a list
of parts that compose the batch and the layout information of a batch, including
the location and orientation of those parts on the surface. These two records
result from the resolution of the allocation problem and the placement problem,
respectively.
– Allocation: Starting from a set of parts already assigned to a printer, we need
to reorganize them into subsets for their manufacture. Each subset corresponds
to one batch. Thus, the output of the allocation problem is a record of the parts
included in each batch.
– Placement: It is the problem of how to properly place multiple parts in the
manufacturing surface (or volume) of a printer so that some production-related
parameter is optimized (e.g., the use rate of the machine) while ensuring the
final quality of parts. The output of the placement problem is the layout distri-
bution of parts on the surface (i.e., the location and orientation of each part on
the surface). For the multi-batch case, there will be as many layout distributions
as batches in the problem.
• Sequencing: This problem is only addressed in the multi-batch case. From a set
of batches, the issue is to decide the sequence in which they will be processed.
Frequently, priority rules are set to help determine the best sequence. This step can
also be addressed jointly with the placement (i.e., the batches are being scheduled
one after the other as their layouts are being determined), or it can be addressed
jointly with the timing problem as well. The output of this problem is the batch
sequence on the machine.
• Timing: After deciding the order in which batches will be processed, the time
when each batch will start to be produced is described as the timing problem. It
is very common that sequencing and timing are determined at the same time in
an integral approach. The output of the timing step is a time schedule for every
batch.
20.4 Conclusions
In this paper, we have provided a preliminary scheme for production planning in

AM. This first approach aims to bring clarity to the myriad of production problems
that have emerged as AM production planning advances toward its maturity. Also,
we intend to complement the current existing reviews and taxonomies of production
problems in AM from an OM perspective.
In the proposed scheme, these problems have been grouped in a two-stage
sequential structure. Starting with the reception of part orders from distributed
customers, the scheme describes the sequence followed by the manufacturer when
he or she addresses the production planning of the 3D factory. We have also proposed
definitions for each problem and summarized their inputs, outputs, and objectives.
We are hopeful that the proposed scheme provides a helpful tool for researchers
and practitioners to better understand the planning process in AM, and to identify
the various interrelated problems to handle. It also opens the opportunity to use this
scheme as the basis for a systematic review and categorization of the heterogeneous
models for nesting and scheduling in AM proposed in the literature. It will allow
pinpointing the best solutions developed so far for the optimization of production
planning in AM.
References
1. Dvorak F, Micali M, Mathieug M (2018) Planning and scheduling in additive manufacturing.

Intel Artif 21:40–52. https://doi.org/10.4114/intartif.vol21iss62pp40-52
2. Zhou L, Zhang L, Laili Y et al (2018) Multi-task scheduling of distributed 3D printing services
in cloud manufacturing. Int J Adv Manuf Technol 96:3003–3017. https://doi.org/10.1007/s00
170-017-1543-z
3. Mehrpouya M, Vosooghnia A, Dehghanghadikolaei A, Fotovvati B (2021) The benefits of
additive manufacturing for sustainable design and production
4. González-Varona JM, Poza D, Acebes F et al (2020) New business models for sustainable spare
parts logistics: a case study. Sustainability 12:3071. https://doi.org/10.3390/SU12083071
5. Majeed A, Zhang Y, Ren S et al (2021) A big data-driven framework for sustainable and
smart additive manufacturing. Robot Comput Integr Manuf 67:102026. https://doi.org/10.1016/
j.rcim.2020.102026
6. Manco P, MacChiaroli R, Maresca P, Fera M (2019) The additive manufacturing operations
management maturity: a closed or an open issue? Procedia Manuf 41:98–105. https://doi.org/
10.1016/j.promfg.2019.07.034
7. Oh Y, Witherell P, Lu Y, Sprock T (2020) Nesting and scheduling problems for additive manu-
facturing: a taxonomy and review. Addit Manuf 36:101492. https://doi.org/10.1016/j.addma.
2020.101492
8. Khorram Niaki M, Nonino F (2017) Additive manufacturing management: a review and future
research agenda. Int J Prod Res 55:1419–1439. https://doi.org/10.1080/00207543.2016.122
9064
9. Kucukkoc I (2019) MILP models to minimise makespan in additive manufacturing machine
scheduling problems. Comput Oper Res 105:58–67. https://doi.org/10.1016/j.cor.2019.01.006
10. Li Q, Kucukkoc I, Zhang DZ (2017) Production planning in additive manufacturing and 3D
printing. Comput Oper Res 83:1339–1351. https://doi.org/10.1016/j.cor.2017.01.013
11. Zhang Y, Bernard A, Harik R, Karunakaran KP (2017) Build orientation optimization for
multi-part production in additive manufacturing. J Intell Manuf 28:1393–1407. https://doi.org/
10.1007/s10845-015-1057-1
12. Araújo LJP, Özcan E, Atkin JAD, Baumers M (2018) Analysis of irregular three-dimensional
packing problems in additive manufacturing: a new taxonomy and dataset. Int J Prod Res
57:5920–5934. https://doi.org/10.1080/00207543.2018.1534016
13. Oh Y, Zhou C, Behdad S (2018) Production planning for mass customization in additive
manufacturing: build orientation determination, 2D packing and scheduling. Proc ASME Des
Eng Tech Conf 2A-2018.https://doi.org/10.1115/DETC2018-85639
14. Canellidis V, Giannatsis J, Dedoussis V (2013) Efficient parts nesting schemes for improving
stereolithography utilization. CAD Comput Aided Des 45:875–886. https://doi.org/10.1016/j.
cad.2012.12.002
15. Chergui A, Hadj-Hamou K, Vignat F (2018) Production scheduling and nesting in additive
manufacturing. Comput Ind Eng 126:292–301. https://doi.org/10.1016/j.cie.2018.09.048
16. Zhang J, Yao X, Li Y (2020) Improved evolutionary algorithm for parallel batch processing
machine scheduling in additive manufacturing. Int J Prod Res 58:2263–2282. https://doi.org/
10.1080/00207543.2019.1617447
17. Kucukkoc I, Li Z, Li Q (2021) 2D nesting and scheduling in metal additive manufacturing.
Commun Comput Inf Sci 1407 CCIS:97–112. https://doi.org/10.1007/978-3-030-76307-7_8
18. Wodziak JR, Fadel GM, Kirschman C (1994) A genetic algorithm for optimizing multiple part
placement to reduce build time. Proc Fifth Int Conf Rapid Prototyping 201:210
19. Zhang Y, Gupta RK, Bernard A (2016) Two-dimensional placement optimization for multi-
parts production in additive manufacturing. Robot Comput Integr Manuf 38:102–117. https://
doi.org/10.1016/j.rcim.2015.11.003
20. Fera M, Macchiaroli R, Fruggiero F, Lambiase A (2020) A modified tabu search algorithm
for the single-machine scheduling problem using additive manufacturing technology. Int J Ind
Eng Comput 11:401–414. https://doi.org/10.5267/j.ijiec.2020.1.001
21. de Antón J, Senovilla J, González JM et al (2020) Production planning in 3D printing factories.
Int J Prod Manage Eng 8:75–86. https://doi.org/10.4995/ijpme.2020.12944
22. Aloui A, Hadj-Hamou K (2021) A heuristic approach for a scheduling problem in additive
manufacturing under technological constraints. Comput Ind Eng 154:107115. https://doi.org/
10.1016/j.cie.2021.107115
23. Li Q, Zhang D, Kucukkoc I (2019) Order acceptance and scheduling in direct digital manu-
facturing with additive manufacturing. IFAC-PapersOnLine 52:1016–1021. https://doi.org/10.
1016/j.ifacol.2019.11.328
24. Wäscher G, Haußner H, Schumann H (2007) An improved typology of cutting and packing
problems. Eur J Oper Res 183:1109–1130. https://doi.org/10.1016/j.ejor.2005.12.047
25. Maravelias CT, Sung C (2009) Integration of production planning and scheduling: overview,
challenges and opportunities. Comput Chem Eng 33:1919–1930. https://doi.org/10.1016/j.com
pchemeng.2009.06.007
26. Maravelias CT (2012) General framework and modeling approach classification for chemical
production scheduling. AIChE J 59:215–228. https://doi.org/10.1002/aic
Chapter 21
Digital Twin for a Zero-defect Operations
Planning in Supply Chain 4.0
Abstract This research project proposes the development of a digital twin (DT)
that simulates the behavior of the zero-defect planning system of a supply chain. The
research will focus on the incorporation of new zero-defect manufacturing (ZDM)
technologies generated from the DT perspective. The production technologies to be
proposed will be oriented toward the development of new models and optimization
algorithms for the ZDM planning problem in the new digitalized supply network
context. The modeling domain will involve up to the second-tier supplier in the
supply chain at the tactical and operational decision levels.
Keywords Supply chain 4.0 · Operations planning · Digital twin · Zero-defect

manufacturing
21.1 Introduction
The industrial sector has been immersed in a stream of profound changes driven by
the latest technological advances, especially those due to the introduction of digital
technologies embodied in the Industry 4.0 paradigm [1]. Supply chain 4.0 (SC4.0)
emerges as the projection of Industry 4.0 in the specific supply chain environment [2].
The implementation of cyberphysical systems (CPS) interconnected by the Industrial
Internet of Things (IIoT) and the Internet of Services (IoS), management of collected
big data, cloud services and the increasing generalization of artificial intelligence (AI)
at all levels, among other enabling technologies, is an opportunity for companies [2].
J. C. Serrano-Ruiz (B) · J. Mula · R. Poler

Centro de Investigación en Gestión e Ingeniería de la Producción (CIGIP), Universitat Politècnica
de València, 03801 Alcoy, Spain
e-mail: jserrano@cigip.upv.es
J. Mula
R. Poler
https://doi.org/10.1007/978-3-031-29382-5_21
It also poses a scenario with a considerable number of issues to be solved. The

scale of change implied by this digitization process requires the generation of new
reference frameworks and models that transfer knowledge to systems and guide their
managers.
The “chain” term conveys a connotation of linearity that was considered appro-
priate in the 1980s, when the supply chain concept became a widely used term in
the business world. This linearity stems from the traditional supply chain view as
the sequence of events that the product or service undergoes from conception to
consumption [3]. This is a vision in which the product is placed at the center of the
model, and the main pursued objective is to achieve its optimal and stable itinerary
all along the production and logistics system. This perspective far from faithfully
reproduces the new and complex situation implied by SC4.0, in which the primary
focus of attention is now paid to customers, and the objective to fulfill is to effi-
ciently meet their demand. A paradigm shift has taken place, and what used to be
“the seller’s market” is now “the buyer’s market”, which means that buyers can define
conditions. This trend leads to increasing product individualization and, in extreme
cases, even to “batch size one” [4]. In this context, speed, flexibility, and adapt-
ability are critical factors to consider. Sustainability emerges as a secondary, but no
less important, focus of attention. Hence, the adoption of management practices for
supply chain systems that consider all aspects of sustainability—economic, social,
environmental—and exploit the digital transformation characteristics that Industry
4.0 represents is a relevant topic that requires study [5].
This research work proposes new operations planning (OP) technologies for the
supply chain that address the aforementioned factors of (1) speed of response, (2) flex-
ibility, (3) adaptability, and (4) sustainability from the general perspective required to
deploy specific enabling technologies to the SC4.0 paradigm, and by placing special
emphasis on the aptitude shown by DT and ZDM technologies to meet the chal-
lenge of successfully addressing the four factors regarded above. The potential of
DT technology to contribute to planning processes is very high. The implementa-
tion of virtual surrogates of processes with which they can be visualized, analyzed,
understood, modeled, simulated, optimized, or predicted [6, 7] allows a simulation
environment to be generated in which the reduction of process times and resource
consumption is favored, and with it the capacity to introduce changes into facilitated
planning to, thus, achieve at faster speed of response, flexibility, and adaptability. It is
worth noting that the expected benefits of successful ZDM introduction include not
only cost reduction, increased efficiency, and more predictable product quality but
also enhanced sustainability, which will come from reduced energy use and resource
consumption [8]. The combined use of DT and ZDM technologies in the supply
chain is moving the supply chain toward the SC4.0 paradigm.
The remainder of this paper is organized as follows. Section 21.2 increases
details on the objectives and methodology. Section 21.3 presents the related liter-
ature. Section 21.4 describes the initial conceptual proposal. Section 21.5 discusses
the main results and contributions. Finally, Sect. 21.6 provides the conclusions and
further research.
21 Digital Twin for a Zero-defect Operations Planning in Supply Chain 4.0 205
21.2 The Main Research Guidelines
The main objective of this project is to generate a digital model that simulates zero-
defect planning processes in the SC4.0 ecosystem. The entire addressed problem will
initially be defined conceptually but, subsequently, for the purposes of this research,
and at the descriptive and experimental levels, it will be limited to the second-tier
supplier in the supply chain, and at the tactical and operational decision levels.
The specific objectives pursued with the research are as to: (i) identify in the
scientific field the current advances and deficiencies in OP DT implementation into
the supply chain, and its various orientations at the tactical and operational levels
toward ZDM, by means of a literature review; (ii) propose a metamodel to support
the automation and cooperative coordination of OP DT in the supply chain based
on the integration of zero-defect planning models that contemplate the characteris-
tics of an I4.0 environment; (iii) put forward optimization, heuristic, metaheuristic,
metaheuristic, and simulation models and algorithms for zero-defect planning based
on the above-proposed conceptual models needed to develop the proposed SC4.0
planning DT; (iv) provide the empirical research of the proposed models and tools.
21.2.1 Research Methodology
This study is based on constructivist research, widely used in areas such as finance
[9], logistics [10], project management [11], or computer science [12]. This research
methodology focuses on the generation of solutions to concrete problems by the
creation of constructs according to the innovative constructivism concept [13]. A
construct can be a new algorithm, a new mathematical model, or a new conceptual
model or framework. The solution-creating constructive process is based on a set of
phases that start with the elicitation of the problem to be addressed and continue to:
(1) obtain exhaustive knowledge about the problem to be solved; (2) construct the
solution to the problem with an appropriate construct; (3) demonstrate the correct
functioning of the generated solution and its benefits; (4) examine the scope of
applying the obtained solution.
The literature that addresses the role of DT in supply chain OP is firstly reviewed.
Thereupon, the focus lies on the literature that has considered the ZDM strategy
within the supply chain to some extent. Subsequently, the literature that has combined
both approaches together is studied with the intention to exploit the mutual synergies
of DT and ZDM in the supply chain. A table defining the main concepts used in this
research is provided in Table 21.1:
Table 21.1 Definition of the main concepts involved

Concept Definition
Industry 4.0 A combination of digital technology with manufacturing that
transforms the industrial production to the next level [14]. Industry
4.0 involves the technical integration of CPS into manufacturing
and logistics, and the use of the IoT and services in industrial
processes [15]
Supply chain 4.0 The supply chain created as a result of the new digital era brought
forth by the fourth industrial revolution [16], Industry 4.0. Several
distinctive terms have been used to describe supply chain 4.0,
including smart supply chain, digital supply chain or intelligent
supply chain [17]
Operations planning A major planning process with important integration power by
connecting different company functions (marketing, financial,
production, etc.) with various points of view, objectives, and
constraints. It supports both vertical integration in relating
strategic and financial plans to operational plans and integration
between companies in the supply chain [18]. A process designed
to help companies to better align customer demand with product
supply [19]
Digital twin A virtual representation of a production system that is able to run
in different simulation disciplines that is characterized by
synchronization between virtual and real systems thanks to sensed
data and connected smart devices, mathematical models, and
real-time data processing [20]
Zero-defect manufacturing A manufacturing strategy which, by assuming that errors and
failures will always exist, focuses on minimizing and detecting
them online so that no production output that is deviated from the
specification advances to the next step [8]
Regarding DT enabling the supply chain OP, dos Santos et al. [21] propose a
continuous decision support system, a DT, that integrates two widely used tech-
niques, namely discrete event simulation and forecasting methods, which can be
used for several operational problems, for instance, OP at its different decision levels.
Wang et al. [22] envision a supply chain planning based on the theoretical founda-
tions and enabling technologies of DT and detail its benefits and potentials in this
specific environment, compared to previous planning approaches in demand forecast,
aggregate planning, and inventory planning terms. Biesinger et al. [23] provide an
approach to tackle the increasing change of production issue that leads to differences
between the current manufacturing condition and planning status, by means of a DT
that enables faster product integration and Industry 4.0 concepts. A case study is
presented by Agostino et al. [24], who firstly discuss the application of simulation
models in production and logistic systems by a DT approach for OP, using current
CPS state data in real time; finally, they evaluate it by means of a real-world scenario
that involves a manufacturer supplying the automotive industry with mechanical
parts. Finally, Maitreesorasuntee et al. [25] discuss how a DT could be used for plan-
ning and scheduling to manage machinery setup complexity, prioritize production,
fulfill inventory without shortage, and conduct what-if analyses. They compare it for
scenarios with different schedules.
To the ZDM strategy in the supply chain environment, contributions from the
scientific community are still lacking to date. For a significant portion of authors,
the zero-defect outcome is not the direct effect of a specific strategy such as ZDM,
but the indirect effect of other different strategies. In addition, many agree about
addressing a topic other than the PO, such as sustainability or quality management.
For example, Thakur and Mangla [26] use the zero-defect concept in the supply
chain as the effect of sustainable operational practices. In contrast, Siddh et al. [27]
consider the zero-defect outcome to be an effect of integrating lean six sigma into the
supply chain as their central idea is that if you know how many defects the process
has, then you can also systematically find out how to eliminate them. Pardamean and
Wibisono [28] also address the impact of six sigma on supply chain performance
by increasing process capability in the value stream, which indirectly leads to a
zero-defect outcome. Finally, Ewald and Schupp [29], focus more on the zero-defect
philosophy and propose a unique approach: they consider that, in order to achieve the
ultimate goal of zero-defects, then managing the customer complaint process should
be investigated and optimized together with a cross-functional team as they argue
that this can generate a positive effect on improving supplier quality.
No literature has been found that addresses a DT enabling ZDM supply chain OP.
From the reviewed literature, it can be concluded that: (i) using a DT as a joint
enabler of OP processes in the supply chain domain has been scarcely addressed to
date; (ii) the zero-defect concept in the supply chain context does not usually appear
as a strategy per se, but as the consequence of applying other strategies; (iii) the joint
use of a DT and ZDM technologies in the general OP domain has not been addressed
by the scientific community because existing contributions by such an approach
have usually focused on a single OP problem (e.g., production planning, production
scheduling, capacity planning, materials planning, job scheduling, or distribution
planning). Hence, a knowledge gap appears for approaches that aggregate the set of
problems as a single superordinate entity.
21.4 Initial Conceptual Proposal
In order to establish the initial reference framework for research, a metamodel based
on a DT is proposed in which the physical plane is defined by both the main processes
making up the supply chain’s OP and the resources that it contemplates and requires
to execute it.
These physical plane processes and resources are virtually replicated in two
different, but complementary, virtual planes: (1) a secondary virtual plane, or a
support plane, in which physical processes are translated, on the one hand, into
the aggregate computational processes needed to solve the posed planning problems
and, on the other hand, into data to feed these computational processes; (2) a primary
virtual plane supported by the secondary, or interface, in which the processes and
Fig. 21.1 Conceptual framework of the DT-based and ZDM-oriented manufacturer’s OP
data from the previous virtual plane are transformed into intelligible information for
the human operator.
Within the framework of this research, the supply chain is assumed as a single
space formed by all the agents involved in it. All the agents use blocks of data
and information, which are personalized to each individual role but, despite being
different, all the data and information have a single common origin, the DT (see
Fig. 21.1), which facilitates the flow of data and information between agents and
allows the existence of a connection and coordination channel for the zero-defect
strategy, thus enabling at least five of the seven characteristic ZDM system areas:
(i) the monitoring of process parameters; (ii) collaborative manufacturing; (iii) data
management optimization; (iv) the reconfiguration and reorganization of production;
(v) the rescheduling of operations.
21.5 Discussion
The implementation of a DT into the supply chain to form a single common space
for all the involved agents raises OP to the level required by collaborative manufac-
turing. A single aggregated planning subject for all supply chain actors, acting in
collaborative manufacturing, guides the supply chain toward better response speed
to disruptive events, greater flexibility and adaptability, and a zero-defect result.
Together, they all contribute to a resilient supply chain.
The DT as a single source of data and information can even lay the foundations
for a customized production system, in which customers can join the supply chain
as agents and participate as required.
SC4.0 is a smart supply chain. It requires the modeling approach for each specific
planning problem to always be optimal and automatically selected so that in a context
like the conceptualized one, it is foreseeable that the DT will use analytical, heuristic,
simulation, or artificial intelligence approaches that have been adapted to the planning
problem type to be solved, the type of agent performing the instance, the problem
size, the required accuracy, and other additional variables that might have to be
considered.
The presented metamodel accepts, as explained above, the involvement of all
supply chain actors, although its development beyond the manufacturer and its
suppliers at the two closest levels is challenging and is relegated to further research
that will consider increasing the number of supplier levels, as well as the wholesale
distributor, the retail distributor, and, whenever required, the customer itself.
21.6 Conclusions and Further Work
In the initial stage of this research, framed within the knowledge area of DT-driven
supply chain OP in ZDM environments, the basic concepts and the general reference
frame that will support this project were established. The presented metamodel favors
attributing to the supply chain, in the first instance, qualities: (1) digital, as this
quality is implicit to using DT technology; (2) fast in response, as it generates a
framework of simulation, analysis, optimization, and prediction for agile planning;
(3) flexible, as it provides the supply chain with a tool like the DT, which allows the
synchronized replanning of the operations of all the intervening agents in the event
of a transitory nature occurring in the short term; (4) adaptable, as it enables the
possibility of reconfiguring planning in circumstances that affect the supply chain
and will continue in the mid or long term; and (5) sustainable, as this is one of the
effects of implementing the ZDM strategy. Secondly, as a result of the above qualities,
the supply chain becomes more robust and resilient. These qualities configure the
supply chain as SC4.0.
The literature review shows that today knowledge gaps exist, and, therefore, spaces
can be explored to guide future research. We highlight the following lines in the scope
of the research to be formulated: (i) from a general perspective, the need to broaden
knowledge throughout the conceptual framework under study; (ii) the study, evalua-
tion, and selection of the digital technologies belonging to the Industry 4.0 spectrum
which will enable the DT to be conceived as a common collaborative supply chain
space; (iii) the cataloging of appropriate modeling and resolution approaches for each
OP process that are compatible with the overall metamodel; (iv) the implementation
of suitable approaches for each process at the tactical and operational decision levels
in the ecosystem delimited by the manufacturer and first- and second-tier suppliers.
Beyond the scope of this project, main further research lines are identified as follows:
(v) the modeling and solving of tactical and operational decision-level planning prob-
lems in ecosystems whose dimension goes beyond the first- and second-tier suppliers;
(vi) the modeling and solving of strategic decision-level process planning problems.
Acknowledgements The research leading to these results received funding from the European
Union H2020 Program under grant agreement No. 825631 “Zero-Defect Manufacturing Platform
(ZDMP)” and under grant agreement No. 958205 “Industrial Data Services for Quality Control in
Smart Manufacturing (i4Q)” and from the Spanish Ministry of Science, Innovation and Univer-
technologies enabling supply chains 4.0 (CADS4.0)”.
References
1. Kagermann H, Wahlster W, Helbig J (2013) Recommendations for implementing the strategic

initiative Industrie 4.0: securing the future of German manufacturing industry. Final report of
the Industrie 4.0 working group. Forschungsunion, p 78
2. Stich V, Adema J, Blum M, Reschke J (2015) Supply chain 4.0: logistikdienstleisterim kontext
der vierten industriellen revolution. In: Logistik – eine Industrie, die (sich) bewegt. Springer
Gabler, Wiesbaden, pp 63–76
3. Blanchard D (2012) Supply chain management best practices, 2nd edn. Wiley
4. Lasi H, Fettke P, Kemper H-G, Feld T, Hoffmann M (2014) Industry 4.0. Bus Inf Syst Eng
6:239–242
5. Cañas H, Mula J, Campuzano-Bolarin F (2020) A general outline of a sustainable supply chain
4.0. Sustainability 12(19):7978
6. Jones D, Snider C, Nassehi A, Yon J, Hicks B (2020) Characterising the digital twin: a
systematic literature review. CIRP J Manuf Sci Technol (Elsevier Ltd) 29(A):36–52
7. Li X, Wang L, Zhu C, Liu Z (2021) Framework for manufacturing-tasks semantic modelling
and manufacturing-resource recommendation for digital twin shop-floor. J ManufSyst (Elsevier
B.V.) 58(B):281–292
8. Lindström J, Kyösti P, Birk W, Lejon E (2020) An initial model for zero defect manufacturing.
Appl Sci (Switzerland), MDPI AG 10(13):4570
9. Kasanen E, Lukka K, Siitonen A (1993) The constructive approach in management accounting
research. J Manage Acc Res (Sarasota) 5:243–264
10. Lukka K (2003) The constructive research approach: case study research in logistics. Turun
Kauppakorkeakoulun Julkaisuja. Publications of the Turku School of Economics and Business
Administration, Series B, vol 1, pp 83–101
11. Oyegoke A (2011) The constructive research approach in project management research. Int J
Managing Projects Bus 4(4):573–595. Emerald Group Publishing Limited
12. Crnkovic GD Constructive research and info-computational knowledge generation. In: Model-
based reasoning in science and technology. Studies in Computational Intelligence, vol 314.
Springer, Berlin, Heidelberg, pp 359–380
13. Meredith J (1993) Theory building through conceptual methods. Int J Oper Prod Manage (MCB
UP Ltd.) 13(5):3–11
14. Alkahtani M, Abidi MH (2019) Supply chain 4.0: a shift in paradigm. In: Proceedings of the
international conference on industrial engineering and operation management. IEOM Society
International, pp 1698–1705
15. Kritzinger W, Karner M, Traar G, Henjes J, Sihn W (2018) Digital twin in manufac-
turing: a categorical literature review and classification. IFAC-PapersOnLine (Elsevier B.V.)
51(11):1016–1022
16. Makris D, Hansen ZNL, Khan O (2019) Adapting to supply chain 4.0: an explorative study of
multinational companies. Supply Chain Forum Int J 20(2):116–131
17. Zekhnini K, Cherrafi A, Bouhaddou I, Benghabrit Y, Garza-Reyes JA (2020) Supply chain
management 4.0: a literature review and research framework. Benchmarking Int J 28(2):465–
501
18. Affonso R, Marcotte F, Grabot B (2008) Sales and operations planning: the supply chain pillar.
Prod Plann Control (Taylor & Francis) 19(2):132–141
19. Ambrose SC, Matthews LM, Rutherford BN (2018) Cross-functional teams and social identity
theory: a study of sales and operations planning (S&OP). J Bus Res (Elsevier) 92:270–278
20. Negri E, Fumagalli L, Macchi M (2017) A review of the roles of digital twin in CPS-based
production systems. Procedia Manuf (Elsevier) 11:939–948
21. dos Santos CH, Lima RDC, Leal F, de Queiroz JA, Balestrassi PP, Montevechi JAB (2020) A
decision support tool for operational planning: a digital twin using simulation and forecasting
methods. Production (Associacao Brasileira de Engenharia de Producao) 30:1–17
22. Wang Y, Wang X, Liu A (2020) Digital twin-driven supply chain planning. Procedia CIRP
(Elsevier) 93:198–203
23. Biesinger F, Meike D, Kraß B, Weyrich M (2019) A digital twin for production planning based
on cyber-physical systems: a case study for a cyber-physical system-based creation of a digital
twin. Procedia CIRP 79:355–360
24. Agostino ÍRS, Broda E, Frazzon EM, Freitag M (2020) Using a digital twin for production
planning and control in industry 4.0. In: Scheduling in industry 4.0 and cloud manufacturing.
International series in operations research & management science, vol 289. Springer, Cham,
pp 39–60
25. Maitreesorasuntee C, Jeenanunta C, Buddhakulsomsiri J, Pannakkong W, Chaysiri R, Masahiro
N, Karnjana J (2020) A steel tube production planning and scheduling with product-dependent
changeover time using digital twin. Int Sci J Eng Technol (ISJET) 4(2):13–19
26. Thakur V, Mangla SK (2019) Change management for sustainability: evaluating the role of
human, operational, and technological factors in leading Indian firms in home appliances sector.
J Cleaner Prod (Elsevier Ltd) 213:847–862
27. Siddh MM, Soni G, Gadekar G, Jain R (2014) Integrating lean six sigma and supply chain
approach for quality and business performance. In: 2014 2nd International conference on busi-
ness and information management (ICBIM), Institute of Electrical and Electronics Engineers
Inc.. IEEE, pp 53–57
28. Gultom GDP, Wibisono E (2019) A framework for the impact of lean six sigma on supply chain
performance in manufacturing companies. IOP Conf Ser Mater Sci Eng (Institute of Physics
Publishing Ltd.) 528:012089
29. Ewald J, Schupp F (2020) Zero shades of gray—reaching zero defects by externalization of the
quality philosophy into the upstream supply chain. In: The nature of purchasing. Management
for professionals. Springer, Cham, pp 89–108
Chapter 22
A Maturity Model for Industry 4.0
Manufacturing Execution Systems
Miguel Á Mateo-Casalí, Francisco Fraile, Andrés Boza, and Raul Poler
Abstract Economic globalization and the increase in consumption by society have

created a need for companies to optimize and improve production processes. Thanks
to new technologies, it is possible to increase their effectiveness to achieve the
required objectives. The degree of automation in factories is already high, so changing
the production process does not generate a significant increase in efficiency. Conse-
quently, it is required to insert new tools that allow a more significant increase of the
factory resources. This is where the concept of “Industry 4.0” is born. The aim of
this work is to stablish an action protocol to implement the status of a Manufacturing
Execution System (MES) in a factory. A maturity model will be proposed to analyze
the state of implementation of the Manufacturing Execution Systems of Industry
4.0 based on three of the three dimensions (technical, operational, and human). The
levels of development in each of them are based on the Capability Maturity Model
Integration ( CMMI).
Keywords Capability Maturity Model Integration · Industry 4.0 · ISA-95 ·

Manufacturing Execution System · Manufacturing Enterprise System Association
M. Á. Mateo-Casalí (B) · F. Fraile · A. Boza · R. Poler

València (UPV), Camino de Vera S/N, 46022 Valencia, Spain
e-mail: mmateo@cigip.upv.es
F. Fraile
e-mail: ffraile@cigip.upv.es
A. Boza
R. Poler
https://doi.org/10.1007/978-3-031-29382-5_22
214 M. Á. Mateo-Casalí et al.
22.1 Introduction
22.1.1 Motivation
The term “Industry 4.0” refers to a new model of organization and control of the
chain throughout the manufacturing systems, supported by information technologies.
Thanks to this technology, we have different measurement devices connected to the
network creating a constant flow of data in the manufacturing, logistics and transport
processes. All these data allow us to monitor the status of the product in real time
[1, 2]. If we combine all of this with cloud storage, dig data, data analytics, or new
architectures based on microservices; they give us a basis for predictive analysis,
facilitating decision-making and promoting automation in any process.
Characteristics of Industry 4.0, such as interoperability, automatization or flexi-
bility, and the relevant technologies to the development of these characteristics, such
as cyberphysical systems, Internet of Things or Smart Data, have been identified in,
but technologies evolve exponentially, and although there are many of them that are
currently revolutionizing industrial processes, their implementation in the organiza-
tion can be complex [3]. For Small and Medium Enterprises (SMEs) that are limited
to invest in research and development, this exponential advancement in technology
may be impossible to achieve [4]. On the other hand, large companies have diffi-
culty implementing a very aggressive technological change in their organization. So,
this digitalization should be a continuous process where the steps forward should be
guided.
Therefore, an analysis tool to check the state of the digital transformation of
the manufacturing system must be defined to guide the company. To do so, the
objective is to establish an action protocol focused on the analysis of the state of
digital transformation in the production chain within a factory.
22.1.2 State of the Art
The use of information technologies has generated software development needs

within manufacturing companies. This phenomenon is known as computer-integrated
manufacturing (CIM). The CIM is a philosophy of approach to an integral organiza-
tion of the factory and its administration. This involves integrating design, manufac-
turing, and management through information systems. The CIM standard is divided
in five levels [5]:
• Level 1. It is made up of the industrial process, the machinery and the necessary
human resources.
• Level 2. It is the integration between the physical part and the most basic control
systems, such as PLCs, sensors.
22 A Maturity Model for Industry 4.0 Manufacturing Execution Systems 215
• Level 3. This level corresponds to the interaction between man and the production
chain. Mainly, we find two methods, the HMI or operator monitors, where we have
operator screens that control a certain part of the process, or SCADA systems,
applications for computers that monitor and manage the factory.
• Level 4. This level has, on the one hand, a database where all the data received
from the plant is stored, from the measurements of the sensors to data from the
PLCs. On the other hand, we find MES, which is the interface between level 3
and level 5. It is the union between the intelligence of the company (business
intelligence) and the processes, being one of the most important parts, since it
allows the interaction in real time, where knowing the demand you can manage
the production flow.
• Level 5. It is the business brain, the part where tools such as ERPs, programs that
manage inventories, billing, logistics are managed.
ISA-95 emerged from the CIM model, which attempts to define the interface
between control functions and business functions. Its objective is to reduce the
number of errors and the cost associated with the implementation, so that the
exchange of information is safe and effective [5].
This standard separates the functionality of the company dividing it into three
layers. A first upper layer (planning), which structures all the business and logis-
tics information, which corresponds to level 4 of CIM. A second intermediate layer
(execution), which integrates all the manufacturing and information control opera-
tions and is located at level 3 of CIM. The last layer (control) that is made up of the
rest of the CIM levels (Fig. 22.1).
The Manufacturing Execution System concept was stipulated in 1992 in Boston
by AMR Research Inc. as the level of execution of manufacturing activities, which,
as we have seen in the ISA-95 model, it is situated between the control systems of
the production chain and the company [6]. The MES is a system that provides all the
Fig. 22.1 ISA-95 model

information required to optimize production from the start of product manufacturing

to its end. MES provides information to workers on how the process is going, helping
them understand the current situation of the plant and how current conditions can be
optimized to improve productivity. In this way, you can work in real time and control
all the elements of the production process.
In 1997, the MESA (Manufacturing Enterprise System Association) industrial
association defined MES as “Guidelines on plant activities when they occur”, empha-
sizing eleven functions, which are: the payment order, personnel and resource
management, traceability of manufacturing orders, products and batches, data acqui-
sition, quality control, procedures management, results analysis, management of
documents and maintenance [7].
22.2 Methodological Approach
Capability Maturity Model Integration (CMMI) is a model for evaluating an orga-

nization’s processes, which was developed by Carnegie-Mellon University (USA)
in 1986 for software implementation processes. This model consists of establishing
key practices in the area of processes and best practices (documenting the process,
providing the organisation with the necessary training and executing in a systematic,
universal and uniform way...) providing a method for assessing a maturity model.
All these practices are grouped into five “maturity levels” [8], so that the company
or organization that accomplishes with all the practices included in one level and its
previous ones will be considered to have reached that level of maturity.
(1) Initial. The organization does not have a stable environment for the development
and maintenance of software. However, they may be using correct techniques,
but they do not affect positively due to poor planning. Almost all the successes
of the company are based on the effort of the workers, but there are always
delays and extra costs. The result is unpredictable.
(2) Repeatable. Minimum project management practices have been generated,
there are metrics and monitoring.
(3) Defined. At this level, the organization already has the correct procedures for
coordinating groups, training of personnel, and the most detailed and advanced
engineering techniques.
(4) Managed. The organization already has a set of measurement processes to
analyze the validity and productivity to make decisions in real time.
(5) Optimizing. The organization constantly improves processes according to the
metrics that are obtained in the production chain.
As we can see, the CMM model establishes a measure of progress as maturity
levels advance. To pass each level, several process areas must be accomplished. That
are identified by the satisfaction or dissatisfaction from several clear and quantifiable
goals. These goals are known in the CMM documentation by the acronym KPA,
which stands for Key Process Area. Each KPA identifies a set of interrelated activities
and practices, that when carried out collectively, allow the fundamental goals of the
process to be achieved.
To develop our maturity model, we will use the “Capability Maturity Model Inte-
gration” model, which is an evolution of the CMM, and appeared in 2001. To develop
our maturity model, we will build a three-dimensional matrix of analysis that we are
going to establish (technique, operation, and human) [9] that refer to innovations,
the maturity of the manufacturing processes, and the roles of the personnel. These
dimensions have been selected because they are the fundamental pillars of digital
transformation.
22.3 Results
The measurement matrix is used to determine the technological implementation

status. This will be the basis to identify all the steps required to move from the
traditional factory to the digital one based on the current analyze. As we have said
before, we will use the CMMI model as a base, establishing five scale levels, to
be able to specify their status within the factory. The three dimensions that we will
analyze will be technical, operational, and human. This snapshot will be the means
for the user to identify the necessary steps to adopt digital automation in a smooth
and phased way. The rows of the matrix will tell us which fields it is important to
enhance for improvement.
The technical, operational, and human dimensions refer to technologies,
processes, and people’s roles, while the columns of the matrix describe the develop-
ment steps for each field of application. The higher the level, the greater its digiti-
zation, so the five columns represent five levels of digital maturity in the production
system [10].
Based on the integrated maturity model (CMMI), these will be the five levels,
each adopting the three dimensions already mentioned before.
• Level 1. The production system does not have the technology, nor does it have
the adequate means to start implementing it. We will call this level Zero.
• Level 2. The production system lacks technologies to monitor and control the
production chain. Decisions are made based on the criteria and experience of the
supervisor. We will call this level ad hoc, which comes from the Latin expression
“What is appropriate, adequate or specially arranged for a certain purpose”.
• Level 3. Restrictions on production system technologies are not fully imple-
mented. Good practices have been added, but they are not well defined, although
there is an intention from the organization. We will call this level Basic.
• Level 4. The architecture used to control the entire production chain is more
sophisticated and optimized, allowing you to collaborate in the change process, as
they are planned. In addition, common standards have been implemented and the
organization uses quantitative analysis of capabilities to predict the development
of the organization. We will call this level defined.
Fig. 22.2 Mapping of three dimensions with CMMI levels
• Level 5. The system has a solid structure based on technology. All systems are
interconnected, and processes are based on rapid calculation of possibilities and
information exchange. We will call this last level optimized.
To identify the current level of the factory, we will make a matrix from the tech-
nical, operational, and human dimension and with these results a position will be
mapped within the matrix. In this way, the matrix provides a short and clear form of
the current state and the desired conditions, showing different alternatives (Fig. 22.2).
22.3.1 Technical Dimension
Through the questionnaire, we will be able to carry out a partial and approximate
evaluation of the state of the dimensions based on our maturity model. The analyzed
elements will be digitization or modeling, agile architecture, security, horizontal
and vertical integration. These dimensions are elaborated based on the Industry 4.0
maturity model of the PWC company [11] (Table 22.1).
22.3.2 Operational Dimension
To define the operational level, we are going to divide operations into seven cate-
gories: detailed production scheduling; execution of production; management of
productive resources; management of the definition of production; collection of
Table 22.1 Technical dimension

Heading level Zero Ad hoc Basic Defined Optimized
Digitalization There is no First Digital Comprehensive Development
or modeling technological technological products and solutions for of new
installation or solutions and services with customers with business
control some isolated software, some models with
applications applications networks and limitations in innovative
data are the supply products and
installed chain. Partners services,
collaborate on making the
digital most of the
transformation technologies
implemented
Agile There is no Fragmented Light The IT Full
architecture data IT connection architecture is functionality
architecture architecture between the implemented in of external
with partners different data the partner data
cubes in network integration
development with
organizations.
The data
exchange is
secure
Security No security Traditional Recognized Risk constantly Optimized
on the security security present with safety in the
network structures challenges, collaborating production
without being but solutions partners chain
focused on not properly
digitization addressed
Horizontal There is no Digitization Vertical Horizontal Partners are
and vertical horizontal or and digitization, integration of fully
integration vertical automation internal processes and integrated,
integrity of threads have processes, data flows with both in
any kind been standardized customers and digitization
installed. and external and in
There is harmonized partners, the processes.
partial data flows integration They access
integration in within the through the all the data in
production or company. network is used almost real
with internal Limited for data use time
and external integration
partners with external
partners
production data; monitoring of production; and analysis of production and perfor-

mance. This dimension is elaborated based on the MESA MON model (Table
22.2).
Table 22.2 Operational dimension

Heading Zero Ad hoc Basic Defined Optimized
level
Detailed There is no Detailed Processes are Process The problems
production scheduling processes are defined across all metrics and detected in the
scheduling drafted organizational management processes and
informally groups, and the control tools used are
organization systems are used to make
follows written in place and improvements
and controlled ensure that and
policies all processes implement
are followed corrective
actions
Execution of Nothing The Production Process The problems
production runs in production processes are metrics and detected in the
production processes are defined in all management processes and
drafted organizational control tools used are
informally groups, and the systems are used to make
organization in place and improvements
follows written ensure that and
and controlled all processes implement
policies are followed corrective
actions
Management There is no Processes vary Responsibilities Process The problems
of productive management across for carrying out metrics and detected in the
resources of organizational activities are management processes and
productive groups, with defined for all control tools are used
resources different organizational systems are to make
processes and groups and in place and improvements
procedures formal lines of ensure that and
used in succession are all processes implement
different defined are followed corrective
groups actions
Management Nothing is The processes Responsibilities Process The problems
of the defined are defined for carrying out metrics and detected in the
definition of about informally activities are management processes and
production production defined for all control tools are used
management organizational systems are to make
groups and in place and improvements
formal lines of ensure that and
succession are all processes implement
defined are followed corrective
actions
(continued)

level
Collection of Production Processes vary The Process Continuous
production data is not across responsibility for metrics and improvement
data collected organizational maintaining the management processes are
groups, with data collection control in place and
different processes is systems are followed.
processes and defined for all in place and Metrics that
procedures organizational ensure that measure
used in groups all processes variances are
different are followed used to make
groups improvements
and
implement
corrective
action plans
Monitoring No The processes Procedures are Process The problems
of production are defined communicated to metrics and detected in the
production tracking informally and all groups and management processes and
are not the policies and control tools used are
formally procedures are systems are applied to
managed available in place and make
ensure that improvements
all processes and
are followed implement
corrective
actions
Analysis of There is no Production There are Process Metrics that
production production and well-documented metrics and measure
and or performance and supported management variances are
performance performance analysis is tools and control used to make
analysis only done methods used for systems are improvements
erratically the production in place and and
and performance ensure that implement
analysis all processes corrective
processes are followed. action plans
Continuous
improvement
processes are
in place and
followed
22.3.3 Human Dimension
Regarding human dimension, refer to all the data that affects the user or worker, we
will only measure two categories: training and data analysis (Table 22.3).
Table 22.3 Human dimension

level
Training There is no There is a Documents All tasks are The
type of document that with documented documentation
information indicates how instructions for is always up to
document or the tasks all applications date. In
training for should be have been addition, there
workers performed, but created is a network
it is not system that
updated to the allows access
new versions to it from any
device with the
appropriate
privileges
Analysis of No data is Some data is All data is All stored data All data stored
data collected or stored in an recorded in a goes through a goes through
analyzed Excel database cleaning analysis
process, so that programs to
only what is extract all the
necessary is relevant
stored information for
the company
22.4 Conclusions
The objective of this paper is to develop a maturity model for the analysis of the
quality of digital transformation, focusing on the implementation of MES within a
production plant. Due to the rapid computer and technological advance, it is appro-
priate to develop a system that allows knowing and facilitating the state in which
a company is in its digital transformation process. To do this, it has developed a
maturity model using the CMMI base for the analysis of the state of implemen-
tation and manufacturing systems. It allows to obtain an image of the state of a
factory at a specific moment within the digital transformation process, focusing on
the three fundamental pillars of production, which are the operational, technological,
and human part. To provide this image of the industry, an analysis matrix has been
developed whose structure is based on the developed maturity model.
In future work, we will implement this matrix to define the different maturity
levels in an application that allows us to easily use this previously created maturity
model within a company’s organization.
References
1. Fraile F, Tagawa T, Poler R, Ortiz A (2018) Trustworthy industrial IoT gateways for interoper-
ability platforms and ecosystems. IEEE Internet Things J 5(6):4506–4514. https://doi.org/10.
1109/JIOT.2018.2832041
2. Ramirez E, Yagüe M, Babé I, Planas A, Estrada J (2020) Estudio smart industry 4.0.
Observatorio 4.0
3. Boza A, Gomez-Gasquet P, Cuenca L, Alarcón F (2020). Proposal of a smart production
planning and scheduling in the industry 4.0 era. In: Ensuring sustainability. Lecture notes in
management and industrial engineering. Springer, Cham, pp 127–135
4. Ricker J (2001) Manufacturing execution system = “empowerment”. In: Proceedings of the
ASME 2001 citrus engineering conference. ASME 2001 Citrus engineering conference. Lake
Alfred, Florida, USA. ASME, pp 47–50. https://doi.org/10.1115/CEC2001-4703
5. Ing. Alicia Junin Duran de Leon MC, Rentería MCJRC, Zamora MCGM, Alva MCSG, Torres
MCLG, Hernández MCZS (2016) Desarrollo de Software basado en el estándar ISA-95. In:
Workshop en computo intelligente en las organizacioned (CIO201), pp 63–68
6. Jürgen K (2007) Manufacturing execution system–MES with 100 figures. Springer-Verlag
Berlin and Heidelberg Gm, p 272
7. Guarín CAN (2010) Sistemas de ejecución de manufactura en la fabricación integrada por
computador y practicas de laboratorio de sistemas scada. Universidad Pontificia Bolivariana
8. Finkelstein A (1992) A software process immaturity model. SIGSOFT Softw Eng Notes
17(5):22–23
9. Fox MS, Chionglo JF, Barbuceanu M (1993) The integrated supply chain manage-
ment system. Retrieved from https://www.researchgate.net/publication/239033756_The_Int
egrated_Supply_Chain_Management_Project
10. Calà ALA, Luder A, Boschi F, Tavola G, Taisch M (2018) Migration towards digital manufac-
turing automation—an assessment approach. In: 2018 IEEE Industrial cyber-physical systems
(ICPS). IEEE, pp 714–719. https://doi.org/10.1109/ICPHYS.2018.8390795
11. Geissbauer R, Vedso J, Schrauf S (2016) Industry 4.0: building the digital enterprise. In: pwc
2016 Global industry 4.0 survey – industry key findings
Chapter 23
Model Experimentation Environment
for Production Planning
Andrés Boza, Pedro Gomez-Gasquet, David Pérez-Perales,

and Faustino Alarcón
Abstract The digital transformation of organizations together with the promotion

of new technologies in the field of Industry 4.0 is offering new possibilities in the
design of production planning models. This paper focuses on this new context to
facilitate model designers and decision-makers the revision of models to adjust them
to these new business contexts. Thus, the design of a model experimentation environ-
ment for production planning is proposed. The proposal includes three subsystems:
Data Modeling, Decision Modeling, and Model Analysis and Investigation: Data
Modeling provides quality datasets for the performance of the experiments. Deci-
sion Modeling manages the models to be analyzed and to be improved. Finally,
Model Analysis and Investigation focus on an experimentation subsystem proposed
to analyze the quality and suitability of the models in a concrete experimental context
using controlled datasets. Thus, the proposed design facilitates the experimentation
of models to later be exploited in business environments.
Keywords Production planning · Mathematical model · Data model
A. Boza (B) · P. Gomez-Gasquet · D. Pérez-Perales · F. Alarcón

Centro de Investigación Gestión e Ingeniería de la Producción (CIGIP), Universitat Politècnica de
València, Valencia, Spain
P. Gomez-Gasquet
D. Pérez-Perales
e-mail: dapepe@cigip.upv.es
F. Alarcón
e-mail: faualva@cigip.upv.es
https://doi.org/10.1007/978-3-031-29382-5_23
226 A. Boza et al.
23.1 Introduction
The planning process determines enterprise objectives and selects a future course of
action to achieve them. This complex decision-making process has to be managed
for the decision-makers as an important part of their management responsibilities
[1]. Model-Driven Decision Support System helps decision-makers to make these
decisions useful for a period of time.
Mathematical models for production planning facilitate the decision-making for
a better organization of the production according to certain criteria and business
restrictions. According to [2], the mathematical model describes the problem by
means of variables that are abstract representations of those elements of the problem
that needs to be considered in order to evaluate the consequences of implementing a
decision.
Thus, Model-Driven Decision Support Systems for production planning systems
have become key elements. However, the digital transformation of organizations
together with the promotion of new technologies in the field of Industry 4.0 means
that these models must be revised to be adapted to this new industrial reality [3].
This paper focuses on this new context to facilitate model designers and decision-
makers the revision of models to adjust them to these new business contexts.
23.2 Information System for Planning Using Mathematical

Modeling
The way in which mathematical programming models have been applied has been
following a series of stages, also known as the mathematical programming modeling
cycle [4]:
(1) Conceptualization: Content and relevant points of the problem without thinking
about mathematical formulation;
(2) Algebraic form: Mathematical formulation of the problem;
(3) Computer-readable form: Numerical representation of the data in rows and
columns;
(4) Translator: Computer tool capable of connecting the algebraic model with the
algorithmic model of the computer;
(5) Solution: The resolution engine includes a resolution algorithm, and this is
capable of interpreting and processing the data matrix, to obtain the optimal
result of the model or to inform that the model is not solvable;
(6) Analysis of the solution. After being processed by the solver engine the results
are stored in a solution file. The information included is the value of the decision
variables, the value obtained from the objective function, and other values related
to the solution.
23 Model Experimentation Environment for Production Planning 227
Deciding which mathematical model best fits the reality (and business need) is not
an easy task. On the one hand, deciding in the conceptualization which elements of
the organization are relevant and should participate in the model is not easy since the
scope of the problem must be limited without leaving out any aspect of interest to it.
On the other hand, establishing the relationships between the elements of the model
as well as establishing the indicator or indicators that allow the decisions obtained
to be compared is not easy either. Thus, we can find controllable factors (which can
be set within a range) and uncontrollable factors. Here we find aspects associated
with a physical perspective of the organization (infrastructures, human resources,
and products), but also organizational/decisional aspects associated with strategic or
tactical policies of the organization that must be properly included (e.g., priority for
VIP clients, minimize distribution costs, or enhance activity in some plants compared
to others).
Also, production planning can be impacted by unexpected events which can
require a change in the released planning, such as broken machines or huge orders.
The quick detection of these unexpected events is essential to avoid bigger troubles.
Thus, new technologies like the Internet of Things can help in this purpose to identify
relevant events and to make a fast analysis of their consequences [5].
All this set of possibilities in the design of the models makes necessary tools for
their analysis.
23.3 Proposal for a Model Experimentation Environment
Information systems designed to define production planning in a company must

answer to the needs of decision-makers. The main focus on these Decision Support
systems is the information provided by them to help manager in their complex
contexts. These systems, supposedly, work with a “tested and quality” mathematical
model. Also, it is necessary that data used in the mathematical model will be accu-
rate, complete, and timely so that the results of the instantiation of the model are
useful for the decision-maker in their decision-making.
However, this approach changes when we look for an experimentation environ-
ment where we want to focus on the analysis of the decision models. That is, the
search of better mathematical models to be used later in production contexts as those
mentioned previously. The focus is not on using timely data in this experimenta-
tion environment, but sufficiently reliable data that cover different scenarios to made
experimentation with models and provide a “tested and quality” model.
Experimenting with the models to achieve higher quality can produce results:
– Reliable. Reliability is about how close repeated measurements are to each other.
– Accuracy. Accuracy is how close the final result is to the correct or accepted value.
Both are affected by the time limit that we let the algorithm or resolution engine
work. Also, it can be affected by the different data instances used.
228 A. Boza et al.
23.4 Design of the Proposal
The proposed design adapts the proposal of [4], also featured in [6], where three
subsystems are established: Data Modeling, Decision Modeling, and Model Analysis
and Investigation:
– The Data Modeling subsystem is responsible for identifying and structuring the
information to be considered in solving the decision problems.
– The Decision Modeling subsystem manages the models proposed for solving
decision problems. These models structure the problem and allow evaluating the
possible decisions that could be made.
– Finally, the Analysis and Investigation subsystem allows the resolution of a deci-
sion problem by instantiating the model with the corresponding data from the
specific context of that decision problem. This independence between the deci-
sion models and the data models allows the resolution for different instances of
the decision problem, that is, solving the decision models for different datasets.
The considerations established in the design of the proposal are:
1. The focus in the Data Modeling subsystem will not be on capturing timely infor-
mation from the organization (e.g., linking it with an ERP system). If not on asso-
ciating data sources with quality datasets for the performance of the experiments
(not necessarily the current data of the organization).
2. It is proposed in the Decision Models Engineering subsystem to extend the set
of tools to be used considering not only Mathematical Models but also heuristic
proposals. In the same way, they will have associated a Data Model (subset of
elements and attributes of the Data Engineering subsystem) necessary for their
resolution.
3. Finally, the operation subsystem to provide information to the decision-maker
by solving the models with the dataset selected by the decision-maker will not be
used. Instead, an experimentation subsystem is proposed to analyze the quality
and suitability of the models in a concrete experimental context using controlled
datasets.
That is, the design seeks a tool that facilitates design of experiments, making
strategic and deliberate changes to produce useful information for the improvement
in the models.
The first approach in the search for quality models includes the following steps
(Fig. 23.1):
1. Model Analysis: Syntax error checking can detect “early” anomalies in the model
formulation.
2. Model and data validation. The analyst checks whether or not the model makes
sense with the model validation.
3. Solution analysis and investigation. After model diagnosis, the analyst may carry
out “what-if” analyses (or scenarios analysis), where the analyst changes the input
Analysis focused on:
Model Analysis Decision Model
Decision Model
Model and data validation
Data Model
Decision Model
Solution analysis Data Model
Data Instantiation
Results
New data scenarios Model adjustments
Fig. 23.1 Steps for quality models
values, using different model data instances. Also, the diagnosis can guide the
introduction of change in the models.
The design of the proposal at a more detailed level includes the Data Modeling
subsystems which structure the data necessary for the decision models and is
connected to the information sources from which to extract the data from different
scenarios through ETL processes. The Decision Modeling for the inclusion and
storage of the decision models according to the Data Modeling, and the Model Anal-
ysis and Investigation subsystem for the analysis of the resolution of the models with
the data instances used in the experimentation environment (Fig. 23.2).
23.5 Conclusions
The need to adapt the production planning models to the new Industry 4.0 envi-
ronments justifies the proposal for the improvement and validation of new decision
models.
The proposed design facilitates the experimentation of models to later be exploited
in business environments. The design allows the model designer together with the
decision-maker to validate the usefulness of the models and adjust them to their
business reality.
The main advantage lies in the existence of a reusable model experimentation
environment for different experiments. Other advantages are: (a) the versatility in the
experimentations to use different data models and decision models, (b) the separation
of data instantiation from the data model design and the decision model, and (c) the
control of the results of the different scenarios proposed in each experiment.
230 A. Boza et al.
Data Decision
Modeling Modeling
Model
Analysis Model Analysis
and
Investigation Model and data validation
Resolution
Data
Process
Sources
Analytical Data Resolution Tool
Extraction, Decision Data
Transformation and
Load (ETL) Process
Solution Analysis
New data scenarios Model adjustments
Fig. 23.2 Model experimentation environment for production planning
Acknowledgements This research is part of the I+D+i project “Integración de la Toma de Deci-
siones de los Niveles Táctico-Operativo para la Mejora de la Eficiencia del Sistema de Produc-
tivo en Entornos Industria 4.0 (NIOTOME)” (Ref. RTI2018-102020-B-I00) funded by MCIN/
AEI/10.13039/501100011033/ ERDF A way of making Europe.
References
1. Duan Y, Ong VK, Xu M, Mathews B (2012) Supporting decision making process with “ideal”
software agents–what do business executives want? Expert Syst Appl 39(5):5534–5547
2. Makowski M (2005) A structured modeling technology. Eur J Oper Res 166(3):615–648
3. Boza A, Alarcón F, Pérez D, Gómez-Gasquet P (2019) Industry 4.0 from the supply chain
perspective: case study in the food sector. In: Technological developments in industry 4.0 for
business applications. IGI Global, pp 331–351
4. Dominguez-Ballesteros B, Mitra G, Lucas C, Koutsoukis N-S (2002) Modelling and solving
environments of mathematical programming (MP): a status review and new directions. J Oper
Res Soc 53(10):1072–1092
5. Boza A, Alarcón F, Alemany MME, Cuenca L (2016) Event classification system to reconsider
the production planning. In: ICEIS 2016: Proceedings of the 18th international conference on
enterprise information systems. SCITEPRESS, pp 82–88
6. Boza A, Ortiz A, Vicens E, Poler R (2009) A framework for a decision support system in a
hierarchical extended enterprise decision context. In: Enterprise interoperability. IWEI 2009.
Lecture notes in business information processing, vol 38. Springer, Berlin, Heidelberg, pp 113–
124
Part VI
Project and Process Management
Chapter 24
BIM Implementation in Construction
Project Management
F. Acebes , R. Testa, J. Alonso, and D. Curto
Abstract The construction industry is one of the world’s most important indus-
tries, but today it is one of the most inefficient and late adopters of technological
advances. Moreover, the development of Building Information Modeling (BIM) tech-
nology has proliferated, which has generated a revolution in working and carrying out
projects, given the many benefits it offers. This work aims to propose an integrative
methodology to complement the current project management in construction based
on both processes of the Project Management Institute (PMBoK) Guide and BIM
methodology. This implementation aims to unify the processes of both method-
ologies, eliminating redundancies and simplifying the work of project managers
and other professionals in the construction industry (architecture, engineering, and
construction [AEC]) to manage their projects simpler and efficiently.
Keywords Project management · PMBoK · BIM · Interoperability · IBCM
24.1 Introduction
One of the key sectors in the global economy that influences social and economic
development is the building sector. Construction becomes more competitive and
F. Acebes (B) · D. Curto

GIR INSISOC - Universidad de Valladolid. Escuela de Ingenierías Industriales, Pº Prado de la
Magdalena s/n, 47011 Valladolid, España
e-mail: fernando.acebes@uva.es
D. Curto
e-mail: david.curto@alumnos.uva.es
R. Testa
Living Werk, Av. Apoquindo 6410, Las Condes, Región Metropolitana, Chile
J. Alonso
Center España, Plaza España 6, 1º, 47001 Valladolid, España
https://doi.org/10.1007/978-3-031-29382-5_24
236 F. Acebes et al.
more accessible when it is carried out more effectively. This makes it possible for
other sectors as well as the industry itself to improve.
But the construction sector is one of the least productive, least open to change,
and least quick to accept new technologies [1]. It can be quite beneficial to update
existing procedures and incorporate new technologies.
On the one hand, the recent growth and development of BIM technology have
increased the need for BIM adoption in the construction sector. On the other hand,
it is now required in many nations [2].
To create an integrated proposal—the IBCM methodology—this work analyzes
the primary BIM Uses with the project management in construction processes. A
single methodology that combines the aforementioned processes will be suggested
after an analysis of the Project Management Knowledge Guide (PMBoK Guide)
[3], the construction extension of the PMBoK Guide [4], and the procedures corre-
sponding to the BIM methodology [5]. The impact of BIM Uses on the knowledge
domains will be examined to reach this goal. Therefore, the processes will be created
to incorporate such BIM Uses.
To achieve the stated objectives, the document is structured as follows. The next
section introduces project management. We present those processes and practices
within the PMBoK Guide which are directly applicable to construction projects and
project management. We continue by describing the BIM methodology, as well as
the different BIM Uses that are essential for goal setting. In this sense, the difference
between traditional processes and collaborative processes is explained. The next
chapter describes the new methodology, developed from integrating the PMBoK
Guide and the BIM methodology processes. The structure, roles, process groups, and
Knowledge Areas are also described. At the end of this work, the conclusions obtained
from the development of the new methodology and the bibliography consulted to
carry out this article will be found.
24.2 Project Management and BIM
24.2.1 Project Management
Applying knowledge, skills, and procedures to manage projects effectively and effi-
ciently is known as project management. Organizations can better compete in their
market by using this strategic competency to link project outcomes to business
objectives [3].
There are three different ways to manage a project: as a stand-alone project (not
included in a portfolio or program), as part of a program, or as part of a portfolio.
Project managers communicate with portfolio and program managers when a project
is a part of either one. Projects, programs, subsidiary portfolios, and operations
managed collectively to accomplish strategic goals are referred to as a portfolio.
24 BIM Implementation in Construction Project Management 237
The set of operations known as project management are those that are focused
on the achievement of a predetermined target with a predetermined scope, within a
predetermined time period, with a predetermined budget, and while preserving the
predetermined expectations and quality.
The Project Management Body of Knowledge Guide (PMBoK Guide) has been
the main focus of this work’s creation [3]. In accordance with this Guide, project
management is accomplished by the appropriate use and integration of logically orga-
nized project management processes. There are other methods to organize processes;
however, this PMBoK Guide divides them into five categories known as “process
groups”.
Project management knowledge areas, which are determined by their knowl-
edge requirements, are another way to categorize processes. Additionally, they are
discussed in terms of their procedures, norms, materials, methods, and techniques.
The PMBoK Guide’s knowledge areas are applicable to construction projects,
including revisions, in accordance with the characteristics, procedures, and applica-
tions that are particular to the process groups and knowledge areas. These knowledge
areas’ resulting behaviors are applicable to the building project at every stage of the
project. This indicates that two new knowledge areas must be added to the PMBoK
Guide’s construction extension [4]:
– Project management for health, security, safety, and the environment.
– Project management for finances.
It should be emphasized that the construction project manager is accountable for
understanding both the project owner’s needs and the best ways to implement the
most typical procedures and particular construction applications.
Integrative project management in the construction industry is in a special posi-
tion because it must address the project as a whole, including stakeholder chal-
lenges, geographic restrictions, and cultural considerations. These factors, along with
project financing, procurement procedures, and risk management, all contribute to
the pressing need to integrate all of these efforts.
Integration, or eliminating differences between the many technical and supporting
disciplines, is the main goal of building project management. When an owner or
developer decides to create a new facility or renovate an existing one, project inte-
gration management gets underway. Construction project delivery may use different
design and construction life cycles due to procedures in all Knowledge Areas,
including project financing, which adds to the complexity.
Contract provisions may include stringent progress and performance reporting
requirements that raise the level of specificity and accuracy required for project
execution monitoring and control. Since changes are frequently viewed as inevitable
in the construction industry, integrated change control is a crucial contractual process.
The improper administration of this activity frequently results in legal issues.
24.2.2 BIM
According to Building Smart International [6], Building Information Modeling

(BIM) is “a collaborative working methodology for the creation and management of
a construction project”.
The BIM approach enables data generation and management by producing
building elements through a single model. Each stakeholder might then analyze,
interpret, and apply the model while contributing to the project [7, 8].
The use of CAD tools for change management is another advantage of the BIM
system. Every graphic must be modified when a change is made to the project, and
all supporting documentation must be reviewed to make sure the change is reflected.
While the BIM system itself automatically and in real-time updates this information
in all the documents (both 2D and 3D) when a change occurs.
Before the introduction of BIM, construction was structured in an individualistic
manner, with each project participant attending to his or her interests while neglecting
the project as a whole. Similar inefficiencies in the project were caused by a lack
of coordination among the numerous stakeholders involved, which increased costs
and lengthened lead times. Since the introduction of BIM, virtual construction has
taken precedence over physical construction. It enables early detection of potential
project risks to offer a less expensive remedy for the discovered issue than on-site
(Fig. 24.1) [9].
The figure makes it evident that each phase in the conventional process is over-
seen by a team, whereas every agent takes part in the project development process
in the collaborative method. The following are the primary benefits of using BIM
technology [10] when completing a project:
– Work is done on a virtual prototype that contains the building components that will
later be physically realized on location with their unique building characteristics
and materials.
– The elements are built in real time and updated in the program’s several windows,
enabling simultaneous work in 2D and 3D.
– It enables quick task verification and coordination among the many project
participants.
– Interferences between the various model components can be examined, allowing
for the prediction of reality and the avoidance of significant site issues.
– Simulations can be performed as the building is being built.
– After construction, an As-Built model can be obtained, enabling management of
the building’s maintenance.
CLIENT CLIENT + ARCHITECTURE + ENGINEERING + CONSTRUCTOR
START DRAFT PLAN CONSTRUCTION DELIVERY
Fig. 24.1 Collaborative process

– The parties engaged in the project can acquire a graphic depiction of the building
that is of a high caliber and meets their expectations.
– Less on-site decision-making as a result of design-phase decisions.
– Extended BIM project development time due to the model being acquired with
all the details that will be added later.
– A decrease in material costs.
– The capacity to adapt to and upgrade to new BIM-compatible technology.
– The ability to access the information in the virtual model.
Because BIM is a huge database, all these benefits are achievable. Environmental,
technical, proprietary, financial, geographical, legal, energy, and other specialized
data are among the most crucial facts that are typically included.
Conflict identification is made possible by BIM in the virtual project phase,
resulting in the elimination of waste and useless tasks (Lean Construction) [11]. It is
caused in part by improved agent integration, knowledge transfer, and communica-
tion (IPD, Integrated Project Delivery) [10, 12, 13]. Additionally, information that is
exchanged early in the project planning phase improves the quality and streamlines
construction activities.
This technology’s significance is in the construction simulation process, which
enables the project to be defined exactly and accurately, resulting in improved
quality, cost management, decreased construction waste, and measurable reductions
in development time.
This information gives rise to the idea of Lean Construction, which essentially
entails minimizing or eliminating all activities and transactions that do not add value
through resource optimization and maximizing the delivery of value to the client To
design and produce at a lower cost, with higher quality, greater safety, and shorter
delivery times, within an environmentally friendly framework [14].
Lean Construction, in this sense, aims to accomplish these goals throughout the
entire life cycle of a building project, involving all social actors who participate in the
design and construction process as well as all individuals and businesses who take
part in the entire supply chain and each value flow, leaving no one out and uniting
everyone under a single objective per with the principles of the Lean system.
Design, supply, and assembly have undergone a revolution in the industrial sector
thanks to Lean production management. Lean transforms how work is carried out
across the board in the delivery process when applied to the end-to-end management
of projects, from design to delivery. Lean Construction applies specific methodolo-
gies to a new process of project delivery and execution, extending from the goals of
a Lean production system—maximizing value and minimizing waste.
These three major pillars—processes, technology, and behavior—form the basis
of a BIM project.
Processes: To successfully deploy BIM, the standard operating procedures must
be changed. These adjustments may result from innovation, which proposes a radical
transformation of the organization, or from continuous improvement, which involves
making modest adjustments while utilizing the existing tools and procedures to
achieve modest improvements.
Technology: Before using BIM, it is crucial to assess how much the current
working environment is improved by this technological advancement. There are
three methods to choose the tool that is best for the business:
– Pile on: Adding the new tool to the task at hand without getting rid of the previous
ones. In this approach, if the tool yields positive results, it can be kept in use until
the old ones become obsolete.
– Swap out: A subpar technology gets changed out with a superior one. Before
implementing the new technology, it is necessary to research its benefits and draw-
backs because the associated workflows must also be changed and are occasionally
unavailable.
– Process first: The workgroup examines the organization’s present process and
considers how it may be enhanced using current technology.
Behavior: To accept this form of technology, it is necessary to be adaptable.
The manual was created by the construction.
A four-step process is outlined by SMART Alliance [6] for improving a thorough
BIM plan:
– Define the proper BIM goals and applications for a project. Defining the broad
objectives for BIM implementation for the project and the team members is the
first stage in creating a BIM project execution plan. Then, based on the project
and team objectives, the most appropriate BIM Uses should be determined. A
specific project activity or project procedure that could gain from the integration
of BIM into that process is referred to as BIM usage.
– Create a plan for implementing BIM. To plan the BIM implementation, a process
mapping approach should be carried out after the phase of defining BIM Uses
is finished. Each process should include the process name, the project phase, the
responsible party, and the “Detailed Map” to which it belongs, as can be seen in
Fig. 24.2. Since multiple processes may share the same detailed map, this notation
is utilized.
– Define the deliverables for BIM. The information exchanges must be identified
so that everyone involved can understand the report’s content when the associ-
ated process maps have been created. The project members, the sender, and the
recipient all engage in these information exchange activities.
Fig. 24.2 Notation of the process map

Three categories of information are represented on the left side of a thorough BIM
Use process map (Fig. 24.3), and the elements are included in the horizontal lines
(referred to as “lanes” in BPMN mapping notation):
1. Reference data: corporate and external structured information resources required

to carry out a BIM Use.
2. Process: a logical flow of actions that make up a certain BIM Use.
3. Information exchange: the BIM process deliverables that can be required as a
resource for additional processes.
The team must be aware of the data that must be delivered and must only identify
the model elements required to implement each BIM Use to define the data trans-
ferred. Although a BIM Use may include several exchanges, only one exchange is
required to document each Use to streamline the process.
The project team must determine who should write this information, when it
should be added to the BIM, and what information exchange requirements apply
to each BIM Use. The project team should choose an element breakdown structure
once the information exchanges (IE) have been established [8].
– Determine the infrastructure needed to support the plan’s implementation. The
identification and definition of the project infrastructure needed to carry out BIM
as intended is the last step in the BIM Project Execution Planning Procedure. The
BIM project execution process is supported by fourteen distinct categories.
It is crucial to make clear that BIM Uses software, even if it is not software. The
software is the tool that enables us to create the model and exchange the various
pieces of information in it, while BIM is the technology, a digital model that creates
a sizable database and makes it possible to manage the components that make up the
infrastructure throughout its life cycle [15].
A virtual model is created to achieve previously agreed-upon goals, and therefore
before modeling begins, it’s important to identify a modeling strategy in line with
the project’s requirements. This will help choose what software to use, what should
be modeled, and what shouldn’t.
The strength of this technology is in the ability to move these databases from one
piece of software to another, with each information application uniquely processing
the data and adding value to the undertaking.
24.3 Proposed Methodology
From the literature review and the analysis of the workflows of both methodologies,
an integrative matrix of BIM Uses and the different Knowledge Areas was created.
After observing that the greatest benefit of the BIM Uses was in the main project
plans (scope, time, and cost), we thought about how to incorporate these Uses into
the workflows (Fig. 24.4).
242
PROJECT
ENVIRONMENTAL
CONDITIONS
PRODUCTIVITY
INFORMATION
PROJECT DELAYS
LEAD TIMES
DOCUMENTS
SET CONSTRUCTION
PREPARA / ADJUST
SEQUENCING AND
SCHEDULE
FLOW
NO NO
LINK 3D ELMENETS TO VALIDATE ACCURACY YES REVIEW 4D MODEL / YES

INITIATING END
ACTIVITIES OF 4D MODEL SCHEDULE
PROCESSES
ESTABLISH Are the model Schedule
INFORMATION correct? optimized?
EXCHANGE
REQUIREMENTS
DRAFT
DRAFT
MODEL
EXCHANGE DOCUMENTS
SCHEDULE
SCHEDULE
AL MODEL
4D MODEL
4D MODEL
STRUCTURAL
ARCHITECTUR
Fig. 24.3 Detailed diagram of BIM process
F. Acebes et al.
Fig. 24.4 IBCM methodology creation flow
This is the process of how this new Integration of BIM in Construction Manage-
ment (IBCM) methodology was developed, which we explain in the following
sections: Organizational roles, Work structure, and IBCM processes.
24.3.1 Organizational Roles
With the implementation of BIM, the traditional system changes from a linear process
to a collaborative process. From a virtual work platform, the different agents make up
the work team exchange information for the optimal development of the project. The
collaborative process obtains improvements in time, cost, and quality (Fig. 24.5).
In this way, the project manager (PM) knows about how the project is carried out.
The PM uses this information to manage the rest of the processes with a global vision
of the project. Therefore, the PM maintains a direct relationship with the promoter,
as well as with all the agents involved in the project. In other words, with all the
stakeholders involved in the project.
This scheme is designed for large-scale projects. However, it could be possible
that the figure of the project manager and BIM Manager could be the same for small
projects [16].
To effectively manage each participant in this collaborative process, it is crucial to
recognize and understand their respective roles. The BIM Manager is a new position
created by the adoption of BIM technology and is essential to the proper operation of
the methodology. The BIM Manager will be responsible for creating the standards
PROJECT OWNER
PROJECT
MANAGER
BIM
MANAGER
ARCHITECTURE BIM ENGINEERING
GENERAL
CONTRACTOR
SUBCONTRACTOR SUBCONTRACTOR SUBCONTRACTOR SUBCONTRACTOR
Fig. 24.5 Working scheme construction with BIM methodology
that will be applied, managing the many stakeholders within the work team, and
keeping the technology up to date during the project’s development.
The skills of a project manager and a BIM Manager are sometimes conflated.
Even though they both have access to the same information, it’s crucial to understand
their respective roles in the project. The BIM Manager is in charge of overseeing all
BIM-related activities in the collaborative setting, exchanging information with other
agents, and maintaining a direct line of communication with the project manager.
24.3.2 Work Structure
The work has a chronological framework via procedures and Knowledge Areas to
accomplish the suggested objectives, similar to the PMBoK. Initiation, Planning,
Execution, Monitoring and Control, and Closure are the process groups that deter-
mine the order of the various operations. In Fig. 24.6, you can see a plan showing
the order of the steps and how they interact.
Although BIM has various uses that are relevant across the project, it is important
to first identify the implementation areas and which processes will benefit from the
use of those particular uses.
EXECUTING
INITIATING PLANNING CLOSING
MONITORING AND
CONTROLLING
Fig. 24.6 Process structure
After conducting research, we concluded that scope, schedule, price, risks, and
stakeholders were the Knowledge Areas that benefited most from the BIM Uses.
These subject areas are referred to as Master Plans (Fig. 24.7).
The other Knowledge Areas, including Quality, Resources, Procurement, HSSE,
and Finance, gain indirect advantages from BIM Uses. Supporting Plans are the name
given to these Knowledge Areas. The management of Communications, for instance,
is not done by any BIM Uses. However, there is a specific sort of BIM Viewer that
enables examination of the models and communication with the other agents via
an online platform, allowing for the submission of interferences or consultation on
specific concerns.
The Support Plans gain benefits from the technique secondarily, whereas the
Main Plans are situated within and directly tied to the BIM environment through the
BIM Uses. Finally, it should be mentioned that all procedures are included in the
INTEGRATION
RESOURCE
MAIN PLANS
QUALITY PROCUREMENT
SCOPE
SECONDARY PLANS
STAKEHOLDERS SCHEDULE
BIM ENVIRONMENT
RISK COST
COMMUNICATIONS FUNDING
INTEGRATION
HSSE
Fig. 24.7 Structure of knowledge areas and BIM

integration procedure. The coordination of the entire project is handled by the final
process.
24.3.3 Integration of BIM in Construction Management

(IBCM) Processes
The methodology’s processes are broken down into three categories. On the one
hand, the Project Initiation Process has all the necessary documents. However, the
majority of the labor and implementation of BIM Uses is concentrated in the Planning,
Execution, Monitoring, and Control Processes.
The project’s activities come to a close, and the Closing Process gathers the
lessons learned. Figure 24.8 illustrates Scope Management, which includes the many
operations outlined.
The information flow is also depicted in the diagram. It displays the documents
created and used throughout the process. The internal procedures and other Knowl-
edge Areas are fed by this knowledge and the exchange paperwork. Finally, it is
possible to see how the BIM Uses integration into the process and this information
exchange’s iterative nature.
Since many BIM Uses are integrated into several processes, they might be used
again during the project’s execution. Additionally, a lot of them exchange information
and acquire fresh information. The model has to be updated because there have been
changes. Another explanation can be that they need to be updated as a result of being
disregarded in an earlier stage.
Start-up Process
The project will begin with the creation of the initial documents, particularly the
conventional Stakeholders Management Plan and Integration Plan from the PMBoK.
They will be consulted throughout their life. Additionally, a document that formally
approves the project and outlines the essential requirements for its success will need
to be created.
To enable the initial ideas of what is to be provided, basic information from
the client or the business should be available. To avoid having to create duplicate
documentation, we chose to combine the Project Constitution Act (PMBoK) and the
PreBEP (BIM).
Additionally, we will begin by locating the stakeholders who have the potential
to directly or indirectly affect the project. To manage their expectations, a Register
of Stakeholders will be created with their interests and requirements included. It will
be finished as the project moves forward. According to their level of involvement
in the project, stakeholders in construction projects can be categorized as direct or
indirect, as shown in Fig. 24.9.
The first step in Stakeholder Management is to gain a comprehensive under-
standing of the project’s goals, advantages, and risks. The entire project lifecycle
EXCHANGE DOCUMENTS PROCESSES DOCUMENTS
PROJECT
CHARTER
REGISTER OF CONTRACT
INITIATING
STAKEHOLDERS
INITIATING
PROJECT PLAN
Fig. 24.8 Process flowchart
SCOPE
PLANNING
MODELLING OF
EXISTING
BIM
CONDITIONS LEGAL
REQUIREMENTS
CODE
BIM
VALIDATION
DESIGN OF THE STAKEHOLDER
CONSTRUCTION
BIM
REQUIREMENTS
DEFINE THE SCOPE
SYSTEM
ANALYSIS OF
BIM
THE LOCATION
PLANNING
DESIGN
BIM
REVIEW
Are the
fulfilled?
NO
conditions
SI
BASELINE
WBS
SCOPE
CREATING
EXECUTING
SCOPE
SCOPE
CONTROL THE
VALIDATE THE
Accepted
NO
deliverables?
3D CONTROL AND
MANAGEMENT
MONITORING AND CONTROLING
BIM
PLANNING
INTEGRATED CHANGE
SI
END
LESSONS
LEARNED
CLOSING
247 24 BIM Implementation in Construction Project Management
Fig. 24.9 Direct and indirect stakeholders
depends on effective communication. The BIM implementation offers a platform for

interaction and teamwork among the many stakeholders.
As a result, it is determined what the primary requirements of the high-level
customer, the project’s primary goals, the deliverables, the expected schedule and
milestones, the likely completion date, and the budget summary are. Even if the
majority of these statistics are somewhat unclear, they still serve as the project’s
starting point [17].
The information exchanges that take place between project participants must be
identified after the pertinent process maps have been created, especially the originator
and recipient of each information exchange transaction so that they may grasp the
information content [18].
The team must comprehend the information that must be delivered to establish
these exchanges, defining only the model elements required to implement each BIM
Use. Although a BIM Use may include several exchanges, only one exchange is
required to document each use to streamline the process.
The project team is in charge of choosing who should write this information, when
it should be added to the BIM, and how information should be exchanged for each
BIM Use.
The project team should choose an element breakdown structure once the
information exchanges (IE) have been developed.
Planning, Implementation, Monitoring, and Control Processes

The project can be realized through the stages of planning, executing, monitoring,
and controlling. Planning is required in all fields of knowledge to choose the best
plan of action for achieving goals. So, from each of them, a particular Management
Plan will be requested. This Plan must be carried out as specified, and it must be
regularly monitored during the project’s development and revised as needed.
Planning should primarily begin with the creation of the Comprehensive Plan,
which outlines all of the work that needs to be done. All project stakeholders involved
in this planning must be aware of it.
On the one hand, the project manager and his team will develop the project
management plan, in which the method for achieving the suggested objectives will
be laid forth. The BIM Execution Plan (BEP), which comprises all the instructions
to execute the BIM, will be elaborated on by the BIM Manager and the remaining
agents.
Additionally, it is suggested to combine both plans (Project Plan + BEP) into a
single document with the project’s essential details. All the agents engaged in the
same project will find that to be helpful. The design of a protocol for how communi-
cations between the many project agents will be formed, as well as the definition and
inclusion of the communication channels and information levels to be transferred,
will also be crucial.
Closure Process
The anticipated work has been completed during this process. Lessons Learned are
created based on the data gathered from the various operations. Future projects must
take this information into account if they are to be successful.
Closing the Project (project activities, phases, or contracts), Closing Resource
Management (project resources, both physical and human, should be transferred
to other tasks or returned), and Closing Procurement Management are the closing
processes (some contracts can be closed, independent of the completion of the overall
project, as is typical of construction projects).
24.4 Conclusion
It is commonly known that procedures are changing as a result of the quick devel-
opment of technology. Understanding how they might affect and enhance construc-
tion project management procedures is crucial. We have seen how different BIM
applications help and enhance project management throughout a project.
Many of the processes could be noted to be similar when comparing the two
techniques (project management and BIM methodology), meaning that information
was being duplicated. Although there are various approaches, including IPD, which
offers a collaborative methodology embracing BIM, it is thought that the project

manager loses his position of leadership and overall control of the project and instead
becomes simply another agent.
While the project manager is in charge of overseeing all processes that are a part of
the project, the idea to develop a new methodology that incorporates BIM Uses into
the workflows permits giving the BIM Manager complete technological management
of the construction processes.
While it is clear that implementing BIM has many wonderful advantages for
scheduling (4D) and cost estimation (5D), it is also crucial to enlarge the scope,
manage risks, and prioritize stakeholders. Successful stakeholder management will
enable us to complete the project.
One more big advantage of project management is communication. In this way,
BIM technology enhances teamwork transparency by enabling all agents to simul-
taneously share the same information. The management of construction projects,
which are getting increasingly complicated, is made simpler by the introduction of
BIM Uses into this new methodology.
Despite the advancement of BIM technology, software is still lacking in many
areas. It provides room for future technological advancements that could lead to
management advancements.
We suggest a future course of action to execute the theoretical model of integration
of BIM and PMBoK techniques once it has been put forth, which would enable us
to increase the efficiency of construction projects.
Acknowledgements This research has been partially financed by the Regional Government of
Castile and Leon (Spain) with Grant, VA180P20.
References
1. Chung B, Skibniewski MJ, Kwak YH (2009) Developing ERP systems success model for the
construction industry. J Constr Eng Manage 135(3):207–216
2. Shou W, Wang J, Wang X, Chong HY (2015) A comparative review of building information
modelling implementation in building and infrastructure industries. Arch Comput Methods
Eng 22(2):291–308
3. Project Management Institute (2017) A guide to the project management body of knowledge:
PMBoK® Guide, Sixth edn. Project Management Institute Inc.
4. Project Management Institute (2016) Construction extension to the PMBOK® Guide.
Pennsylvania, USA: Project Management Institute, Inc.
5. M. y A. urbana (MITMA) Ministerio de Transportes (2020) BIM. Retrieved from https://cbim.
mitma.es/. Accessed on 10 Dec 2020
6. buildingSMART alliance® (2010) BIM project execution planning guide. Retrieved from
https://www.nibs.org/page/bimc. Accessed on 12 Dec 2020
7. Bryde D, Broquetas M, Volm JM (2013) The project benefits of building information modelling
(BIM). Int J Proj Manage 31(7):971–980
8. Peterson F, Hartmann T, Fruchter R, Fischer M (2011) Teaching construction project
management with BIM support: experience and lessons learned. Autom Constr 20(2):115–125
9. The Computer Integrated Construction Research Group (2010) Building information modeling
execution planning guide. Pennsylvania State University, vol 53, p 160
10. Rodríguez AMR, Cordero P, Candelario A (2016) BIM. Diseño y gestión de la construcción.
Madrid, España: Ediciones Anaya Multimedia
11. Pons JF (2014) Introducción a Lean Construction. Madrid, España: Fundación Laboral de la
Construcción
12. Lean Construction Institute (2020) Retrieved from https://www.leanconstruction.org/.
Accessed on 15 Dec 2020
13. American Institute of Architects (2007) California council integrated project delivery: a guide.
Am Inst Archit 1(1):62
14. Ballard G, Howell G (2003) Lean project management. Build Res Inf 31(2):119–133
15. Kang TW, Hong CH (2015) A study on software architecture for effective BIM/GIS-based
facility management data integration. Autom Constr 54:25–38
16. Ma X, Xiong F, Olawumi TO, Dong N, Chan APC (2018) Conceptual framework and
roadmap approach for integrating BIM into lifecycle project management. J Manage Eng
34(6):05018011
17. Travaglini A, Radujković M, Mancini M (2014) Building information modelling (BIM) and
project management: a stakeholders perspective. Organ Technol Manage Const Int J 6(2):1058–
1065
18. Heigermoser D, de Soto BG, Abbott ELS, Chua DKH (2019) BIM-based last planner system
tool for improving construction project management. Autom Constr 104:246–254
Part VII
Strategy, Innovation, Networks
and Entrepreneurship
Chapter 25
Airspace Operations Research Supported
by EU Funds and Industry 4.0 Practices
J. A. Calvo-Fresno , J. Morcillo-Bellido , and B. Rodrigo-Moya
Abstract Horizon Europe is the financial tool to deliver the EU contribution to

research, development, and innovation actions under the next multiannual financial
framework (MFF); among them, those related to airspace operations. The budget of
Horizon Europe is distributed across areas of intervention which are similar to the
characteristics of the “Industry 4.0” concept. To support the design and planning
of the future airspace operations research projects, it is relevant to analyze to what
extent the use of EU funds in this domain is in line with the distribution of funds
expected under the new MFF. The parameters for comparison used are the airspace
operations performance objectives, which present some correspondence with both
Horizon Europe and Industry 4.0 characteristics. The result of the analysis indicates
that the use of funds in the last 20 years follows a distribution similar to the planned
one, although EU funds in support of some objectives would be increased at the
expense of others to achieve a full alignment with it.
Keywords Air navigation system · Air traffic control · Multiannual financial

framework · Aviation research
J. A. Calvo-Fresno (B)
SESAR Joint Undertaking, 1160 Brussels, Belgium
e-mail: jacfpropio@gmail.com
J. Morcillo-Bellido
Universidad Carlos III. Av. de La Universidad, 30, 28911 Madrid, Spain
e-mail: morcillo@ing.uc3m.es
B. Rodrigo-Moya
UNED. Paseo Senda del Rey, 11, 28040 Madrid, Spain
e-mail: brodrigo@cee.uned.es
https://doi.org/10.1007/978-3-031-29382-5_25
256 J. A. Calvo-Fresno et al.
25.1 Introduction
Since it was first mentioned in 2011, many descriptions of the concept of “Industry
4.0” have been elaborated [1]. In this article, the authors consider as relevant char-
acteristics of Industry 4.0 it’s enabling key technologies [2], its features [3], and its
triggers [4].
The concept of Industry 4.0 is frequently associated with the manufacturing
industry. Regarding the air navigation sector, one of its most relevant elements
is through the provision of services to the users of airspace. Notwithstanding the
foregoing, this article focuses on establishing a correspondence between the above-
mentioned characteristics of Industry 4.0 and the air navigation system high-level
objectives which are used to describe its operational performance. These high-level
objectives are the ones targeting safety, capacity, cost efficiency, and environmental
impact, established by the European Commission [5]; and the security objective intro-
duced by the Advisory Council for Aviation Research in Europe after the 11-S events
[6]. All these objectives are equally relevant. Given the fact that there is no direct
relation between Industry 4.0 characteristics and these high-level airspace opera-
tions objectives, the elements that describe the EU financial tool “Horizon Europe”
provide the bridge to establish such correspondence. In this study, the authors iden-
tify the areas of intervention in Horizon Europe and the characteristics of Industry
4.0, in order to establish the relations between these two and the high-level objectives
mentioned. Finally, the EU financial contribution to these high-level objectives in
the period 1995–2020 is obtained, to perform a comparison between the figures from
this period and the expected use of EU funds in Horizon Europe.
25.2 Industry 4.0, Horizon Europe, and the Performance

Objectives of the Air Navigation System
Horizon Europe is the financial tool for European research, development, and inno-
vation under the 2021–2027 multiannual financial framework, having as predeces-
sors Horizon 2020 and the Framework Programs. The European Parliament and the
Council have established Horizon Europe in 2021 by means of a regulation that
structures this framework program in three pillars. As a part of the second pillar, six
clusters have been defined [7] focusing on different societal, industrial, and environ-
mental aspects of life in the European Union. Clusters 3, 4, and 5 describe some areas
of intervention with relevance to future research in airspace operations management.
Cluster 3 (“Civil Security for Society”) corresponds to cybersecurity and protec-
tion against security threats, in particular those affecting critical infrastructures and
air traffic. Cluster 4 (“Digital, Industry, and Space”) corresponds to the development
of new digital technologies and digitalization, artificial intelligence and automation,
the next generation of Internet, and advanced computing capabilities enabling the
analysis of big data [8, 9]. The use of space systems through the value chain and
25 Airspace Operations Research Supported by EU Funds and Industry 4.0 … 257
the achievement of smarter services and a carbon neutral circular industry are also
mentioned. In cluster 5 (“Climate, Energy, and Mobility”), the main driver is “making
[…] the transport sectors more climate and environment-friendly, more efficient and
competitive, smarter, safer and more resilient” [7]. Under this cluster, it is foreseen
the creation of several European partnerships for specific sectors, being one of them
the successor of the SESAR Joint Undertaking.
Based on the description of these areas of intervention, the constitutive items of
Industry 4.0 characteristics, and the definition of the high-level objectives for the
air navigation system operational performance, the authors have deducted, firstly
which of these characteristics of Industry 4.0 would correspond to each of the areas
of intervention; and in a second step, to what high-level operational performance
objectives each area of intervention is making a significant contribution. As a result
of this analysis, correspondences and relations between the mentioned characteris-
tics, areas, and objectives are found. Table 25.1 shows these correspondences and
relations.
25.3 Methodology and Objective
Once established the correspondences and relations above indicated, it is possible to

make a comparison between how the EU funds used in airspace operations research
are distributed, and what is the distribution of funds that would be expected from the
definition of Horizon Europe. To that end, the authors make use of the information
on the funding of air navigation projects developed under the 4th, 5th, 6th , and
7th Framework Programs, and under SESAR 1 and SESAR 2020 programs. This
information is extracted from the EC webpages Transport Research and Innovation
Monitoring and Information System (TRIMIS) [10] and Community Research and
Development Information Service (CORDIS) [11]; from the records of the SESAR
Joint Undertaking on the financial information of the projects managed by this EU
body, and from the EC tool SYGMA/COMPASS, which is open to the beneficiaries
of Horizon Europe funds. Those are considered to be reliable secondary databases
in accordance with the criteria of Ajayi [12].
This information corresponds to financial contributions received during the period
1995–2020. The apportionment of the EU financial contribution to each performance
objective is done following the methodology applied in a doctoral thesis developed
in this specific field [13]. This methodology is described as follows.
For the projects developed under the 4th, 5th, 6th , and 7th Framework Programs,
the fundamental research, flight demonstrations or complementary projects devel-
oped under SESAR 1, and the project executed under Horizon 2020, their description
provides information on the performance objectives to which the project is targeted,
and on the relative weight of each objective in relation to the use of resources required
by the project. The total amount of EU funds used in the project is then apportioned
to each objective proportionally to their estimated relative weight.
Table 25.1 Industry 4.0 characteristics, Horizon Europe areas of intervention, and airspace
operations objectives. Source: Authors based on references
Cluster Horizon Europe area of Industry 4.0 characteristics Airspace operations
intervention objective
3 Security of airspace AIDC, cybersecurity Security
operations
Security of critical AIDC, cybersecurity Security
infrastructure
Cybersecurity Cybersecurity Security
4 Key digital technologies Digitalization Safety, capacity
Emerging air navigation Augmented reality, M2M Safety, capacity, cost
technologies efficiency
Artificial intelligence Automation, drones Safety, capacity, cost
efficiency, environmental
impact, security
Next-generation Internet Transparency, networking Safety, capacity, security
Advanced computing Miniaturization, simulation Safety, capacity
Big data Big data analytics Safety, capacity
Smart air navigation Business intelligence, pay Capacity, cost efficiency
services per use
Carbon-free airspace Resource efficiency Environmental impact
operations
Space technologies for air Digitalization, Cost efficiency
navigation Miniaturization
5 Environmentally friendly Resource efficiency Environmental impact
airspace operations
Efficient and competitive Business intelligence Cost efficiency
airspace operations
Smart airspace operations Business intelligence Capacity, cost efficiency
Safe airspace operations Business intelligence, big Safety
data analytics
Resilient airspace Business intelligence, big Safety, capacity, security
operations data analytics
The industrial research projects developed under SESAR 1 are grouped in

Solutions, whose description provides information on:
• Which deliverables of each project are used to build each solution.
• What is the maturity level of each of these deliverables.
• What are the performance objectives addressed by each deliverable.
• What is the relative weight of each objective in each deliverable.
The total amount of EU funds used in each solution is estimated based on the funds
apportioned to each project distributed proportionally across their deliverables, on the
Table 25.2 Ratios of use of EU funds in airspace operations research. Real values for the period
1995–2020 and expected Horizon Europe values
Safety Capacity Cost efficiency Environmental impact Security
1995–2020 (KEUR) 310.567 292.195 225.810 119.175 43.170
1995–2020 (%) 31.34 29.49 22.79 12.03 4.36
Expected HE (%) 25.05 22.95 24.88 15.77 11.36
estimated weight of their objectives, and taking into consideration the ratio between
funds used for results of different maturity levels.
In terms of these maturity levels, and using as a reference the technology readiness
levels, or TRL [14], it can be mentioned that for the Framework Programs 4th and
7th and the SESAR fundamental research projects, the results are predominantly
oriented to low TRLs (1 and 2); for the 6th Framework Program and the SESAR
1 solutions the predominant TRLs are higher (3 to 7); the projects in Framework
program 5th and the complementary SESAR 1 projects show results across most
TRLs. Finally, flight demonstrations aim at TRLs in the highest range (8 and 9).
Regarding the distribution of use of funds expected under Horizon Europe and
excluding the financial contributions to the European Defence Fund and the potential
use of resources from the European Union Recovery Instrument, the budget available
for the execution of the actions [7] is set up at 86.123 millions of Euros, out of which
clusters 3, 4, and 5 receive an overall financial contribution of 28.826 millions of
Euros, with a split of 1.560 (5.4%), 13.633 (47.3%), and 13.633 millions of Euros,
respectively.
For comparison purposes, and in the absence of more accurate information, this
split is assumed to be applicable also when considering the distribution of the EU
budget for airspace operations research projects under Horizon Europe. Conse-
quently, these percentages are apportioned across the performance objectives with the
assumption that, in each of the areas of intervention, and considering the performance
objectives for which a relation with the given area of intervention has been identified
(as presented in Table 25.1), each of these identified performance objectives has an
equal relative weight. Similarly, the relative weight of all areas of intervention in
each cluster is assumed to be equal. The results of these calculations are summarized
in Table 25.2.
Figure 25.1 shows the comparison of the ratios of use of EU funds in airspace oper-
ations research, for each of the high-level objectives. Since 1995, the predominant
aim of the air navigation research projects has been facilitating the achievement of
the safety and capacity objectives, being those the two core aspects of the air naviga-
tion services. The expected distribution in the next multiannual financial framework
Fig. 25.1 Comparison between the expected funding ratio for airspace operations research in
Horizon Europe and the real funding ratio for the period 1995–2020
maintains those two objectives as recipients of most of the future EU contribution,

although a reduction of a 20% in their ratio is found. This reduction allows to increase
the ratios of the objectives of cost efficiency and environmental impact, which are
expected to experience, respectively, increases of 17 and 14%. And, and above all,
the reduction is made for the benefit of the security objective, for which the expected
ratio raises to more than double of its current real value.
The increasing interest of the Union in supporting the areas of activity related to the
environmental impact is consistent with the expected positive effect that Industry 4.0
would have on environmental sustainability [15]. Similarly, economic sustainability
is expected to be an early benefit of Industry 4.0 developments [16].
The increase in the expected ratio for the security objective is explained by Vaidya,
Ambad, and Bhosle [17]: “With the increased connectivity and use of standard
communications protocols that come with Industry 4.0, the need to protect critical
systems and system data from cyber security threats increases dramatically”. This
fact is even more relevant in the case of the air navigation sector. First, because
the integration of the different constituents of the air navigation system under an
“Industry 4.0” concept would correspond to a horizontal integration and, in a future
scenario [18], to an end-to-end engineering integration. Both models require the
use of powerful software tools, and an efficient digitalized ecosystem, to ensure the
integration of the different networks and activities, incorporating the needs of the
airspace users [19]. Second due to the close relation between airspace management
for civil use, and the needs of the national air defence, that generates a strong inter-
dependency of the civil and the military systems, even at the level of automation and
artificial intelligence technologies [20].
25.5 Conclusions
The results of the comparative analysis show that the approach taken for the drafting
of the regulation establishing Horizon Europe is compatible with the concepts of
Industry 4.0, for which it is possible to find correspondences between the charac-
teristics of this last one and Horizon Europe areas of intervention. Consequently,
the operational performance objectives established for the air navigation system can
also be linked to these characteristics, as Horizon Europe areas of intervention will
be materialized among others through airspace operations research projects aimed
at the achievement of such objectives.
The ratios per objective of use of funds with which the Union is expected to
contribute to those research projects during the next multiannual financial framework
are similar to the ratios found for the 1995–2020 period. It can be inferred that this
distribution of funds is in general adequate for the execution of airspace operations
research projects under Horizon Europe, and in line with Industry 4.0 concepts.
Nevertheless, it is opportune to introduce some refinement of the funding distri-
bution, in order to better match the future needs of the air navigation system in view
of its foreseeable evolution. This refinement would consist in a moderate increase of
the ratios of funds for projects aiming at the objectives of cost efficiency and envi-
ronmental impact, and a significant increase of the ratio associated with the security
objective, to ensure as much as possible an adequate response to the challenges in
the cybersecurity domain of the future air navigation system, with the occasion of
the development of the new research and development programs in air navigation.
References
1. Pfeiffer S (2017) The vision of “industrie 4.0” in the making—a case of future told, tamed,
and traded. NanoEthics 11:107–121. https://doi.org/10.1007/s11569-016-0280-3
2. Tjahjono B, Esplugues C, Ares E, Pelaez G (2017) What does industry 4.0 mean to supply
chain? Procedia Manufact 13:1175–1182
3. Pfohl H-C, Yahsi B, Kurnaz T (2015) The impact of industry 4.0 on the supply chain. In:
Proceedings of the Hamburg international conference of logistics (HICL), vol. 20. Epubli
GmbH, Berlin, pp 31–58
4. Lasi H, Fettke P, Kemper H-G, Feld T, Hoffmann M (2014) Industry 4.0. Bus Inf Syst Eng
6:239–242. https://doi.org/10.1007/s12599-014-0334-4
5. European Commission COM (2008) 750 final. Communication from the commission to the
council and to the European parliament—The European ATM Master Plan
6. ACARE (2002) Strategic research agenda 1. Retrieved from http://www.acare4europe.org/doc
uments/archive/acare-sra-1
7. European Parliament and Council (2021) Regulation (EU) 2021/695 of 28 April 2021 estab-
lishing Horizon Europe—the framework programme for research and innovation, laying down
its rules for participation and dissemination, and repealing regulations (EU) 1290/2013 and
1291/2013. Official J Eur Union, L 170, dated 12.5.2021
8. European Commission (2019a) European partnerships under Horizon Europe: results of the
structured consultation of member states. Draft report for the meeting of the shadow config-
uration of the strategic programme committee on 27 June 2019. Retrieved from https://www.
era-learn.eu/documents/results_structured_consultation_ms
9. European Commission (2019b) Orientations towards the first strategic plan for Horizon Europe.
European research and innovation days, EARTO. Brussels, 24–26 September 2019. Retrieved
from https://ec.europa.eu/info/sites/info/files/research_and_innovation/strategy_on_research_
and_innovation/documents/ec_rtd_orientations-he-strategic-plan_122019.pdf
10. European Commission (2021a) TRIMIS. Transport research and innovation monitoring and
information system, EC DG MOVE. Retrieved from https://trimis.ec.europa.eu/
11. European Commission (2021b) CORDIS. Community research and development information
service, EU Publications Office. Retrieved from http://cordis.europa.eu/
12. Ajayi VO (2017) Primary sources of data and secondary sources of data. https://doi.org/10.
13140/RG.2.2.24292.68481
13. Fresno JAC (2020) Análisis del uso de fondos comunitarios en proyectos de navegación aérea.
Vías para la mejora de su impacto y evaluación. Programa De Doctorado EN Unión Europea,
UNED
14. European Commission C(2014)4995 final. Amending implementing decision C(2013)8631
adopting the 2014–2015 work programme in the framework of the specific programme imple-
menting horizon 2020—the framework programme for research and innovation (2014–2020),
22 June 2014. Retrieved from https://ec.europa.eu/research/participants/data/ref/h2020/wp/
2014_2015/annexes/h2020-wp1415-annex-g-trl_en.pdf
15. Oláh J, Aburumman N, Popp J, Khan MA, Haddad H, Kitukutha N (2020) Impact of industry
4.0 on environmental sustainability. Sustainability 12(11):4674. https://doi.org/10.3390/su1
2114674
16. Ghobakhloo M (2020) Industry 4.0, digitization, and opportunities for sustainability. J Cleaner
Prod 252:119869. https://doi.org/10.1016/j.jclepro.2019.119869
17. Vaidya S, Ambad P, Bhosle S (2018) Industry 4.0 – A glimpse. Procedia Manufact 20:233–238.
https://doi.org/10.1016/j.promfg.2018.02.034
18. SESAR Joint Undertaking (2019) A proposal for the future architecture of the European
airspace. https://doi.org/10.2829/309090
19. Sony M (2018) Industry 4.0 and lean management: a proposed integration model and research
propositions. Prod Manufact Res 6(1):416–432. https://doi.org/10.1080/21693277.2018.154
0949
20. Birdal EO, Üzümcü S (2019) Usage of machine learning algorithms in flexible use of airspace
concept. In: 2019 IEEE/AIAA 38th digital avionics systems conference (DASC), San Diego,
CA, USA, pp 1–5. https://doi.org/10.1109/DASC43569.2019.9081654
Chapter 26
Dealing with the Externalities
of the Sharing Economy: Effect
of Airbnb in Long-term Rental Prices
in the City of Madrid
R. Marque, G. Morales-Alonso, Y. M. Núñez, and A. Hidalgo
Abstract The rapid rise of accommodation platforms such as Airbnb has helped
to democratize tourism as a leisure activity, due to the lower prices offered when
compared to traditional hotels. However, this rising disruptive industry is affecting the
social structures of the hosting cities, which can be understood as an externality of the
industry. Prices in the long-term rental market have been stated to rise when Airbnb
reaches a city. This is attributed to the switch of part of the properties traditionally
offered for long-term rental to the new touristic accommodation market. In this
research we focus on how the long-term rental market in the city of Madrid is affected,
in terms of prices, by the presence of Airbnb (with a special focus on multilistings) and
the distance of the proprieties to the city center. We use artificial neural networks to
show that long-term rental prices are directly correlated with the number of properties
offered on Airbnb, with the number of multilistings and the distance to the center of
the city ranking in second and third position, respectively.
Keywords Gentrification · Touristification · Collaborative economy · Platform

economy
R. Marque · G. Morales-Alonso (B) · Y. M. Núñez · A. Hidalgo

Department of Industrial Engineering, Business Administration and Statistics, Universidad
Politécnica de Madrid, C/ de José Gutiérrez Abascal, 2, 28006 Madrid, Spain
e-mail: gustavo.morales@upm.es
R. Marque
e-mail: r.marque@alumnos.upm.es
Y. M. Núñez
e-mail: ym.nunez@upm.es
A. Hidalgo
e-mail: antonio.hidalgo@upm.es
https://doi.org/10.1007/978-3-031-29382-5_26
264 R. Marque et al.
26.1 Introduction
In recent years, the term sharing or collaborative economy has been heard more and
more frequently. This concept refers to the new economic paradigm of the collab-
orative commons, which is gaining importance throughout the world. It has been
described as an alternative economic path for capitalism [1]. Some authors consider
that this new collaborative system would introduce a general change in economic
life, providing it with a more democratic, ecological, and equitable approach [2, 3].
Contrarily, some authors defend that the sharing economy is the ultimate form of
neoliberalism [4, 5].
In spite of these confronting views, the working principles of the collaborative
economy are clear: there is a search for efficiency in the exchange of goods or
services. Supported by advances in information technologies, this system aims to
facilitate access to underused assets through online platforms, which enable the
exchange of goods and services between equals with a marginal cost close to zero.
The rapid emergence of the sharing economy has produced sudden changes in
multiple sectors. Various studies show the very high growth rates of intermediary
platforms in these business models [6, 7]. In many cases, this growth exceeds that of
traditional agents by multiplicative factors, which generates controversy with them
and opens the debate on whether legislation is necessary in certain sectors of activity.
Some of these intermediary platforms, in addition to causing tension with competi-
tors of their industries, produce externalities in certain markets that indirectly affect
a large part of the population. The most current and representative examples are the
Uber and Airbnb platforms, which have forced a dramatic turn in the mobility and
housing sectors, respectively.
This study focuses on the analysis of the activity of the Airbnb digital platform
and its external effects on the real estate market in the city of Madrid. Airbnb is
a digital platform that enables its users to offer and rent properties for short-term
leases, generally of a tourist type. The real value obtained by the platform is reduced
by charging a small commission to both the host and the guest for each transaction.
In other words, the platform is fed back through the growth in its business volume,
taking advantage of network effects and their externalities, by connecting more and
more hosts and guests in a wheel that is difficult to regulate.
But this growth affects the long-term rental market of the city of Madrid, altering
both its economic and social nature. These changes lead to the appearance and
rapid development of gentrification and touristification processes in the most central
neighborhoods of the city.
26 Dealing with the Externalities of the Sharing Economy: Effect of Airbnb … 265
26.2 Theoretical Framework
The sharing economy can be defined as an economic system whose intermediary

companies use online platforms to facilitate and reduce the cost of for-profit transac-
tions and give temporary access (without transfer of ownership) to idle resources of
consumers [8]. Due to its nature of using assets that do not belong to the company,
the growth potential of collaborative platforms is very high, due to the scalability of
their business model. This high growth impacts at several levels, which are usually
referred to as market externalities [9–11]. Two of these externalities are gentrification
and touristification.
Gentrification can be defined as the change that occurs in the population of a
territory by replacing the inhabitants of a central area of an urban nucleus with users
of a higher socioeconomic status. As a consequence, the original inhabitants of these
central areas are relegated to residential areas with a lower economic level [12].
The classic gentrification process is divided into four interlinked phases: abandon-
ment, stigmatization, regeneration, and commodification. The process begins with
the abandonment and divestment in a specific area, by multiple economic and polit-
ical interests. Subsequently, with the emergence of large speculative movements, the
revaluation of that area begins that previously suffered deterioration, stigmatization,
and falling property prices. In the phase of regeneration and subsequent commer-
cialization, credit institutions and real estate giants enter, whose activity is reduced
to making large investments in the purchase and rehabilitation of buildings, with the
aim of raising the level of the neighborhood to the standards of new buyers. It should
be noted that, with the economic recomposition after the last financial crisis, the
gentrification process has become a more complex mechanism [13]. To delve into
the gentrification process, Neil Smith’s Rent-Gap Theory, formulated in 1979, is an
inevitable reference. This theory explains the gentrification process in urban centers,
focusing on the cities of New York and Philadelphia [14].
The term touristification is used to describe the consequences of tourism in the
most central neighborhoods of cities. It is defined as the impact that tourist over-
crowding has on the commercial and social fabric of certain areas within the cities.
Its consequences are that traditional services, facilities, and shops are substituted by
others fully oriented to meet the needs of the waves of tourists [13].
The irruption of P2P tourist accommodation platforms has created new speculative
opportunities for specific segments of real estate in central city neighborhoods, which
are moving away from traditional rental toward this new form of exploitation, leading
to a sharp rise in housing prices, especially in the rental market [13]. This movement
favors the replacement of the local population by a foreign population in constant
circulation, that is, a flow of tourism in search of leisure. For this reason, some
experts speak of a possible future restructuring of these urban areas, turning them
into themed spaces that hardly meet the minimum requirements for the traditional
resident.
This rental style allows the rotation of real estate stock with great ease, by
managing rentals of a few days with essentially tourist purposes. In addition, the char-
acteristics of these platforms enable their rapid growth, since they facilitate commer-
cialization for days, make it easier for owners to include offers in the market, and
provide great accessibility to international tourists through online access to properties
[6].
Tourism has been stated to rise long-term rental prices within a city from 4–5%
[15] up to 13.5% [16]. As highlighted in [11], 18% of hosts of sharing accommodation
platforms in Boston are multilistings, and they account for 46% of the properties
offered through Airbnb in that city. For the city of Madrid, we find for 2019 that
55% of the total offer belongs to hosts that offer more than one property. For our
study, the total offer of properties is divided into hosts that offer only one property
(singlelister) and hosts with more than one (multilister), to account for their influence
in long-term rental prices in a separate manner.
26.3 Methodology
To study the impact of Airbnb on the real estate ecosystem in Madrid, the relationship
between the increase in long-term rental prices and the increase in the platform offered
in each of the 21 districts of the city is evaluated. To this end, the databases of the
idealista.com real estate website and the InsideAirbnb platform have been managed.
The real estate website idealista.com is an intermediary agent between landlords
and tenants (or buyers and sellers) of real estate. It has thousands of offers in Spain,
both for rental housing and for sale; therefore, it serves as a point of reference to
study the fluctuations of the real estate market in recent years. It provides access to
reports on the sale or rental price of the home, which can be filtered by region, city,
and district.
InsideAirbnb database is an independent website that provides the tools and data
to explore the status of all Airbnb platform listings in major cities around the world.
Key data can be accessed to support the analysis of the impact of this platform and
its competition with the residential housing market.
The predictive model extreme learning machine (ELM) algorithm within the
multilayer perceptron (MLP) of artificial neural networks (ANNs) is used to analyze
the relationship between the variables under study. The MLP allows the approxima-
tion of a continuous function between a group of input and output variables, making
a general function estimation [17].
In the case under study, an attempt is made to demonstrate how the incursion of
a platform of collaborative origin, such as Airbnb, influences the socioeconomic
aspects of the city of Madrid. This platform has entered the real estate market as an
intermediary in the promotion of short-term tourist accommodation services. In our
case, the impact of this incursion into traditional long-term rental will be analyzed.
There is an open debate on whether the transactions carried out through this type
of disruptive business respect their collaborative origin or embrace business models
based on the platform economy [18].
To that end, we take the long-term rental prices of the 21 districts of the city of
Madrid in 2019 as the independent variable of our study. The dependent variables
under study are (i) the total number of properties offered in Airbnb for each district,
(ii) the number of those properties offered that belong to the same owner (referred
to as multilisting), and (iii) the distance to the city center.
Several transformations of the dependent variables have been sought after, from
which the corrected normalized has proven to give the best approximation to the
average real prices, see Table 26.1, where the average real price for the city is 13.75
e/m2 while the predicted is 13.61 e/m2 .
Once the corrected normalized transformation has been chosen, the MLP calcu-
lation has been conducted. To build the MLP model, the sample was divided into
three: 71.5% training, 19% testing, and 9.5% for the reserve. The neural network
presents an architecture with an input layer with the three covariates with a corrected
normalized scale, a hidden layer can also be obtained with a number of units in the
hidden layer of four and the activation function was the hyperbolic tangent. Finally,
Table 26.1 Real and predicted prices for the 21 districts of the city of Madrid
District Arganzuela Barajas Carabanchel Centro Chamartín Chamberí Ciudad
Lineal
Real price 15.65 12.29 11.75 18.6 16.33 18.13 13.58
[e/m2 ]
Predicted 17.43 10.8 13.69 18.63 13.43 17.84 13.47
price
[e/m2 ]
District Fuencarral Hortaleza Latina Moncloa Moratalaz Puente De Retiro
Vallecas
Real price 12.35 13.51 12.26 14.55 10.98 11.75 16.25
[e/m2 ]
Predicted 11.26 11.6 13.74 14.25 11.01 13.56 15.05
price
[e/m2 ]
District Salamanca San Blas Tetuan Usera Vicálvaro Villa De Villaverde
Vallecas
Real price 18.88 12.26 15.6 11.6 10.85 11.15 10.63
[e/m2 ]
Predicted 18.08 12.06 15.27 12.37 10.62 10.65 11
price
[e/m2 ]
Fig. 26.1 Relative importance of the dependent variables on the long-term rental prices of the 21
districts of the city of Madrid
we have an output layer that corresponds to the dependent variable with standardized
scale and identity activation function, where the error function is the sum of squares.
The results obtained point toward to the offer of home-sharing properties being
the most determining factor in the long-term rental prices, while the presence of
multilistings has a relative importance of 40%. The least importance corresponds to
their distance to the city center; see Fig. 26.1.
26.5 Conclusions
The purpose of this research is to shed light on the current situation in the tourist
accommodation market provided by the sharing or collaborative economy. In partic-
ular, we deal with its possible positive and negative consequences on the different
agents that participate in the real estate market.
From a theoretical point of view, we anchor on the theory of Rifkin [1] of the
zero marginal cost society, according to which the collaborative economy allows
working in an environment of almost zero marginal costs, being more effective than
the operation of the regular companies.
On the other hand, when this theory is applied to tourist accommodation, a more
efficient functioning of the market is achieved, putting underused resources to work.
This in turn has two effects. On the one hand, access to tourist accommodation is
democratized, through an increase in supply, a decrease in prices, and a more varied
offer. Thus, it has been found that the presence of Airbnb causes an increase in
demand for tourist accommodation [9]. On the other hand, families who have looked
to real estate as a value storage for their savings now find a way to monetize them
more successfully than long-term renting.
But this discussion cannot be made without looking at the externalities that the
collaborative economy applied to tourist accommodation has on the context (city)
in which they operate. The fact that short-term rentals come from apartments that
were previously offered for long-term rental causes a reduction in the supply in this
market, which puts upward pressure on prices in it. Various researchers have already
found a correlation between the long-term rental price boom and the Airbnb supply,
in cases such as Paris, New York, or Los Angeles [19].
The present work provides data on the case of Madrid, showing that this correlation
also occurs. In particular, we look at the effect multilistings have on this phenomenon.
We define multilisting as an entity (person or company) that has more than one
property offered on the platform. Our research thesis is multilistings are operating at
the edge of the sharing economy, since they are not actually using an underutilized
asset (a fact that is at the core of the definition of sharing economy), and that its effect
on the long-term rental price is relevant, although not the most relevant one.
The results obtained show, by means of artificial neural networks that the rental
price in the long term is directly correlated with the number of properties offered on
Airbnb, with the number of multilistings and the distance to the center of the city
ranking in second and third positions, respectively.
References
1. Rifkin J (2014) The zero marginal cost society: the internet of things, the collaborative
commons, and the eclipse of capitalism. Palgrave Macmillan, New York
2. Böcker L, Meelen T (2017) Sharing for people, planet or profit? Analysing motivations for
intended sharing economy participation. Environ Innov Societal Transitions 23:28–39. https://
doi.org/10.1016/j.eist.2016.09.004
3. Boar A, Bastida R, Marimon F (2020) A systematic literature review. Relationships between the
sharing economy, sustainability and sustainable development goals. Sustainability 12(17):6744.
https://doi.org/10.3390/su12176744
4. Martin CJ (2016) The sharing economy: a pathway to sustainability or a nightmarish form
of neoliberal capitalism? Ecol Econ 121:149–159. https://doi.org/10.1016/j.ecolecon.2015.
11.027
5. Murillo D, Buckland H, Val E (2017) When the sharing economy becomes neoliberalism on
steroids: unravelling the controversies. Technol Forecast Soc Chang 125(C):66–76. https://doi.
org/10.1016/j.techfore.2017.05.024
6. Biswas R, Pahwa A, Sheth M (2015) The rise of the sharing economy: the Indian landscape.
Ernst & Young Global Limited, Delhi, India
7. PriceWaterhouseCoopers LLP (pwc) (2015) The sharing economy
8. Ranjbari M, Morales-Alonso G, Carrasco-Gallego R (2018) Conceptualizing the sharing
economy through presenting a comprehensive framework. Sustainability (Switzerland)
10(7):2336. https://doi.org/10.3390/su10072336
9. Coello-Vilarino G, Morales-Alonso G, Carrasco-Gallego R (2019) Externalities of the sharing

economy : effect on employment of holiday accommodation platforms. The case of Madrid.
In: 13th International conference on industrial engineering and industrial management, XXIII
Congreso de Ingeniería de Organización, Gijón, Spain
10. Martín MG, Álvarez AP, Ordieres-Meré J, Villalba-Díez J, Morales-Alonso G (2021) New
business models from prescriptive maintenance strategies aligned with sustainable development
goals. Sustainability 13(1):216. https://doi.org/10.3390/su13010216
11. Horn K, Merante M (2017) Is home sharing driving up rents? Evidence from Airbnb in Boston.
J Hous Econ 38:14–24. https://doi.org/10.1016/j.jhe.2017.08.002
12. Glass R (1964) Aspects of change. MacGibbon & Kee, London
13. Sorando D, Ardura A (2016) First we take Manhattan: La destrucción creativa de las ciudades.
Madrid, Spain: Libros La Catarata
14. Wachsmuth D, Weisler A (2018) Airbnb and the rent gap: gentrification through the sharing
economy. Environ Planning A Econ Space 50(6):1147–1170. https://doi.org/10.1177/030851
8X18778038
15. Paramati SR, Roca E (2019) Does tourism drive house prices in the OECD economies? Evidence
from augmented mean group estimator. Tour Manage 74:392–395. https://doi.org/10.1016/j.
tourman.2019.04.023
16. de Arenaza DR-P, Hierro LÁ, Patiño D (2019) Airbnb, sun-and-beach tourism and residential
rental prices. The case of the coast of Andalusia (Spain). Curr Issues Tourism 25(20):3261–
3278. https://doi.org/10.1080/13683500.2019.1705768
17. Morales-Alonso G, Guerrero YN, Aguilera JF, Rodríguez-Monroy C (2020) Entrepreneurial
aspirations: economic development, inequalities and cultural values. Eur J Innov Manage
24(2):553–571. https://doi.org/10.1108/EJIM-07-2019-0206
18. Torrent-Sellens J (2019) ¿Economía colaborativa o economía de plataforma? Más allá de un
debate inacabable. Harvard Deusto Bus Rev 289:58–69
19. Garcia-López M-À, Jofre-Monseny J, Martínez-Mazza R, Segú M (2020) Do short-term
rental platforms affect housing markets? Evidence from Airbnb in Barcelona. J Urban Econ
119:103278. https://doi.org/10.1016/j.jue.2020.103278
Chapter 27
Cognitive Ergonomics Perspective
to Boost Human-centered Innovations
in Industry 4.0
Juan Antonio Torrecilla-García , María Carmen Pardo-Ferreira ,

and Juan Carlos Rubio-Romero
Abstract This paper aims to comprehend the human-centered innovations oppor-

tunities that might emerge from the Cognitive Ergonomics scope within Industry
4.0. The present work can be considered an exploratory study, starting with a litera-
ture review to identify the guiding principles and concepts of Industry 4.0 in order to
develop a technological mapping of potential innovations areas and analyze the poten-
tial implementation challenges. Industry 4.0 networked approach rethinks persons’
working environments and daily work processes. Persons in the smart factory:
workers on the level of operators and managers face in this automatized and digi-
talized system, a variety of challenges. The Cognitive Ergonomics seems to be the
next leading paradigm in safety management of Industry 4.0, closely related to the
embedded technologies use and workforce well-being to ensure optimal levels of
productivity and efficiency.
Keywords Cognitive Ergonomics · Industry 4.0 · Innovation · Human-centered

approach · Occupational health and safety
27.1 Introduction
In recent years, the Industry 4.0 paradigm has been increasingly gaining space as a
one with strong potential to create new value within the industrial and technological
scope. Also, it is considered the turning point for new industry-based business models
with significant impact on societies. So, Industry 4.0 may be considered currently
J. A. Torrecilla-García (B) · M. C. Pardo-Ferreira · J. C. Rubio-Romero

School of Industrial Engineering, Universidad de Málaga, 29071 Málaga, Spain
e-mail: juantorrecilla@uma.es
M. C. Pardo-Ferreira
e-mail: carmenpf@uma.es
J. C. Rubio-Romero
e-mail: juro@uma.es
https://doi.org/10.1007/978-3-031-29382-5_27
272 J. A. Torrecilla-García et al.
as a preferred system for the companies to bring more efficiency and data-based
decision-making processes to the core of manufacturing companies. This approach
gives emergent opportunities for new development on both technology and human–
machine relations scale.
The close interdependence between production processes and data management
within the Industry 4.0 improves many intrinsic internal factory processes as data
processing and transfer is carried out in a particular way [1–4]. The interoperability,
real-time capability, service orientation, and modularity are considered key assump-
tions of Industry 4.0. Hence, new organizational forms and interrelations emerge
across the entire organization. All of these based on the scope of the dynamic,
real-time optimized, self-organizing infrastructure, are considered as a milestone
in the evolution of the value-added network within the industrial environments.
This network structure and interdependence give industrial organizations the ability
to optimize different parts of the productive process according to the established
strategy, availability, and resources utilization. This capacity is clearly based on the
potential of Industry 4.0 to detect internally and externally all affecting factors and to
process in real-time all available data. Consequently, Industry 4.0 may be considered
as an advanced stage of digitization in the manufacturing sector. On the other hand,
this network-based approach and emergence of automatized analytical and business
intelligence capabilities enhance the spreading of new forms of human–machine
interaction all around the industrial organization.
The data display interfaces such as touch screens, data tablet devices, augmented
reality touchpoints alongside with 3D printing based on digital instructions; all these
are becoming new ordinariness for factory workers of all organizational levels. In
addition, it is convenient to stress the importance of close interrelations among all
working principles of Industry 4.0; cyberphysical systems, big data, the Internet
of Things (IoT), virtual reality, and cybersecurity. These essential parts of Industry
4.0 make the manufacturing company a smart factory based on automation, intelli-
gence and connectivity. In the context of an ever-changing environment and tech-
development, the smart manufacturing organization is able to improve its ability to
optimize processes and to innovate the soft capacities. Thanks to the connectivity and
automation, the datasets are distributed down to components, machines, and plants
procedures even if human supervision and control are permanently required.
Thus, Industry 4.0 is gradually becoming a new level of manufacturing organiza-
tion and control over the entire value chain of the lifecycle of products [5]. This inno-
vativeness is to be the determinant driver to reach new levels of connections between
persons (workers but also digitally connected product users), objects, and systems
in Industry 4.0 [6]. These interrelations should become more common and efficient
to reach decentralized production and self-regulation approach of value creation in
manufacturing [7]. The conditions of industrial 4.0 plants and environments will
impact directly on productivity and efficiency of workers.
Consequently, the Industry 4.0 networked approach rethinks persons’ working
environments and daily work processes. Persons in the smart factory: workers on
the level of operators and managers face in this automatized and digitalized system,
a variety of challenges [8, 9]. Particularly, the requirement of human end-to-end
27 Cognitive Ergonomics Perspective to Boost Human-centered … 273
control of the increasing presence of automated machines, production lines, and

cobots impacts strongly on the workstation proceedings and possible workload levels.
Within this scope, even if the key determinant of Industry 4.0 is technology, the impli-
cations for the human ecosystem of smart factories go much deeper [10–13]. Thus,
not only operational improvements of Industry 4.0 systems will enhance innovation
on the strategic business level, the “soft” approach will impact it as well. The strategic
changes embedded in the Industry 4.0 implementation aim to correspond to alterna-
tions in the organizational culture and management style. In general, manufacturing
systems are safety–critical environments because of their potential negative impact
on the life and health of workers.
As such, also Industry 4.0 has significant safety risks associated with the autom-
atized processes and the necessity of new professional profiles take-up to be able to
operate within the data-dependent factory environment. All Industry 4.0 processes
require programming, assessment, analysis, early warning, and control. On the level
of the performance of the tasks, Industry 4.0 determines the dynamics and routines
of each job post; as well as it imposes new constraints to workers’ well-being and
new requirements of competencies in every area.
In order to comprehend all implications of Industry 4.0 to the workers’ perfor-
mance, occupational health and safety (OHS) must be considered. The OHS as a set
of strategic and functional processes aimed at reducing all employee-related risks
draws a new horizon perspective when Industry 4.0 is considered [14, 15]. Even
being the OHS a proactive and multidisciplinary field, its operational premises and
processes are closely industry or branch related. At this scope, in Industry 4.0 specific
OHS determinants must be taken into account. In particular, the impact of highly
digitalized systems and human–machine interaction must be analyzed to be able
to establish an appropriate framework of safety management. One of Industry 4.0
challenges is to be able to keep pace with safety changes at a time the technological
advances are improving efficiency and optimizing production.
Ergonomics has been increasing its place within the OHS in the industrial envi-
ronment in recent decades [16]. Some new approaches have been developed in close
alignment with the progressive digitalization of industry. One of them can be Partic-
ipatory Ergonomics [17] to be able to define optimal adaptation of the work environ-
ment to men on the collaborative ground. But another one, still increasingly gaining
space, is Cognitive Ergonomics. In particular, it occurs in face of the increase of the
human–machine interactions and data display screens as the main interface. This
work is based on the premise of the necessity of a change of the paradigm of both the
innovation opportunities and Cognitive Ergonomics, to be actively part of Industry
4.0 within the company’s strategies. The main purpose of this research is to provide a
prospective revision for the future to combine innovation and Cognitive Ergonomics
linked to the processes and multiactor systemic approach of Industry 4.0.
27.2 Theoretical Background and Method
A new mode of production of Industry 4.0 strengthened by the use of digital tech-
nologies aims to potential hybridizations between the physical world and the virtual
world called cyberphysical system. These hybrid environments have a strong impact
on the cognitive capacities of workers [18–20]. Even if some authors [21] address
the statement that research in Industry 4.0 is strictly related to machine and tech-
nology focuses on machines and does not examine the human workforce’s role for
the strategic planning and implementation of smart factories, the importance of the
correct evaluation of human performance is necessary for designing intelligent manu-
facturing systems [22, 23]. It is relevant to emphasize that one of the main goals of
Industry 4.0 is to reach a better system performance as well as a better workplace
for humans [24]. Nevertheless, the majority of key Industry 4.0 technologies have
some control and command issues to be resolved when complex systems of smart
factory considered [25]. It rises the challenge of delimiting the role of the human
operator of all integrated and digital interfaces [26]. These questions related to so
called Operator 4.0 have been the basis of different studies and research approaches
[27–30].
As exposed by Romero et al. [31], the Operator 4.0 fulfilling the assigned task is
permanently exposed to a wide range of factors that demand a high state of alert and
awareness. The Industry 4.0 environments claim constant maximization of decision-
making, perception, and analysis skills to be able to control diverse data display
interfaces and to perform in a flexible way to adjust the requirements of the technology
[24]. Operator 4.0, highly skilled and trained at the digital level, performs more
cognitive than physical tasks [32]. The mental and cognitive abilities, as well as the
mastery and efficient operational performance of different technologies, workflow,
and processes’ analysis are already demanded as a regular daily basis from the
Operator 4.0 to boost productivity and accelerate real-time decision-making. It relates
closely with the necessity of processing significant amounts of data and information
during the work shift. Both these strands of work performance generate significant
impact on cognition and might lead to mental workload [32]. The mental workload
of Industry 4.0 can be considered psychosocial risks on the individual level of each
worker [28]. However, in the highly interconnected systems as Industry 4.0, the high
level of attentiveness and mental workload might lead to a wider range of risks: system
failures, more frequent maintenance breaks, resources wasted. Hence, it increasingly
becomes a major strategic goal of Industry 4.0 to assess mental workload in order to
reduce or avoid the high load imposed on the operator [33–35].
Cognitive ergonomics focuses on the interaction between tools and users; the
mental human mechanisms of information processing and understanding; and the
impact of environmental and task factors on the cognitive processes of the capacity
of reasoning and taking an action [36]. The Cognitive Ergonomics in the situation
as Industry 4.0 systems aims to define optimal forms of human interaction with
several elements of automatized and digitalized manufacturing systems. The Cogni-
tive Ergonomics approach within the OHS of Industry 4.0 provides the analysis of
how Operator 4.0 work affects the mental processes and how mental processes may
affect (in particular, negatively) the work [37]. Some authors [38] have provided
studies based on the cognitive work analysis (CWA) to detect and define assistance
systems to support workers in the control of Intelligent Manufacturing System in
Industry 4.0. Although this type of analysis is necessary, the real challenge lies in
the correct definition of ranges of effectiveness when the mental workload is consid-
ered. The appropriate measuring of mental workload In the case of Operators 4.0
contributes to a more accurate recognition of worker performance due to the neces-
sity of detecting situations when cognitive factors delay objectives achievement with
task demand increase [39]. The basic premise of Cognitive Ergonomics to provide
safe and efficient operations of complex systems [40] is to ensure the workstation
environment and mental workload best adjusted to each worker capacity [41].
The human–computer interactions widely spread within the Industry 4.0 organiza-
tion bring a substantial change within the use of information systems of the factory as
well as the need for more frequent workstation rotation to guarantee the optimal level
of mental perception due to data display interfaces centricity of Operator 4.0 perfor-
mance. These cognitive considerations are closely related to the design strategic
framework of new approaches focused on the possibility to give the Operator 4.0, an
active position to modify the system configuration (supervision and control) or carry
out maintenance and diagnostic operations [38].
This research has conducted a scoping review of concepts of Cognitive
Ergonomics in Industry 4.0. The search methodology used in this research was based
on the combination of keywords in the WOS and SCOPUS databases. The search
on this database was based on the use of keywords that were combined with each
other and was carried out in February 2021. The keywords chosen for this research
were: “Ergonomics”, “Cognitive Ergonomics”, and “Industry 4.0”. For this research,
the model developed by Aria and Cuccurullo [42] called Bibliometrix was used.
The prospective analysis of human-centered is related to boost innovations process
of Cognitive Ergonomics in Industry 4.0 and it has been developed applying the
Research through Design (RtD) method. This method has been applied to the extrac-
tion of potential fields of future innovations and prospective correlations with the
cognitive performance of Industry 4.0 workers. RtD is a method that is used increas-
ingly both inside designs, architectural- and engineering-design research [43] With
clear limitations as a theory-building approach [44], the RtD can provide overall
delimitation of new conceptualizations; thus, it can borne sufficient as an initial
proposal of any framework within the not-fully empirically studied contexts. The
present research, as an initial part of a bigger applied study, complies RtD approach
as it attempts to revalue common phenomena within a limited context [45] and it
proposes a new conceptual approximation [46].
27.3 Human-centered Innovation in Industry 4.0

in Cognitive Ergonomics Scope
Cognitive Ergonomics presents the expandable potential to enhance innovations

within the Industry 4.0 paradigm. The emergent technologies that constitute Industry
4.0 may take the role of attraction poles to design and develop a specific solution to
give an answer to increasing demands in the cognitive risks and prevention manage-
ment. To propose some examples–however, the range of possibilities is much wider—
the virtualization of Industry 4.0 processes becomes both the cause of cognitive risks
and the area of opportunities to create innovative devices, services, or processes.
Some steps toward Virtual Ergonomics centered both on physical and cognitive
factors, have been undertaken [47]. According to this approach the evaluation of the
human factors in virtual, Industry 4.0 environments is achievable to simulate the man
interaction with complex machine and digital systems.
Also, the trend of cognitive manufacturing emerges from the Industry 4.0
challenges based on Cognitive Ergonomics. Cognitive manufacturing automatizes
workers reactions according to the real-time and applied analytics of 4.0 systems
and facilitates actionable updated knowledge to ease the stress put on human perfor-
mance [48, 49]. As an initial finding of this research, the overall framework of
opportunities for human-centered innovations is proposed as follows in Fig. 27.1.
Fig. 27.1 Areas of human-centered innovations’ opportunities according to Cognitive Ergonomics

perspective of Industry 4.0 (Source Authors’ elaboration)
27.4 Conclusion and Future Work
The interest in the Industry 4.0 is rapidly increasing from business, society, and
academia. Switching from the reactive model of business innovations to more
emergent fields, the Cognitive Ergonomics scope must be taken into considera-
tion. However, the relationship between occupational health and safety (OHS), the
cognitive perspective of workers and, above all, the “real” impact on operations and
businesses is still to be analyzed, the preliminary finding of the present research,
based on existing scientific research, provides areas of potential opportunities to
activate the innovative development generation. This finding is closely related to
the fact that human-centered innovations within the internal industry environments
often requires in-depth adaptations in firms’ business models or even the creation of
new ones. This Cognitive Ergonomics challenge may produce new streams of R&D
development aligned with OHS strategies or regulations for Industry 4.0. But also,
it can scale out and provide a new approach of innovation for Industry 4.0 auxiliary
business models. The results of the study, even at this exploratory and conceptual
scope till now, might have an important impact on potentialities for Industry 4.0 wider
ecosystems. Hence, the well-being and effectiveness of the Industry worker are still
at the very core of any human-centered innovations based on Cognitive Ergonomics
performance within the OHS strategy of each smart factory. In future works the vali-
dation instruments of Cognitive Ergonomics-related innovations will be developed
to become an effective support for executive decision-making. These proposed areas
of opportunities for innovation, although many shortcomings are envisioned within
the initial version, will also require the future in-depth strategic analysis to be able
to draw the roadmaps of implementation in real Industry 4.0 organizations.
References
1. Macit I (2017) Industry 4.0 benefits of enterprise resource planning: proposal a framework
model. Yönetim Bilişim Sistemleri Dergisi 2(1):50–60 or KURUMSAL kaynak planlamasinin
endüstri 4.0 kazanimlari: bir yapisal çati modeli önerisi. Yönetim Bilişim Sistemleri Dergisi
2(1):50–60
2. Jerman A, Bach MP, Aleksic A (2020) Transformation towards smart factory system: examining
new job profiles and competencies. Syst Res Behav Sci 37(2):388–402
3. Liao Y, Deschamps F, de Freitas Rocha Loures E, Ramos LFP (2017) Past, present and future
of industry 4.0 - a systematic literature review and research agenda proposal. Int J Prod Res
55(12):3609–3629
4. Reinhardt IC, Oliveira JC, Ring DT (2020) Current perspectives on the development of industry
4.0 in the pharmaceutical sector. J Ind Inform Integr 18:100131
5. Ganzarain J, Errasti N (2016) Three stage maturity model in SME’s toward industry 4.0. J Ind
Eng Manage (JIEM) 9(5):1119–1128
6. Haleem A, Javaid M (2019) Additive manufacturing applications in industry 4.0: a review. J
Ind Integr Manage 4(4):1930001–1930024
7. Brettel M, Friederichsen N, Keller M, Rosenberg M (2014) How virtualization, decentralization
and network building change the manufacturing landscape: an industry 4.0 perspective. Int J
Mech Ind Sci Eng 8(1):37–44
8. Birkel HS, Veile JW, Müller JM, Hartmann E, Voigt K-I (2019) Development of a risk
framework for industry 4.0 in the context of sustainability for established manufacturers.
Sustainability 11(2):384–411
9. Moktadir MA, Ali SM, Kusi-Sarpong S, Shaikh MAA (2018) Assessing challenges for imple-
menting industry 4.0: implications for process safety and environmental protection. Process
Saf Environ Prot 117:730–741
10. Ivanov LA, Kapustin IA, Borisova ON, Plsarenko ZV (2020) Nanotechnologies: a review of
inventions and utility models. Part II. Nanotechnol Constr Sci Internet-J 12(2):71–76
11. Türkes, MC, Oncioiu I, Aslam HD, Marin-Pantelescu A, Topor DI, Căpus, neanu S (2019) Drivers
and barriers in using industry 4.0: a perspective of SMEs in Romania. Processes 7(3):153–173
12. Thames L, Schaefer D (2016) Software-defined cloud manufacturing for industry 4.0. Procedia
CIRP 52:12–17
13. Xu LD, Duan L (2019) Big data for cyber physical systems in industry 4.0: a survey. Enterp
Inf Syst 13(2):148–169
14. Badri A, Boudreau-Trudel B, Souissi AS (2018) Occupational health and safety in the industry
4.0 era: a cause for major concern? Saf Sci 109:403–411
15. Kocaay F (2019) Possible effects of industry 4.0 to work life, occupational health and safety.
Int Refereed Acad J Sports Health Med Sci 31:71–83
16. Christy V, Duraisamy S (2020) Ergonomics and employee psychological well being. Int J
Manage 11(3):435–438
17. Burgess-Limerick R (2018) Participatory ergonomics: evidence and implementation lessons.
Appl Ergon 68:289–293
18. Sun S, Zheng X, Gong B, Paredes JG, Ordieres-Meré J (2020) Healthy operator 4.0: a human
cyber–physical system architecture for smart workplaces. Sensors 20(7):1–21
19. Planke LJ, Lim Y, Gardi A, Sabatini R, Kistan T, Ezer N (2020) A cyber-physical-human
system for one-to-many UAS operations: cognitive load analysis. Sensors 20(19):5467
20. Jiao J, Zhou F, Gebraeel NZ, Duffy V (2020) Towards augmenting cyber-physical-human
collaborative cognition for human-automation interaction in complex manufacturing and
operational environments. Int J Prod Res 58(16):5089–5111
21. Parrott A, Warshaw L (2017) Industry 4.0 and the digital twin – manufacturing meets its match.
In: Deloitte.Insights. Retrieved from https://www2.deloitte.com/us/en/insights/focus/industry-
4-0/digital-twin-technology-smart-factory.html
22. Fantini P, Pinzone M, Taisch M (2020) Placing the operator at the centre of industry 4.0 design:
modelling and assessing human activities within cyber-physical systems. Comput Ind Eng
139:105058
23. Jones AT, Romero D, Wuest T (2018) Modeling agents as joint cognitive systems in smart
manufacturing systems. Manuf Lett 17:6–8
24. Longo F, Nicoletti L, Padovano A (2017) Smart operators in industry 4.0: a human-centered
approach to enhance operators’ capabilities and competencies within the new smart factory
context. Comput Ind Eng 113:144–159
25. Kadir BA, Broberg O (2021) Human-centered design of work systems in the transition to
industry 4.0. Appl Ergon 92:103334
26. Romero D, Stahre J, Taisch M (2020) The operator 4.0: towards socially sustainable factories
of the future. Comput Ind Eng 139:2–5
27. Gorecky D, Schmitt M, Loskyll M, Zühlke D (2014) Human-machine interaction in the industry
4.0 era. In: 2014 12th IEEE International conference on industrial informatics (INDIN). IEEE,
pp 289–294
28. Papcun P, Kajáti E, Koziorek J (2018) Human machine interface in concept of industry 4.0.
In: 2018 World symposium on digital intelligence for systems and machines (DISA). IEEE,
pp 289–296
29. Wittenberg C (2016) Human-CPS interaction—requirements and human-machine interaction
methods for the industry 4.0. IFAC-PapersOnLine 49(19):420–425
30. Kaasinen E, Schmalfuß F, Özturk C, Aromaa S, Boubekeur M, Heilala J, Walter T et al (2020)
Empowering and engaging industrial workers with operator 4.0 solutions. Comput Ind Eng
139:105678
31. Romero D, Bernus P, Noran O, Stahre J, Fast-Berglund Å (201) The operator 4.0: human cyber-
physical systems & adaptive automation towards human-automation symbiosis work systems.
In: Advances in production management systems. Initiatives for a sustainable world. APMS
2016. IFIP Advances in information and communication technology, vol 488. Springer, Cham,
pp 677–686
32. Pascual DG, Daponte P, Kumar U (2019) “Operator 4.0” handbook of industry 4.0 and smart
systems, pp 239–285
33. Sweller J (1988) Cognitive load during problem solving: effects on learning. Cogn Sci
12(2):257–285
34. Gevins A, Smith ME (2003) Neurophysiological measures of cognitive workload during
human–computer interaction. Theor Issues Ergon Sci 4(1–2):113–131
35. Illankoon P, Tretten P, Kumar U (2018) Identifying significance of human cognition in future
maintenance operations. In: Intelligent human systems integration. IHSI 2018. Advances in
intelligent systems and computing, vol 722. Springer, Cham, pp 550–556
36. Pesante JA, Williges RC, Woldstad JC (2001) The effects of multitasking on quality inspection
in advanced manufacturing systems. Human Factors Ergon Manuf Serv Ind 11(4):287–298
37. Bommer SC, Fendley M (2018) A theoretical framework for evaluating mental workload
resources in human systems design for manufacturing operations. Int J Ind Ergon 63:7–17
38. Pacaux-Lemoine M-P, Berdal Q, Guérin C, Rauffet P, Chauvin C, Trentesaux D (2022)
Designing human–system cooperation in industry 4.0 with cognitive work analysis: a first
evaluation. Cogn Technol Work 24:93–111
39. Colombi JM, Miller ME, Schneider M, McGrogan MJ, Long CDS, Plaga J (2012) Predictive
mental workload modeling for semiautonomous system design: implications for systems of
systems. Syst Eng 15(4):448–460
40. Wu X, Feng C, Wanyan X, Liu S, Ding L, Miao C, Wang Y, He X () How shared screen affected
team collaboration task, a case study of ergonomics experiment on team situation awareness.
In: Engineering psychology and cognitive ergonomics. EPCE 2018. Lecture notes in computer
science, vol 10906. Springer, Cham, pp 241–249
41. Hertzum M, Holmegaard KD (2013) Perceived time as a measure of mental workload: effects
of time constraints and task success. Int J Hum-Comput Interact 29(1):26–39
42. Aria M, Cuccurullo C (2017) bibliometrix: an R-tool for comprehensive science mapping
analysis. J Informet 11(4):959–975
43. Verbeke J, Pak B (2013) Knowing by designing. KU Leuven, Leuven
44. Zimmerman J, Stolterman E, Forlizzi J (2010) An analysis and critique of research through
design: towards a formalization of a research approach. In: DIS’10: Proceedings of the 8th
ACM conference on designing interactive systems, Aarhus, Denmark, pp 310–319
45. Grand S, Jonas W (2012) Mapping design research. Birkhauser, Basel, Switzerland
46. Stappers PJ, Giaccardi E (2017) Research through design. In: The encyclopedia of human-
computer interaction. The Interaction Design Foundation, Denmark, pp 1–94
47. Lanzotti A, Vanacore A, Tarallo A, Nathan-Roberts D, Coccorese D, Minopoli V, Papa S et al
(2020) Interactive tools for safety 4.0: virtual ergonomics and serious games in real working
contexts. Ergonomics 63(3):324–333
48. IBM Corporation (2017) Cognitive manufacturing: an overview and four applications that are
transforming manufacturing today. IBM Corporation, Armonk, NY, USA
49. Thorvald P, Lindblom J, Andreasson R (2019) On the development of a method for cognitive
load assessment in manufacturing. Robot Comput-Integr Manuf 59:252–266
Chapter 28
Business Model Patterns: A Systematic
Literature Review
D. Ibarra , A. M. Valenciano, and J. I. Igartua
Abstract Recent studies have shown that the application of business model patterns
facilitates business model innovation. However, the literature is still quite fragmented,
and there is no commonly accepted approach to characterize business models based
on patterns. To fill this gap, this paper conducts a systematic literature review, aiming
to address three research questions: (1) How are BM patterns defined and operational-
ized? (2) What are the application domains in which BM patterns are explored? and
(3) What methodological approaches are followed in defining BM patterns? The
results of the review show that there is a lack of clarity between the concepts of
business model archetypes and business model patterns. Furthermore, few studies
address business model patterns from a generic point of view; there is a lack of inte-
grative approaches that include business model patterns based on different domains;
and some overlap of patterns from one study to another has also been identified.
Finally, in terms of methodological approach, the use of morphological analysis and
the BM pattern generation methodology suggested by (Amshoff et al. in Int J Innov
Manag 19:1540002, 2015) is recommended for the analysis and identification of
business model patterns.
Keywords Business model · Business model innovation · Business model

archetypes · Business model patterns
28.1 Introduction
In an ever-changing competitive environment increasingly challenged by digitaliza-

tion and sustainability, business model innovation (BMI) is becoming key to creating
a sustainable competitive advantage [1]. BMI refers to purposeful changes to the
value delivery, value creation and value capture dimensions of a firm’s business
model (BM) and/or to the architecture linking them [2]. Through BMI, companies
D. Ibarra (B) · A. M. Valenciano · J. I. Igartua

Mondragon Unibertsitatea, Loramendi 4, 20500 Mondragón, Spain
e-mail: dibarra@mondragon.edu
https://doi.org/10.1007/978-3-031-29382-5_28
282 D. Ibarra et al.
can create new ways to deliver more value to customers and stakeholders, compre-
hensively optimize their resources and capabilities for value creation, and identify
new ways to reduce costs and generate new sources of revenue. However, companies
still struggle to think in terms of BMs, and there is a lack of systematic tools to
facilitate the creativity and divergent thinking needed to innovate them [3].
In response, some authors recommend using patterns or archetypes as analogies
for creative imitation. These tools are intended to serve as ideal examples of BM
types (i.e., archetypes) based on pattern recognition in the structure of existing BMs
[4].
Patterns describe how to configure design options to transform BMs based on
strategic objectives [1]. They incorporate empirical findings from real cases and are
usually described with a meaningful title, a short description, and an overview of
the BM components that play a key role in the pattern [3]. One of the reasons for
the popularity of patterns could be associated with their ability to provide simplified
solutions to complex processes [5]. Thus, several contributions can be found in the
academic literature that propose the use of BM patterns, as tools to understand and
learn from existing solutions and to generate new business opportunities [4, 6, 7].
This article is part of an ongoing research project called NEBA. The main objective
of the project is the characterization of the industrial fabric of Gipuzkoa from the
perspective of BMs. Due to the extent of BM literature, the first work package of
NEBA has focused on developing a systematic literature review on BM patterns.
This article, therefore, describes the review process followed, the results found, and
the conclusions drawn to define the methodology to be adopted in the project for the
characterization of BMs based on patterns.
28.2 Research Methodology
To develop the research, a systematic review was conducted following the five
steps proposed by Denyer and Tranfield [8]. First, three research questions were
formulated:
1. How are BM patterns defined and operationalized?
2. What are the application domains in which BM patterns are explored?
3. What methodological approaches are followed in the definition of BM patterns?
Next, the location of articles was addressed by using two databases, Scopus and
Web of Science. The search for articles was carried out in November 2020 with the
following search string: “business model pattern*” OR “business model archetype*”
OR “business model typolog*” OR “business model taxonom*” in titles, abstracts,
and keywords. Only scientific articles and reviews published in English were included
to ensure quality and comprehensibility. The total number of articles identified was
110, which came down to a total of 66 articles published between 2011 and 2020,
after eliminating duplicates.
28 Business Model Patterns: A Systematic Literature Review 283
To select and evaluate the articles, the titles and abstracts of the 66 articles were
analyzed in detail to identify those that were relevant to the research. To this end,
two inclusion criteria were established:
• Articles had to be focused on business cases. Thus, articles that explored BMs of
educational centers, universities, or public institutions were excluded.
• Articles had to contain BM patterns. Articles that did not directly address BM
patterns’ definition, identification, or analysis were excluded (i.e., development
of BM frameworks or business modeling tools).
Considering these criteria, 28 articles were excluded as they did not meet the
inclusion criteria. It should be clarified that the application of these criteria meant
that no articles from 2011, 2012, and 2013 were included. The remaining 39 articles
were carefully read to identify those that potentially answered the purpose of the
study.
28.3 Results of the Review
The results of the review show a wide range of approaches in BM patterns literature
(Table 28.1). The studies identified differ in the number of patterns identified, their
scope, and their methodological approach. In the following lines, the results of the
review are described according to the research questions established in Sect. 28.2.
28.3.1 How Are BM Patterns Defined and Operationalized?
Patterns defined in the studies range from a single archetype [10] to 194 patterns
[6]. Twenty studies operationalize the patterns based on BM components, using
previously identified BM frameworks [e.g., 15, 31, 34, 39, 42]. On the other hand,
nineteen authors do not operationalize the BM patterns but rather provide a brief
description of the patterns [e.g., 1, 35, 12] or they classify identified patterns based
on the BM dimensions on which they have an impact [e.g., 4, 26].
There is often a lack of clarity in the constructs used, with the terms archetypes and
patterns often being used interchangeably. Some scholars make a distinction between
the two concepts but, even so, the definition of each one, as well as their relationship,
remains unclear [7, 34]. Several authors make a distinction between prototypical
business models, which represent holistic business models, and solution patterns,
which focus on patterns affecting certain building blocks, such as razor and blade
[4, 7]. From the articles reviewed, Lüdeke-Freund et al. [29] suggest adopting the
approach of Alexander et al. [43], who argue that a pattern “describes a problem
which occurs over and over again in our environment and then describes the core of
the solution to that problem in such a way that you can use this solution a million
times over without ever doing it the same way twice” [29, 43]. In this sense, business
Table 28.1 Articles selected for evaluation from the systematic literature review
Reference Nº/Op Scope Methodological approach
Abdelkafi and Hansen [9] 4/N Eco-entrepreneurship; Case studies
e-mobility
Amshoff et al. [7] 10/Y Disruptive technologies Mixed method
Andreassen et al. [10] 1/Y Two-sided BM; mobility Case studies
Ansong y Boateng [11] 16/Y Digital firms; emerging Descriptive statistics
economies
Birkie [12] 9/N Sustainability; Case studies
manufacturing
Bocken et al. [1] 8/N Sustainability Literature review
Bohnsack et al. [13] 4/Y Sustainability; electric Case studies
vehicles
Brown et al. [14] 7/N Renewable energy; Mixed method
prosumers
Burger and Luke [15] 3/Y Distributed energy resources Descriptive statistics
D’Amato et al. [16] 8/Y Bioeconomy; CE Case studies
Facchinetti and Sulzer [17] 3/Y Energy management; energy Conceptual
hubs
Frankenberger et al. [18] 4/N Open innovation Case studies
Garbuio and Lin [19] 10/N Health start-ups; AI Conceptual
Giovani [20] 4/N Open data; pharmaceuticals; Descriptive statistics
biotech
Gyimóthy [21] 3/N Collaborative economy; Conceptual
tourism
Holzmann et al. [22] 2/N 3D printer manufacturers Review; cluster analysis
Hora et al. [23] 10/N Sustainable mass Conceptual
customization
Kortmann and Piller [24] 9/N Open innovation; Conceptual
manufacturing
Kowalkowski [25] 3/Y Innovation in services Conceptual
Kwon et al. [26] 55/N PSS Literature review
Laudien and Pesch [27] 4/Y Digital services; Case studies
digitalization
Linton and Öberg [28] 4/N Digitalization; tourism Conceptual
Lüdeke-Freund et al. [29] 6/Y CE Literature review
Lüdeke-Freund et al. [30] 45/N Sustainability Review; delphy
Mosig et al. [31] 3/Y Mass customization; textile Case studies
industry
Peppou [32] 5/N Biotechnology Content analysis
Peters et al. [33] 3/Y Telemedicine Case studies
(continued)

Reference Nº/Op Scope Methodological approach
Pieroni et al. [5] 20/N CE Literature review
Pieroni et al. [34] 162/Y CE; manufacturing; Mixed method
multisectorial
Reinhardt et al. [35] 9/N Sustainability; e-vehicles; Case studies
batteries
Remane et al. [4] 182/Y Generic BM patterns Literature review
Trabucchi et al. [36] 4/N Collaborative economy Literature review
Ulvenblad et al. [37] 8/Y Agri-food production; Case studies
sustainability
Weking et al. [6] 194/N Generic BM patterns Literature review
Weking et al. [38] 13/Y Industry 4.0 Case studies
Whalen [39] 3/Y CE Literature review
Yang and Evans [40] 4/Y PSS; sustainability Case studies
Zeleti and Ojo [41] 15/Y Open data Conceptual
Zufall et al. [42] 7/Y Sustainability; smartphones Case studies
Notes Nº/Op. Number of patterns identified/operationalization of patterns (Y: Yes; N: No); PSS:
Product-service system; CE: Circular economy; AI: Artificial intelligence
model patterns can be understood as the combination of various business model

design choices repeatedly observed across the business models of different unrelated
companies.
Overall, the patterns identified during the review often overlap or are repeated in
several studies, sometimes with different names. Therefore, it is difficult to compare
the patterns with each other, due to different conceptualizations and interpretations
of the concepts business model, pattern, and archetype.
28.3.2 What Are the Application Domains in Which BM

Patterns Are Explored?
As for the scope of the selected articles, fourteen authors identified and analyzed
BM patterns through the lens of sustainability [1, 9, 12, 13, 23, 30, 35, 37, 42] and,
more specifically, from a circular economy approach [5, 16, 29, 34, 39]. Five studies
focus on patterns derived from digital transformation and disruptive technologies
[7, 22, 28, 31, 38]. Some authors describe BM patterns based on servitization and
product-service systems [25–27, 40]. Others, in turn, have particularly focused on
energy hubs and renewable energy [14, 15, 17]. Four studies explore BM patterns
in the context of open innovation [18, 20, 24, 41]. Three studies address patterns
in the collaborative economy [21, 36] and multisided markets [10]. Another three
articles analyzed BM patterns in the health sector [19, 32, 33]. Finally, only two of
the articles analyzed provide generic BM patterns [4, 6]. These results emphasize
the increasing need to understand and characterize potential new business models in
the context of the climate change and digital transformation challenges. Moreover,
in recent years, there is mainly a trend toward the study of circular economy business
model patterns.
28.3.3 What Methodological Approaches Are Followed

in the Definition of BM Patterns?
In terms of the methods applied, most of the research is based on case studies (37%).
Some authors carried out a literature review to define the BM patterns of their studies,
drawing on patterns previously defined by other authors (26%). Some of the articles
were conceptual (21%). Finally, some authors apply descriptive statistics (8%) and
mixed methods (8%) to identify BM patterns.
From the articles analyzed, two main methodological approaches have been iden-
tified as relevant for the characterization of BMs based on patterns: the morphological
analysis [26, 29, 38] and the methodology for BM pattern generation suggested by
Amshoff et al. [7] and recently adopted by Pieroni et al. [34].
Morphological Analysis Morphological analysis is a problem-solving approach in

which a solution, in this case the BM, is decomposed into smaller dimensions (BM
dimensions). These dimensions can in turn be further evaluated independently, spec-
ifying the components and related characteristics that comprise each BM dimension
[26]. This technique is useful for qualitative analyses of multidimensional objects,
such as BMs, as it reduces the complexity and number of design choices that are rele-
vant for their characterization [29]. In the articles analyzed in this research, authors
usually define the BM dimensions addressed in the morphological analysis based on
data extracted from BM patterns existing in the literature or from the results of case
studies.
Methodology for BM Pattern Generation The first step of this methodology is the
selection of the companies to be analyzed. Secondly, a business model framework
is selected. Thirdly, the business models of the selected companies are described by
means of a series of related variables and configuration options. For this purpose,
company websites, industry portals, academic and gray literature, interviews, etc.
can be used as sources of information. Fourthly, a list of binary characteristics is
drawn up indicating the configuration options that each company uses in its business
model. To determine which recurring combinations exist, the authors recommend
the use of a similarity matrix, showing which configuration options are used together
in a large number of business models [7, 34]. To this end, multidimensional scaling
(MDS) is recommended, where configuration options with a high similarity value
are placed in proximity within a two-dimensional map. This makes it possible to
identify patterns of business models in an easy and comprehensible way. Finally, a

second matrix that allows assessment of how often patterns are combined with each
other for each company needs to be developed.
28.4 Conclusions
The systematic review of the literature presented in this article illustrates the diversity
of approaches used to study business model patterns. Returning to our three research
questions, the following conclusions can be drawn:
1. There is a lack of clarity between the concepts of archetype and business model
patterns. Therefore, before starting any investigation, we recommend clarifying
the construct to be used, its definition, scope, and dimensions.
2. The scope of the selected articles varies considerably, most of them being focused
on a particular research domain (e.g., sustainability or circular economy). In
turn, few studies address business model patterns from a generic view. There is
a lack of integrative approaches including business model patterns based on
different scopes. Moreover, overlaps between patterns from different scopes
should be explored. For instance, patterns within circular economy, product-
service systems, and digitalization domains refer sometimes to the same patterns
with different nomenclatures.
3. The common method for identifying and classifying patterns is observation and
case studies, which provide meaningful information on business model patterns,
but are subject to the opinion of the authors, are difficult to replicate and do
not allow for generalization of results. Morphological analysis is considered
useful for defining and visualizing the different design options underlying busi-
ness model patterns, thereby reducing the complexity associated with them and
considering only those options relevant to their characterization. Finally, the
methodology suggested by Amshoff et al. [7] is considered appropriate when
the aim of the research is the identification of new business model patterns in
multiple sectors [34]. It is therefore concluded that, with a view to developing the
BM characterization tool, the ongoing NEBA project will follow the BM pattern
generation methodology.
References
1. Bocken NMP, Short SW, Rana P, Evans S (2014) A literature and practice review to develop
sustainable business model archetypes. J Cleaner Prod 65:42–56
2. Foss NJ, Saebi T (2018) Business models and business model innovation: between wicked and
paradigmatic problems. Long Range Plan 51(1):9–21
3. Gassmann O, Frankenberger K, Csik M (2014) The business model navigator: 55 models that
will revolutionise your business, FT Publishing (Financial Times). Pearson, p 400
4. Remane G, Hanelt A, Tesch JF, Kolbe LM (2017) The business model pattern database-a tool
for systematic business model innovation. Int J Innov Manage 21(1):1750004
5. Pieroni MPP, McAloone TC, Pigosso DCA (2020) From theory to practice: systematising and
testing business model archetypes for circular economy. Resour Conserv Recycl 162:105029
6. Weking J, Hein A, Böhm M, Krcmar H (2018) A hierarchical taxonomy of business model
patterns. Electron Markets 30(3):447–468
7. Amshoff B, Dülme C, Echterfeld J, Gausemeier J (2015) Business model patterns for disruptive
technologies. Int J Innov Manage 19(3):1540002
8. Denyer D, Tranfield D (2009) Producing a systematic review. In: The sage handbook of
organizational research methods. Sage Publications Limited, pp 671–689
9. Abdelkafi N, Hansen EG (2018) Ecopreneurs’ creation of user business models for green tech:
an exploratory study in e-mobility. Int J Entrepreneurial Ventur 10(1):32–55
10. Andreassen TW, Lervik-Olsen L, Snyder H, Van Riel ACR, Sweeney JC, Van Vaerenbergh
Y (2018) Business model innovation and value-creation: the triadic way. J Serv Manage
29(5):883–906
11. Ansong E, Boateng R (2019) Surviving in the digital era—business models of digital enterprises
in a developing economy. Digital Policy Regul Governance 21(2):164–178. https://doi.org/10.
1108/DPRG-08-2018-0046
12. Birkie SE (2018) Exploring business model innovation for sustainable production: lessons from
Swedish manufacturers. Procedia Manuf 25:247–254
13. Bohnsack R, Pinkse J, Kolk A (2014) Business models for sustainable technologies: exploring
business model evolution in the case of electric vehicles. Res Policy 43(2):284–300
14. Brown D, Hall S, Davis ME (2019) Prosumers in the post subsidy era: an exploration of new
prosumer business models in the UK. Energy Policy 135:110984
15. Burger SP, Luke M (2017) Business models for distributed energy resources: a review and
empirical analysis. Energy Policy 109:230–248
16. D’Amato D, Veijonaho S, Toppinen A (2020) Towards sustainability? Forest-based circular
bioeconomy business models in Finnish SMEs’. Foresh Policy Econ 110:101848
17. Facchinetti E, Sulzer S (2016) General business model patterns for local energy management
concepts. Front Energy Res 4(7):1–12
18. Frankenberger K, Weiblen T, Gassmann O (2014) The antecedents of open business models:
an exploratory study of incumbent firms. R&D Manage 44(2):173–188
19. Garbuio M, Lin N (2019) Artificial intelligence as a growth engine for health care startups:
emerging business models. Calif Manage Rev 61(2):59–83
20. Giovani B (2017) Open data for research and strategic monitoring in the pharmaceutical and
biotech industry. Data Sci J 16:18
21. Gyimóthy S (2017) Business models of the collaborative economy. In: Collaborative economy
and tourism. Tourism on the verge. Springer, Cham, pp 31–39
22. Holzmann P, Breitenecker RJ, Schwarz EJ (2019) Business model patterns for 3D printer
manufacturers. J Manuf Technol Manage 31(6):1281–1300
23. Hora M, Hankammer S, Canetta L, Sel SK, Gome S, Gahrens S (2016) Designing business
models for sustainable mass customization: a framework proposal. Int J Ind Eng Manage
7(4):143–152
24. Kortmann S, Piller F (2016) Open business models and closed-loop value chains: redefining
the firm-consumer relationship. Calif Manage Rev 58(3):88–108
25. Kowalkowski C (2016) Service innovation in industrial contexts. In: Service innovation.
Translational systems sciences, vol 6. Springer, Tokyo, pp 235–249
26. Kwon M, Lee J, Hong YS (2019) Product-service system business modelling methodology
using morphological analysis. Sustainability 11(5):1376
27. Laudien SM, Pesch R (2019) Understanding the influence of digitalization on service firm
business model design: a qualitative-empirical analysis. Rev Manage Sci 13(3):575–587
28. Linton G, Öberg C (2020) A conceptual development of a business model typology in tourism:
the impact of digitalization and location. Technol Innov Manage Rev 10(7):16–27
29. Lüdeke-Freund F, Gold S, Bocken NMP (2019) A review and typology of circular economy
business model patterns. J Ind Ecol 23(1):36–61
30. Lüdeke-Freund F, Carroux S, Joyce A, Massa L, Breuer H (2018) The sustainable busi-
ness model pattern taxonomy—45 patterns to support sustainability-oriented business model
innovation. Sustain Prod Consumption 15:145–162
31. Mosig T, Grafmüller LK, Lehmann C (2017) Business model patterns of B2B mass customizers:
the case of German textile SMEs. Int J Ind Eng Manage 8(3):99–110
32. Peppou G (2018) Biotechnology-driven business model archetypes: sustainability, innovation
and commercial viability. J Commerical Biotechnol 24(3):41–56
33. Peters C, Blohm I, Leimeister JM (2015) Anatomy of successful business models for complex
services: insights from the telemedicine field. J Manage Inf Syst 32(3):75–104
34. Pieroni MPP, McAloone TC, Pigosso DCA (2021) Circular economy business model
innovation: sectorial patterns within manufacturing companies. J Cleaner Prod 286:124921
35. Reinhardt R, Christodoulou I, García BA, Gassó-Domingo S (2020) Sustainable business model
archetypes for the electric vehicle battery second use industry: towards a conceptual framework.
J Cleaner Prod 254:119994
36. Trabucchi D, Muzellec L, Ronteau S (2019) Sharing economy: seeing through the fog. Internet
Res 29(5):996–1013
37. Ulvenblad P-o, Ulvenblad P, Tell J (2019) An overview of sustainable business models for
innovation in Swedish agri-food production. J Integr Environ Sci 16(1):1–22
38. Weking J, Stöcker M, Kowalkiewicz M, Böhm M, Krcmar H (2020) Leveraging industry 4.0 – a
business model pattern framework. Int J Prod Econ 225:107588
39. Whalen KA (2019) Three circular business models that extend product value and their
contribution to resource efficiency. J Cleaner Prod 226:1128–1137
40. Yang M, Evans S (2019) Product-service system business model archetypes and sustainability.
J Cleaner Prod 220:1156–1166
41. Zeleti FA, Ojo A (2017) The 6-values open data business model framework. In: Government
3.0–next generation government technology infrastructure and services. Public administration
and information technology, vol 32. Springer, Cham, pp 219–239
42. Zufall J, Norris S, Schaltegger S, Revellio F, Hansen EG (2020) Business model patterns
of sustainability pioneers - analyzing cases across the smartphone life cycle. J Cleaner Prod
244:118651
43. Alexander C, Ishikawa S, Silverstein M (1977) A pattern language: towns, buildings,
construction. Oxford university press, New York
Part VIII
Supply Chain Management and Logistics
Chapter 29
The Potential of Industry 4.0 in Lean
Supply Chain Management
John Reyes , Josefa Mula , and Manuel Díaz-Madroñero
Abstract This paper aims to determine the relations between Industry 4.0 (I4.0)
technologies and lean manufacturing (LM) practices to provide a lean supply chain
management 4.0 (LSCM 4.0) framework. First, a systematic review of the scientific
literature on LSCM 4.0 is presented to examine its content and degree of contex-
tualization. Next a general LSCM 4.0 construct is proposed, based on identified
theoretical evidence. Ten waste types that impact the performance of today’s compa-
nies are indicated. The use of lean tools supports the change in the organizational
culture toward a flexible resilient organization. I4.0 technologies, such as IoT, cloud
computing, artificial intelligence and simulation, among others, are fundamental for
the digital transformation of supply chains (SCs) and well support the implemen-
tation of LM tools like Kanban and just-in-time. For SC users and researchers, the
results contribute a decision-making approach in a digitization context and, at the
same time, to reduce waste, even when facing possible disruptions.
Keywords Lean manufacturing · Supply chain management · Industry 4.0
J. Reyes · J. Mula · M. Díaz-Madroñero (B)

València, Alcoy, Spain
e-mail: fcodiama@cigip.upv.es
J. Reyes
e-mail: jreyvas@doctor.upv.es; johnpreyes@uta.edu.ec
J. Mula
J. Reyes
Universidad Técnica de Ambato, Ambato, Ecuador
https://doi.org/10.1007/978-3-031-29382-5_29
294 J. Reyes et al.
29.1 Introduction
In today’s world, industrial SCs face formidable challenges to efficiently establish

tools that lower costs and are competitive in a digitalized environment. Supply chain
management (SCM) encompasses the flow of goods from the supplier through manu-
facturing and distribution chains to end users [1]. In line with this integrated SCM
vision, the digital supply chain (DSC) concept implies adopting sophisticated and
intelligent technological capabilities to make SCs more connected, collaborative, and
efficient [2]. In this context, a number of approaches like LM and, more recently, I4.0,
have been developed to help manufacturers to fulfill these objectives [3]. The I4.0
idea is based on advances in information and communication technologies (ICT)
and data warehousing to increase productivity because they support greater opti-
mization and simulation capabilities [4]. The number of studies that have explored
the integration between LM practices and I4.0 technologies has increased in recent
years [5]. Although the need to implement both LM and I4.0 is clear for many SC
managers, they are not sure about how to combine these two elements to achieve
their convergence and to avoid contradictions between operational performance and
the integration of I4.0 technologies into those LM practices [6]. Nevertheless, several
researchers [7–9] have implemented LM tools into SCM activities in an I4.0 environ-
ment and reported improved organizational outcomes, although further studies are
required to validate the proposed conceptual frameworks in various manufacturing
environments to improve practical validity and broaden the scope of their application.
To date, very few authors have studied the I4.0 technologies that most favor the
implementation of LM tools to improve organizational performance, especially with
disruption risks like pandemics or other unexpected crises. This digital revolution is
forcing industrial companies to review their strategies and to possibly revise whether
their previous lean strategy should be adapted or reconsidered to prioritize the deploy-
ment of I4.0 technology [10]. During a pandemic, firms must respond to drastic
changes in supply and demand; for example, during the COVID-19 pandemic, the
lean philosophy was questioned due to supply shortages [11]. Regarding I4.0 tech-
nologies, systems dynamics simulation can help to predict possible points of failure
in SCs, along with the overall impact of the ripple effect on performance [7]. In
this context, the present article aims to analyze the I4.0 technologies that most favor
the implementation of LM tools into current SCs. This literature review emphasizes
existing theoretical discussion on lean SC implementation, which provides a deeper
understanding of new waste inherent to the digitization process and its impact on
organizational performance. So this study contributes to the state-of-the-art litera-
ture on DSCs by proposing a general construct based on the implementation of I4.0
technologies that involve lean thinking as an organizational performance philosophy.
This analysis also provides theoretical arguments that can help researchers and prac-
titioners to develop resilient SCs in situations with disruptive risks because they may
affect performance. Finally in the advent of I4.0, abundant data availability, high
computing power, and large storage capacity have made machine learning (ML)
approaches an attractive solution for addressing manufacturing challenges [12].
29 The Potential of Industry 4.0 in Lean Supply Chain Management 295
The rest of the paper is structured as follows. Section 29.2 describes the proposed
method. Section 29.3 offers a brief description of the literature on SCM (LSCM) and
I4.0 lean technologies with the results of their interaction or LSCM 4.0. Section 29.4
closes the article with conclusions and identified future research lines.
The systematic literature review strategy proposed by Denyer [13] was applied in this
study. It is a specific methodology that locates existing studies, selects and evaluates
contributions, analyzes and synthesizes data, and reports evidence to reach conclu-
sions on the questions posed to fulfill the research objective. This paper reviewed
articles about SCM and LM, combined with I4.0 to develop a flexible and resilient
organization. First, research questions were defined: What is the current LSCM
knowledge state? which I4.0 technologies provide the most support for the implemen-
tation of LM tools?. The Scopus and Web of Science (WoS) databases were chosen
for the review. The search and analysis of articles were performed by combining
several keywords. The search query in the title was: “Supply chain” and “lean” and
“Industry 4.0”. Thus, in the databases, we obtained Scopus–51, WoS–39. Likewise, a
series of exclusion criteria were determined: documents that were not in English; the
last 5 years of publication of the found references; documents that were not aligned
with the research topic; journals that did not appear in the Scimago Journal Rank
for Scopus or in the Journal Citation Report for WoS. Finally, the abstracts of all the
papers were verified, and 27 were selected. Of the selected publications, 22 articles
discuss LM tools related to I4.0 technologies. Several authors refer to SC disruptions
and resilience aspects [7, 11, 14].
29.3.1 Supply Chain Management
SCM has been used for planning and controlling physical and information flows,
internal and external logistics activities, and processes with other companies, and
also for addressing the relationship developed and the processes shared with both
customers and suppliers [15]. Several authors mention SC components as organiza-
tional, information, process functional, technological, and financial [16]. However,
the SC structure is herein classified in terms of: management components, SC
processes, SC flows, and network structure [10]. Indeed, management components
have been the subject of practical studies on a variety of environments and sectors.
Control and planning methods [1–3, 5, 7, 16–23], ICT [1, 3–7, 15–25], and orga-
nizational structure [15–18, 22, 24] are described, while other documents address
knowledge management aspects, such as [2, 4, 24]. We also find SC processes that
296 J. Reyes et al.
represent planning [18, 19, 25], sourcing [17, 20, 23], production [8, 10, 16], and
delivery [2, 7, 19]. Moreover, SCs strive to create a flow of resources from the begin-
ning of the chain (the raw materials end) to the end of the chain (customers) [2].
Virtual goods/services and returns flows are detailed by 15 references [2–8, 15–
19, 21, 22, 25], and virtual value and real-time information also come over in the
reviewed papers [3, 4, 6, 8, 15, 20, 23, 26]. Finally, the network structure involves
vertical and/or horizontal integration [22] and collaborative relationships between
SC processes [1, 22, 23].
29.3.2 Technological Structure
Núñez [15] identified 11 trends in digital technology, which have been supported by
other studies in SCM, and these I4.0 technologies include IoT [19], big data [25],
cloud computing [3], blockchain and social media [2], simulation [7] and digital
twin [4], tracking and tracing systems (TTS) [17], autonomous vehicles[27], artificial
intelligence (AI) or ML [12], radiofrequency identification (RFID) [24], cybersecu-
rity [5], additive manufacturing (AM) or 3D printing [6], and cyberphysical systems
(CPS) [18]. Therefore, I4.0 proposes using several technologies for decision-making,
which can be used to support the control of anomaly identification [3]. For example,
during the COVID-19 pandemic, AM and 3D printing have been used to provide
medical supplies, and the market size of these products has been estimated to grow
[7], while simulation can help to predict possible points of errors in SCs [11, 14]. In
addition, technologies, such as IoT, big data, and AI, are affecting every aspect of
how companies organize and manage their SCs, and strongly influence sustainability
[4, 5, 18].
29.3.3 Lean Manufacturing Tools
Lean thinking, based on the improvement of the Toyota Production System (TPS),
is based on two main pillars [24]: just-in-time (JIT) and automation (jidoka). These
pillars are also the basis of LM [18]. A lean SC must enable a waste-free flow of
goods, services, and technology from suppliers to customers [1].
Therefore, the overall objective of implementing LSCM is to eliminate waste from
non-value added activities [24], including transportation, inventory, motion, waiting,
overproduction, overprocessing, and defects [8]. However, some I4.0 properties may
increase some new waste types: non-utilized talent, poor information management,
poor supplier quality [2, 21, 24, 25]. Moreover, efficient and resilient SCs that offer
the advantages of both lean and risk-resistant SCs have been studied by Ivanov [16].
On the lean tools supported by I4.0 technologies, the results provided by [9, 27]
show that a good consensus and empirical support are found for the interdependence
between two concepts. Table 29.1 provides the relation identified in the reviewed
papers between using I4.0 technologies and LM practices for SCM. On LM tools, 14
references correspond to Kanban and 13 include JIT. Several authors present value
stream mapping (VSM) and productive/preventive maintenance (TPM) with nine
references each, seven references address 5S (sort, straighten, shine, systematize,
sustain), Kaizen (6 references), and five references focus on Poka-yoke and total
quality management (TQM) in their models from a theoretical point of view. At the
same time, another group of four articles [3, 18, 21, 26] addresses Andon. Only three
articles [3, 8, 26] are based on SMED.
Table 29.1 Technology support and tools for LM 4.0

LM tools Description I4.0 technology support and references
Kanban Production control system that uses Cloud computing [3, 5, 6, 15, 19–21, 26],
material flow signaling IoT [3, 5, 6, 15, 18, 19, 21], AM [3, 5, 6,
15, 21, 26], simulation [5, 8, 15, 20, 21,
26, 27], big data [3, 5, 6, 15, 26, 27], AI
[3, 15, 26]
JIT Maximizes efficiency with minimum Big data [2, 3, 5, 15, 27], simulation [3, 5,
waste 7, 15, 21, 23, 27], RFID [2, 15, 17, 22,
24, 26], TTS [2, 5, 17], cybersecurity [5,
15, 22], blockchain [2, 15]
VSM Graphic representation of a SC’s Simulation [3, 8, 21, 27], big data [2, 3,
information flow 21, 25–27], IoT [2, 3, 18, 21, 25], AI [3,
26]
TPM Reduces the frequency with which AM [2, 3, 5, 6, 18, 21, 27], simulation
failures appear in systems and digital twin [3, 5, 8, 21, 26], big data
[2, 3, 5, 6, 26, 27], CPS [2, 3, 5, 18, 21]
5S It aims to achieve permanently cleaner IoT [3, 19, 26], simulation [3, 8, 27], big
and better organized workplaces for data [3, 26, 27], autonomous vehicles [26,
higher productivity 27], AI [3, 26]
Kaizen Continuous processes improvement CPS [2, 3, 5, 18], cloud computing and
big data [2, 3, 5, 27], simulation [5, 8, 15,
27], TTS [2, 5]
TQM Determines the fulfillment of quality IoT, cloud computing and big data [3, 5,
specifications 18, 21], simulation [3, 5, 8, 21],
cybersecurity [5]
Poka-yoke Avoids defects appearing in processes Autonomous vehicles [26, 27], RFID [2,
24, 26], cloud computing [3, 26, 27], AI
[3, 26]
Jidoka Automatically detects errors in IoT [3, 19, 22], big data, and simulation
production processes [3, 22]
Andon Shows production notifications in real AI [3, 9, 22, 26], cloud computing [3, 21,
time 26], big data [3, 21, 26], CPS [9, 18, 21]
Heijunka Improves the flow of a process to better IoT [3, 9, 18], big data, cloud computing,
meet customer demand and AI [3, 26]
SMED This means “single minute exchange of Simulation [3, 8], AM [3, 26]
dies”, to reduce setup time
298 J. Reyes et al.
29.3.4 Lean Supply Chain Management 4.0
On the one hand, SCM has been used to plan and control physical and information
flows and, internal and external logistics activities, and to address relationships and
shared processes with customers and suppliers [15].
On the other hand, the implementation of lean thinking in SCM has become a
priority. This is especially true because it leads to improved quality, reduced costs
and increased flexibility in companies [15, 16]. According to Veile [22], a model that
incorporates LM practices into the operational management of processes accepts
different implementation strategies in organizations to sustainably reduce waste.
The literature has identified two main strategies: (i) pull production and (ii) create
customer value [17, 23]. Consequently, entrepreneurs currently face the challenge
to integrate external partners into organizations to create value for customers using
technological support, such as IoT, cloud computing, big data and data analytics,
blockchain and simulation, and by tracking and localization like RFID, AM and
autonomous vehicles, among others. Figure 29.1 shows the I4.0 technologies as
supports for applying LM tools in SCM. This framework is named LSCM 4.0 and
illustrates a formalization that is summarized as four dimensions that interact with
the 10 lean waste types. It includes the main benefits for companies to improve their
performance in terms of lean tools and methodologies to support managers [2, 4, 16,
18–20, 23, 26], process optimization across the value chain with I4.0 technologies
[4, 5, 15, 20, 23, 25] and collaborative relationships across supply networks [1, 4, 7,
11, 14].
Fig. 29.1 LSCM 4.0 general construct

29.4 Conclusions
Despite the growing popularity of I4.0, to date few studies have compiled and
presented the scattered literature on how I4.0 relates to LM tools [9, 27]. This article
not only proposes a general LSCM 4.0 construct but provides an overview of current
findings and research opportunities. Therefore, it highlights the influence of key I4.0
technologies on an SC’s planning. In this context, technological innovation allows
SCs to continue with a constant process of continuous improvement because AI and
blockchain applied with enterprise social networks are being implemented for better
real-time visibility, predictive outage analyses, cost reduction and customer satis-
faction. For example, some companies are implementing internal social networking
sites like Facebook to promote employee engagement and to encourage knowledge
sharing to enhance collaboration and innovation [2]. One recommended practice
in all I4.0 implementation stages is to apply lean management as a requirement to
adopt I4.0 technologies, especially considering the organizational perspective and
implications on the value proposition for virtual goods/services and return flows
[16, 18].
The evolution of LM tools to reduce waste involves the importance of the role of
the human factor and culture, as defined in the TPS [24]. In the industrial sector, talent
is considered the most valuable asset to manage production. As one of the 10 lean
waste types, unused human talent due to the effect of organizational culture has an
impact on the level of success of implementing an LSCM system [3, 22]. Therefore,
training programs based on e-learning scenarios in relation to new technologies
and occupational safety become more important with increasing of man–machine
interactions [12]. This fact is aligned with the results of the other two supplier-
related LM operational constructs, JIT delivery and developing suppliers, which
reveal a higher level of interaction between their practices and the same set of I4.0
technologies [22]. The results of this study confirm that SCs’ digitization improves
the five explored lean principles. The findings of this study contribute to the relevant
existing literature to identify particular aspects of how SCs’ digitization positively
enhances the adoption of Kanban, JIT, VSM, TPM, and the 5S lean operations
practices. The present findings offer a valuable theoretical contribution to identify
ways to integrate I4.0 technologies. SC survival issues were not studied in-depth, but
are recognized as crucial issues after the spread of the COVID-19 pandemic [16].
The insertion of design processes for I4.0, such as planning based on simulation
data and digital twins, can help business sustainability and to develop a flexible
resilient organization. In practical terms, the benefits for performance improvement,
process optimization, and collaborative relationships across supply networks are
herein addressed. Financial viability constraints have been identified [8]. Finally, the
quantitative validation of the LSCM 4.0 conceptual proposal is a future research line.
300 J. Reyes et al.
Acknowledgements This research leading to these results received funding from the: European
Union H2020 Program under grant agreement No 958205 “Industrial Data Services for Quality
Control in Smart Manufacturing (i4Q)”; Spanish Ministry of Science, Innovation and Univer-
technologies enabling supply chains 4.0 (CADS4.0)”; and PhD grant from Technical of Ambato
University.
References
1. Frazzon E, Tortorella GL, Dávalos R, Holtz T, Coelho L (2017) Simulation-based analysis of

a supplier-manufacturer relationship in lean supply chains. Int J Lean Six Sigma 8(3):262–274
2. Haddud A, Khare A (2020) Digitalizing supply chains potential benefits and impact on lean
operations. Int J Lean Six Sigma 11(4):731–765
3. Pagliosa MM, Tortorella GL, Ferreira JCE (2019) Industry 4.0 and lean manufacturing: a
systematic literature review and future research directions. J Manuf Technol Manage 32:543–
569
4. Nascimento DLM, Alencastro V, Quelhas OLG, Caiado RGG, Garza-Reyes JA, Lona LR,
Tortorella G (2018) Exploring industry 4.0 technologies to enable circular economy practices
in a manufacturing context: a business model proposal. J Manuf Technol Manag 30(3):607–627
5. Kamble S, Gunasekaran A, Dhone NC (2020) Industry 4.0 and lean manufacturing practices
for sustainable organisational performance in Indian manufacturing companies. Int J Prod Res
58(5):1319–1337
6. Tortorella G, Sawhney R, Jurburg D, de Paula IC, Tlapa D, Thurer M (2021) Towards the
proposition of a lean automation framework: integrating industry 4.0 into lean production. J
Manuf Technol Manage 32(3):593–620
7. Ivanov D, Dolgui A (2020) OR-methods for coping with the ripple effect in supply chains
during COVID-19 pandemic: managerial insights and research implications. Int J Prod Econ
232(1):107921
8. Zhu X-Y, Zhang H, Jiang Z-G (2019) Application of green-modified value stream mapping
to integrate and implement lean and green practices: a case study. Int J Comput Integr Manuf
33:716–731
9. Fortuny-Santos J, López PR-de-A, Luján-Blanco I, Chen P-K () Assessing the synergies
between lean manufacturing and industry 4.0. Dir y Organ 71:71–86
10. Garay-Rondero CL, Martinez-Flores JL, Smith NR., Morales SOC, Aldrette-Malacara A (2019)
Digital supply chain model in industry 4.0. J Manuf Technol Manage 31(5):887–933
11. Craighead CW, Ketchen Jr DJ, Darby JL (2020) Pandemics and supply chain management
research: toward a theoretical toolbox*. Decis Sci 51(4):838–866
12. Cadavid JPU, Lamouri S, Grabot B, Pellerin R, Fortin A (2020) Machine learning applied in
production planning and control: a state-of-the-art in the era of industry 4.0. J Intell Manuf
31(6):1531–1558
13. Denyer D, Tranfield D (2009) Producing a systematic review. In: The sage handbook of
organizational research methods. Sage Publications Ltd, Thousand Oaks, CA, pp 671–689
14. van Remko VH (2020) Research opportunities for a more resilient post-COVID-19 supply
chain—closing the gap between research findings and industry practice. Int J Oper Prod.
Manage 40(4):341–355
15. Núñez-Merino M, Maqueira-Marín JM, Moyano-Fuentes J, Martínez-Jurado PJ (2020) Infor-
mation and digital technologies of industry 4.0 and lean supply chain management: a systematic
literature review. Int J Prod Res 58(16):5034–5061
16. Ivanov D (2022) Viable supply chain model: integrating agility, resilience and sustainability
perspectives—lessons from and thinking beyond the COVID-19 pandemic. Ann Oper Res
319:1411–1431
17. Sanders A, Elangeswaran C, Wulfsberg J (2016) Industry 4.0 implies lean manufacturing:
research activities in industry 4.0 function as enablers for lean manufacturing. J Ind Eng Manage
9(3):811–833
18. Buer S-V, Strandhagen JO, Chan FTS (2018) The link between industry 4.0 and lean manu-
facturing: mapping current research and establishing a research agenda. Int J Prod Res
56(8):2924–2940
19. Ben-Daya M, Hassini E, Bahroun Z (2019) Internet of things and supply chain management:
a literature review. Int J Prod Res 57(15–16):4719–4742
20. Novais L, Maqueira JM, Ortiz Á, Bruque S (2019) Strategic simulation models as a new
methodological approach: an application to information technologies integration, lean/just-in-
time and lead-time. Cent Eur. J. Oper Res 29:1185–1218
21. Ejsmont K, Gładysz B (2020) Lean industry 4.0—wastes versus technology framework. In: The
10th international conference on engineering, project, and production management. Springer,
ISBN 978-981-15-1909-3, pp 537–546
22. Veile JW, Kiel D, Müller JM, Voigt KI (2019) Lessons learned from industry 4.0 implementation
in the German manufacturing industry. J Manuf Technol Manage 31(5):977–997
23. Rossini M, Portioli A (2018) Supply chain planning: a quantitative comparison between lean
and info-sharing models. Prod Manuf Res 6(1):264–283
24. Khorasani ST, Cross J, Maghazei O (2020) Lean supply chain management in healthcare: a
systematic review and meta-study. Int J Lean Six Sigma 11(1):1–34
25. Gupta S, Modgil S, Gunasekaran A (2020) Big data in lean six sigma: a review and further
research directions. Int J Prod Res 58(3):947–969
26. Mayr A, Weigelt M, Kühl A, Grimm S, Erll A, Potzel M, Franke J (2018) Lean 4.0-a conceptual
conjunction of lean management and industry 4.0. Procedia CIRP 72:622–628
27. Valamede LS, Akkari ACS (2020) Lean 4.0: a new holistic approach for the integration of lean
manufacturing tools and digital technologies. Int J Math Eng Manage Sci 5(5):851–868
Chapter 30
Enablers and Barriers to Industry 4.0
Implementation
Blanca Guerrero, Josefa Mula , and Guillermina Tormo
Abstract The objective of this article is to study the main factors that enable
transformation toward Industry 4.0 (I4.0), along with the barriers involved in
its implementation. It is a first step toward a more wide and empirical study on
I4.0 development in selected Spanish companies. Here a detailed analysis of 23
antecedent works is carried out. It then proposes a classification of the general
dimensions to which both enablers and barriers refer: strategical, managerial and
organizational, technological and sustainability. This study contributes to the
literature analysis about implementing I4.0 by presenting both the main barriers and
enablers to implement this structure.
Keywords Supply chain · Industry 4.0 · Sustainability · Barriers · Enablers
30.1 Introduction
Industry 4.0 (I4.0) is a concept that involves organizational and technological

changes, along with integrating the value chain and developing new business models
driven by customer needs and mass customization requirements, and enabled by the
integration of innovative information and communication technologies (ICT) [1].
Thus, I4.0 is a digital container of different technologies, principles, and manage-
ment systems that fosters these strategies: cost reduction, product customization,
B. Guerrero · J. Mula (B)

València, Valencia, Spain
B. Guerrero
e-mail: bguerrero@cigip.upv.es
G. Tormo
Centre of Business Management Research CEGEA, Universitat Politècnica de Valencia, Valencia,
Spain
e-mail: gtormo@upv.es
https://doi.org/10.1007/978-3-031-29382-5_30
304 B. Guerrero et al.
time-to-market reduction, flexible and agile production implementation, supply chain

development and manufacturing companies integration [2]. The main objectives of
implementing I4.0 implementation are: lowering the error rate [3]; improving lead
times (compliance with market conditions) [3]; improving efficiency [3, 4]; ensuring
reliable operations (e.g., less downtime) [3]; load balancing and stock reduction [4];
reducing environmental impacts [4]; cutting costs incurred when developing and
engineering new products [2]; cost savings from quality products [4, 5]; reducing
costs, e.g., human resources, inventory management, and operating costs [3, 6].
The objective of this paper is to identify the key enablers and barriers that are
found in I4.0 implementation processes. To do so, 23 related literature works were
analyzed. I4.0 enablers and barriers were classified in terms of strategical, managerial
and organizational, technological and sustainability issues. The reminder of the paper
is organized as follows. Section 30.2 presents the related works adopted as the basis
of this work. Section 30.3 presents and classifies the identified key enablers and
barriers for I4.0 implementation. Section 30.4 discusses the main findings of this
work. Finally, Sect. 30.4 provides the conclusions and further research lines.
30.2 Antecedent Works
Chiarini et al. [2] investigate I4.0 technologies (big data, digital supply chain, Internet
of Things (IoT), cloud, robotics, additive 3D, autonomous vehicles) adopted in Italy,
and how they foster specific manufacturing strategies: ICT integration; lean; servi-
tization, i.e., the ability to provide customers with value-added services with the
physical product [2]; supply chain integration, design-to-cost and green. In order to
implement the European factory of the future (FoF) and I4.0, Pessot et al. [7] iden-
tify the challenges, drivers and opportunities, which they group into four dimensions:
strategy, organization, management, and technology. Based on multinational compa-
nies (MNEs), Makris et al. [8] investigate the impact of big data, cloud computing,
and 3D printing on the supply chain 4.0. Ghadge et al. [4] evaluate the impact
of I4.0 implementation [9] on supply chains by considering potential drivers and
barriers according to organizational, legal and ethical, strategic and technological
business dimensions. Horváth and Szabó [3] study the driving forces and barriers
regarding I4.0 implementation in terms of human resources, financial resources and
profitability, market conditions and competitors, management expectations, produc-
tivity and efficiency, management reality, organizational and technological factors
for MNEs, and small- and medium-sized enterprises (SMEs). Pirola et al. [10] also
focus on SMEs’ digital readiness levels. Bosman et al. [11] investigate the role of
firm size, access to funds and industry type on decisions to invest in I4.0 technolo-
gies. In line with procurement 4.0 [12] and logistics [13], the impact of digitiza-
tion on organizational performance, I4.0 implementation principles, key barriers and
technologies are studied. Ivanov et al. [14] carry out a survey on I4.0 topics with
researchers in industrial engineering, operations management, operations research,
control and data science areas. I4.0 has been related to sustainable, green and circular
30 Enablers and Barriers to Industry 4.0 Implementation 305
economy supply chain practices. Yadav et al. [15] identify 28 sustainable supply chain
management (SSCM) challenges and 22 solution measures based on I4.0 and circular
economy, and they develop and test a framework to improve SSCM adoption in an
automotive case study. Medina-Serrano et al. [16] address collaborative and sustain-
ability practices for the supply chain design problem. Hong et al. [17] investigate
the impact of SSCM practices on supply chain dynamic capabilities and economic,
environmental and social performance, in Chinese enterprises. Khanzode et al. [18]
evaluate eight I4.0 barriers for sustainable production in Indian micro- and SMEs
by the DEMATEL technique. Kayikci [19] proposes a set of criteria to evaluate the
sustainability impact of digitization on logistics in terms of economy, environment,
and society dimensions. Regarding key enabling technologies for sustainable supply
chain 4.0, Ramirez-Peña et al. [20] highlight them in the shipbuilding sector, specif-
ically big data, cloud computing, blockchain, cybersecurity, artificial intelligence
and simulation. Kumar et al. [21] identify 21 barriers for implementing the green
lean six sigma (GLS) concept in the product development process of the Indian
automotive sector. Ghadimi et al. [6] identify enablers related to implement green
manufacturing practices in Irish SMEs. Kumar et al. [5] rank the different barriers
that impede a supply chain’s sustainable operations in the I4.0 and circular economy
context. Nascimento et al. [22] explore how I4.0 technologies are integrated with
circular economy practices, whereas Rajput and Singh [23] provide I4.0 enablers
and barriers by establishing a link between circular economy and I4.0 in designing
supply chains.
30.3 Enablers and Barriers
According to the general dimension to which they belong, the following subsections
list the different enablers of and barriers for I4.0 implementation.
30.3.1 Strategic
This dimension presents the drivers and challenges related to long-term I4.0
implementation (Table 30.1).
30.3.2 Managerial and Organizational
This dimension distinguishes between operational barriers and enablers (Table 30.2),
and those related to human resources (Table 30.3).
Table 30.1 Enablers and barriers for the strategic dimension

Enablers Literature support Barriers Literature support
Supplier commitment and [15] Inappropriate reverse [15]
involvement for logistics system
sustainability adoption
Adoption of advanced [15] Loss of return material in [15]
quality improvement transit
techniques
Building brand image [15] Lack of effective [15]
based on circular economy communication with
and I4.0 suppliers
Internal reconstruction [15] Unavailable effective [15]
ability framework for SSCM
adoption
Servitization [2] Complexity in the supply [15]
chain configuration
Supply chain integration [2] Using outdated auditing [15]
standards
Sustainable resource [17] Ineffective performance [15]
management measurement system
Stakeholder management [6] Inappropriate execution of [15]
flexibility sustainability practices
To have a well-defined and [7] Lack of access to the [18]
overall I4.0 strategy with market
clear goals and benefits
To identify key roles and [7] Workforce’s non-readiness [4, 18]
assign clear responsibilities
To introduce new [7] Supply chain resilience, [14]
third-party partners to ripple effect and risk
expand knowledge and to management
enhance connectivity
New business models and [3, 4] Insufficient strategy for [5]
value offers for enhanced integrating I4.0 and
competitiveness circular economy
Coordination and [6, 7, 9, 17, 23] Insufficient market [5]
collaboration among demand
supply chain members
(collaborative model)
Effective facility planning [6, 8, 9] Short-term goals [5]
(logistic facilities),
infrastructure building, and
standardization
Replicating sustainability [9] Ineffective performance [5]
adoption strategies framework
(continued)

Agility [4] Improper defining and [21]
analysis toward green lean
six sigma
Lack of a clear digital [4, 7]
vision/strategy
Need to perform gradual [7]
and incremental changes
To underestimate data [7]
security vulnerabilities
To assess disruption in [7]
business models
Lack of proper common [3]
thinking
Lack of research and [4]
development
Supplier unwilling to [21]
transform toward GLS and
poor raw material quality
30.3.3 Technological
The technological dimension indicates the different technologies that lead to I4.0,
as well as the problems that may arise to achieve or implement these technologies
(Table 30.4).
30.3.4 Sustainability
Finally, the sustainability dimension distinguishes among economic, legal, political,

social, and environmental factors (Tables 30.5, 30.6, and 30.7).
30.4 Discussion and Conclusions
In summary, we have presented an exhaustive list of the main enablers and barriers for
I4.0 implementation extracted from 23 antecedent works. These have been classified
into five main categories: strategical, managerial, organizational, technological, and
sustainability. Here many analyzed papers take into account the automotive sector
[7, 12, 15, 21] or the electronic and automatic industry [2, 3, 7, 16, 17]. The number
of studies carried out in India is highlighted [4, 5, 15, 18, 21, 23]. Along within
Table 30.2 Enablers and barriers for the operational dimension

Strong interdepartmental [7, 15] Poor management [15]
ICT linkage system commitment to adopt
sustainability
To implement lean [2] Conflict between product [15]
management and supply sustainability policy and
chain agility principles free trade provisions
To improve green logistics [5] Non-uniform alignment of [15]
sustainability,
organizational goal, and
customer expectation
Supply chain risk [17] Ineffective linkage of [15]
management and sustainability with an
continuity existing process structure
To adopt FoF solutions for [7] Inappropriate part [21]
enriching customer service standardization and
and experienced scheduling
management
To leverage new product [7] Ineffective time [21]
introduction rates by management
digitizing the product
portfolio
To continuously monitor [3] Inappropriate identification [21]
company performance of activities/areas to be
“leaned and greened” and
unreliable “data collection
and retrieval system”
To follow market trends [3] Lack of conscious [3]
planning: defining goals,
steps, and needed resources
Innovation ability [17] Inadequate organizational [3]
structure and process
organization
To adopt advanced [15, 23] Lack of a unified [3]
predictive maintenance communication protocol
systems and recovery
To leverage digital [7] Complex network systems [4]
technologies to improve
flexibility toward mass
customization
Europe, the contribution of Italy stands out in some studies [2, 7, 10, 14]. For Spanish
studies, readers are referred to [16, 20]. A considerable number of articles focuses on
sustainability [5, 6, 15–23]. Among the general dimensions identified in this review,
we highlight certain enablers and barriers that are the most representative insofar as
the number of times they appear. In the strategical dimension, as enablers we find
Table 30.3 Enablers and barriers for the human resources dimension
Knowledge acquisition [17] Ineffective employee [3, 5, 15, 21]
and absorptive capacity training for sustainability,
lack of skilled workforce
Social network [17] Resistance of culture [4, 5, 15]
relationship ability change, lack of digital
culture
To reduce human work [3, 4] Lack of effective [3, 4, 15]
and allocate the employee engagement and
workforce to other areas empowerment. Resistance
to change by employees
and middle management
To find/hire competent [14] Human factors not [15, 21]
staff considered. Poor quality
of human resources
Workers’ reskilling [7, 14] To evaluate the impact of [7]
disruption of technologies
on how people work
To invest in [7] Longer learning time [3]
comprehensive training
schemes by integrating
different means and
methods
Demand for greater [3] Lack of necessary [3, 7]
control talented/skilled people
and leadership from top
management
Lack of understanding the [4, 5, 14, 15, 18, 21]
importance of I4.0 at top
management levels
Poor team management [3, 4, 15, 21]
and lack of cross function
team. Lack of effective
inter-departmental
communication.
Contradictory interest
the collaborative model [6, 7, 9, 17, 23] and logistics facilities [6, 8, 9]. For the
operational dimension, there are still no well-defined enablers or barriers, mainly
because they are not repeated in the studied articles. On human resources, we find
that the main barrier is related to lack of interest and training in I4.0 by managers
and employees [3–5, 14, 15, 18, 21]. In general, the technological dimension is,
on the whole, more enabling than disabling, and the most considered technologies
are big data and analytics, cloud computing and manufacturing [2, 8, 14, 20, 23],
and collaborative robots or AGV [2, 14, 20, 23]. In the sustainability dimension, we
highlight the high cost of implementing I4.0 as a barrier [3, 4, 14, 15, 18, 21, 23]
Table 30.4 Enablers and barriers for the technological dimension

To adopt smart factory [15, 23] Security/hacking [3, 14]
components concerns, unsafe data
storage systems
To integrate digital and [14, 15, 23] Incompetent technology [21]
physical systems and inferior facilities for
manufacturing
To adopt advanced [15] Inappropriate [21]
machine learning communication system
algorithms and lCT enablement
To digitize supply chain [2, 15, 18] To organize to implement [7]
activities and embed data analyses
across company business
and functions
Cybersecurity [14, 18, 20] Lack of information about [7]
potentialities of newer
FoF technologies
Data capturing, sensors, [14, 23] Low level of importance [7]
monitoring and control for and utilization and
humans, products, solutions for the
machines, and equipment human–machine
interaction
Big data and analytics [2, 8, 14, 20, 23] Lack of back-end systems [3]
for integration
Cloud computing and [2, 8, 14, 20, 23] Lack of standards [3]
manufacturing including technology and
processes
Additive manufacturing [14, 20] The need for large [3]
amounts of storage
capacity
Horizontal/vertical digital [14, 20] Poor data quality and [4]
integration management
software/system
Artificial intelligence [14, 20] Poor resource [4, 5, 21]
Augmented reality [14, 20] /infrastructure quality or
utilization
Cyberphysical systems [14]
Virtual reality [11, 14]
Collaborative robots, [2, 14, 20, 23]
automated guided vehicles
(AGV)
3D printing [2, 8]
IoT [2, 20]
Visual computing [23]
(continued)

Self-configure and routable [23]
Process digitization and [23]
automation
Simulation [20]
System integration of all [2, 23]
the production process
elements and
interoperability
Blockchain [20, 23]
Semantic interoperability [23]
Reliability, scalability, [23]
modularity, quality of
service, flexibility, and
value of networks of ICT
Table 30.5 Enablers and barriers for the economical dimension

Enablers Literature Barriers Literature support
support
To understand [15] High cost of sustainability [3, 4, 14, 15, 18, 21, 23]
socioeconomic adoption/difficulty to access
benefits credit/lack of access to capital
for I4.0
Design-to-cost [2] Lack of available resources [15]
techniques
Product quality [4, 5] Strong perception of low [15]
economic returns
Functional service [23] Inappropriate SSCM cost [15]
economy estimation
To invest in [6] High disposal costs [15]
innovation and
technology
Pressure/competition [3, 6] No significant perceived [4, 14]
from market competitive benefits or unclear
economic benefits
Risk of misinvestment or setting [5, 7]
the proper amount of investments
according to strategic objectives
Shortcomings in tendering [3]
systems
Long evaluation period for [3]
tenders
Table 30.6 Enablers and barriers for the political, legal, and social dimension
Improved employees and [5] Social security of [18]
community health employees/fear of job loss
To adopt safety standards [15] Unsupportive culture and [21]
lack of motivation and
encouragement
Laws and policy [23] Competition and [21]
uncertainty
Government promotion [6] Low level of education, [7]
and regulation high average age of
employees and managers
Unavailability of [15]
sustainability standard and
regulations
Lack of government policy [4, 5, 14, 18]
frameworks/support
Insufficient legislation and [5]
control
Lack of standardization, [7]
norms and certification for
FoF
Legal issues [4]
Data privacy and security [4]
issues
Customers not involved in [21]
greening, quality program,
and customer unawareness
of GLS
among the economic factors. Of the legal factors, lack of a political framework and
support for this initiative are the main barriers [4, 5, 14, 18]. Finally, no barriers stand
out in the environmental factors, but the main enablers are waste recovery, reducing,
monitoring, and controlling pollution [5, 6, 17, 23]. Further research is oriented
to the analysis of these enablers or barriers and about their use or exploitation for
implementing I4.0 solutions in real-world companies. Thus, a forthcoming work will
provide a new empirical study into the current trends on implementations of I4.0 in
Spanish textile and automobile manufacturing companies by assessing the different
level of effect of each of these variables and exploring the relation between them.
Table 30.7 Enablers and barriers for the environmental dimension

Rewards and incentives for [5, 15] Lack of awareness of [15]
greener activities sustainable standards for
raw materials
To adopt 6 R’s in the [5, 15] Design complexity for [15]
organization and industrial reducing energy use
ecology initiatives
To educate customers for [15] Unoptimized and [21]
recycling practices and non-green material
suppliers to use recyclable management and logistics
materials
Environmental product [15] Lack of waste [5]
design and lifecycle management
analysis
To implement green [2, 6]
manufacturing and green
logistics principles
To use materials as energy, [5, 23]
recover energy and reduce
emissions
Resource circularity [5, 17]
Waste recovery, reduction, [5, 6, 17, 23]
and pollution monitoring
and control
Process design for resource [5, 7]
and energy efficiency in
operations management
To leverage environmental [7]
and social trends for the
shift toward FoF beyond
technological and business
ones
To persist in reducing the [7]
use of traditional energy
sources in favor of
renewable ones
Acknowledgements This work was supported by the Spanish Ministry of Science, Innovation,
and Universities project entitled “Optimisation of zero-defects production technologies enabling
supply chains 4.0 (CADS4.0)” (RTI2018-101344-B-I00).
References
1. Nosalska K, Pi˛atek ZM, Mazurek G, Rz˛adca R (2020) Industry 4.0: coherent definition frame-
work with technological and organizational interdependencies. J Manuf Technol Manage
31(5):837–862
2. Chiarini A, Belvedere V, Grando A (2020) Industry 4.0 strategies and technological develop-
ments. An exploratory research from Italian manufacturing companies. Prod Planning Control
31(16):1385–1398
3. Horváth D, Szabó RZ (2019) Driving forces and barriers of industry 4.0: do multinational and
small and medium-sized companies have equal opportunities? Technol Forecast Soc Change
146:119–132
4. Ghadge A, Er Kara M, Moradlou H, Goswami M (2020) The impact of industry 4.0
implementation on supply chains. J Manuf. Technol. Manage 31(4):669–686
5. Kumar P, Singh RK, Kumar V (2021) Managing supply chains for sustainable operations in
the era of industry 4.0 and circular economy: analysis of barriers. Resour Conserv Recycl
164:105215
6. Ghadimi P, O’Neill S, Wang C, Sutherland JW (2020) Analysis of enablers on the successful
implementation of green manufacturing for Irish SMEs. J Manuf Technol Manage 32:85–109
7. Pessot E, Zangiacomi A, Battistella C, Rocchi V, Sala A, Sacco M (2021) What matters in
implementing the factory of the future: insights from a survey in European manufacturing
regions. J Manuf Technol Manage 32(3):795–819
8. Makris D, Hansen ZNL, Khan O (2019) Adapting to supply chain 4.0: an explorative study of
multinational companies. Supply Chain Forum Int J 20(2):116–131
9. Veile JW, Kiel D, Müller JM, Voigt K-I (2019) Lessons learned from industry 4.0 implemen-
tation in the German manufacturing industry. J Manuf Technol Manage 31(5):977–997
10. Pirola F, Cimini C, Pinto R (2019) Digital readiness assessment of Italian SMEs: a case-study
research. J Manuf Technol Manag 31(5):1045–1083
11. Bosman L, Hartman N, Sutherland J (2019) How manufacturing firm characteristics can influ-
ence decision making for investing in industry 4.0 technologies. J Manuf Technol Manage
31:1117–1141
12. Bienhaus F, Haddud A (2018) Procurement 4.0: factors influencing the digitisation of
procurement and supply chains. Bus Process Manage J. 24(4):965–984
13. Hermann M, Bücker I, Otto B (2020) Industrie 4.0 process transformation: findings from a
case study in automotive logistics. J Manuf Technol Manage 31(5):935–953
14. Ivanov D, Tang CS, Dolgui A, Battini D, Das A (2021) Researchers’ perspectives on industry
4.0: multi-disciplinary analysis and opportunities for operations management. Int J Prod Res
59(7):2055–2078
15. Yadav G, Luthra S, Jakhar SK, Mangla SK, Rai DP (2020) A framework to overcome sustainable
supply chain challenges through solution measures of industry 4.0 and circular economy: an
automotive case. J Clean. Prod 254:120112
16. Medina-Serrano R, Gonzalez R, Gasco J, Llopis J (2020) Collaborative and sustainable supply
chain practices: a case study. J Enterprising Communities People Place Glob Econ 14(1):3–21
17. Hong J, Zhang Y, Ding M (2018) Sustainable supply chain management practices, supply chain
dynamic capabilities, and enterprise performance. J Cleaner Prod 172:3508–3519
18. Khanzode AG, Sarma PRS, Mangla SK, Yuan H (2021) Modeling the industry 4.0 adoption
for sustainable production in micro, small & medium enterprises. J Cleaner Prod 279:123489
19. Kayikci Y (2018) Sustainability impact of digitization in logistics. Procedia Manuf 21:782–789
20. Ramirez-Peña M, Fraga FJA, Salguero J, Batista M (2020) Assessing sustainability in the
shipbuilding supply chain 4.0: a systematic review. Sustainability Switzerland 12(16):6373
21. Kumar S, Luthra S, Govindan K, Kumar N, Haleem A (2016) Barriers in green lean six sigma
product development process: an ISM approach. Prod Plan Control 27(7–8):604–620
22. Nascimento DLM, Alencastro V, Quelhas OLG, Caiado RGG, Garza-Reyes JA, Lona LR,
Tortorella G (2018) Exploring industry 4.0 technologies to enable circular economy practices in
a manufacturing context: a business model proposal. J Manuf Technol Manage 30(3):607–627
23. Rajput S, Singh SP (2019) Connecting circular economy and industry 4.0. Int J Inf Manage
49(C):98–113
Chapter 31
Blockchain Impact on Supply Chain
Performance
Jesús Morcillo-Bellido , Lucía Romero Fernández-Cuartero ,

and Jesús Morcillo-García
Abstract Blockchain has become more relevant in the efficient and secure manage-
ment of the supply chain, due to the grown importance that companies give to ensuring
the transparency and traceability of all the movements made in each of the links of
the supply chain. In this article, a study of the blockchain practices carried out by
different companies from various sectors, come to infer that traceability, security,
and cost reduction are the main objectives when companies decide to tackle projects
of this nature. Another issue that can be deduced is that companies generally look
for partners specialized in this sector to help them implement these projects, thus
sharing both the knowledge generated and the development costs of the projects.
Keywords Blockchain · Supply chain management · Supply chain traceability
31.1 Introduction
The so-called fourth industrial revolution [1, 2] refers to the application of new
technologies to businesses transforming the way in which products and services
are processed, distributed, and commercialized [3]. Among the technologies that
allow this transformation process and whose impact has been growing exponentially
J. Morcillo-Bellido (B)
Área de Ingeniería de Organización. Escuela Politécnica Superior, Universidad Carlos III.
Avenida de La Universidad, 30. Leganés, 28911 Madrid, Spain
L. Romero Fernández-Cuartero
Máster Universitario en Ingeniería Industrial. ETSII. Universidad Politécnica de Madrid, Calle de
José Gutiérrez Abascal, 2, 28006 Madrid, Spain
e-mail: luciaromerofc@gmail.com
J. Morcillo-García
Facultad de Ciencias Económicas Y Empresariales. Departamento de Organización de Empresas,
Universidad Nacional de Educación a Distancia (UNED), Senda del Rey, 11, 28040 Madrid, Spain
e-mail: jmorcillo@cee.uned.es
https://doi.org/10.1007/978-3-031-29382-5_31
318 J. Morcillo-Bellido et al.
in recent years, are the Internet of Things, big data, artificial intelligence, additive
manufacturing, machine learning, and blockchain networks [2].
The popularization of the blockchain concept arose in 2008 when a group of
people, anonymously and under the pseudonym of Satoshi Nakamoto, published the
article “A peer-to-peer electronic cash system” [4] where they introduced the concept
of virtual currencies or cryptocurrencies. The article mentions the three fundamental
aspects on which these cryptocurrencies are based: the blockchain platform, the
operating mechanism of these platforms, and the cryptocurrency itself (bitcoin). The
name “blockchain” references the way the information is stored in the network, as
a chain of blocks. A block is a data structure composed by a list of transactions
that have been created by “peers” of the blockchain network, modifying the state of
the blockchain [5]. Once the validation process is performed, the validated data is
grouped into blocks whose concatenation will form the chain. The validation process
of each block depends on the previous block, since a unique alphanumeric specific
code of the previous block (called hash value) is used in the process. Furthermore,
each of the blocks of the chain stores part of the information of the previous block:
its hash value. As a result, the chain of blocks is sequential, and any modification in
the information of a block would affect the blocks that are registered later, since their
hash values would have changed [4, 6]. A minimal modification in the information
of a block would generate a different hash value. Since blockchain platforms are
distributed networks, each user has an identical copy of the complete record [7], so
that in the event of being altered, one of the copies would be easily identified as it
would not coincide with the rest. On the other hand, the combination of consensus
protocols together with cryptographic methods make fraud and data manipulation
difficult [8], creating an immutable information record. These characteristics are
the main strength of this technology, allowing entities that lack of trust to reach a
consensus that leads to a safe and accurate information record [9].
The attraction of companies to blockchain technology has supported the develop-
ment of different platforms, thus emerging other types oriented to more reduced envi-
ronments where privacy plays an important role. Three large groups of blockchain
platforms are distinguished: public, private, and hybrid, which are mainly differenti-
ated by the existence or not of a permission to be part of them [10]. Public networks
are designed in a way that all the participants have access to visualize and edit the
information. All users may propose transactions, as well as take part in their valida-
tion through the aforementioned consensus protocols, among which the best known
are: “proof of work” used in bitcoin, and “proof of stake” used in Ethereum. Contrary
to public networks, private and hybrid platforms are designed to be used in certain
environments such as companies or supply chains [8], where the number of users is
restricted. Access limitation to the network is achieved by requiring a permission or
participation without which the user will not be able to be part of it [11]. Furthermore,
in these platforms, the identity of all the participants is known and different degrees
of access to information can be established according to the user.
Processes such as those that make up supply chains, are relevant application areas
for blockchain. In a supply chain, it is important to have a true and secure informa-
tion record, which allows efficient coordination between the different companies. A
31 Blockchain Impact on Supply Chain Performance 319
blockchain network is, in its most basic concept, a distributed record of information
characterized by the transparency, security and immutability of the stored data [12,
13]. Among the types of blockchain networks that exist, a hybrid-type platform is
ideal to meet the needs of a supply chain. Although initially blockchain platforms
were related to finance issues, in recent years, they have ceased to be exclusively
associated with the financial field, becoming a new paradigm in the organization and
storage of information [7] on a large scale. The inherent characteristics of blockchain
networks make this technology optimal for this new approach to use, which has
attracted the interest of numerous companies that have invested in the development
of new platforms [14]. Blockchain platforms are characterized by being decentralized
distributed networks, which implies that there is no central authority figure or user
who has to approve the information to be registered [8]. Instead, all users are able to
participate in the validation process to record information. This is done through the
so-called consensus protocols. These mechanisms make it possible to dispense with
intermediary agents [14], whose task was to ratify the validity of this information,
as affirm companies that use blockchain in their operations, such as Maersk [15].
Blockchain facilitates the validation and measurement of the effectiveness of the
supply chain. Companies that apply blockchain in their supply chain can monitor
shipments, transportation, quality, etc. All these provide security to the entire supply
chain, which in addition with eliminating middleman auditors, could increase the
efficiency and reduce costs [13]. Thus, incorporating blockchain will allow to ensure
that quality conditions during transportation and storage were maintained for that
product [11]. It is a key objective of this study to identify the impact of blockchain
in the supply chains of some companies, trying to identify the areas affected by the
implementation of blockchain as well as the expected or already obtained impacts.
31.2 Objectives and Methodology
The application of blockchain outside of finance is still, to some extent, experimental.

Some of the most promising non-financial applications are related to the supply chain
areas of sectors such as food, pharmaceutical, and agriculture [16]. The approach
made by the authors to the blockchain can be described as an attempt at “theory
creation” based on case studies [17, 18]. It is the priority objective of this study
to deepen the knowledge about the application of blockchain in the industry, for
which the cases of several companies will be studied, trying to identify: (i) if these
companies are applying blockchain mechanisms in their supply chains, (ii) what
types of mechanisms do they apply and, (iii) if possible, trying to start identifying
the influence of blockchain mechanisms on the results of the organizations where
they are applied, in terms of cost reduction and improvement customer service.
This document is the result of an inductive study of the cases of a sample of
companies. In a study developed by Hassini et al. [19], where a review of the litera-
ture published on the supply chain is made, it is concluded that the second most used
method is the case study. This contrasts somewhat with the general research trend,
where the case study is not always well recognized as a good research tool. In this,
research could be explained by the fact that the topics of blockchain and industry 4.0
represent a relatively new research space and researchers still need to further expand
the knowledge of the situations described in real cases to deepen their knowledge
of what is happening and better understand the factors involved, something that can
be achieved through the use of case studies [20]. Therefore, given the nature of the
topics to be investigated, it was decided to carry out a case study, a method that
according to Eisenhardt [21] is suitable for topics that have to do with strategic deci-
sions of business management. Yin [22] advises using the study of cases where the
boundaries between the context and the phenomenon to be observed are not evident.
The collection of information was carried out through a search of the information
published by these companies. Regarding the cases, care has been taken to maintain
the consistency of the data used. An attempt has been made to maintain coherence
between the data by comparing and avoiding distortion and bias in case management
[23].
31.3 Case Studies
31.3.1 Case 1: Walmart
Walmart has topped the Fortune Global 500 ranking in the last two years. Several
times this company has tried to implement new management systems in its supply
chain to improve the traceability of its products [24]. Faced with an incident, such as
the appearance of a bacterium (i.e., salmonella and listeria) that has harmful effects
on consumers, the origin of product could be traced immediately.
Walmart launched two blockchain pilot tests, for pork filets from China, and
mangoes from South America. They recorded all the information on the platform,
from raw materials origin to the final product in the market, linking each of the
elements with a QR code [25]. This information tracked various data, such as the
storing and transport temperature and humidity level. They also tracked fertilizers
and pesticides used for their growth, and the storage location of each unit along every
step of the supply chain. If any of the conditions measured did not meet the values
previously set in the platform (i.e., excessive temperature for conservation), an oper-
ator receives an alert. The access and subsequent analysis of all data helped them
improving their working protocols in transport, storage, and growth, preventing more
damaged units [25] and therefore reducing risks and cost involved. Also, the imple-
mentation of this project contributed to increase confidence in the products thanks
to the certificates that were permanently added to the network, which corroborated
different aspects related to quality, health, and safety inspections.
After the trial of this pilot project, Walmart concluded that the blockchain platform
facilitated and speeded up the information management and control. According to
Walmart, determining the origin of one of its products took about a week (through
e-mails and supplier contacts) whereas after the implementation of the blockchain
platform, this time was drastically reduced to a few seconds [24]. This is due to the
transparency offered by the platform to all the supply chain members.
In 2017, Walmart, Nestlé, and Unilever collaborated with IBM to create a
blockchain platform specially designed for food supply chains. This platform is
called IBM Food Trust and is currently used by many companies, such as Carrefour,
Smithfield or Tesco [26]. By the end of 2018, Walmart used this platform to keep trace
of 25 products such as fruits, vegetables, dairy products, and meat, and announced
that for the end of 2019, they would require their fresh-food suppliers to use it as
well [24]. This allowed Walmart to keep record of every step of their products along
the supply chain before they were exposed in their shelves to end consumers.
31.3.2 Case 2: Maersk
In maritime trade, Maersk stands out as the world’s leading company in container
transport. This sector, especially for the companies that trade internationally, handles
a high volume of administrative work, in which numerous companies and interme-
diary agents converge (customs and port authorities, for example) in the process of
transporting the merchandise. Each of the containers that is transported is associated
with a large amount of documentation belonging to the different organizations that
participate in the supply chain. To achieve a secure storage of all this information,
Maersk created together with IBM a hybrid blockchain platform known as Trade-
lens, with the aim of reducing and eliminating possible fraud and errors in the process
(that lead to delays in deliveries), minimize transportation time, improve inventory
management, and reduce costs [27]. This platform has been implemented today in
numerous ports and customs around the world, in countries such as Canada, New
Zealand, and Brazil, as well as in the Spanish ports of Algeciras, Bilbao, Valencia,
and Barcelona [28]. The integration of Tradelens at these points makes it possible to
speed up the inspection procedures that are carried out on maritime transport, which
cause significant delays in the delivery times of the merchandise. Since information
is recorded in real time, customs inspectors will be able to see when the cargo ship
is expected to arrive so they can have most of the documents checked by then, being
only left to make the control´s procedures that need to be done in person.
A key factor this technology offers is the high level of security in the information
registered, preventing the modification or manipulation of the data stored. When a
user wants to access a certain document in the platform, cryptography verifies if
the document has been modified since it was added to the network. This is done by
a comparison of their hash values [29]: when the document is first uploaded, it is
assigned a unique combination of characters (the hash value) that will change if any
modification is made on the document. When it is later downloaded by another user,
the system will check if the hash value coincides with the one stored previously.
If it does not occur, a message will be displayed on the screen notifying the user
that the document has been modified. This verifying method brings transparency
and promotes confidence among the users of the supply chain, who will rely on
the information they are working with. Following up with the previous scenario,
the customs operators will trust the information related to the shipment they are
inspecting, without doubting if it is fraudulent.
In addition to the high security system these platforms incorporate, Tradelens
is designed in a way that only those users who participate in any of the shipment
process could access to information. Access varies depending on the role of the user
along the supply chain [30]. This way Tradelens offers a single shared view of the
information that is recorded while respecting the privacy and confidentiality of the
data, which will not be accessed by users who do not have the necessary permissions
granted [30, 31].
31.3.3 Case 3: Pfizer, AbbVie, and Genentech
The supply chain in the pharmaceutical industry is complex because of all the actors
that take part in it, from the raw materials producer to the end consumer. Traceability
of products is a key factor in this industry because of the health-related risks that are
involved. And yet, it is one of the main handicaps of this industry, being especially
relevant in the case of the withdrawal of a certain batch of a product [32, 33]. The
characteristics of blockchain technology make it an ideal solution to mitigate the
problems identified in the management of these supply chains [33, 34].
USA enacted the law Drug Supply Chain Security Act (DSCSA) in 2013 which
established the need to develop monitoring systems for prescription pharmaceu-
tical products during all phases of the supply chain [35]. In 2017, the MediLedger
project connected the different pharma supply chain members to explore the possible
use of blockchain technology for the compliance with the requirements established
in the DSCSA [36]. The main objective of the solution proposed by MediLedger
is to establish the equivalent of a standardized and certified barcode for manufac-
turers to enter data in the blockchain of the supply chain, where only authorized
companies can store and view the data [36, 37]. In 2019, the Food and Drug Admin-
istration (FDA) requested the preliminary results of the pilot projects that could test
the interconnectivity requirements developed in the DCSA law for the monitoring
and traceability of each pharmaceutical product [35–37]. MediLedger DSCSA Pilot
Project [36] preliminary results draw the following conclusions: (i) it is confirmed
that the blockchain technology complies with the requirements established by the
FDA, (ii) speed, transaction throughput and an appropriate cost can be accomplished
to meet the stakeholder needs, (iii) no confidential information is shared and privacy
is maintained, while ensuring the immutability of the transactions, (iv) it allows full
traceability of the product from its originating manufacturer to end customer, (v) the
implementation could be complex and would require a stabilization period, (vi) the
long-term success of this technology implies the long-term commitment of all the
supply chain members.
31.4 Conclusions
In the analysis carried out, the companies studied have used blockchain in different
ways. Walmart used blockchain to improve product traceability in order to ensure
food safety. According to data provided by the company, the result is reduction in
process time and costs. Maersk valued in blockchain that the company can provide
its customers with a comprehensive transport model. This enables full transparency
and traceability, which contribute to generate significant cost reductions for Maersk´s
customers and enables them to better their inventory management.
Regarding the implementation of the blockchain in the pharmaceutical industry,
companies are working within a platform organized by a sector association
(MediLedger) to achieve the objectives of traceability, confidentiality, security and
transparency established by the FDA. However, the pharma companies participating
in the project state that it is a highly complex project with a long implementation
period. As the preliminary results pointed out on the MediLedger DSCSA Pilot
Project [36], there are relevant challenges that must be overcome in the following
years to achieve a complete implementation.
In all the cases studied, there is a common requirement, and this is the need
to organize collaborated networks in order to implement blockchain standards. It
is mandatory to think about blockchain with a full integrated vision of the supply
chain. Interesting enough, blockchain will be most widely applicable by supply
chain companies of the same sector that share similar requirements. This collabora-
tion would imply a reduction in implementation and maintenance costs due to the
achievement of economies of scale, on top of increasing its application areas.
References
1. Lasi H, Fettke P, Kemper HG, Feld T, Hoffmann M (2014) Industry 4.0. Bus Inf Syst Eng
6(4):239–242
2. Bagnoli C, Dal Mas F, Massaro M (2019) The 4th industrial revolution: business models and
evidence from the field. Int J e-Serv Mobile Appl (IJESMA) 11(3):34–47. https://doi.org/10.
4018/IJESMA.2019070103
3. Schwab K (2016) La cuarta revolución industrial. World Economic Forum
4. Nakamoto S (2008) A peer-to-peer electronic cash system. Bitcoin 4
5. Wüst K, Gervais A (2018) Do you need a blockchain? In: 2018 crypto valley conference
on blockchain technology (CVCBT) IEEE, pp 45–54. https://doi.org/10.1109/CVCBT.2018.
00011
6. Narayanan A (2016) Bitcoin and cryptocurrency technologies: a comprehensive introduction
(Illustrated ed.). Princeton University Press
7. Swan M (2015) Blockchain: blueprint for a new economy, 1st edn. O’Reilly Media Inc,
Sebastopol, CA
8. Zheng Z, Xie S, Dai H (2017) An overview of blockchain technology: architecture, consensus,
and future trends. https://doi.org/10.1109/BigDataCongress.2017.85
9. Ernst & Young Global (2019) Three ways blockchain can help implement smart contracts.
https://www.ey.com/en_gl/tax/three-ways-blockchain-can-help-implement-smart-contracts
10. Buterin V (2015) On public and private blockchains. https://blog.ethereum.org/2015/08/07/on-

public-and-private-blockchains/
11. Jayachandran P (2017) The difference between public and private blockchain. https://www.
ibm.com/blogs/blockchain/2017/05/the-difference-between-public-and-private-blockchain/
12. Abeyratne S, Monfared R (2016) Blockchain ready manufacturing supply chain using
distributed ledger. Int J Res Eng Technol 05(09):1–10
13. Kshetri N (2018) 1 Blockchain’s roles in meeting key supply chain management objectives.
Int J Inf Manage 39:80–89. https://doi.org/10.1016/j.ijinfomgt.2017.12.005
14. Cachin C, Vukolic M (2017) Blockchain consensus protocols in the wild
15. Hackius N, Petersen M (2017) Blockchain in logistics and supply chain: trick or treat? Proceed-
ings of the Hamburg international conference of logistics (HICL), vol 23. Epubli GmbH, Berlin,
pp 3–18. https://doi.org/10.15480/882.1444
16. Bünger M (2017) Blockchain for industrial enterprises: hype, reality, obstacles and outlook.
Retrieved June, pp 705–716
17. Eisenhardt KM, Graebner ME (2007) Theory building from cases: opportunities and challenges.
Acad Manag J 50(1):25–32
18. Kshetri N (2016) Creation, deployment, diffusion and export of Sub-Saharan Africa-originated
information technology-related innovations. Int J Inf Manage 36(6):1274–1287. https://doi.org/
10.1016/j.ijinfomgt.2016.09.003
19. Hassini E, Surti C, Searcy C (2012) A literature review and a case study of sustainable supply
chains with a focus on metrics. Int J Prod Econ 140(1):69–82. https://doi.org/10.1016/j.ijpe.
2012.01.042
20. McCutcheon DM, Meredith JR (1993) Conducting case study research in operations manage-
ment. J Oper Manag 11(3):239–256. https://doi.org/10.1016/0272-6963(93)90002-7
21. Eisenhardt KM (1989) Building theories from case study research. Acad Manag Rev 14(4):532–
550. https://doi.org/10.2307/258557
22. Yin RK (1994) Discovering the future of the case study. Method in evaluation research. Eval
Pract 15(3):283–290. https://doi.org/10.1016/0886-1633(94)90023-X
23. Eppler MJ (2006) Managing information quality: increasing the value of information in
knowledge-intensive products and processes. Springer Science & Business Media. https://
doi.org/10.1007/3-540-32225-6
24. HyperLedger (2019) How Walmart brought unprecedented transparency to the food supply
chain with Hyperledger Fabric
25. Kamath R (2018) Food traceability on blockchain: Walmart’s Pork and mango pilots with IBM.
J Br Blockchain Assoc 1(1):1–12. https://doi.org/10.31585/jbba-1-1-(10)2018
26. IBM (2018) IBM food trust—fact sheet. https://newsroom.ibm.com/download/IBM+Food+
Trust+-+Ecosystem+Fact+Sheet+Dec+2018.pdf
27. IBM (2018) Maersk and IBM introduce TradeLens blockchain shipping solution
28. Tradelens: The tradelens ecosystem. https://s3.us.cloud-object-storage.appdomain.cloud/tradel
ens-web-assets/TradeLens-Ecosystem.pdf
29. Tradelens (2018) Tradelens—shipment manager. https://docs.tradelens.com/how_to/shi
pment_manager/
30. Tradelens (2020) TradeLens data sharing specification: data sharing model. https://docs.tradel
ens.com/reference/DSS_Data_Sharing_Model_V4.0.pdf
31. Valenciaport (2018) Valenciaport se integra en el proyecto TradeLens, una plataforma
tecnológica de ámbito internacional desarrollada por IBM y Maersk
32. Queiroz MM, Telles R, Bonilla SH (2019) Blockchain and supply chain management integra-
tion: a systematic review of the literature. Supply Chain Manag 25(2):241–254. https://doi.org/
10.1108/SCM-03-2018-0143
33. Schöner MM, Kourouklis D, Sandner P, Gonzalez E, Förster J (2017) Blockchain technology
in the pharmaceutical industry. Frankfurt School Blockchain Center, Frankfurt, Germany
34. Angeles R (2018) Blockchain-based healthcare: three successful proof-of-concept pilots worth
considering. J Int Technol Inf Manage 27(3):47–83
35. U.S Food & Drug Administration.: Drug supply chain security act (DSCSA). https://www.fda.
gov/drugs/drug-supply-chain-integrity/drug-supply-chain-security-act-dscsa
36. MediLedger DSCSA Pilot Project (2020) MediLedger DSCSA pilot project
37. Alles M, Gray GL (2020) The first mile problem: deriving an endogenous demand for auditing
in blockchain-based business processes. Int J Account Inf Syst 38:100465. https://doi.org/10.
1016/j.accinf.2020.100465
Chapter 32
Proposal of a Methodology
and Associated Techniques for the Design
and Management of the Global Supply
Chain Operations Strategy According
to a Circular Economy Criterion
Darwin Aldas-Salazar , Manuel Díaz-Madroñero , and Josefa Mula
Abstract Given the demands and social awareness of the environment, today it is
important to link operations in a company with processes that allow sustainability
circles to be created. The main research objective of this article is to propose a
methodology for the design and management of operations strategies in global supply
chains with a circular economy and sustainability approach. To fulfill this objective,
a constructive research methodology is proposed in which each specific objective is
fulfilled with the sequential activities and tasks set out in the time planning scheme.
The future results to be obtained by the proposed methodology would be for industrial
companies belonging to global supply chains to optimize their operations strategies
in cost, quality, delivery, and flexibility terms in a sustainable manner.
Keywords Supply chain · Operations strategy · Global · Circular economy ·

Sustainability
32.1 Introduction
Despite being a mature discipline, supply chain (SC) management has significantly
evolved in the last two decades [1]. From 1980 onward, the supply chain management
D. Aldas-Salazar · M. Díaz-Madroñero (B) · J. Mula

Research Centre On Production Management and Engineering (CIGIP), Universitat Politècnica de
València, Alcoy, Spain
e-mail: fcodiama@cigip.upv.es
D. Aldas-Salazar
e-mail: darwinsaldas@uta.edu.ec
J. Mula
D. Aldas-Salazar
Universidad Técnica de Ambato, Ambato, Ecuador
https://doi.org/10.1007/978-3-031-29382-5_32
328 D. Aldas-Salazar et al.
term was developed to express the need to integrate key business processes from end
users to original suppliers [2]. SCs are evolving and changing in size, shape, and
configuration, and in the way they are coordinated, controlled, and managed [3]. SC
management has traditionally gone through three stages: creation, integration, and
globalization. It is characterized by the globalization of management in organiza-
tions in order to increase their competitive advantage and reduce costs by global (G)
outsourcing [4, 5]. Due to globalization and internationalization of businesses, the
global supply chains (GSC) term is adopted whose large-scale unbundling began
in the USA. However from the late 1980s to the start of this century, Asia’s total
trade has more than doubled by the efficient implementation of GSC and is the
world’s pioneer in this field today [6, 7]. The design of the production and logis-
tics system has to be aligned with the company’s operational strategy. Traditionally,
the operations strategy (OS)has focused on the decision areas of capacity, supply
network, process technology and organization and development [8, 9]. Critical deci-
sions made to approach the design and configuration of a GSC include: location of
supply sources; strategic role of plants, suppliers and warehouses; integration or frag-
mentation of production and logistics operations; service delivery strategies (supply
strategy, manufacturing strategy, purchasing strategy); global network of operations
(distribution network, manufacturing network, supplier network) [10].
Circular economy (CE) principles are based mainly on optimizing the use of avail-
able resources, materials and products, while maintaining their value in the economy
as a whole for as long as possible, minimizing waste generation, and focusing not
only on cost-effectiveness but also on environmental consequences [11, 12]. Indus-
trial companies seek to achieve positive outcomes across environmental, social and
economic dimensions where governmental aspects are key elements [13, 14]. The
aim of this article is to present the doctoral thesis research that is being carried out
to propose a methodology and the development of associated techniques that allow
the analysis and design of the OS of global supply chains (GSC) by an approach in
which the effective application of the CE concept prevails.
The remainder of the article is structured as follows. Section 2 presents the descrip-
tion of the problem. Section 3 shows the research methodology. Section 4 proposes
the working plan. Finally, Sect. 5 provides conclusions and further research.
32.2 Problem Description
The need to develop international markets with more diverse and sophisticated
customer requirements, and to implement global OS (offshoring and back shoring),
has made it necessary to set up and manage increasingly complex production and
logistics networks [15, 16]. This implies that GSC actors (e.g., distributors, manu-
facturers, suppliers, logistics operators, etc.) must develop new supply and OS to
reach performance targets in quality, flexibility, reliability, speed, and cost terms [8],
as well as agility, strategic sourcing, and efficient information exchange [17]. The
internationalization process is one of the most difficult decisions to make because
32 Proposal of a Methodology and Associated Techniques for the Design … 329
it inherently involves many risks associated with global economy [18–20], mainly
for companies with limited resources [21–24]. During this process, various factors,
such as collaboration along the SC, top management commitment, and the presence
of environmentally oriented policies or regulations, must be considered [25]. This is
why some choose to work with lean processes and others with agile processes [26].
However, no work has been identified to date that quantitatively addresses the
overall OS design beyond proposing existing methodologies to address some decision
areas in isolation [27]. Moreover, current models have not considered disruption risk
constraints, or the domino and resilience effects that are addressed in GSC [28, 29].
Therefore, the approach of this thesis is based on proposing a methodology supported
by quantitative models for the design process of the OS by considering all its decision
areas in an integrated manner.
The shift from the traditional SC to the sustainable supply chain has recently taken
place in different sectors via the CE concept [30–34], which encompasses complexity
theories, transaction cost economics, and information theories that have made this
adoption in the last decade essential for industries to continue in the global market-
place [35, 36]. The transformation toward business models based on CE affects the
OS because the management of new product flows and reuse and remanufacturing
processes requires making changes in areas that range from product development
to production and SC management by considering environmental performance and
human well-being [37–40]. Therefore, the application of CE principles to SC func-
tioning requires organizations making a systemic and holistic change [41, 42] by
redesigning their production and logistics networks with a new approach that inte-
grates the OS reformulation from all decision-making areas, as well as global ST
principles. Therefore, this thesis aims to incorporate the CE criterion, which is an
evolving model of economic and sustainable development, into these decision areas
[43]. Given the complexity of designing and managing GSC , there is a growing need
for not only quantitative models and tools but also for international manufacturing
systems frameworks, to help managers to design and manage their networks[22, 44]
in a CE context [45–47].
OS must demonstrate improved economic performance while promoting the
reduction the use of existing resources by focusing on both profitability and environ-
mental consequences. For all these reasons, the following general research question
is posed:
GRQI: What would be the most suitable conceptual and analytical method-
ology for the design and management of the global supply chain operations strategy
according to a CE approach?
The following specific research questions derive from this general research
question:
RQ1: Are companies aligned to GSC with a tendency to manage OS according
to sustainability (ST) and CE principles?
RQ2: What are the OSs with a CE and ST approach used by companies in GSC ?
RQ3: What conceptual and technical requirements are necessary to develop a

methodology for designing operations strategies in sustainable global supply
chains?
RQ4: What quantitative tools, models, and algorithms are needed to design and
validate the conceptual approach to OS in environmentally sustainable GSC ?
RQ5: What methods, algorithms, or combination of techniques are suitable to
generate an effective and efficient prototype decision tool?
RQ6: Are the proposed models and decision tools for OS design in GSC useful
for companies and do they address their need to integrate sustainable processes?
In order to answer the questions herein posed, the following overall objective is
proposed to be the developed within the doctoral dissertation:
OO: Propose and validate a methodology for the design and management of OSs
in GSC that focuses on CE and ST.
The specific objectives of the thesis proposal are to:
SO1: In the scientific and professional field, identify companies’ requirements
to improve the design and management of developing the OS in GSC by a CE
approach.
SO2: Formulate the conceptual and technical requirements for the new method-
ology to design the OS in GSC by a CE approach as a basic criterion.
SO3: Propose the models, indicators, tools/techniques, and resolution algorithms
needed to develop and validate the conceptual proposal for the design and
management of the OS in environmentally sustainable GSC .
SO4: Integrate the different models and algorithms making up the proposal into
a prototype of an effective and efficient decision-making tool.
SO5: Demonstrate the perceived usefulness for companies, and the objects of
applying the proposed models and decision tool.
32.3 Research Methodology
This doctoral thesis is based on constructivist research that is widely used in areas
such as finance [48], logistics [49], project management [50], or computer science
This research methodology focuses on generating solutions to concrete problems
by the creation of constructs according to the innovative constructivism concept
[51, 52]. A construct can be a new algorithm, a new mathematical model, or a new
conceptual model or framework. The solution-creating constructive process is based
on a set of phases that start by eliciting the problem to be addressed and continuing
with: (a) obtaining exhaustive knowledge about the problem to be solved through the
literature and case studies reviews; (b) constructing the solution to the problem via
an appropriate construct; (c) demonstrating the correct functioning of the generated
solution and the contribution of the solution; (d) examining the scope of applying
the obtained solution. Figure 32.1 shows the fulfillment of each research question
Fig. 32.1 Research methodology
and objective in the different methodology stages. Analytical models, effective and
informed decision support tools [53], systems thinking, hybrid models, simulation
models [54], fuzzy analytical hierarchy models [55], and non-linear programming,
among others, can be integrated for model building.
32.4 Work Plan
In order to fulfill the overall objective and the specific objectives set out, five stages
were developed while the doctoral thesis was being written (see Fig. 32.2), which
are divided into five chapters which are, in turn, divided into several activities (A),
broken down into several tasks (T).
32.5 Conclusions
In order to fulfill the objectives set out in this research work, the conceptual and
analytical modeling of GSC will be addressed to represent the different strategies and
scenarios to be considered. This methodology is expected to help managers of compa-
nies and business units of industrial groups to achieve greater effectiveness in imple-
mentations in terms of the effectiveness of both start-up times and in reducing non-
contemplated costs. The extension of conceptual models and working methods with
the support of optimization and simulation tools to help critical decision-making will
enable the incorporation of distributed and multilocalized production configurations,
restrictions of the supplier network, and the environmental, reactivity, scalability and
rapid adjustment of the existing system, efficient production, robustness and safety
criteria, which are all so necessary to guarantee the ST of industrial companies in
international environments. Lack of initiatives to fulfill industrial companies’ objec-
tives in terms of internationalization of operations renders it necessary to provide
the support generated by this thesis for the local OS the move toward the global
company both successfully and sustainably. Further research is oriented to develop,
Fig. 32.2 Outline of the doctoral thesis schedule
apply, and validate the proposed methodology for the design and management of the
GSC operations strategy under a CE criterion.
Acknowledgements This work was supported by the Spanish Ministry of Science, Innovation and
Universities project entitled “Optimisation of zero-defects production technologies enabling supply
chains 4.0 (CADS4.0)” (RTI2018-101344-B-I00).
References
1. Cannella S, Dominguez R, Framinan JM, Ponte B (2018) Evolving trends in supply chain
management: complexity, new technologies, and innovative methodological approaches.
Complexity 2018:7916849
2. Longoni A, Luzzini D, Guerci M (2018) Deploying environmental management across func-
tions: the relationship between green human resource management and green supply chain
management. J Bus Ethics 151:1081–1095
3. MacCarthy B, Constantin B, Jan O, Singh SJ, Xiande Z (2016) Supply chain evolution—theory,
concepts and science. Int J Oper Prod Manag 36:1696–1718
4. Ballou R (2007) The evolution and future of logistics and supply chain management. Eur Bus
Rev 19:332–348
5. Carter C, Liane E (2011) Sustainable supply chain management: evolution and future directions.
Int J Phys Distrib Logist Manag 41:46–62
6. Luck P (2019) Global supply chains, firm scope and vertical integration: evidence from China.
J Econ Geogr 19:173–198. https://doi.org/10.1093/jeg/lbx044
7. Alfaro L, Antràs P, Chor D, Conconi P (2019) Internalizing global value chains: a firm-level
analysis. J Polit Econ 127:508–559. https://doi.org/10.1086/700935
8. Slack N, Lewis M (2017) Operations strategy
9. Gong Y (2015) Global operations strategy
10. Monczka RM, Handfiled RB, Giunipero LC, Patterson JL (2009) Purchasing and supply chain
management
11. Xu Z, Elomri A, Pokharel S, Mutlu F (2019) The design of green supply chains under carbon
policies: a literature review of quantitative models. Sustain 11:1–20
12. Li G, Wu H, Sethi SP, Zhang X (2021) Contracting green product supply chains considering
marketing efforts in the circular economy era. Int J Prod Econ 234
13. Koberg E, Longoni A (2019) A systematic review of sustainable supply chain management in
global supply chains. J Clean Prod 207:1084–1098
14. Kumar S, Raut RD, Nayal K, Kraus S, Yadav VS, Narkhede BE (2021) To identify industry
4.0 and circular economy adoption barriers in the agriculture supply chain by using ISM-ANP.
J Clean Prod 293
15. Stentoft J, Olhager J, Heikkilä J, Thoms L (2016) Manufacturing backshoring: a systematic
literature review. Oper Manag Res 9
16. Johansson M, Olhager J (2018) Manufacturing relocation through offshoring and backshoring:
the case of Sweden. J Manuf Technol Manag 29
17. Kim M, Chai S (2017) The impact of supplier innovativeness, information sharing and strategic
sourcing on improving supply chain agility: global supply chain perspective. Int J Prod Econ
187:42–52
18. Tang O, Nurmaya Musa S (2011) Identifying risk issues and research advancements in supply
chain risk management. Int J Prod Econ 133
19. Revilla E, Saenz MJ (2017) The impact of risk management on the frequency of supply chain
disruptions: a configurational approach. Int J Oper Prod Manag 37
20. Varzandeh J, Farahbod K, Jake Zhu J (2016) Global logistics and supply chain risk management.
J Bus Behav Sci 28:124–130
21. Ferdows K (1997) Making the most of foreign factories. Harv Bus Rev 75
22. Mediavilla M, Errasti A (2010) Framework for assessing the current strategic factory role and
deploying an upgrading roadmap: an empirical study within a global operations network. In:
Proceedings of APMS 2010—international conference on advances in production management
systems
23. Deresky H (2017) International management managing across borders and cultures TexT and
cases. Pearson Educ, India, p 9
24. Yip G, Johnson G (2007) Transforming strategy
25. Giuffrida M, Mangiaracina R (2020) Green practices for global supply chains in diverse
industrial, geographical, and technological settings: a literature review and research agenda
26. Qi Y, Huo B, Wang Z, Yeung HYJ (2017) The impact of operations and supply chain strategies
on integration and performance. Int J Prod Econ 185
27. Behzadi G, O’Sullivan MJ, Olsen TL, Scrimgeour F, Zhang A (2017) Robust and resilient
strategies for managing supply disruptions in an agribusiness supply chain. Int J Prod Econ
191
28. Aydin G, Cattani K, Druehl C (2014) Global supply chain management. Bus Horiz 57:453–457
29. Um J, Han N (2020) Understanding the relationships between global supply chain risk and
supply chain resilience: the role of mitigating strategies. Supply Chain Manag 26
30. Gardas BB, Raut RD, Narkhede B (2019) Determinants of sustainable supply chain manage-
ment: a case study from the oil and gas supply chain. Sustain Prod Consum 17
31. Masoumi SM, Kazemi N, Abdul-Rashid SH (2019) Sustainable supply chain management in
the automotive industry: a process-oriented review
32. Rajeev A, Pati RK, Padhi SS (2019) Sustainable supply chain management in the chemical
industry: evolution, opportunities, and challenges
33. Ünal E, Urbinati A, Chiaroni D (2019) Managerial practices for designing circular economy
business models. J Manuf Technol Manag 30
34. Grafström J, Aasma S (2021) Breaking circular economy barriers
35. Junjun L, Yunting F, Qinghua Z, Joseph S (2018) Green supply chain management and the
circular economy: reviewing theory for advancement of both fields. Int J Phys Distrib Logist
Manag 48:794–817
36. Yadav G, Luthra S, Jakhar SK, Mangla SK, Rai DP (2020) A framework to overcome sustainable
supply chain challenges through solution measures of industry 4.0 and circular economy: an
automotive case. J Clean Prod 254:120112
37. Batista L, Bourlakis M, Liu Y, Smart P, Sohal A (2018) Supply chain operations for a circular
economy
38. Dev NK, Shankar R, Qaiser FH (2020) Industry 4.0 and circular economy: operational
excellence for sustainable reverse supply chain performance. Resour Conserv Recycl 153
39. Centobelli P, Cerchione R, Chiaroni D, Del Vecchio P, Urbinati A (2020) Designing business
models in circular economy: a systematic literature review and research agenda. Bus Strateg
Environ 29
40. Lopes de Sousa Jabbour AB, Rojas Luiz JV, Rojas Luiz O, Jabbour CJC, Ndubisi NO,
Caldeira de Oliveira JH, Junior FH (2019) Circular economy business models and operations
management. J Clean Prod 235
41. Paraskevopoulou C, Vlachos D (2020) A circular economy perspective for dairy supply chains.
Presented at the
42. Bressanelli G, Perona M, Saccani N (2019) Challenges in supply chain redesign for the circular
economy: a literature review and a multiple case study. Int J Prod Res 57:7395–7422
43. Bai C, Sarkis J, Yin F, Dou Y (2019) Sustainable supply chain flexibility and its relationship
to circular economy-target performance. Int J Prod Res 0:1–18
44. Vereecke A, Van Dierdonck R (2002) The strategic role of the plant: testing Ferdows’s model.
Int J Oper Prod Manag 22
45. Ormazabal M, Prieto-Sandoval V, Puga-Leal R, Jaca C (2018) Circular economy in Spanish
SMEs: challenges and opportunities. J Clean Prod 185
46. Prieto-Sandoval V, Jaca C, Ormazabal M (2018) Towards a consensus on the circular economy.
J Clean Prod 179
47. Suárez-Eiroa B, Fernández E, Méndez-Martínez G, Soto-Oñate D (2019) Operational principles
of circular economy for sustainable development: linking theory and practice
48. Kasanen E, Lukka K, Siitonen A (1993) The constructive approach in management accounting
research. J Manag Account Res 5:243–264
49. Lukka K (2003) The constructive research approach. Case Stud Res Logist Ser B
50. Oyegoke A (2011) The constructive research approach in project management research
51. Meredith J (1993) Theory building through conceptual methods. Int J Oper Prod Manag 13
52. Mediavilla M, Errasti A, Domingo R, Martínez S (2012) Value chain based framework for
assessing the Ferdows’ strategic plant role: an empirical study. In: IFIP advances in information
and communication technology
53. Fahimnia B, Sarkis J, Talluri S (2019) Editorial design and management of sustainable and
resilient supply chains. IEEE Trans Eng Manag 66:2–7
54. Rebs T, Brandenburg M, Seuring S (2019) System dynamics modeling for sustainable supply
chain management: a literature review and systems thinking approach. J Clean Prod 208:1265–
1280
55. Zimmer K, Fröhling M, Breun P, Schultmann F (2017) Assessing social risks of global supply
chains: a quantitative analytical approach and its application to supplier selection in the German
automotive industry. J Clean Prod 149:96–109
Chapter 33
Industry 4.0 Practices Applied in Pharma
Sector Supply Chain
Jesús Morcillo-Bellido and Ramón Merino-Fuentes
Abstract The term Industry 4.0 refers to a new industrial model concept based
on the digitization and monitoring of the certain operational processes included in
the business model. The objective of this study is to deepen the knowledge of real
business application of Industry 4.0 tools in a sample of large worldwide compa-
nies within the pharma sector. From the analysis carried out, it can be inferred that
companies in the pharma sector are already far away from applying different Industry
4.0 technologies as a way to increase process efficiency and traceability, especially
“Internet of Things” and “additive manufacturing”.
Keywords Industry 4.0 · Pharma sector · Industry 4.0 in pharma
33.1 Introduction
According to a study led by the well-known consulting firm McKinsey [1], the
technologies known as “Industry 4.0” are playing a decisive role in articulating the
response of company’s response to the situation emerged by the COVID-19 pandemic
in their supply chains. Studies identified that aspects such as greater transparency
in the supply chain, increased productivity, and agility in operations management to
respond to changes in demand are becoming more relevant. The term Industry 4.0
(I4.0) refers to a new industrial model based on the digitization and monitoring of
the operational processes [2–4]. Like previous industrial models, I4.0 emerges linked
to an industrial revolution, driven by the consumer’s needs and supported by tech-
nological and technical innovations [5]. There are several specific factors that lead
I4.0 development, the main one could be a change in the demand behavior according
J. Morcillo-Bellido (B)
Universidad Carlos III, Av. de La Universidad, 30, 28911 Madrid, Spain
R. Merino-Fuentes
GMS Management Solutions, S.L. Pl. Pablo Ruiz Picasso, 1, 28020 Madrid, Spain
e-mail: ramon.96.merino@gmail.com
https://doi.org/10.1007/978-3-031-29382-5_33
336 J. Morcillo-Bellido and R. Merino-Fuentes
to which the consumer is asking for greater specialization and personification on

products and services. In addition, as a consequence of the increase in labor costs
in the main manufacturing countries, many companies are starting to consider the
possibility of moving their production centers to consumption countries, thus being
closer to the end customer [6]. This movement implies the use of more advanced
technological means that make it possible to minimize the impact of labor on the cost
of products. Authors have considered as main pillars: Internet of Things, additive
manufacturing, big data, and artificial intelligence [7]. The I4.0 model pursues envi-
ronmental sustainability through the circular economy that could be the fifth pillar,
emphasizing the maximization of the recirculation of materials and energy, at the
same time that it tries to keep the products in their greatest range of use in the largest
possible time [8] trying to minimize the waste generated. In the case of the so-called
Internet of Things (IoT), it is based on a highly distributed network made up of a
large number of elements of information/communication technologies (like sensors
or actuators) which seeks to connect several devices without the need for cables. This
pillar plays an essential role in the new industrial model as it enables the digitization
of processes, offering a decentralized control model, and improving communication
and efficiency of company processes [2]. “Additive manufacturing (AM)” is a rather
new manufacturing methodology based on the addition and superposition of material
layer, as opposed to subtractive manufacturing that is based on the elimination of
material [9]. This new type of manufacturing is close to the objectives that drove the
birth of I4.0, since it allows obtaining products with a high degree of personaliza-
tion and moving the production centers close to the end customers. In addition, AM
offers the possibility of manufacturing on demand, providing products with greater
added value and better consumer expectations fulfillment [10]. The term “big data
(BD)” refers to the immense set of heterogeneous, unstructured, and dynamic data
that exists, and to the elements that perform the treatment of this complex informa-
tion base [11]. The last pillar studied by authors as part of I4.0 is called “artificial
intelligence (AI)”, which consists of the development of algorithms that offer greater
efficiency when solving problems or making decisions, as well as facilitating the reso-
lution of increasingly complex problems. It is a tool with great potential for obtaining
dynamic and more efficient predictive models that could already bring benefits to
the organizations [12]. Despite not being considered within this study scope, it is
worth it to say that experts include “blockchain” as part for I4.0, as a technology
that secures recordkeeping capabilities satisfactorily offering a trustworthy forensic
trail which is quite useful [13]. However, like in any disruptive change, there are a
number of barriers (Table 33.1).
33.2 Objectives and Methodology
This document is the result of an inductive study of different pharma company cases.
Given the nature of the topics to be investigated, it was decided to carry out a case
33 Industry 4.0 Practices Applied in Pharma Sector Supply Chain 337
Table 33.1 Industry 4.0 concept evolution

Years Authors Concept
2012 Boyes et al. [14] General Electric proposed a model for connecting industrial
machines sensors/actuators to Internet. Idea tries to connect
several industrial networks to generate value.
2013 Bidet-Mayer [15] Montebourg proposes the “Industrie 4.0” concept, with the
aim that French manufacturers invest in the renewal and
technological updating of their factories.
2014 Radziwon et al. [16] Ubiquitous Factory concept. Manufacturing model based
on: information transparency, autonomous control, and
sustainable manufacturing.
2016 Sanders et al. [17] Industry 4.0 as a driver to achieve intelligent manufacturing
plants through advanced communication and information
systems implementation, based on future technologies.
2017 Rymaszewska et al. [18] New customer relationship model. IoT allows unlocking the
potential of innovative systems of products and services on
a large scale.
2018 Vaidya et al. [19] Disruptive organization control of the entire value chain of
the product life cycle, in order to satisfy the increasingly
specific and individualized customer needs.
2019 Agostini and Filippini [20] Industrial revolution was based on the interconnection of all
the elements along the value chain and the creation of smart
grids that allow autonomous control.
2020 Machado et al. [21] The plan is to make use of technologies and getting business
processes fully integrated.
2021 Caiado et al. [22] Relationship between the most disruptive technologies and
production systems. Connection between intelligent
operations and supply chain management.
study, a method that according to Eisenhardt [23] is suitable for topics that have
to do with business management strategic decision [23]. Yin [24] advises using the
study of cases where the boundaries between the context and the phenomenon to
be observed are not evident [24]. The information collection was carried out from
information published in different information sources (web pages, reports, etc.).
A selection of the companies was made following the company’s relevance criteria
and considering the feasibility of access by the authors to relevant information. Six
leading companies in the worldwide pharmaceutical sector have been studied (Pfizer,
Roche, Novartis, Johnson & Johnson, Merck and Sanofi).
33.3 Case Study Analysis
33.3.1 Pharma Sector Introduction
Large pharmaceutical companies have currently business models based on mass

production, managing large batches of products in such a way that companies can
get some economies of scale. On the other hand, the development of genomic infor-
mation and the aging of the population, with its consequent increase in the number
of older people with chronic diseases, lead to a change in the demand behavior and
as consequent change on supply chain and operations management, which tends to
become more personalized [25].
33.3.2 Case Studies
Following it is analyzed several case practices related to I4.0 in a sample of six

important companies within worldwide pharma sector. A global search approach
has been carried out, since these practices are often identifiable as a centralized
strategy (Table 33.2).
33.4 Discussion and Conclusions
Analyzing the different practices included on the cases, it is possible to infer that
I4.0 is being broadly implemented at pharma industry; nevertheless, projects are
still at an initial implementation phase (most of them still could be considered at
pilot projects). It is possible to realize that all the sample’s companies are currently
working on several projects linked to Industry 4.0, and Table 33.3 shows the main
results initially planned as most relevant from I04 current projects.
Pharma companies are focused on the implementation of programs related to IoT
and AI, big data, and AM manufacturing. As described in Table 33.3, each company
applied I4.0 for different purposes, but in all cases projects are focused on operations
and supply chain improvement. Given that the group of companies included in the
sample are among the largest companies in terms of volume and, in practice, and
they are considered as the most innovative ones (both in scientific and management
terms) could be expected that in the coming years this group of companies will be
acting an important engine that extends I4.0 practices widely in the sector, serving
as reference model for many other companies at pharma industry and potentially in
other sectors with similar business requirements.
Table 33.2 Industry 4.0 projects identified

Company Industry 4.0-related projects
Pfizer Project with IBM to transform and improve the efficiency of diagnostic
tests and data capture for Parkinson’s treatment. A system of sensors
and mobile devices monitors the symptoms of the disease continuously
on real time [26]. In addition, greater efficiency in data acquisition
allows to shorten the clinical testing phase reducing the cost associated
time to market [27]. Still it could be considered a pilot, in which all the
processes are automated and monitored. The results show an
operations throughput time reduction, an improvement in the final
product quality, and an inventories’ level reduction [27]
Roche Company is managing an automated warehouse in Kaiseraugst,
Switzerland, based on a SAP environment, controlled by a material
flow and storage management system, and supported by a TCP/IP
(Ethernet) network that ensures an uninterrupted flow of information
between the sensors and management control system. The pharma
“cold chain” required by certain medicines is never interrupted, and
material flows become faster and more efficient [28]. Another example
is related to Warfarin control device (based on IoT) which allows the
monitoring of the blood coagulation level of patients and serves to
prevent potential heart attacks [29]. Moreover, this company developed
an alliance with Sensyne Health company to improve trial planning by
using AI models which identify potential patient populations, matching
information from existing patients [30]. This company also uses 3D
printing technology developed by MarketBot for an experiment with a
new range of drugs for the rheumatoid arthritis treatment [31]
Novartis It has developed Chimeric Antigen Receptor T cell therapy project,
based on artificial intelligence. It consists of extracting cancer cells
from the patient, genetically reprogramming them, and introducing
them back into the patient’s body, so that they can recognize specific
markers such as cancer cells and fight them more effectively [32].
Obtaining human tissue through 3D printing to use it in experimental
tests. The artificially obtained tissue that perfectly replicates the human
one makes unnecessary the use of animals or even the humans
involvement in this type of test [33]
Johnson & Johnson Through the use of sensors, devices, and protocols based on the IoT,
information collection is carried out both from patients and from
manufacturing processes in real time. The real-time “patient–doctor”
interconnection increases the degree of patient/consumer satisfaction.
On the other hand, the pharmaceutical company knows the state of its
processes and facilities on real time, solving inefficiencies or failures
faster and even predicting them before they happen [34]. Company has
developed an integration of all the necessary instruments in a surgical
intervention in a single device, obtained by 3D printing [35]
(continued)

Company Industry 4.0-related projects
Merck Sharp & Dohme It has applied Hadoop, a cloud-based tool, to find an efficient solution
to low yields quantities in vaccine manufacture. This solution based on
big data technology requires test-based information to be dumped and
the results should be analyzed, but resources and time are not
consumed in managing and storing data [36]. Project belongs to
Numerate platform, which is focused on the development of medicines
through algorithms and computing resources in the cloud. The benefits
achieved have been linked to increased development speed and
reduced associated costs [37]. This company has also applied
EXVIVE3DTM, an artificial human tissue obtained through additive
manufacturing whose allows to reduce the lead time of the research
processes for new drugs [38]
Sanofi This company manages a pilot of six “smart” production plants in the
USA, Canada, Ireland, France, China, and Brazil. These plants
combine various technical and technological initiatives, such as the
incorporation of collaborative robots to manage daily tasks. Plans are
fully monitored, by multitude of sensors distributed throughout all the
production lines that send information on their status, the data is
collected and the control of the entire plant is managed centrally. These
data are used to develop predictive models that help to avoid failures
and improve processes efficiency [39]
Table 33.3 Industry 4.0 pharma main practices and expected results
Company Main I4.0 practices Results
Pfizer IoT for diagnostic Throughput time reduction
Shorten clinical testing
Roche IoT for warehousing and Faster materials flow
distribution Improved trials planning
Novartis IA for cell therapy Less animals and human-based
Artificial tissue using AM experiments
Johnson & Johnson IoT for patients and AM Connection/integration
manufacturing process Patient’s needs prediction
Merck Big data application Development speed
Cost reduction
Sanofi AI and robots in factories Predictive models
Processes efficiency
References
1. Agrawal M, Dutta S, Kelly R, Millán I (2021) COVID-19: an inflection point for Industry 4.0.
McKinsey & Company
2. Rojko A (2017) Industry 4.0 concept: background and overview. Int J Interact Mobile Technol
11(5):77–90
3. Thahjono B, Esplugues C, Ares E, Pelaez G (2017) What does industry 4.0 mean to supply
chain?. Procedia Manuf (13):1175–1182
4. Fatorachian H, Kazemi H (2021) Impact of Industry 4.0 on supply chain performance. Prod,
Planning Control 32(1):63–81
5. Lasi H, Fetke P, Kemper H, Feld T, Hoffman M (2014) Industry 4.0. Bus Inform Syst Eng
6(4):239–242
6. Arbulu I, Vivek L, Mancini M, Alpesh P, Tonby O (2018) Industry 4.0: reinvigorating ASEAN
manufacturing for the future. In: Digital capability center. McKinsey & Company, Singapore
7. Baur C, Wee D (2015) Manufacturing’s next act. McKinsey & Company, Munich, p 6
8. Müller JM, Kiel D, Voigt KI (2018) What drives the implementation of industry 4.0? The role
of opportunities and challenges in the context of sustainability. Sustainability 10(1):247
9. Wohlers T, Gornet T (2014) History of additive manufacturing. Wohlers Rep 24:118
10. Mellor S, Hao L, Zhang D (2014) Additive manufacturing: a framework for implementation.
Int J Prod Econ 149:194–201
11. Demchenko Y, Membrey P (2014) Defining architecture components of the big data ecosystem.
In: International conference on collaboration technologies and systems. IEEE s.l., pp 104–112
12. McKinsey Analytics: global AI survey: AI proves its worth, but few scale impact McKinsey &
Company, http://www.mckinsey.com/home, Accessed 01 2020
13. Oracle: emerging technologies: driving financial and operational efficiency. Research
conducted in partnership with ESG (2020)
14. Boyes H, Hallaq B, Cunningham JM, Watson T (2018) The industrial internet of things (IIoT):
an analysis framework. Comput Ind 101:1–12
15. Bidet-Mayer T (2016) L’industrie du futur: une compétition mondiale. La Fabrique de
L’industrie s.l.
16. Radziwon A, Bilberg A, Bogers M, Madsen ES (2014) The smart factory: exploring adaptive
and flexible manufacturing solutions. Procedia Eng 69:1184–1190
17. Sanders A, Elangeswaran C, Wulfsberg JP (2016) Industry 4.0 implies lean manufacturing:
research activities in industry 4.0 function as enablers for lean manufacturing. J Ind Eng Manage
(JIEM) 9(3):811–833
18. Rymaszewska A, Helo P, Gunasekaran A (2017) IoT powered servitization of manufacturing—
an exploratory case study. Int J Prod Econ 192:92–105
19. Vaidya S, Ambad P, Hosle S (2018) Industry 4.0—a glimpse. Procedia Manufact (20):233–238
20. Agostini L, Filippini R (2019) Organizational and managerial challenges in the path toward
Industry 4.0. Eur J Innov Manage 1460–1060
21. Machado GC, Winroth MP, Ribeiro da Silva EHD (2020) Sustainable manufacturing in Industry
4.0: an emerging research agenda. Int J Prod Res 58(5):1462–1484
22. Caiado RGG, Scavarda LF, Gavião LO, Ivson P, de Mattos Nascimiento DL, Garza-Reyes
JA (2021) A fuzzy rule-based industry 4.0 maturity model for operations and supply chain
management. Int J Prod Econ (231):107883
23. Eisenhardt KM (1989) Building theories from case study research. Acad Manag Rev 14(4):532–
550
24. Yin RK (1994) Case study research: design and methods, applied social research. Methods Ser
5
25. Stegemann S (2016) The future of pharmaceutical manufacturing in the context of the scientific,
social, technological and economic evolution. Eur J Pharm Sci 90:8–13
26. Mullin E. Pfizer and IBM launch ambitious ‘Internet of things’ for Parkinson’s research. Forbes
Homepage, https://www.forbes.com/sites. Accessed 02 2020
27. Zimoch J (2016) Virtual clinical trials: Pfizer’s digital transformation of pharmaceutical R&D.
Technology and operations management. At Harvard Business School
28. Jonischkei B. Industry 4.0: Pfizer opens continuous manufacturing plant in Freiburg. healthcare
industry BW. https://www.gesundheitsindustrie-bw.de. Accessed 03 2020
29. Reimann H (2016) Advanced cold storage logistics for sensitive drugs. Roche s.l., Kaiseraugst
30. Steiner J. A connected future. PharmaTimes, http://www.pharmatimes.com. Accessed 04 2020
31. Smith S. 43 pharma companies using artificial intelligence in drug discovery. BenchSci. https://
blog.benchsci.com. Accessed 03 2020
32. Novartis Global Hompage, https://www.novartis.com: manufacturing CAR-T cell therapies:
The Novartis approach. Accessed 04 2020
33. O’Brien T. 3-D printing: scientists reinvent research tools. Novartis Global Hompage, https://
www.novartis.com. Accessed 03 2020
34. Bradbury D. 3 ways Johnson & Johnson is harnessing digital innovation to better deliver
medicine to you. Johnson & Johnson Hompegae, https://www.jnj.com. Accessed 04 2020
35. Brewster S. The power of 3-D printing: how this technology is blazing new medical frontiers.
Johnson & Johnson Hompegae, https://www.jnj.com. Accessed 04 2020
36. Henschen D. Merck optimizes manufacturing with big data analytics. InformationWeek
Homepage, https://www.informationweek.com. Accessed 05 2020
37. Numerate Homepage, http://www.numerate.com: numerate forms drug discovery collaboration
with Merck to utilize numerates in silico drug design technology. Accessed 03 2020
38. Molitch-Hou M. Organovo signs multi-year 3D bio printing deal with pharma giant Merck. 3D
printing industry homepage: https://3dprintingindustry.com/news/organovo-signs-multi-year-
3d-bioprinting-deal-with-pharma-giant-merck-47415/. Accessed 04 2021
39. Sanofi Homepage, https://www.sanofi.com: the fourth industrial revolution. Accessed 05 2020
Chapter 34
A Conceptual Framework
of a Blockchain Application
in a Manufacturing Supply Chain
Erick Ponce, Josefa Mula , and David Peidro
Abstract This paper presents a conceptual framework to apply blockchain tech-

nology to two implementation areas in a supply chain. It specifically intends to
provide product tracking information to all stakeholders in the product development,
assembly and subsequent delivery phases, manage information of supplied compo-
nents to perform the assemblies where it is included, and track injected components.
Information flow and exchange are supported by blockchain. This article contributes
to this emerging technology by providing an overview of blockchain and its appli-
cation to an industrial supply chain by presenting the real problems encountered
in the chain, examines the implications of information centralization, traceability,
and transparency, and identifies some potential challenges in the work performed
by collaborating companies in quality, receipt, and shipment of goods terms, among
others.
Keywords Blockchain · Manufacturing · Supply chain
34.1 Introduction
An industrial supply chain involves many actors. These systems currently present
several inefficiencies, which range from the traceability of the components making
up products to their delivery to end customers. Blockchain has become an allied
technology to solve such problems in supply chains and to also connect, relate, and
centralize scattered information [1]. This means that blockchain offers permanent and
E. Ponce · J. Mula · D. Peidro (B)

València, C/ Alarcón, 1, 03801 Alcoy, Alicante, Spain
e-mail: dapeipa@cigip.upv.es
E. Ponce
e-mail: erpon@epsa.upv.es
J. Mula
https://doi.org/10.1007/978-3-031-29382-5_34
344 E. Ponce et al.
immutable transactions and access to distributed data. In turn, this technology has the
potential to facilitate data exchange and to reduce fraud or adulteration opportunities,
which transmit more transparency and trust to consumers [2]. Blockchain can be
defined as a database that is distributed among different users and is cryptographically
protected and structured in transactional and mathematically related information
blocks. Its main use is framed within a system where several parties interact and do
not trust one another [3]. Readers are referred to [4–6] for blockchain reviews or
conceptual frameworks.
A traditional supply chain is driven by planning and communication, while future
demand is estimated based on the past and present. All information is sent to the
implicated parties, which expect to obtain details about it in time to respond to any
changes, delays, or errors related to the cost involved in its management. Among
blockchain characteristics, it is worth stressing data transparency based on the types
of participants’ access to each piece of information in the supply chain domain [7].
As privacy is an important property of any information system, and each company
handles certain information that cannot be shared, inherent tension appears between
privacy and transparency. Both are related and inherent to form part of a group of
collaborating companies, and therefore, trust is necessary for them to work with one
another.
The use of an enterprise resource planning (ERP) system allows each company
to more efficiently manage its resources (orders, internal logistics, invoicing, etc.).
Working with partner companies requires managing the supply chain at a level at
which participants need to visualize the transactions made during certain processes.
Blockchain can support the different actors involved in a supply chain by improving
processes and operations by more secure, transparent, and efficient transactions, and
by providing trust and reliability in all the transactions and information shared by
each network participant [8]. Implementing a blockchain in an industrial supply chain
involves understanding the behavior of intercompany relationships, and also what this
behavior is like as it is significantly influenced by all the relationships involved in the
supply chain network. The objective of this paper is to use blockchain technology to:
(i) conceptualize and design an industrial supply system through shared information
of unique inventories of companies working collaboratively; (ii) design a supply
chain traceability system to monitor and control assembled and injected products to
improve the traceability of components and processes performed with the product.
The remainder of the article is organized as follows. Section 34.2 presents the
related works. Section 34.3 describes the problem being addressed. Section 34.4
proposes the conceptual framework. Finally, Sect. 34.5 provides conclusions and
further research.
34 A Conceptual Framework of a Blockchain Application … 345
34.2 Related Works
In the industrial supply chain field, it is worth mentioning the work of Queiroz and
Wanba [8], whose exploratory study analyzes implementing blockchain in logistics
and supply chains in India and the USA. It concludes that relationships between stake-
holders improve thanks to cooperation and trust, among others. Product traceability
can also significantly improve by allowing customers to acquire the information they
need at all times, which would improve the logistics service level. Their study shows
that adopting this technology varies from one country to another if the particularities
of each one and the infrastructure where it is applied are contemplated. Thus, imple-
menting this technology is related to the influence of facilitating conditions and the
trust they confer on companies in any country.
Another application of this technology is presented in a food retailer supply chain
through e-commerce [9]. The actors in this chain include the producer, the retail
store, the distributor, and the end user. The proposal suggests using blockchain to
improve the process. One of the main problems presented by the retailer is fighting
against counterfeit and expired products, which result in waste due to improper
preservation or unsafe storage, inaccurate quantity information, delays, and errors in
incoming goods. Employing this technology would provide reliable data about the
production method, origin and certification, increased visibility to processes, inven-
tory status, knowledge in forecasting future product orders, and reduced returns and
paperwork costs. In another supply chain in the food sector, Arena et al. [10] propose
a blockchain-based application for the traceability and certification of extra virgin
olive oil and involve the farmer, manufacturer, transport, and sellers. One reason why
this system was considered for this product was counterfeiting in origin and quality
terms, among others. The application provides a system that tracks the whole produc-
tion process from plantation to points of sale. This proposal enables product data to
be collected and certified in all the operation phases, which are provided by the chain
actors and sensors used in transport, storage, and production. The implementation
of a blockchain enables end users to access the entire product history, namely culti-
vation, harvest, production, packaging, preservation, and transport processes, which
includes indications of any information manipulation occurring.
One of the contributions that blockchain offers green supply chain manage-
ment is that it supports environmental sustainability. Kouhizadeh and Sarkis [11]
discuss aspects such as supplier selection and environmental performance measure-
ments. Other green supply chain management-related activities include eco-design
and material handling. Packaging can be reused and tracked so that blockchain-
enabled traceability can prolong the packaging material life span through efficient
management. Hazardous waste tracking is critical given issues associated with poorly
managed landfills. The need to have a permanent record and to track waste disposal
can help environmental problems to be avoided.
Another work in the health field stands out [12], which designs an intelligent
system for supervising a vaccine supply chain to achieve vaccine traceability. By
346 E. Ponce et al.
using smart contracts, both inoculation records and vaccine circulation, and trace-
ability of vaccine operation records, are queried. It also prevents product information
adulteration by applying RFID technology together with blockchain. Another feature
of this system is that it can detect expired vaccines by querying their production date
and the quality guarantee period.
Other blockchain applications can be found in [13–15]. One conclusion is that
conducting more conceptual, descriptive, and empirical research into blockchain
technology applications in the industrial supply chain domain is necessary to exploit
all the benefits that this technology can provide.
34.3 Problem Description
The approach to blockchain applications in an industrial supply chain (Fig. 34.1) may
differ depending on the industry type to which it will be applied. In this case, applying
this technology in a supply chain is proposed where a group of companies working
collaboratively in the plastic parts injection and furniture assembly area is interre-
lated. The addressed problem specifically focuses on both the supply management of
different actors and component demand management. In the considered supply chain
management, the working environment is a client company, two assembly companies
and several firms supplying metal, wood, and plastic parts. The relationship linking
the customer, assembler, and suppliers is limited by not managing a single database
in which information is updated and flows. The customer is a company in the furni-
ture manufacturing and marketing sector, and it is also a supplier because it is in
charge of managing the shipments of some components to assemble finished prod-
ucts, while all the other components are provided by suppliers in two ways: directly
to assemblers and passing the customer’s quality controls before components are
sent to assemblers.
The companies that assemble furniture use the parts sent by the customer and
suppliers. These parts are incorporated into an inventory in ERP. All the assembly
companies have an independent ERP as they are different companies but, at the same
time, the customer/supplier has another ERP with its own inventory. This implies
a first problem because components are constantly shipped and received so that
assemblers can complete daily loads. Suppliers of parts organize deliveries to the
customer, as well as deliveries sent directly to assemblers. In the latter case, the
customer cannot confirm the assembly of finished products if it does not know which
of these components have been correctly received by assemblers.
Assembly is carried out with the orders placed by the customer with assemblers so
that they are progressively shipped as produced (daily committed orders). Another
detected disadvantage is no system determines a priority order, assembly begins
according to the order in which orders arrive, which can lead to important loads or
long-distance shipments to be delivered late. Another defect in the current system
occurs when a component is changed in orders which are updated or changed. This
Fig. 34.1 Flowchart of an industrial supply chain
change is communicated verbally and the assembly considers the updated compo-
nents, but the inventory discount in ERP takes the original list of components into
account. Thus, in the customer’s inventory, the component considered in the original
order is discounted and the component placed in the finished product is discounted in
the real inventory. Therefore, the current system allows neither flexibility given the
variations in order characteristics nor the visualization of any made changes. When
furniture is assembled, the customer cannot see that it is complete until it arrives at
its warehouse because assembly completion and the subsequent transport of assem-
bled products cannot be confirmed. However, the assembly company can confirm
that assembly is complete and is awaiting loading so that efficient and real transport
planning can take place.
If end customers find a defect or flaw in the purchased product, they place a
claim. This can be done in two ways: sent to the marketing company or the customer
company. To deal with this claim, a replacement of the defective part or a total product
change can be carried out. However, while tracking the part with the defect, following
its traceability is complicated because the current system tracks the sold product and
which firm assembled it, but not which components or which batch of parts it was
made with. So not knowing the defective production lot means that the components
or affected finished products cannot be blocked, and traceability is incomplete. This
generates expenses and loss of time for returns because information is not organized,
segmented, and unified. Assemblers use basic technology to measure productivity
and to fulfill orders and committed loads, which can pose problems when confirming
deliveries, personnel overtime, and product cost overruns. The cost overruns caused
by this deficiency in assemblers do not reach the customer given the set prices per
product, which means that assemblers make less profit and, hence, their profitability
is lower. Each supplier’s quality levels are not optimally managed because the main
348 E. Ponce et al.
people involved in detecting these defects are assemblers, who detect them in parts
during assembly, but have no direct channel to these suppliers for two reasons:
the customer makes purchases; assemblers have updated and accumulated records
with time, which does not allow suppliers’ performance to be assessed. The plastic
injection molding company acts as a supplier by delivering parts to the customer. It
is noteworthy that this injection molding company has an assembly section, and it
assembles furniture for this same customer.
With the order management generated by the customer, operations commence with
the product order. Purchase orders can be periodic or urgent depending on the stock
situation and forecast accuracy. Once the customer confirms a purchase order, and
with the quantities and delivery date according to an assembly forecast and a finished
product order forecast, it can place orders with the plastic injection molding company.
Currently, orders are placed and requested by the customer’s ERP and sent by e-mail.
This means that there is no constant real-time line of communication, which makes
human errors possible as the necessary orders for the subsequent assembly might
not be correctly managing, while the injection molding company manages the order
in its own ERP. One factor to consider is the possibility of a non-delivery of injected
parts, which creates a bullwhip effect on the other processes that follow, as well as
cost overruns and non-deliveries.
34.4 Conceptual Proposal
The blockchain application in this industrial supply chain aims to solve most of
the aforementioned shortcomings by considering the benefits of its implementation
(Fig. 34.2). The use of blockchain in the relationships of collaborating companies
with a supply sourcing network means linking the information that needs to be
shared. In the situation, in which each company has independent ERPs and different
inventories with distinct codifications, the benefits of blockchain lie in it centralizing
this information, making it known to all the participants in the chain and each made
transaction is validated, which allows it to be reliable and honest because these data
are immutable. Thus, the end customer would acquire knowledge of the inventory
held by each assembler and could better manage the arrival of goods from parts
suppliers. So blockchain could ensure the supply chain’s end-to-end integration and
allow the transfer of information flows on items and batches by providing inter-
operability within existing systems and ERPs to, thus, facilitate the scheduling of
shipments of finished products to foreign customers with a narrower non-delivery
error margin, which would improve service levels. The orders placed by the customer
would be registered in blockchain. Then, the production planning system will give
them an order of priority according to the load of that day, which depends directly on
transport planning. If any variation or deviation takes place in planning, it would be
updated directly in blockchain by the customer and would be received and accepted
by the assembler to meet daily delivery schedules. Assemblers with manufacturing
execution system (MES) technology would confirm that the assembly of the ordered
Fig. 34.2 Blockchain conceptual framework
finished products was complete by changing their status to await the carrier’s loading.
This visualization would allow the customer to make adjustments to order delivery
planning on an ongoing basis, which would reduce cost overruns due to incomplete
finished product shipments.
The inventory information of each assembler and supplier in a blockchain allows
the components used to manufacture each product to be tracked. As this information
is registered in blockchain, in the event of a quality claim, traceability can be followed
from the order with which it was delivered to the end customer, the transport means,
the assembler that produced it, the suppliers of the used components and the batch
to which they belong. All this facilitates claim management and allows batches of
finished products and defective components to be blocked with consequent savings
in costs and time. Each assembly company’s capacity depends on its productivity and
internal performance. Hence, if it manages a productivity control system, it will be
able to measure its possibilities of accepting orders and will, thus, allow the customer
to acquire knowledge and facilitate decision-making to assign more orders, or not to
a given assembler so that it can meet the request. By being centralized in blockchain
and validated by both the assembler and supplier upon returns, registering defective
parts would facilitate a more accurate supplier evaluation for the customer because
currently defects are detected only by assemblers, and the customer is limited to
deduct from its inventory what the assembler indicates because no history of each
350 E. Ponce et al.
supplier’s quality defects exists. Blockchain’s transparency can improve the cooper-
ation of supply chain members, which would bring about a major transformation in
industrial collaborative relationships. Working in a network is an essential variable
for most organizations, which implies entities having to develop coordination tools to
support this interaction. Blockchain implementation to manage injection orders and
components allows the customer to generate component requirements by planning
and forecasting finished products. Consequently, the injector can optimize produc-
tion and planning processes to anticipate out-of-stock parts or out-of-stock material,
and to reduce the bullwhip effect, which would result in lower logistic costs, and
higher profit margins and profitability. The transportation management of injected
products would also be a benefit because the production of a batch of parts would be
registered in the inventory and displayed in the blockchain customer service.
34.5 Conclusions
As herein defined, the adoption of blockchain technology starts from making a deci-
sion that must be related to a search for elements to solve any problems that emerge
and arise in an industrial supply chain. Therefore, decisions about product tracking
mechanisms, accuracy of the information flow along the supply chain, visibility
process, and inventory status would guarantee improved efficiency and agility in
activities, and would also contribute to optimization in production planning deci-
sions. This approach allows a system to be adopted that demonstrates increased
visibility in inbound and outbound processes and knowledge in order to support
forecasting future product orders, as well as improved traceability, reduced returns
and costs due to lost time, paperwork and returns, among others. This technology can
help to build effective trust mechanisms between collaborating companies. Benefits
come from the network effect and can be obtained by proper generated information
exchange based on earning actors’ trust, with the support of technology where no
third party is involved and decisions are made based on reaching smart contracts.
In this sense, sharing information has plenty of advantages, but there are also risks
that should be managed. The proposed conceptual framework can serve as a starting
point for future implementations to consider the large number of components used for
assembly. RFID technology can be integrated with blockchain to facilitate incoming
and outgoing finished goods without having to resort to barcodes or QR. This would
also help to avoid errors given the wide range of colors and similar component
characteristics, which would be minimized by being identified by this technology.
Acknowledgements This work was supported by the Spanish Ministry of Science, Innovation and
Universities project entitled “Optimisation of zero-defects production technologies enabling supply
chains 4.0 (CADS4.0)” (RTI2018-101344-B-I00).
References
1. Ivanov D, Dolgui A, Sokolov B (2019) The impact of digital technology and Industry 4.0 on
the ripple effect and supply chain risk analytics. Int J Prod Res 57:829–846
2. Borrero JD: Agri-food supply chain traceability for fruit and vegetable cooperatives using
blockchain technology. CIRIEC-Espana Rev Econ Publica, Soc y Coop 71–94 (2019)
3. Preukschat A (2017) Blockchain: la revolución industrial de internet. Gestión 2000
4. Zheng Z, Xie S, Dai H, et al (2017) An overview of blockchain technology: architecture,
consensus, and future trends. Proc—2017 IEEE 6th Int Congr Big Data, BigData Congr
2017:557–564
5. Pilkington M (2016) Blockchain technology: principles and applications. In: Research
handbook on digital transformations. Edward Elgar Publishing
6. Dolader C, Bel J, Muñoz JL (2017) La blockchain: fundamentos, aplicaciones y relación con
otras tecnologías disruptivas. Econ Ind 405:33–40
7. Wüst K, Gervais A (2018) Do you need a blockchain? In: 2018 crypto valley conference on
blockchain technology (CVCBT), pp 45–54
8. Queiroz MM, Wamba SF (2019) Blockchain adoption challenges in supply chain: an empirical
investigation of the main drivers in India and the USA. Int J Inf Manage 46:70–82
9. Perboli G, Musso S, Rosano M (2018) Blockchain in logistics and supply chain: a lean approach
for designing real-world use cases. IEEE Access 6:62018–62028
10. Arena A, Bianchini A, Perazzo P, et al (2019) BRUSCHETTA: an IoT blockchain-based frame-
work for certifying extra virgin olive oil supply chain. Proc—2019 IEEE Int Conf Smart Comput
SMARTCOMP 2019:173–179
11. Kouhizadeh M, Sarkis J (2018) Blockchain practices, potentials, and perspectives in greening
sup-ply chains. Sustain 10
12. Yong B, Shen J, Liu X, et al (2019) An intelligent blockchain-based system for safe vaccine
supply and supervision. Int J Inf Manage
13. Sikorski JJ, Haughton J, Kraft M (2017) Blockchain technology in the chemical industry:
machine-to-machine electricity market. Appl Energy 195:234–246
14. Yang A, Li Y, Liu C et al (2019) Research on logistics supply chain of iron and steel enterprises
based on block chain technology. Future Gener Comput Syst 101:635–645
15. Leng K, Bi Y, Jing L, Fu HC, Van Nieuwenhuyse I (2018) Research on agricultural supply chain
system with double chain architecture based on blockchain technology. Futur Gener Comput
Syst 86:641–649
Chapter 35
Mathematical Programming Model
for Collaborative Replenishment
Between Competitive Supply Chains
in the Footwear Sector
Mario J. Seni and David Peidro
Abstract This work proposes a mathematical programming model to model the

collaborative replenishment process between competitive supply chains. The basic
objective is to reduce the costs associated with replenishment, production, invento-
ries and customer services by putting to good use volume discounts in a production
setting with variable capacity, along with the deferred demand possibility. The collab-
orative mathematical model is compared to a non-collaborative model, and the cost
savings made with collaboration are analyzed. The proposed models are applied to
a case study based on real data acquired from three supply chains in the footwear
sector in Colombia. The collaborative model’s results indicate cost savings for the
participating collaborative enterprises regardless of them being small or large.
Keywords Collaborative procurement · Supply chain · Mathematical

programming
35.1 Introduction
Collaborative replenishment processes between competitor enterprises are relatively

new to industry, and one example of such is the alliances present in the automo-
bile sector [1]. Contemplating them as an alternative to solve reduced purchase and
inventory costs is studied by a collaborative replenishment approach that is funda-
mentally based on joint replenishment models. The first joint replenishment notion
dates back to the 1960s and 1970s [2, 3], when a mathematical model was proposed
that assumed deterministic demand known as the Joint Replenishiment Problem
(JRP). This model seeks to minimize purchase costs by placing joint orders of many
M. J. Seni (B)
e-mail: mjsenmol@etsii.upv.es
D. Peidro
Centro de Investigación en Gestión E Ingeniería de Producción, Alcoy, Spain
https://doi.org/10.1007/978-3-031-29382-5_35
354 M. J. Seni and D. Peidro
articles with only one supplier to make full use of discounts obtained for volume.
Many algorithms and heuristics allow these models to seek an efficient solution by
contemplating numerous joint lot configuration scenarios and options [4]. Collabora-
tive replenishment as a process dates back to when the Just-In-Time philosophy was
implemented in Toyota’s model in the 1980s [5] and also in other models that allow
supply chain (SC) objectives to be aligned, such as VendorManaged Inventory [6].
Joint replenishment by means of a collaborative process among different companies
can be modeled according to the two previously set out approaches [7–11]. However,
integrating it with spatio-temporal SC decisions can be restricted, especially when
production time constraints, assigning resources capacity in the production plan and
managing the bill of materials [12], and the option of deferring demand during each
period [13], have to be taken into account.
Collaborative replenishment models between competitor enterprises are not only
a mechanism that makes full use of discounts but can also help to balance power
between a large size supplier and smaller enterprises [14]. This situation is common
in the footwear industry in Colombia where these power relations cause buyers inven-
tory problems, which is why finding collaboration mechanisms between competitor
SCs can be beneficial in certain scenarios [15]. The progress made thanks to informa-
tion technologies helps to make the intermediation between parties easier, but without
compromising “delicate data” which, thus, opens a window to develop collaborative
economy between competing networks [16]. The objectives of the present work
were to: (i) develop a collaborative mathematical programming model to study the
benefits of extending SC planning by the vertical integration of competition to make
good use of suppliers’ discounts and to reduce costs; (ii) apply the model devel-
oped in a case study based on real data from the footwear sector in Colombia where
different sized SMEs compete; (iii) compare the results obtained with the model to
a non-collaborative replenishment approach.
The remainder of the article is as follows: Sect. 35.2 indicates what the
model contemplates as a mixed integer linear mathematical programming model.
Section 35.3 applies the contemplated model to a case study from the footwear
sector in Colombia to compare the collaborative model results to a non-collaborative
model. Finally, Sect. 35.4 presents the conclusions and future research lines.
35.2 What the Model Contemplates
Three production plants from the footwear sector in Colombia that manufacture a
basic shoe type (black leather moccasin) for some of their own customers. As these
customers are located close to one another, delivery times are almost immediate.
All these plants need to share the same type of leather to make the upper shoe parts
(upper) that are stuck to shoe soles. The production capacity of each factory varies in
time due to workers’ special contract type, which is known as “piecework”. This work
type involves rendering service whereby only the quantity of product that a worker
can make in an 8-h working day is paid. The factories have a maximum number of
35 Mathematical Programming Model for Collaborative Replenishment … 355
workers because of restricted workspaces on plants. A minimum lot size is needed to

start production on each plant. The capacity of the raw material inventories that can
be stored on plants is limited and the maximum purchases budget per month cannot
be exceeded.
Each supplier offers discount packages depending on the purchase lot size, and
each lot size must fall within a (minimum and maximum) range to obtain a given
discount, while supplier delivery times have to remain constant. Customers’ monthly
demand is stable, and there are warehouses with a maximum capacity where goods
are stored. Customers receive deliveries on a weekly basis, and delays are allowed
provided they do not exceed a given percentage of monthly demand. The last week
in every month is set as a top limit to meet any demand delayed from previous weeks.
35.2.1 Collaborative Model
The collaborative replenishment model between competitor SCs from the footwear
sector is presented as a mixed integer linear mathematical programming model for
which the following can be assumed:
– The data in each SC is completely visible.
– The quality and color of the raw material to be used are the same.
– Demand is considered to be certain.
– The delivery times to suppliers are the same and constant.
– The delivery times to suppliers are immediate.
– Purchases of other raw materials are worthless.
Table 35.1 shows the model’s nomenclature, including the sets, data, and variables
required to set it up. All the variables, except those indicated as being binary in this
table, are considered to be integer variables from the nature of the problem.
The mathematical model is formulated as follows:
(35.1)
Subject to:
⎡ ⎤

J
T

I Ckm = I Ckm−1 + ⎣ E jkt − L jkm−1 + L jkm − dkm ⎦∀m, k (35.2)
j∈PCk t∈S E m

J
[L jkm − L jkm−1 ] ≤ dkm ∗ ck ∀m, k (35.3)
j∈PCk
Table 35.1 Nomenclature

Sets
I Suppliers (i = 1..I)
J Production plants (j = 1..J)
T Periods of time (weeks) (t = 1..T )
K Customers (k = 1..K)
A Volume discounts (a = 1..A)
M Periods of time (months) (m = 1..M)
Data
PC k Plants that can supply customer k
SE m Weeks in month m
Nj Customers supplied per plant j
d km Demand of customer k in month m
aj Productivity per worker and week on plant j
µ Lead time
lminj Minimum lot size for plant j
tdes ia Purchasing lot size for volume discount a
ICC k Maximum storage level of finished product in customer k
ck Maximum percentage of the delayed demand for customer k
nj Unit cost of storing raw material on plant j
oj Set cost of launching purchase orders on plant j
cd jk Cost of delaying demand
des ia Price of the raw material for supplier i and discount a
mpj Maximum level for storing the raw material on plant j
fj Purchasing budget for plant j
capwj Maximum number of workers available on plant j
bomj Raw material needed to produce one finished product unit
mpij Maximum level of storing the finished product on plant j
ct j Unit cost of storing the finished product on plant j
cwj Unit production cost on plant j
LC k Minimum lot size to be delivered to customer k
Variables
L jkm Delayed demand for customer k in month m from plant j
IC km Inventory level of the finished product for customer k in month m
E jkt Quantities sent from plant j to customer k in week t
(continued)

Variables
Pjt Quantity produced on plant j in week t
R ijat Raw material ordered from j to supplier i with discount a
IMPjt Inventory level of the raw material on plant j during period t
IPjt Inventory level of the finished product on plant j during period t
W jt Workers employed for the production on plant j in week t
Yati Purchase indicator with discount a (binary variable) for supplier i during t
Ʈi jt Purchase indicator (binary variable) for supplier i on plant j during t
F jt Production indicator (binary variable) on plant j in week t
FC jkm Minimum delivered lot size indicator (binary variable) sent to customer k in month m

J
L jkm ≤ dkm ∗ ck ∀k, m|m = M (35.4)
j∈PCk

J
T

FCkm ∗ LCk ∗ dkm ≤ E jkt ∀k, m (35.5)
j∈PCk t∈S E m

J
T

E jkt ≤ FCkm ∗I CCk ∀k, m (35.6)
j∈PCk t∈S E m
I Ckm ≤ I CCk ∀k, m (35.7)
Pjt = W jt ∗a j ∀ j, t (35.8)
W jt ≤ capw j ∀ j, t (35.9)
F jt ∗lmin j ≤ Pjt ∀ j, t (35.10)
Pjt ≤ M∗F jt ∀ j, t (35.11)

K
I P jt = I P jt−1 + P jt − E jkt ∀ j, t (35.12)
k∈N j
I P jt ≤ mpi j ∀ j, t (35.13)

I
A

I M P jt = I M P jt−1 + i
Rjat−µ − W jt ∗bom j ∗ a j ∀ j, t (35.14)
i a
I M P jt ≤ mp j ∀ j, t (35.15)

J
Yati ∗tdesa−1
i
≤ i
Rjat ∀i, a, t (35.16)
j

J
i
Rjat ≤ Yati ∗tdesai ∀i, a, t (35.17)
j

A
Yati ≤ 1∀i, t (35.18)
a
(35.19)

I
A
T
i
Rjat ∗desai ≤ f j ∀ j, m (35.20)
i a t∈S E m
The purpose of objective function (35.1) is to minimize costs, including the vari-
able and fixed costs associated with launching purchasing orders, raw material and
finished product storage, production and deferred demand. Constraint (35.2) corre-
sponds to the customer’s inventory balance. Equations (35.2) and (35.3) manage
deferred demand, which obliges all the demand deferred in the last weekly period to
be covered. Constraints (35.5) and (35.6) ensure that a minimum lot is delivered to
the customer. Equation (35.7) determines the maximum storage capacity. Constraints
(35.8) and (35.9) are associated with personnel management and cannot exceed the
maximum threshold. Constraints (35.10) and (35.11) ensure the minimum production
lot.
Equations (35.12)–(35.15) define the inventory balance and the maximum storage
capacity of both finished products and raw materials on plants. Constraints (35.16)
to (35.19) control the collaborative purchases made so they do not exceed the defined
purchasing volume limits, apart from ensuring that only the purchasing volume is
selected, as well as a discount for each supplier and period of time. Finally, Eq. (35.20)
manages the maximum purchase budget.
35.2.2 Model with Collaboration
In order to set the collaborative model comparison framework presented in the

previous section, a scenario is used in which the different factories are not willing to
collaborate in their replenishment process. This implies amending three constraints

of the collaborative models, (35.16), (35.17) and (35.18), and the binary variable of
the purchasing indicator to produce the model without collaboration:
Yati j ∗tdesa−1
i
≤ Rjat
i
∀i, a, t, j (35.21)
i
Rjat ≤ Yati j ∗tdesai ∀i, a, t, j (35.22)

A
Yati j ≤ 1∀i, t, j (35.23)
a

J
A
R ij át ≤ tdes iA ∀i, t (35.24)
j á
Equations (35.21) to (35.23) allow the same purchase discounts as the collabora-
tive model, but individually. For this reason, the summations per plant are removed
and constraints are defined for every plant. Finally, a new constraint is added (35.24)
to respect suppliers’ maximum capacity, which continues to be the same as for the
collaborative model.
35.3 Analyzing the Results
Both the collaborative and non-collaborative mathematical programming models

were solved for the three different sized manufacturing plants: two large-sized plants
(Bellino and Ortiz) and a small one (Madrid). A 12-week (3-month) horizon was
considered, with six alternative suppliers and two discount levels each, plus six end
customers.1 The mathematical models were developed in the Python programming
language and were solved with the GUROBI solver using equipment with an I5
9600 K processor and 16 GB of RAM.
The optimum results generated for each model show that the scenario which
most reduces the total costs was the collaborative model with $91.169 as opposed
to $94.797 for the non-collaborative scenario. This difference was distributed as
described in Table 35.2 according to the various contemplated costs.
The replenishment costs (sum of purchases costs and launching orders) lowered
in the collaborative scenario, which allowed the different plants to sum purchase
requirements to obtain more favorable discounts. More orders with better discounts
were placed and, at the same time, the inventory costs significantly lowered because
the non-collaborative scenario needs to place large individual orders to obtain the
best discounts, which led to a higher individual inventory level. In the collaborative
1 Details of the other data used in the models can be supplied according to requirements.
Table 35.2 Comparison of

Type of cost Better scenario Savings (%)
savings per cost between
collaborative and Replenishment Collaborative 3
non-collaborative scenarios Inventory Collaborative 69
Deferred demand Collaborative 17
Production – 0
Table 35.3 Comparison of savings per cost and plant between collaborative and non-collaborative
scenarios (values rounded off)
Enterprise Replenishment (%) Inventory (%) Deferred demand (%) Total (%)
Madrid 23 −100 31 23
Bellino 2 27 0 2
Ortiz 1 97 27 4
model, a higher response capacity was produced, which managed to lower the costs
of deferring demand by 17%, which improved weekly customer services. Production
costs were the same in both models because delaying demand at the end of the month
was not permitted.
In the collaborative model, the distribution of these costs differently affected
each company individually because the difference in company size influenced both
replenishment costs and raw material storage. The smallest firm (Madrid) made
more profits thanks to it collaborating with the other enterprises, despite the worse
inventory values (basically of raw materials). Table 35.3 compares the savings for
each type of cost and plant to the costs generated in the non-collaborative scenario.
The two larger enterprises (Bellino and Ortiz) managed to comparatively lower
all their costs thanks to the collaboration mechanism. In particular, they managed
to significantly reduce their inventory maintenance costs compared to Madrid. The
smallest enterprise had to place bigger orders to obtain more economic discounts
comparatively to its size. Nonetheless, this small enterprise managed to reduce
its overall total costs more (23%) thanks to the profits made with collaborative
replenishment.
35.4 Conclusions
Joint vertical replenishment collaboration between SCs that share different suppliers
is a paradigm that offers good results compared to non-collaborative scenarios.
When discounts for quantities help substantial savings to be made in replenishment
costs, these collaborative processes can lead to better results for the companies that
apply them. This work presents a mathematical programming model for collabora-
tive replenishment which was applied to a case study based on real data from the
footwear sector in Colombia. The results revealed that cost savings were made for
the set of companies, which were obtained regardless of the size of the firms partic-
ipating in collaboration. The collaborative approach can help small enterprises to
obtain volume discounts which they would never otherwise attempt to benefit from
on their own, and those enterprises that place large orders individually obtain more
possibilities of obtaining discounts.
Future research lines offer the possibility of acquiring a distributed collaboration
model, one that contemplates dealing with the uncertainty found in some of the
model’s data by means of fuzzy mathematical programming and perform a sensitivity
analysis to measure the robustness of the proposed model using system dynamics by
modifying certain critical parameters, such us the purchasing lot size, among others.
Acknowledgements This work was supported by the Ministry of Science, Innovation, and Univer-
sities project entitled “Optimisation of zero-defects production technologies enabling supply chains
4.0 (CADS4.0)” (RTI2018-101344-B-I00).
References
1. Financial times, https://www.ft.com/content/ac3aa4ca-62f0-11e1-9245-00144feabdc0.

Accessed 01 March 2021
2. Starr MK, Miller DW (1962) Inventory control: theory and practice. Prentice Hall, Englewood
Cliffs
3. Shu FT (1971) Economic ordering frequency for two items jointly replenished. Manage Sci
17:99–110
4. Goyal SK, Satir AT (1989) Joint replenishment inventory control: deterministic and stochastic
models. Eur J Oper Res 38:2–13
5. Spekman R, Davis E (2003) The extended enterprise: gaining competitive advantage through
collaborative supply chains. Prentice Hall, Upper Saddle River
6. Derrouiche R, Neubert G, Bouras A (2008) Supply chain management: a framework to
characterize the collaborative strategies. Int J Comput Integr Manuf 21:426–439
7. Hsu SL (2009) Optimal joint replenishment decisions for a central factory with multiple satellite
factories. Expert Syst Appl 36:2494–2502
8. Otero C, Amaya R, Yie R (2019) A stochastic joint replenishment problem considering trans-
portation and warehouse constraints with gainsharing by Shapley value allocation. Int J Prod
Res 57:3036–3059
9. Moon IK, Cha BC, Lee CU (2011) The joint replenishment and freight consolidation of a
warehouse in a supply chain. Int J Prod Econ 133:344–350
10. Wang L, He J, Zeng YR (2012) A differential evolution algorithm for joint replenishment
problem using direct grouping and its application. Expert Syst 29:429–441
11. Vanvuchelen N, Gijsbrechts J, Boute R (2020) Use of proximal policy optimization for the
joint replenishment problem. Comput Ind (119)
12. Spitter JM, Hurkens CAJ, de Kok AG, Lenstra JK, Negenman EG (2005) Linear programming
models with planned lead times for supply chain operations planning. Eur J Oper Res 163:706–
720
13. Guptas A, Maranas CD (2003) Managing demand uncertainty in supply chain planning. Comput
Chem Eng 27:1219–1227
14. Li L (2019) Cooperative purchasing and preactive inventory sharing—channel balancing and
performance improvement. Eur J Oper Res 278:738–751
15. Mehrjerdi Y, Shafiee M (2021) A resilient and sustainable closed-loop supply chain using
multiple sourcing and information sharing strategies. J Clean Prod 289:125–141
16. Zehui G, Qiying H, Chon-Huat G, Rui Z (2021) Action-dependent commitment in vertical
collaborations: the effect of demand-creating innovations in a supply chain. Transp Res Part E
147:147–164
Chapter 36
Moving Toward the Physical Internet:
A Model that Moves Toward
Sustainability Against a Necessary
Backdrop of Industrial Transformation
Carlos Alonso de Armiño , Roberto Alcalde Delgado ,

Luis Santiago García Pineda , and Manuel Manzanedo
Abstract A new idea is emerging more and more strongly in the European Union
(EU), of the so-called Physical Internet, a model that aims to handle physical goods in
their storage and transport processes in a parallel way to how the Internet handles data,
which involves efficiency and sustainability. Some steps and considerations will be
required, and in the background, we perceive an appreciable necessary transformation
in the volumes and dimensions of the goods to be transported and therefore in the
industrial and consumer products.
Keywords Physical Internet · European Union · Supply chain management ·

Logistics · Efficiency · Sustainability
36.1 Introduction
In a liquid modernity and changing world, disruptive ideas begin to take on a certain
dreamlike character. This seems to be the case when we are introduced to the concept
of the so called Physical Internet; the idea that the logistics of goods could have a
principle of activity equivalent to that of the Internet.
But there must be something solid in this idea when it is one of the fundamental
development axes for Alliance for Logistics Innovation through Collaboration in
C. Alonso de Armiño (B) · R. Alcalde Delgado · L. S. García Pineda · M. Manzanedo

Universidad de Burgos, Burgos 09001, España
e-mail: caap@ubu.es
R. Alcalde Delgado
e-mail: radelgado@ubu.es
L. S. García Pineda
e-mail: lsgpineda@ubu.es
M. Manzanedo
e-mail: mms0133@alu.ubu.es
https://doi.org/10.1007/978-3-031-29382-5_36
364 C. Alonso de Armiño et al.
Europe (ALICE) [1], the European platform for technological innovation created
in 2013 and which currently integrates 149 important organizations and companies
linked to the development of European logistics under the prism of efficiency and
sustainability. Its working axes are: (i) The sustainability of logistics, (ii) Freight
Corridors, Warehouses, and Synchronization, (iii) Technological Interconnection
Systems, (iv) Coordination and Collaboration of Supply Chains, and (v) Urban Logis-
tics. Nowhere are all these frameworks better united and linked than in their central
project of study; the so-called Physical Internet whose Roadmap [2] has been defined
in 2020.
36.2 Background
The necessary link between economic activity and transport is well known since the
beginning of our civilizations, as well as the harmful effects linked to its activity and
the need for infrastructures to support it.
36.2.1 The EU at a Crossroads in the Transport Sector
We could say that a large part of the activity carried out by the EU to develop the
objective set by its founding framework, the Treaty of Rome of 1957, regarding the
adequate development of transport systems, has been to seek: (i) the development
of adequate and interconnected transport infrastructures in its territories, (ii) the
improvement in efficiency and sustainability of the means of transport execution in
each of its different modes, (iii) the search for the exchange of goods toward more
efficient modes, and (iv) the regulation of activities related to transport in search of
efficiency and safety.
These objectives have been developed in various initiatives and their more specific
aims have been set out in the so-called EU Transport White Papers, which have been
published every five years since 1992. Underlying all of this is the Union’s continuous
search for sustainable development.
On this path, there have been many disappointments, perhaps the most notable
being the relative failure in the decided commitment to transfer freight transport
activity from road to rail, which after four plans of firm commitment to this initiative,
ended up forcing the EU to complement its commitment with the support of the
maritime mode through Short Sea Shipping. Perhaps the most notable successes
in this search for modal interchange have come from the intermodality linked to
containerized transport, whose activity has shown significant growth rates in all the
territories of the Union.
36 Moving Toward the Physical Internet: A Model that Moves Toward … 365
36.2.2 Increases in Freight Transport Efficiency
One of the main axes that have stood out, therefore, in the deployment of transport
policies by the Union has been the search for increased efficiency in transport means
and infrastructures.
On the one hand, it is worth considering that means of transport have not ceased to
evolve in terms of increasing their capacities and improving their energy efficiency,
which is directly linked to their reduction in harmful emissions. On the one hand,
this trend has been framed in the trend that has received the name of Longer and
Heavier Vehicles for Freight Transport (LHVFT), which encompasses the continuous
evolution of means of freight transport, and whose most notable recent aspect is the
irruption of the so-called Mega-Trucks of more than 25 m in length and with a
Maximum Authorized Mass (MAM) of 60 tons [3].
On the other hand, it is also worth mentioning the progressive creation of high-
capacity European corridors for freight and passenger transport, and their integration
in the so-called Trans-European Transport Networks (TEN-T), which has constituted
a rational basis for the integration of its Member States [4].
36.2.3 The Engine of Private Initiative
The truth is that any development in the field of transport will ultimately be supported
by the entities that support their economic relationship on it. Not only the companies
specialized in transport and logistics will present a fundamental criterion in the use
and choice of the different modes of transport, but also the supply chains consti-
tuted by their customers will develop for them a framework of demands and service
requirements on their geographical deployment of supply and distribution and the
provision of storage centers and consolidation of goods along the same. We are
moving in the well-known framework of supply chain management (SCM), which
throughout its spatial deployments will have an important set of warehouses—Hub
own or subcontracted.
36.3 The Physical Internet Model
The conjunction of the background described in the previous section, with the func-
tional abstraction of the Internet communications model, results in the conceptual
model of the so called Physical Internet (PI). Precisely this acronym PI will become
the symbol of this conceptual model that is materializing in the form of multiple EU
projects (Fig. 36.1).
Fig. 36.1 Internet physics logo. (Project/trademark by ALICE: alliance for logistics innovation
through collaboration in Europe [1]; reproduced with permission)
36.3.1 Central Idea of the General Physical Internet Model
The PI is “a global logistical system based on the interconnection of logistical

networks through a standardized set of protocols for smart collaboration, containers,
and interfaces in order to increase efficiency and sustainability” [2].
The conceptual extrapolation between the “real Internet” model and the Physical
Internet is intended to be almost straightforward [3]:
• On the one hand, the suppliers of raw materials and products to be distributed
through the logistics network, together with the clients who receive them, would
be the equivalent of the information systems that would be connected in the “real
Internet” by means of transport increasingly equipped with greater capacities [4].
• On the other hand, transport infrastructures, insofar as they are connected in
the form of conventional transport networks, or efficient high-capacity transport
networks, would become equivalent to data communication systems [5].
• Finally, the information systems that support the temporary data servers, necessary
for the efficient exchange of data, as well as their accessible repository, in that
abstract and global concept that has come to be called “the cloud”, would be the
warehouses of goods connected through these infrastructures by the means of
transport (Fig. 36.2).
36.3.2 A Complex Contractual World
Some of the most complex and long-standing contractual protocols in the legal
systems of the world regulate the process of transporting goods. Especially when
this process affects transits between different member states of the European Union.
Aspects such as the clear establishment of sender and consignee, temporary holders,
Fig. 36.2 Real internet and physical internet simile
payment of rates and customs charges, payment of goods, responsibilities, and

commitments acquired in the execution of the process, make up a good battery
of possible alternatives that try to limit to a subset of reasonable possibilities in the
so-called International Commercial Terms (Incoterms). It must be taken into account
that the legal security of these contracts is very high, being admitted as a guarantee
in many credit mechanisms and being able to be assigned as rights to a third party.
The minimum expression of these contractual commitments is nowadays the obli-
gation to accompany the transit of goods with appropriate documentation. For tran-
sits in the most common mode of European goods movements, road transport, the
compulsory Contrat de transport international de Marchandise par Route (CMR)
for international transport and the so-called Carta de Porte for the transport of
goods within the same Member State, which in Spain is established by the Ley de
Ordenación de los Transportes Terrestres (LOTT) and its corresponding regulation
Reglamento de Ordenación del Transporte Terrestre (ROTT).
Fortunately, in this sense, progress has been made in recent months thanks to the
new e-FTI Regulation of the European Union on the Exchange of Electronic Docu-
mentation in the Transport of Goods, and the recent regulatory framework approved
by the Directorate General of Transport of the Ministry of Transport, Mobility and
Urban Agenda that allows the use of electronic support for the generation of the
Road Transport Control Document [6].
36.3.3 Strengthen and Monitor the System
For the vast majority of us the “real Internet” would be to conceive moderately
the functioning of what is described in the previous point, under a basis of global
accessibility and proper functioning, but in our eyes and our understanding essential
elements in its operation would escape.
From the outset, it should be considered that information systems communication

environments present redundant and alternative structures and strong security proto-
cols with reinforcement mechanisms. Likewise, PI will have to have strict operation
and verification protocols that will rely heavily on:
1. Alternative transport routes and modes. Perhaps not as efficient or sustainable as
the initial options, but they constitute a viable and reasonable alternative for the
adequate flow of elements throughout the transport system.
2. Process compliance protocols. Monitoring of guidelines and requirements to be
covered in a certain flow throughout the system. Related to origins, destinations,
times, storage, and handling conditions and incorporation of possible added value
processes in these flows.
3. Process verification mechanisms. Every day with greater possibilities of support
in new technologies such as radio-frequency identification of goods (RFID),
geographic positioning tracking through Global Positioning System (GPS),
mechanisms for communication of the status of the elements of the system
through the so-called Internet of Things (IoT), and all this with absolute inte-
gration of real-time systems and the guarantee of data integrity provided by the
blockchain [7].
36.3.4 The Devil is in the Details
Despite the considerations of the previous point, something essential of the “real
Internet” on which its development and expansion are based continues to be hidden
from our eyes; the use of data communication protocols in packets of limited length
and format in the so-called Department of Defence )DoD) or Transmission Control
Protocol/Internet Protocol (TCP/IP). In them, so to speak, the series of data that
constitutes the transmission of a file in the system is divided into a series of data
packets of determined format and length, constituting in themselves a real guarantee
of standardization of processes in the network of the “real Internet”, and of the
protocols for its temporary or definitive storage. As an absolute reflection of this
circumstance appears the concept of the so-called PI-Containers, whose development
deserves a specific section.
36.4 PI-Containers: An Idea that Encompasses

Transformation
In addition to the need to create the equivalent of data packages, the aforementioned
success of intermodal containerized transport has been added, a conjunction that has
given rise to the idea of PI-Containers, as atomic elements to be transported within
the PI [8, 9] (Fig. 36.3).
Fig. 36.3 Study of possible dimensions of M-boxes for the development of PI-containers based on
the europallet
However, unlike what happens with data, most goods cannot be broken down into
subassemblies of identical dimensions and characteristics that are then regrouped to
restore the original product.
Thus, the PI-Containers are conceived as subdivisions of a commonly accepted
volume, taking as a measure that of the unitizing element of goods par excellence in
the EU; the europallet, with a logical limitation of gauge by the means of transport
in 2.40 m, which would force to design the products according to this set of possible
dimensions that yields a total of 440 possibilities. This is pointed out by Ballot et al.
[10] who refer to the packaging of these elements to be transported as M-Boxes on
the basis of the MODULUSHCA study (Modular Logistics Units in Shared Comodal
Networks) [11] and the deployment of technical possibilities regarding the develop-
ment of PI-Containers by Landschützer [8] in which a reasonable implementation
of five typologies is limited, with a target time horizon for transformation in 2030
according to their previous work [9] (see Fig. 36.4).
36.5 Conclusions
The physical Internet seems conceptually a valuable and sustainable model, but in
order to make it work it will require a major overhaul of the design and packaging
of the products it is applied to.
Fig. 36.4 Time horizon of the development of the M-boxes capacitive of PI-containers occupations
References
1. ALICE. www.etp-logistics.eu. [Online]. Available: https://www.etp-logistics.eu/. Accessed: 09

March 2021
2. Ballot E, Montreuil B, Meller R (2014) The physical internet—the network of logistics
networks. La documemtation Francaise/Predit
3. Ballot EM et al (2020) Road map to the physical internet
4. Longer and heavier vehicles for freight transport | EU science hub. [Online]. Avail-
able: https://ec.europa.eu/jrc/en/publication/eur-scientific-and-technical-research-reports/lon
ger-and-heavier-vehicles-freight-transport. Accessed: 09 March 2021
5. Richardson T (1997) The trans-European transport network. Eur. Urban Reg. Stud. 4(4):333–
346
6. Consejo-Europeo BOE (2020) ES-DOUE-L-2020-81225 Reglamento (UE) 2020/1056 del
Parlamento Europeo y del Consejo, de 15 de julio de 2020, sobre información electrónica
relativa al transporte de mercancías. España, p 16
7. Saberi S, Kouhizadeh M, Sarkis J, Shen L (2019) Blockchain technology and its relationships
to sustainable supply chain management. Int J Prod Res 57(7):2117–2135
8. Landschützer C, Ehrentraut F, Jodin D (2015) Containers for the physical internet: requirements
and engineering design related to FMCG logistics. Logist Res 8(1):1–22
9. Landschützer C, Jodin D, Ehrentraut F (2014) Modular boxes for the physical internet—
technical aspects. undefined
10. Montreuil B, Ballot E, Tremblay W (2014) Modular design of physical internet transport,
handling and packaging containers. Prog Mater Handl Res 13
11. Final report summary—MODULUSHCA (Modular logistics units in shared co-modal
networks)|Report summary|MODULUSHCA|FP7|CORDIS|European commission. Quebec,
(2014)
Part IX
Sustainability, Eco-efficiency and Quality
Management
Chapter 37
Machine Learning Approaches to Predict
the Use of Share Bicycles According
to Weather Conditions
Alejandro Escudero-Santana , Andrea Beltrante,

Elena Barbadilla-Martín , and María Rodríguez-Palero
Abstract Bike sharing services are a reality that is developing more and more every
day, contributing to reduce private car use. A bike sharing system is not limited to
the fleet and the stations, but requires important support of internal office services,
cyclical maintenance of the bikes, and their continuous redistribution. The various
supporting services should be organized according to the number of circulating bicy-
cles, and thus accurate demand previsions can provide considerable help in opti-
mizing the costs bear by the service provider. The use of bicycles follows a cyclical
pattern, but it also depends highly on the weather conditions. This work aims to
adapt and apply different machine learning algorithms to predict this demand. It uses
a real database, containing data on two years of bicycle rentals in London. The results
obtained validate the methodology.
Keywords Share bicycle · Machine learning · Forecasting · Neural networks ·

Random forest
37.1 Introduction
Bike sharing services are an ecological reality that is developing more and more
every day, contributing to reduce private car use. This phenomenon is happening in
all kinds of cities. The motivations for using the bicycle as an alternative to the car
vary from the need for more agile means of transportation to avoid congestion, to the
situations in which the quality of the public transport is not sufficient. Bike sharing
services also contribute to solve the “last mile problem”, which is the difficulty that
public transport users might encounter to reach their destination from the nearer stop;
A. Escudero-Santana (B) · E. Barbadilla-Martín · M. Rodríguez-Palero

Escuela Técnica Superior de Ingeniería, Universidad de Sevilla, Avd. de los Descubrimientos S/N,
41092 Seville, Spain
e-mail: alejandroescudero@us.es
A. Beltrante
Politecnico Di Milano, Milano, Italy
https://doi.org/10.1007/978-3-031-29382-5_37
374 A. Escudero-Santana et al.
this inconvenience often imposes the use of the car for long routes that, if it were
not for that last mile, it could have been done conveniently with buses and trains.
Finally, a good bike sharing service is a valuable alternative for tourists’ mobility,
thus contributing to decongest the public transport network in cities’ central areas,
the ones that suffer the most from overcrowding.
A bike sharing system is not limited to the fleet and the stations, but requires
important support of internal office services, cyclical maintenance of the bikes, and
the continuous redistribution of these between the different city’s areas. The various
supporting services should be organized according to the number of circulating bicy-
cles, and thus, their programming is based on demand forecasts: Accurate demand
previsions can provide considerable help in optimizing the costs borne by the service
provider [1–4].
The use of bicycles, and mobility in general, is a phenomenon that follows a
pattern: The demand is higher on working days, especially during peak hours, while
on holidays the demand is lower and more stable throughout the day. Among the
different means of transport, the bicycle is the one that most is influenced by the
weather [5], so it is interesting to take advantage of weather data to obtain more
precise predictions on the bike sharing’s use. Using new machine learning tools, it
is possible to increase the reliability of the predictions, obtaining interesting results
also with average computer power. These forecasts, integrated into the programming
of the support services, allow optimizing the management of this important public
service.
This work aims to study, adapt, and apply machine learning algorithms to a real
database, consisting of two years of hourly observations on the number of shared
bikes rented in London and the weather conditions of the city at that time. Studying
and interpreting the correlation between weather conditions and the rents’ number
allow to estimate bikes’ use on the basis of weather forecasts. Better predictions
improve the efficiency in scheduling all the support services, including the redistri-
bution of bicycles and the customer support office. The scope includes the analysis of
the available data, with special emphasis on the outliers that could negatively affect
the algorithms’ performances, and the development of tools to modify and manage
their different natures. Among machine learning algorithms, a greater focus is placed
on the random forest regressor and the neural network.
37.2 Methodology
The objective of the study is the discovery of the relationship between the use of
shared bicycles and the weather conditions of the day, aiming at the development of
algorithms able to predict future demand on the basis of weather forecasts.
The tool used is supervised learning, a branch of machine learning science that
trains an algorithm on how to infer an output given a set of inputs. The training is
done by supplying historical examples, made up of pairs of inputs and outputs: the
algorithm acquires experience from the vision of historical data, which are the starting
37 Machine Learning Approaches to Predict the Use of Share Bicycles … 375
point of the entire process. In the case of this work, the inputs are the daily weather
conditions and the characteristics of the day, while the output is the estimation of the
bicycles use for each hour. Since the output is a continuous numerical value, this is
a regression problem, and the output is usually defined as a label.
Among all the existing methodologies, the work will study two tools: random
forests (RF) and neural networks (NN).
There is a set of training data used to fit the regression models and train the neural
network. Each stage of the project includes one or more cross-validation phases,
which consist of checking the predictive capabilities of the model without using the
test set. Cross-validation is based on dividing the training set into n data subsets, of
which all, except one, are used to fit/train the model under analysis, which is later used
to make forecasts about the last package. The procedure is repeated several times,
leaving an always different batch out of the training stage, and in each iteration
the deviation between the forecasts and the labels of the last package is measured;
the mean of these errors is a good approximation of the error expected from the
model against new data, providing a measure of the effectiveness of continuous
improvement.
For the implementation of the different procedures, the Scikit-Learn and Tensor-
Flow libraries have been used.
37.3 Case Study
The case studied is framed in information on two years of bike sharing use in the
city of London, from 01/04/2015 to 01/04/2017, with hourly measurements of the
number of bikes rented and the weather conditions at that time. The database has been
obtained through the “Kaggle” platform. (https://www.kaggle.com/hmavrodiev/lon
don-bike-sharing-dataset).
37.3.1 Data Characteristics
The database is complete, meaning that for each observed hour there is no null value.
The data collected in the database are:
• timestamp: It collects information on the date and time of each observation. It
allows to uniquely identify each data, even those that might be added to the
database in the future.
• cnt: It is the number of rentals started in each hour. It is the “label”, the objective
to be estimated.
• t1 and t2: They are the average temperature and the average perceived temperature
for each hour.
• hum and wind_speed: It information on average humidity and wind speed for each
hour.
• weather_code: It indicates the weather situation in coded form:
– 1 = Clear/mostly clear
– 2 = Scattered clouds/Few clouds
– 3 = Broken clouds/Partially cloudy
– 4 = Cloudy/fog
– 7 = Rain/Light rain shower/Light rain
– 10 = Rain with thunderstorm
– 26 = Snowfall
– 94 = Freezing fog
• is_holiday and is_weekend: They take the value “1” if the day to which the
observation belongs is a holiday/belongs to the weekend.
• season: A value between 0 and 3 (included) identifies each season.
The histograms that represent the data are shown in Fig. 37.1. It is possible to
observe that the numerical attributes have a certain tendency of a normal distribu-
tion, even when some of them present certain asymmetry. Regarding the categorical
attributes, it is observed that the vast majority of the observations are concentrated
in five classes of weather_code and that the number of holidays in the sample, iden-
tified by the attribute is_holiday, is small. The label under study, cnt, is strongly
unbalanced toward low values.
Figure 37.2 shows the evolution of the variable cnt, with Fig. 37.3 showing an
enlargement of a few days. In Fig. 37.2, it is possible to observe the existence of
Fig. 37.1 Histogram of the database

Fig. 37.2 Time series of the variable cnt
Fig. 37.3 Detail of the variable cnt
atypical values, while Fig. 37.3 shows the daily cyclical reiteration of the variable,
and the difference between working days and holidays.
37.3.2 Data Preprocessing
In order to adapt the existing data to the object of study, a series of adjustments have
been made to them. On the one hand, the timestamp attribute has been transformed to
an hour (0:23), a day of the week (weekday), and a month. This allows to characterize
the average days (holidays, working days, and weekends), as shown in Fig. 37.4.
To investigate if there is any strong dependence between any variable and the label
under study, a correlation analysis of the variables has been carried out (Table 37.1).
It can be seen that no variable is highly correlated with the label, although the most
influencing variables are the variables t1, t2, and hour (positively), and the variable
hum (negatively).
Likewise, it is interesting to check the existence of outliers of the data. The analysis
has been carried out according to the variable hour and the type of day (see Fig. 37.5).
On holidays the number of outliers is very limited, and these are not far from the
upper limit, while on business days there is a more significant presence of both
positive and negative outliers. There is also an interesting pattern of pairs of values
well above the average. Analyzing these data in particular, it has been observed that
Fig. 37.4 Patterns of the daily average demand
Table 37.1 Correlation of

Variable Correlation index
variables with cnt
cnt 1,000,000
t1 0.391911
t2 0.371496
hour 0.325219
wind_speed 0.129930
month 0.062791
is_holiday −0.057081
weekday −0.066630
is_weekend −0.096213
season −0.118855
weather_code −0.161517
hum −0.472819
they refer to two specific days in which the London Underground was totally shut
down due to a strike, and therefore, it was decided to eliminate those days from the
database.
The convenience of incorporating new variables or merging some attributes has
also been analyzed. The results of the experimentation advise the creation of a new
variable not_workday, merging the variables is_holiday and is_weekend.
37.4 Experimentation and Results
In the first stage, three regression methodologies with standard parameters were
studied. Results are shown in Table 37.2.
Subsequently, a refinement of the algorithm was carried out on the most promising
regression model, the random forest, reaching a Typical Prediction Error of 262.95.
This result was obtained with the following parameterization (n_estimators = 1000,
Fig. 37.5 Boxplots of “cnt” label for each hour of workdays
Table 37.2 Comparison of

Linear Tree regressor Random
regression techniques
regression forest
Prediction 561.98 3.34 85.63
error on
training set:
Typical 568.27 373.92 282.45
prediction
error:
Standard 61.21 76.64 67.75
deviation:
min_samples_split = 6, min_samples_leaf = 3, max_features = "auto”, max_depth

= None, bootstrap = True).
In the case of neural networks [6], a fine adjustment was made for the number
of layers, the number of neurons for each layer, and the activation function. The
parameters chosen were five layers of 37 neurons, being the activation function of
the ReLU type.
Once all the cross-validation tests to make a fine adjustment of the proposed
models have been carried out, it is possible to proceed to test the algorithms on the
percentage of data reserved for the final validation.
In the case of the random forest, a mean prediction error equal to 260.61 is
obtained, a result consistent with the mean error obtained with the cross-validation,
262.95, thus confirming the good behavior of the model and the absence of overfitting.
The model developed using neural networks obtains a mean prediction error equal
to 249.92; which improves the results of the random forest. In this case, it is important
to highlight that the result is somewhat worse than the one obtained in the last cross-
validation test. Figure 37.6 shows the predictions made, together with the value of
the label and the error made in the estimation for the neural network model.
Fig. 37.6 Estimation of the neural network
In both models, it is observed that the greatest errors occur on the days corre-
sponding to Christmas holidays. Observing the chart, the unusualness of this period
can be verified: It is characterized by working days with low demand and holi-
days, such as the 25th itself, in which it is unusually high, being the bicycle service
practically the only public transport system available in the city.
37.5 Conclusions
The work presented illustrates a proposed methodology for forecasting the use of
bike sharing services, mainly focused on the weather characteristics of the day. To
develop this methodology, the work was done on two models, one based on random
forests and the other based on neural networks.
The results obtained with the test set show that the selected models manage to
forecast the demand with good precision, with a mean squared error comparable to
that obtained in the cross-validation stage.
It is important to note that these models, when atypical days appear in the sample,
such as the Christmas period, are not capable of making good forecasts. To solve the
problem, it would be necessary to train the models with a sample greater than two
years.
To ensure the proper functioning of the models, it would be necessary to contin-
uously update the database to intercept future changes in the pattern of bicycle use
(e.g., an increase in daily rentals due to a growth in the number of users and available
stations).
Although the work is focused on the city of London, the developed algorithms
are easily adaptable to other cities.
References
1. Lin L, He Z, Peeta S (2018) Predicting station-level hourly demand in a large-scale bike-sharing

network: a graph convolutional neuronal network approach. Transp Res Parc C: Emerg Technol
97:258–267
2. Xu C, Ji J, Liu P (2018) The station-fee sharing bike demand forecasting with a deep learning
approach and large-scale datasets. Transp Res Part C: Emerg Technol 95:47–60
3. Chan PC, Wu JL, Xu Y, Zhang M, Lu XY (2019) Bike sharing demand prediction using artificial
immune system and artificial neuronal network. Soft Comput 23(2):613–626
4. Pan Y, Zheng RC, Zhang J, Yao X (2019) Predicting bike sharing demand using recurrent
neuronal networks. Procedia Comput Sci 147:562–566
5. Rudloff C, Leodolter M, Bauer D, Brög W, Kehnscherper K (2015) Influence of weather on
transport demand: case study from Viena, Austria. Transp Res Rec 2482:110–116
6. Bishop CM (1995) Neuronal networks for pattern recognition. Clarendon Press
Chapter 38
Green Aspects on Value Stream Mapping
Estefania Pilaloa-Morales and Pilar I. Vidal-Carreras
Abstract The use of Lean methodology today has expanded from the industrial
sector to other very diverse sectors. A key tool for optimizing processes is Value
Stream Mapping (VSM) as it allows for identifying activities that do not add value to
production. Due to its great applicability, the VSM could consider not only industrial
aspects to generate a global vision with a social, economic, and environmental focus
of the activities of the industry. This work focuses on the environmental approach
of the VSM, conducting a bibliographic review on this topic. When conducting the
analysis, it is identified that few cases consider these aspects mentioned. There are
companies that carry out a current VSM analyzing their energy systems, present
improvement solutions, and propose a future VSM to avoid waste. There are compa-
nies that also include social and economic variables to take greater advantage of the
benefits of Energy Value Stream Mapping (EVSM). It was concluded that VSM with
an environmental focus can and should be implemented to improve the sustainability
of productive activities. In addition, by using performance indicators, activities that
are measurable can be established and constantly propose improvement actions. For
future research, the concept of the sustainable circular economy could be intro-
duced, emphasizing the reduction to the minimum of waste in the sanitary process
and giving value to the current resources, materials and products used, proposing
significant improvements.
Keywords Lean · Value stream mapping · EVSM · Green manufacturing
E. Pilaloa-Morales
Departamento de Organización de Empresas, Universitat Politècnica de València, Valencia, Spain
e-mail: aepilmor@doctor.upv.es
P. I. Vidal-Carreras (B)
Departamento de Organización de Empresas, Grupo ROGLE, Universitat Politècnica de València,
Valencia, Spain
e-mail: pivicar@omp.upv.es
https://doi.org/10.1007/978-3-031-29382-5_38
384 E. Pilaloa-Morales and P. I. Vidal-Carreras
38.1 Introduction
Today, concern for the environment has taken on special relevance to develop produc-
tive activities that mark the starting point to develop the environmental responsibility
of companies in various sectors, promoting economic growth in a sustainable way
and with greater social cohesion around the environment. But this need is linked to
the proposal of relevant activities that allow optimizing the efficient use of resources
and proposing continuous improvement actions. For this, various methodologies
have been developed and improved over the years to generate a significant impact in
the socioeconomic area of the industries. Thus, it was born the Lean methodology,
as an innovative proposal, which considers the adequate management of resources
within any company or sector to reduce or eliminate activities that do not generate
added value. But for this process to occur, tools are required to obtain these opti-
mized results. One of them is the Value Stream Mapping (VSM) that not only allows
analyzing the material flows in the economic aspect. Instead, the proposal is born
to consider addressing the use of this tool considering social, economic, and envi-
ronmental indicators that generate sustainability [1]. Therefore, studies have been
proposed that show the generation of value in the productive processes of different
industries from a social perspective, considering people and the planet as its main
actors to generate activities that significantly impact the environment and imple-
menting tools to manage the resources consciously [2]. Likewise, the implementa-
tion of this tool in any sector allows obtaining valuable information on the resources,
technology, and machinery to be optimized, but it is necessary to awaken the interest
in the people (human factor) of any industry with which it is intended to work to
develop and implement key strategies for continuous improvement, achieving true
environmental integration that lasts in the long term [3, 4].
The order of this work is as follows. Section 38.2 presents the theoretical frame-
work and the research questions to be solved. Section 38.3 describes the materials
and methods used to develop the work. Section 38.4 addresses the relevant results of
the investigation, identifying important differences of each work analyzed. Finally,
Sect. 38.5 presents the conclusions and the proposed future lines of research.
38.2 Framework and Research Questions
The primary goal of lean manufacturing was to reduce the cost and to improve
productivity by eliminating wastes or non-value-added activities [5]. This optimiza-
tion methodology has been implemented in many industries over the years. Some
reviews and studies have focused on discovering the effect of lean on the environ-
ment, determining that the performance of the industry is achieved by reducing waste
and pollution, as well as the consideration of social needs [6]. Thus, lean and green
strategies were born as compatible initiatives by focusing both on the reduction of
waste and the efficient use of resources in companies [7].
38 Green Aspects on Value Stream Mapping 385
According to Gregori et al. [8], the VSM represents the entire production flow in
terms of value. The flow of value is an issue that considers all the actions necessary
to transform raw materials into a final product that can be properly marketed. The
mapping of the value chain is considered a business planning or communication tool.
The interpretation of the developed representation allows to create and define a real
state, managing to identify residues to get rid of them based on planned actions. As
Huang and Tomizuka [9] mention in their analysis of production processes, industries
are currently trying to include the VSM tool in their analysis of the value chain. Thus,
it is possible to analyze what activities are being carried out in the production lines
and identify opportunities for improvement focused on reducing costs, saving time,
and reducing inventory. According to Verma and Sharma [10], Energy Value Stream
Mapping (EVSM) is a tool based on the Value Stream Methodology. This has been
done by adding energy components in addition to the cost in VSM, and the same has
been analyzed with respect to time. Then, the EVSM identifies the level of energy
used and the waste at each production stage; therefore, it determines the opportunities
to promote energy conservation. Additionally, possible results are proposed that
consider the improvement options, establishing future scenarios using the EVSM.
The suggested model is used to establish an energy budget and establish saving
measures, expanding the diagnostic analysis of production processes. In addition, it
is a very versatile graphical tool for industries [11] that also allows simulations of the
initial information to propose a Future VSM in a simple and optimal way [12]. As
Bogdanski et al. [13] mention in their paper, EVSM is a powerful method that takes
into account more realistic situations and provides valuable information for both
manufacturing engineers and product designers. This is achieved due to the clear
composition of the energy demand established in accordance with all the relevant
subsystems of the factory and its close relationship with the operating states of the
equipment involved. A relevant criterion is the one presented by Schillig, Stock, and
Müller [14] in their analysis when referring to the contribution of time and energy
as added value criteria, indicating that if the VSM should be extended to an EVSM
with respect to the consumption of energy in production processes, it is not pertinent
to consider cycle times as an added value.
The questions to be solved in this work are the following:
RQ1. How much research on VSM with green aspects has been published and in
which areas?
RQ2. What green variables were considered?
RQ3. What are the future lines of research proposed by the studies developed so
far?
38.3 Materials and Methods
The protocol for the systematic literature review (SLR) has been generated including
the following steps: (a) conceptual discussion of the problem; (b) literature review
purpose; (c) search strategy; (d) paper selection criteria; (e) single paper analysis;
(f) descriptive analysis of the extracted database; (g) synthesis and content analysis
[15]. Summary of the phases (a) and (b) of this SLR is shown in the above section,
named framework.
According to the search strategy, the academic database searched was Web of
Science (ISI), concretely Science Citation Index Expanded (SCI-EXPANDED), and
Social Sciences Citation Index (SSCI). The search strategy was TS = (“green value
stream” OR “energy value stream” OR “sustain* value stream” OR “environmental
value stream” OR “clean value stream”) without time restriction. The term TS
searches in the title, abstract, author keywords, and keywords plus. The time span
was all the years. From this search, 39 papers were obtained, which after reading
their abstract, it was decided to include all of them. Therefore, there were 39 papers
that went to phase (e) carried out. Note that the references of the 39 papers cannot be
incorporated into the work due to space limitations, but are available upon request
to the authors. The stages of the method (f) descriptive analysis of the extracted
database; (g) synthesis and content analysis are related in the subsequent sections.
38.4 Results
This section focuses on solving the previously established research questions.

RQ1. How much research on VSM with green aspects has been published and in
which sectors?
As can be seen (Fig. 38.1), the publication of articles from the established areas
has had slight variations until it reached its maximum peak in 2016, identifying 7
published articles. From that moment on, there has been a trend in the publication
of the proposed topic that varies between 4 and 6 articles per year.
In Fig. 38.2, you can see the countries in which the research publications were
made, identifying that Germany is the country with the highest number of refer-
ences (eight references), followed by Indonesia (seven references), the USA (four
references), UK (three references), and the rest of countries with less than three
references.
8 7
6
No. References
6 5 5
4 4
4
2 2 2
2 1 1
0 0
0
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
Year
Fig. 38.1 Distribution of references by year of publication

10 8 7
4 3 3
5 2 2 2 2 1 1 1 1 1 1
0
INDIA
KINGDOM
CHINA
GERMANY
INDONESIA
ITALY
MALAYSIA
TURKEY
AUSTRALIA
AUSTRIA
BRAZIL
RUSSIA
SPAIN
SWEDEN
UNITED
STATES
UNITED
Fig. 38.2 References by country of publication
The research and experimentation sectors of the articles presented were analyzed
and identified. Table 38.1 shows the results of the comparative analysis. As can
be seen in Table 38.1, the largest number of investigations has been carried out
in manufacturing processes in general, presenting six references. The automotive
industry follows with five references. Next, three papers have been made in relation
to the electronics industry. Likewise, from the food industry, metal industry, and
electrical manufacturing, three investigations have been presented. In relation to the
railway industry and the furniture industry, two articles have been submitted. Finally,
for the rest of the industries, there is only one article published for each one.
RQ2. What green variables were considered?
Due to the large number of variables used in the analyzed research articles, it was
decided to group the variables depending on the main indicator to which they belong.
Thus, the following environmental indicators were taken into account: energy, air,
water, noise, materials that are part of the process (waste, garbage, raw material),
biodiversity, time, fuel indicators, and other variables that they cite in the research.
Figure 38.3 shows the number of indicators used in the research articles depending
on the group to which they belong.
Table 38.1 Comparison of the sectors of the articles analyzed

Total No. of references Sector/reference number
6 General manufacturing
5 Automotive
3 Electronic
3 Electrical
3 Food
3 Metal
2 Railway
2 Furniture industry
1 Metro cards, alcohol and sugar, mining, grinding balls, substrate, glass,
rubber, plastic injection molding, energy, milling, electroless nickel
plating
Energy 40
Time 25
Materials 15
Water 11
Other 5
Fuel 4
Air 4
Biodiversity 2
Noise 1
0 5 10 15 20 25 30 35 40 45
No. Green variables
Fig. 38.3 Number of green variables used
Additionally, it was decided to present a more detailed analysis of the green vari-
ables showing the group to which each one belongs and their units of measurement.
The results are shown in Table 38.2. In total, 78 green variables are presented.
RQ3. What are the future lines of research proposed by the studies developed so
far?
Table 38.2 Comparison between the green variables used in the analyzed articles
Group Green variables
Energy Lighting load, ventilation load, heating load, electricity, uninterrupted power
supply (UPS) load, plug load, cooling, energy consumption total and part, energy
cost part, power required for production, power required for holding, electrical
power process, electrical power idle, electrical power ramp-up, total energy
demand per day, total energy demand per PCB, energy waste, energy
consumption, electric energy, renewal energy, avg. processing load, avg. ramp-up
load, avg. transport load, shared transp. load shared, avg. third ord-dev. load,
theoretical absorption, actual absorption, specific absorption, value. add energy,
non-val. add energy, energy value-adding, energy non-value-adding, energy
transport, coefficient of energy consumption for holding function, thermal energy,
electrical driving force, GHG emission from energy consumption of the line, ratio
of renewable energy used, HVAC pressure
Transport Transport time coming, transport time going, time transport
Materials Material, waste, scrap, garbage, raw material waste, material utilization rate, mass
of restricted disposals, gaseous wastes generation, solid waste generation
Water Water, cooling water, water consumption, water cost, effluent treatment cost,
water treatment cost, water reused, total water consumption, water eutrophication,
process water waste
Other Chemical consumption, eutrophication potential, land contamination
Fuel Gas, emissions (el), emissions (gas), fuel, Heat
Air Air acidification, air pollution, air quality, compressed air
Biodiversity Biodiversity
Noise Noise
An interesting idea is proposed by Alvandi et al. [16] considering that more

research can be carried out taking into account the size of the company. Thus, the
proposed solutions can be grouped together and environmental performance and
productivity can be increased, as well as the value of GPI, green productivity index.
Complementing this idea, Shahbazi et al. [17] state that in addition to the size of the
company, the type of industry, type of product, and the types of auxiliary and residual
materials that intervene in the processes should be analyzed. In this way, it would
be possible to consider all value-added activities and non-value-added activities and
optimize resources. On the other hand, Jamil et al. [18] present a proposal that is
based on developing more studies for different industrial sectors that could favor the
creation of a list of sustainability metrics for each sector. In addition, these studies
could consider the DMAIC approach when using the Sus-VSM tool, managing to
develop maps of future states and studies to validate the exposed methodology. An
interesting perspective is the one presented by Mishra et al. [19] commenting that
future lines of research could use lean maturity assessments. Such assessments could
help organizations measure their current state in terms of strengths, opportunities,
weaknesses, and where they should be in the future. Similarly, Muñoz-Villamizar
et al. [20] indicate that future work should explore more support techniques and tools
(e.g., monitoring) to test new environmental metrics based on the interest of decision-
makers. Finally, as well as what Tasdemir and Gazo [21] comment in their article, it is
of vital importance to consider the extensions and immersions that the VSM has in the
industries since a good analysis leads to new developments. Therefore, future lines of
research should channel energy into the development of new methodologies, macros,
and tools that help achieve truly sustainable organizations and supply chains Also,
according to Sunk et al. [22] can rationalize the experiences and achieve proposals
for improvement, e.g., the ergonomics and maintenance value stream, achieving a
deeper study of times and the application of VSM in the first stages of product
development to optimize resources and eliminate non-value-added activities.
38.5 Conclusions
Today energy is one of the most used and necessary resources for any industry.
Its conscious use allows optimizing processes and generating savings focused on
sustainability. Green Value Stream Mapping offers an opportunity to eliminate
non-value-added activities that not only focus on productive development, but also
consider a sustainable aspect. Future research can focus analysis on sectors that are
constantly under pressure and focus solutions on environmental sustainability, for
example, the health and services, transport, logistics, and public sectors.
References
1. Septifani R, Astuti R, Akbar RN (2020) Green productivity analysis of tempeh chips production.
In: IOP conference series: earth and environmental science. Institute of Physics Publishing
2. Fischer J, Weinert N, Herrmann C (2015) Method for selecting improvement measures for
discrete production environments using an extended energy value stream model. In: Procedia
CIRP. Elsevier, pp 133–138
3. Garza-Reyes JA, Torres Romero J, Govindan K, Cherrafi A, Ramanathan U (2018) A PDCA-
based approach to environmental value stream mapping (E-VSM). J Clean Prod 180:335–348
4. Faulkner W, Badurdeen F (2014) Sustainable value stream mapping (Sus-VSM): methodology
to visualize and assess manufacturing sustainability performance. J Clean Prod 85:8–18
5. Jasti NVK, Kodali R (2014) A literature review of empirical research methodology in lean
manufacturing. Int J Oper Prod Manag 34:1080–1122
6. King AA, Lenox MJ (2001) Lean and green? An empirical examination of the relationship
between lean production and environmental performance. Prod Oper Manag 10:244–256
7. Mollenkopf D, Stolze H, Tate WL, Ueltschy M (2010) Green, lean, and global supply chains.
Int J Phys Distrib Logist Manag 40:14–41
8. Gregori F, Luzi A, Papetti A, Germani M, Camorani F (2016) An innovative tool to monitor and
represent energy value stream of a production system. In: Proceedings of the ASME interna-
tional design engineering technical conferences and computers and information in engineering
conference, 2016, vol 4. American Society Mechanical Engineers, New York
9. Huang Y, Tomizuka M (2017) Production flow analysis through environmental value stream
mapping: a case study of cover glass manufacturing facility. In: Procedia CIRP. Elsevier B.V.,
pp 446–450
10. Verma N, Sharma V (2016) Energy value stream mapping a tool to develop green manufac-
turing. In: Procedia engineering. Elsevier Ltd, pp 526–534
11. Keskin C, Kayakutlu G (2012) Value stream maps for industrial energy efficiency. In:
PICMET’12: proceedings—technology management for emerging technologies, pp 2824–
2831
12. Marimin, Darmawan MA, Martini S, Rahmanto AD (2015) Green productivity improvement
model for pre-processed rubber (bokar): case study at rubber small holders plantation. In:
Proceedings of 2014 2nd international conference on technology, informatics, management,
engineering and environment, TIME-E 2014. Institute of Electrical and Electronics Engineers
Inc., pp 135–140
13. Bogdanski G, Schoenemann M, Thiede S, Andrew S, Herrmann C (2013) An extended energy
value stream approach applied on the electronics industry. In: Advances in Production manage-
ment systems: competitive manufacturing for innovative products and services, AMPS 2012,
PT I. pp 65–72
14. Schillig R, Stock T, Müller E (2015) Energy value-stream mapping a method to visualize waste
of time and energy. IFIP advances in information and communication technology. Springer,
New York LLC, pp 609–616
15. Vidal-Carreras Pilar I, Garcia-Sabater Julio J, Marin-Garcia Juan A, Garcia-Sabater JP (2016)
Value stream mapping on healthcare. In: Proceedings 2015 international conference industrial
engineers system management IEEE IESM 2015, pp 272–276
16. Alvandi S, Li W, Schönemann M, Kara S, Herrmann C (2016) Economic and environmental
value stream map (E2VSM) simulation for multi-product manufacturing systems. Int J Sustain
Eng 9:354–362. https://doi.org/10.1080/19397038.2016.1161095
17. Shahbazi S, Kurdve M, Zackrisson M, Jönsson C, Kristinsdottir AR (2019) Comparison of
four environmental assessment tools in Swedish manufacturing: a case study. Sustainability
11:2173. https://doi.org/10.3390/su11072173
18. Jamil N, Gholami H, Saman MZM, Streimikiene D, Sharif S, Zakuan N (2020) DMAIC-based
approach to sustainable value stream mapping: towards a sustainable manufacturing system.
Econ Res Istraz 33:331–360
19. Mishra AK, Sharma A, Sachdeo M, Jayakrishna K (2019) Development of sustainable value
stream mapping (SVSM) for unit part manufacturing: a simulation approach. Int J Lean Six
Sigma 11:493–514. https://doi.org/10.1108/IJLSS-04-2018-0036
20. Muñoz-Villamizar A, Santos J, Garcia-Sabater JJ, Lleo A, Grau P (2019) Green value stream
mapping approach to improving productivity and environmental performance. Int J Product
Perform Manag 68(608–625):6
21. Tasdemir C, Gazo R (2018) A systematic literature review for better understanding of lean
driven sustainability. Sustain 10
22. Sunk A, Kuhlang P, Edtmayr T, Sihn W (2017) Developments of traditional value stream
mapping to enhance personal and organisational system and methods competencies. Int J Prod
Res 55:3732–3746
Index
A Competency-based interview, 3, 4, 6–8

Accidents, 57–60, 62–66 Construction Sector, 57, 59, 60, 63, 65
Additive manufacturing, 193, 196, 197 Control, 102–105
Advanced Management, 11–13, 15, 16 Cooling schedules, 19, 23, 25, 27
Air navigation system, 256, 257, 260, 261
Air traffic control, 256
Annualized hours, 111, 114, 115 D
Artificial Intelligence, 67, 129, 130, 134, Data model, 228, 229
135, 147, 148, 151, 156 Deep learning, 102
Automotive industry, 78, 80, 90 Deployment, 67–70, 72, 73
Aviation research, 256 Design of experiments, 137, 138, 143
Digital marketing, 147, 151, 156
Digital twin, 173–178, 180, 181, 203–209
B
Barriers, 303–313
Bibliometric analysis, 49
E
Blockchain, 41–46, 317–323, 343–346,
Efficiency, 363–366
348–350
Enablers, 303–313
Building Information Modeling (BIM),
Energy Value Stream Mapping (EVSM),
235, 236, 238–246, 248–250
383, 385
Business Model, 281, 283, 285–287
Entrepreneurship Education (EE), 29–32,
Business model archetypes, 281, 282
36
Business Model Innovation, 281
Entrepreneurship intention, 31, 36
Business model patterns, 281, 282,
European Union, 363, 366, 367
285–287
C F
Capability Maturity Model Integration, Flowshop, 129–132, 134
213, 216, 217, 222 Forecasting, 147, 380
Circular economy, 327–330, 332
Cognitive Ergonomics, 271, 273–277
Collaborative Economy, 264, 268, 269 G
Competence, 3–8 Gentrification, 264, 265
Competence-based assessment, 5 Global, 327–332
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer 393
Nature Switzerland AG 2023
https://doi.org/10.1007/978-3-031-29382-5
394 Index
I NetLogo, 19, 22–24, 27

Industry 4.0, 11–17, 30, 31, 41, 42, 45, 46, Network analysis, 54
119–121, 123, 127, 173–175, 213,
214, 218, 271–277, 293, 295, 303,
335, 337–340 O
Industry 4.0 in pharma, 340 Occupational health and safety, 273, 274,
Information Systems, 58, 60, 65 277
Innovation, 271, 273, 275–277 Online advertising, 147, 148, 151–157
Integration of BIM in Construction Operations management, 102–105
Management (IBCM), 243, 246 Operations planning, 204, 206
ISA-95, 215 Operations strategy, 327–332
Optimization, 19–21, 27, 119–122, 127,
183–185, 188, 189
J Overtime, 109–115
Job Shop, 93, 96, 97, 99, 129, 130, 133, 134
P
Packaging, 77, 79, 80, 82–86, 88, 90
K
Pharma sector, 335, 338
KNIME platform, 58
Physical internet, 363–367, 369
Platform Economy, 267
Predictive Quality, 67, 72, 73
L
Production planning, 119–121, 123, 125,
Labor flexibility, 109–111, 114, 115
127, 183, 184, 188, 193–197, 199,
Layout redesign, 165
200, 225–227, 229, 230
Lean, 383, 384, 389
Production scheduling, 137–141, 143
Lean manufacturing, 164, 183–186, 188,
Project Management, 235–237, 249, 250
189, 293, 296
Project Management Body of Knowledge
Logistics, 363–366, 369
(PMBoK), 235–237, 244, 246, 250
Lot-sizing, 77–80, 82, 86, 88, 90
Pyomo, 119–124, 126, 127
M R
Machine learning, 67–73, 130, 147, 148, Random Forest, 374, 375, 378–380
151–157, 373, 374 Raw materials, 77–80, 82, 84–86, 88, 90
Manufacturing, 67, 68, 72, 73, 346, 348 Reinforced learning, 102
Manufacturing Enterprise System Retail, 109–111, 113–115
Association, 216, 219
Manufacturing Execution System, 213,
215, 216, 222 S
Manufacturing Sector, 57, 58, 60, 64 Scheduling, 77–80, 82, 86, 88–90, 129,
Mathematical model, 226–228 133, 134, 173–178, 180, 181,
Mathematical programming, 353–355, 193–197, 200
359–361 Sequencing, 93, 94, 97–99
Maturity model, 11, 14–16 Share Bicycle, 374
Mixed Integer Linear Programming, 77, 79, Simulated annealing, 19–22, 25, 27
80, 82, 86, 87, 90 Simulation, 93–99
MPL, 119–124, 126, 127 Smart Contracts, 41, 42, 44, 45
Multiannual financial framework, 255, 256, STEM, 30–32
259, 261 Supply chain, 304–306, 308, 310, 327–332,
Multiskilling, 111, 114 343–348, 350, 353, 354, 361
Supply chain 4.0, 203, 206
Supply chain management, 293–295, 298,
N 317, 322, 365
Nesting, 194–198, 200 Supply chain traceability, 320
Index 395
Sustainability, 303–309, 311, 312, 327, University, 30, 31, 36

329–331, 363, 364, 366
Systematic Mapping Study, 147–149, 151,
156 V
Value Stream Mapping, 383–386, 389
T
Teaching resource, 20, 26 W
Teamwork, 3–8 Waste reduction, 170
3D printing, 196, 197
Time Driven Activity Base Costing, 49–51,
53–55 Y
Touristification, 264, 265 Youth Entrepreneurship (YE), 29, 36
Training, 11–17
Z
U Zero-defect manufacturing, 173–175, 177,
Uncertainty, 109–111, 183–185, 188, 189 180, 181, 203–205, 207–209

978 3 031 29382 5

Uploaded by

Copyright:

Available Formats

978 3 031 29382 5

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

978 3 031 29382 5

Uploaded by

Copyright:

Available Formats

Lecture Notes in Management and Industrial

Industry 4.0: The Power

Juan José Lavios Virginia Ahedo

ISSN 2198-0772 ISSN 2198-0780 (electronic)

Management and Industrial Engineering constitutes a broad and diverse discipline

for Industrial Management (AIM) and Asociación para el Desarrollo de la Inge-

Burgos, Spain Luis R. Izquierdo

Part I Education in Organizational Engineering

Part II Management Information Systems and Knowledge Management

7 Using Data Mining to Analyze Occupational Accidents

Part III Operations Research, Modelling and Simulation

Part IV Product Design, Industrial Marketing and Consumer Behaviour

Part V Production Planning and Control

Part VI Project and Process Management

Part VII Strategy, Innovation, Networks and Entrepreneurship

26 Dealing with the Externalities of the Sharing Economy: Effect

Part VIII Supply Chain Management and Logistics

Part IX Sustainability, Eco-efficiency and Quality Management

Amable Juarez-Tarraga, Cristina Santandreu-Mascarell,

A. Juarez-Tarraga · C. Santandreu-Mascarell · M. V. Fuster-Estruch

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 3

Keywords Teamwork · Competence · Competence-based assessment ·

1.1.1 Detected Need and Proposal for Action

There is an inherent difficulty in assessing transversal competences (TCs) [1–4]. This

1.1.2 Context Description

According to Aguilar Botija [16], “teamwork implies creating and developing a

1.2 Proposal for the Development of the Good Practice

It is proposed to develop the experience in the course Business Management. This

Specifically, it has been considered appropriate to develop the experience in the

1.3 Contribution/Relevance of the Developed Protocol

• Detecting and valuing TC Teamwork in a colleague or classmate.

1.4 Action Plan/Protocol

1.4.2 Analysis of Information and Evaluation of the Results

• Clarify your strategy for greater personal effectiveness.

1.5 Limitations, Future Research

València [cited 2021 Feb 20], pp 1012–27. Available from: http://ocs.editorial.upv.es/index.

J. I. Igartua , J. Retegi , and J. A. Eguren

Keywords Industry 4.0 · Training · Advanced management · Maturity model

J. I. Igartua (B) · J. Retegi · J. A. Eguren

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 11

The ability of company managers to assimilate these opportunities and changes

2.2 Industry 4.0 Training

2.3 Proposed Approach for Industry 4.0 Training

Fig. 2.1 Industry 4.0 training approach

3. Innovation through Industry 4.0: Development of innovations through Industry

2.4 Training Process and Results

Fig. 2.2 Industry 4.0 training process

Compliance with Ethical Standards

8. Mourtzis D (2018) Development of skills and competences in manufacturing towards education

José Ignacio Santos , María Pereda , Virginia Ahedo ,

Abstract Simulated annealing is one of the most popular metaheuristic optimization

Keywords Simulated annealing · Optimization · Teaching resource · NetLogo ·

J. I. Santos (B) · V. Ahedo · J. M. Galán

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 19

Equation (3.1) simulates a thermodynamic system from a sequence of states at a

The simulated annealing algorithm implements the exact Metropolis acceptance

3.2 Annealing Schedule Configuration

• Exponential or geometric decay, where the temperature is decreased in successive

• Logarithmic cooling schedule or Boltzmann annealing [9], an extraordinarily slow