LOGISTICS MANAGEMENT

Theory Section – Alessandro Perego – Year 2019-20

Annotations
The parts coloured in grey like this one are the notes written during the
lessons, while the black ones are taken from the slides given by the
professor. Inside this document there are the theoretical topics of the course.
The information written in this document are not enough to pass the exam,
there are also some books suggested by the professor, some additional
activities such as seminars and visits and tutorship sessions, and obviously
the content of the course may change during the years.

Michela Beraldo
beraldomichela97@gmail.com
 Summary

01. Objectives, Teaching methods, Assessment ................................................................. 1

0 AGENDA ...................................................................................................................................... 1
1 OBJECTIVES, TEACHING METHODS, ASSESSMENT ................................................................................ 1
ILOS (INTENDED LEARNING OUTCOMES) ................................................................................................. 1
TEACHING METHODOLOGIES ................................................................................................................. 1
TEACHING METHODOLOGIES AND ILOS.................................................................................................... 2
ASSESSMENT ..................................................................................................................................... 2
1.4.1 Critical Thinking................................................................................................................................................................................ 2
ASSESSMENT AND ILOS ....................................................................................................................... 3
2 PROGRAMME AND CONTENTS ........................................................................................................... 3
ASSUMPTIONS ................................................................................................................................... 3
BASICS OF LOGISTICS MANAGEMENT ....................................................................................................... 3
2.2.1 Logistics execution ........................................................................................................................................................................... 3
2.2.2 Logistics planning ............................................................................................................................................................................. 3
2.2.3 Performance measures ................................................................................................................................................................... 4
THE PROGRAMME OF THE LOGISTICS MANAGEMENT COURSE .......................................................................... 4
LOGISTIC MANAGEMENT ...................................................................................................................... 4
2.4.1 Logistic System Design and Strategy............................................................................................................................................... 4
2.4.2 Logistics Planning (advanced) ......................................................................................................................................................... 4
2.4.3 Logistics Execution (recap of basics + advanced topics) ................................................................................................................ 5
SUPPLY CHAIN COLLABORATION ............................................................................................................. 5
3 SCHEDULE OF THE COURSE .............................................................................................................. 5
OUT-OF-CLASSROOM LESSONS ............................................................................................................... 5
3.1.1 Tutorship sessions............................................................................................................................................................................ 5
3.1.2 Virtual visits to Distribution Centers ............................................................................................................................................... 5
3.1.3 The Esselunga Logistics Challenge .................................................................................................................................................. 6
3.1.4 Talks on “Sustainable and Fair Logistics" ........................................................................................................................................ 6
TEACHING MATERIAL ........................................................................................................................... 6
3.2.1 Suggested Bibliography(*):.............................................................................................................................................................. 6
3.2.2 How to find us .................................................................................................................................................................................. 6

02. Introduction to Logistics and Supply Chain Management ............................................. 7

1 INTRODUCTION............................................................................................................................. 7
INTENDED LEARNING OBJECTIVES ........................................................................................................... 8
AGENDA........................................................................................................................................... 8
2 LOGISTICS MANAGEMENT ............................................................................................................... 8
2.1.1 Council of Supply Chain Management Professionals (www.cscmp.org) ...................................................................................... 8
THE EVOLUTION OF LOGISTICS .............................................................................................................. 9
THE EVOLUTION OF LOGISTICS: DISTRIBUTION .......................................................................................... 9
THE EVOLUTION OF LOGISTICS: INTEGRATED LOGISTICS .............................................................................. 9
2.4.1 NCPD (National Council of Physical Distribution Management), 1976 ......................................................................................... 9
2.4.2 CLM (Council of Logistics Management), 1985 ............................................................................................................................ 10
2.4.3 CLM (Council of Logistics Management), 1992 ............................................................................................................................ 10
THE EVOLUTION OF LOGISTICS: SUPPLY CHAIN MANAGEMENT ..................................................................... 10
2.5.1 CSCMP (Council of Supply Chain Management Professionals), nowadays................................................................................. 10
SUMMARY OF RADAR CHART ................................................................................................................ 11
3 MAIN TRENDS............................................................................................................................. 11
1. FROM THE OPTIMISATION OF SINGLE ACTIVITIES TO AN INTEGRATING FUNCTION ........................................... 11
3.1.1 Physical Distribution/Materials Management.............................................................................................................................. 13
3.1.2 Integrated Logistics ........................................................................................................................................................................ 13
2. FROM PHYSICAL DISTRIBUTION MANAGEMENT TO SUPPLY CHAIN MANAGEMENT .............................................. 14
3.2.1 The Supply Chain............................................................................................................................................................................ 14
3.2.2 The Food Supply Chain .................................................................................................................................................................. 14
3.2.3 The Pharma Supply Chain .............................................................................................................................................................. 14
3.2.4 The evolution of Logistics: Supply Chain Management ............................................................................................................... 14
3.2.5 Supply Chain Based Competition .................................................................................................................................................. 15
3.2.6 Performances of the FMCG Supply Chain .................................................................................................................................... 15

I
 Summary

3. FROM EXECUTION TO PLANNING TO DESIGN OF LOGISTICS SYSTEM ............................................................ 17

3.3.1 The 3 “stacks” of logistics management ....................................................................................................................................... 17
3.3.2 1. Logistics Execution ..................................................................................................................................................................... 17
3.3.3 2. Logistics Planning ....................................................................................................................................................................... 18
3.3.4 3. Logistic System design and strategy ......................................................................................................................................... 18
3.3.5 The performance measurement system ...................................................................................................................................... 19
4. FROM COST REDUCTION/OPTIMIZATION TO CUSTOMER SERVICE AND PROFIT OPTIMIZATION ............................. 20
3.4.1 (1) Cost reduction/optimization.................................................................................................................................................... 20
3.4.2 Importance of Logistics and Supply Chain Management ............................................................................................................ 20
3.4.3 The incidence of Logistic costs: An analysis by industry .............................................................................................................. 21
3.4.4 Cost of Logistics as % of sales ........................................................................................................................................................ 21
3.4.5 Globalization: Export grows more than GDP ................................................................................................................................ 21
3.4.6 The evolution of the market: trends in factors of production in Europe ................................................................................... 21
3.4.7 Cost of Logistics as % of sales in different industries................................................................................................................... 22
3.4.8 The incidence of Logistic costs: An analysis by Supply Chain – the textile/apparel ................................................................... 22
3.4.9 (2) Customer service and profit optimization .............................................................................................................................. 22
3.4.10 Relation between logistic costs and service level ........................................................................................................................ 22
5. FROM DIRECT FLOWS OF GOODS TO DIRECT/INDIRECT (REVERSE) FLOWS OF INFORMATION, SERVICES, MONEY AND
REVERSE LOGISTICS OF GOODS ............................................................................................................ 23
3.5.1 Information .................................................................................................................................................................................... 23
3.5.2 Customer service ........................................................................................................................................................................... 24
3.5.3 Reverse logistics ............................................................................................................................................................................. 24
3.5.4 The “weight” of marketing mix variables ..................................................................................................................................... 25
3.5.5 Service level expected by the customers of the Logistic Service Providers (LSPs) ..................................................................... 25

4 CONCLUSIONS ............................................................................................................................ 26
4.1.1 Supplementary Readings ............................................................................................................................................................... 26
4.1.2 Relevant websites .......................................................................................................................................................................... 26
4.1.3 Challenge questions....................................................................................................................................................................... 26

03. Logistics Management and Strategy ........................................................................... 27

0 INTRODUCTION........................................................................................................................... 27
LEARNING OBJECTIVES ...................................................................................................................... 27
AGENDA......................................................................................................................................... 27
1 LOGISTICS AND THE MARKETING MIX .............................................................................................. 27
LOGISTICS MANAGEMENT: OBJECTIVES .................................................................................................. 27
LOGISTICS AND THE SOURCES OF VALUE ................................................................................................. 28
2 THE IMPACT ON ROA (RETURN ON ASSETS) .................................................................................... 28
(1) REDUCING LOGISTICS COSTS .......................................................................................................... 29
(2) INCREASING THE REVENUES -> SERVICE LEVEL .................................................................................. 29
2.2.1 Classes of indicators ...................................................................................................................................................................... 29
2.2.2 Availability related - Fill Rate ......................................................................................................................................................... 29
2.2.3 Time-related indicators ................................................................................................................................................................. 30
2.2.4 Quality/Accuracy related ............................................................................................................................................................... 30
2.2.5 Compound indicators .................................................................................................................................................................... 31
(3) INCREASING THE PROFIT: THE COST-SERVICE TRADE-OFF ...................................................................... 31
2.3.1 The “margin-service” relationship ................................................................................................................................................ 31
(4) THE WORKING CAPITAL ................................................................................................................. 31
2.4.1 Reducing the current assets .......................................................................................................................................................... 31
2.4.2 The Cash to Cash cycle (C2C) ........................................................................................................................................................ 32
2.4.3 Examples of C2C (Farris & Hutchison, 2002) ................................................................................................................................ 32
(5) REDUCING THE INVESTMENT IN FIXED ASSETS..................................................................................... 32
(6) MANAGING THE TRADE-OFFS .......................................................................................................... 33
EXERCISE ....................................................................................................................................... 33
2.7.1 Questions ....................................................................................................................................................................................... 33
2.7.2 Solution .......................................................................................................................................................................................... 33

II
 Summary

3 THE MAIN LOGISTIC STRATEGIES ................................................................................................... 34

1ST SUPPLY CHAIN STRATEGY (BASED ON HAU L. LEE’S FRAMEWORK) ............................................................ 34
3.1.1 Define the Supply Chain strategy - Method ................................................................................................................................. 34
3.1.2 1. Know your Supply Chain -> supply chain profile ...................................................................................................................... 34
3.1.2.1 Supply Chain Profile in different sectors Demand-side uncertainty ................................................................................................................................................ 35
3.1.3 2. Supply Chain Strategies (Strategic Approaches) ...................................................................................................................... 36
3.1.4 3. Consistency between Supply Chain profile and Supply Chain strategy .................................................................................. 37
3.1.5 1st Supply chain Strategy: a recap ................................................................................................................................................. 38
2ND SUPPLY CHAIN STRATEGY .............................................................................................................. 38
3.2.1 1. Time based strategies ................................................................................................................................................................ 38
3.2.2 2. Asset productivity strategies ..................................................................................................................................................... 38
3.2.3 3. Technology based strategies ..................................................................................................................................................... 39
3.2.4 4. Relationship based strategies ................................................................................................................................................... 39

4 CONCLUSIONS ............................................................................................................................ 39
SUPPLEMENTARY READINGS ................................................................................................................ 39
CHALLENGE QUESTIONS ..................................................................................................................... 40
CRITICAL THINKING .......................................................................................................................... 40
4.3.1 Rank in decreasing order of logistics costs (as % of revenues) ................................................................................................... 40
4.3.2 What is the difference between Order Cycle Time and Supply Chain Lead Time? .................................................................... 40
4.3.3 What is the relation between Order Fill Rate (OFR) and Inventories? ....................................................................................... 41
4.3.4 How are service level indicators chosen? And the target values for these indicators?............................................................. 42
4.3.5 What is the correct use of the ROA model regarding Logistics Strategy? .................................................................................. 42
4.3.6 Is a logistics outsourcing policy more consistent with a lean or an agile supply chain strategy? ............................................. 43

04. Warehouses: general overview .................................................................................. 45

0 INTRODUCTION........................................................................................................................... 45
LEARNING OBJECTIVES ...................................................................................................................... 45
AGENDA......................................................................................................................................... 45
1 ROLE OF WAREHOUSES ................................................................................................................. 45
THE DISTRIBUTION SYSTEM ................................................................................................................ 45
THE LOGISTIC CHANNEL ..................................................................................................................... 45
BASIC WAREHOUSING DECISIONS.......................................................................................................... 46
WAREHOUSES: THE BASIC FUNCTIONS ................................................................................................... 47
2 TYPES OF WAREHOUSES ............................................................................................................... 48
TRANSIT POINTS .............................................................................................................................. 49
2.1.1 The role of transit point................................................................................................................................................................. 49

3 WAREHOUSE OPERATIONS ............................................................................................................ 49

THE MAIN “FLOW MANAGEMENT” ACTIVITIES ........................................................................................... 49
3.1.1 The inbound flow ........................................................................................................................................................................... 49
3.1.2 The outbound flow ........................................................................................................................................................................ 49
3.1.3 The “new” value-adding activities ................................................................................................................................................ 50
THE MAIN FUNCTIONAL AREAS OF THE WAREHOUSE ................................................................................... 50
3.2.1 The chart of the main flows .......................................................................................................................................................... 50
3.2.2 Warehouse Management System ................................................................................................................................................ 51

4 STORAGE SYSTEMS ...................................................................................................................... 51

PARAMETERS................................................................................................................................... 51
4.1.1 Design parameters ......................................................................................................................................................................... 51
4.1.2 Descriptive parameters ................................................................................................................................................................. 51
4.1.3 Cost indicators ............................................................................................................................................................................... 51
4.1.4 Impact of storage/handling systems on performances ............................................................................................................... 52
PALLET HANDLING TRUCKS .................................................................................................................. 52
4.2.1 Walkie stackers .............................................................................................................................................................................. 52
4.2.2 Front loading forklift trucks ........................................................................................................................................................... 52
4.2.3 Side loading Forklift trucks ............................................................................................................................................................ 52
4.2.4 AGV (Automated Guided Vehicles) ............................................................................................................................................... 53
TYPOLOGIES OF STORAGE SYSTEMS ....................................................................................................... 53
4.3.1 1. Block stacking ............................................................................................................................................................................. 53
4.3.2 2. Drive-in or drive-through systems (Rail tracks) ........................................................................................................................ 53
4.3.3 3. Selective pallet rack systems ..................................................................................................................................................... 54
4.3.4 4. Flow rack systems ...................................................................................................................................................................... 55
COMPARISON BETWEEN TRUCKS SERVING SINGLE DEEP SELECTIVE RACKS ........................................................ 55

III
 Summary

5 CONCLUSIONS ............................................................................................................................ 55
SUPPLEMENTARY READINGS ................................................................................................................ 55
CHALLENGE QUESTIONS ..................................................................................................................... 55

05. Management Policies and Design Methodology ......................................................... 57

0 INTRODUCTION........................................................................................................................... 57
LEARNING OBJECTIVES ...................................................................................................................... 57
1 MANAGEMENT POLICIES ............................................................................................................... 57
STORAGE SYSTEM MANAGEMENT POLICIES............................................................................................... 57
OPERATIVE CYCLE TYPES – CONVENTIONAL TRUCKS ................................................................................... 57
1.2.1 Comparison between single and dual command cycles .............................................................................................................. 58

2 RETRIEVING AND ACCESS INDEX .................................................................................................... 58

ACCESS INDEX EXAMPLES.................................................................................................................... 58
2.1.1 Example 1 ....................................................................................................................................................................................... 58
2.1.2 Example 2 ....................................................................................................................................................................................... 58
2.1.3 Example 3 ....................................................................................................................................................................................... 58
2.1.4 Allocation of the UL depending on the AI .................................................................................................................................... 59
ALLOCATION POLICIES ....................................................................................................................... 59
2.2.1 Dedicated storage .......................................................................................................................................................................... 59
2.2.2 Class Based storage ....................................................................................................................................................................... 59
ALLOCATION OF THE ITEMS IN A CLASS BASED STORAGE SYSTEM FOLLOWING AI ............................................... 59
2.3.1 ABC curve “Accesses-UL locations” .............................................................................................................................................. 60
CRITERIA TO FORM THE ZONES ............................................................................................................ 60
2.4.1 Shape of the zones......................................................................................................................................................................... 60
2.4.2 Number of classes.......................................................................................................................................................................... 60
PROS/CONS OF THE ALLOCATION CRITERIA ............................................................................................. 61
2.5.1 Example of SC required ................................................................................................................................................................. 61

3 STORAGE SYSTEM DESIGN PRINCIPLES ............................................................................................ 61

4 STORAGE SYSTEM DESIGN METHOD ................................................................................................. 62
ASSUMPTIONS ................................................................................................................................. 62
THE DESIGN FRAMEWORK ................................................................................................................... 62
1. DETERMINATION OF THE DESIGN PARAMETERS ..................................................................................... 62
4.3.1 TC - Analysis of the physical flows ................................................................................................................................................ 62
4.3.2 SC - Inventory analysis ................................................................................................................................................................... 63
4.3.3 SC assessment - The forecasting factor ........................................................................................................................................ 64
2. LAYOUT DESIGN ........................................................................................................................... 64
4.4.1 1. Layout typologies ....................................................................................................................................................................... 64
4.4.2 2. Determination of the storage area ........................................................................................................................................... 65
4.4.3 3. Optimal shape of the storage area ........................................................................................................................................... 66
4.4.4 4. Storage area design ................................................................................................................................................................... 67
3. THROUGHPUT CAPACITY ASSESSMENT ................................................................................................ 67
5 CONCLUSIONS ............................................................................................................................ 68
SUPPLEMENTARY READINGS ................................................................................................................ 68
CHALLENGE QUESTIONS ..................................................................................................................... 68

06. Automated Warehouses ............................................................................................ 69

0 INTRODUCTION........................................................................................................................... 69
LEARNING OBJECTIVES....................................................................................................................... 69
AGENDA......................................................................................................................................... 69
1 TECHNICAL FEATURES .................................................................................................................. 69
AUTOMATED WAREHOUSES ................................................................................................................. 69
1.1.1 Aisle width ...................................................................................................................................................................................... 69
1.1.2 Area Utilization Rate ...................................................................................................................................................................... 69
1.1.3 S/R machine features..................................................................................................................................................................... 70
1.1.4 S/R machine kinematics................................................................................................................................................................. 70
MAIN DIFFERENCES WITH THE TRADITIONAL STORAGE SYSTEMS.................................................................... 70
1.2.1 Comparison among pallet racks storage systems ........................................................................................................................ 70

IV
 Summary

RACK REPRESENTATION ..................................................................................................................... 71

1.3.1 From “spatial” to “time” coordinates ........................................................................................................................................... 71
1.3.2 Example: from spatial to time coordinates .................................................................................................................................. 71
1.3.3 Rack shape factor ........................................................................................................................................................................... 71
SYSTEM CONFIGURATIONS .................................................................................................................. 72
1.4.1 Investment costs ............................................................................................................................................................................ 72

2 MANAGEMENT POLICIES ............................................................................................................... 72

OPERATIVE CYCLES TYPES .................................................................................................................. 73
2.1.1 Single command cycles – AS/RS .................................................................................................................................................... 73
2.1.2 Dual command cycles – AS/RS ...................................................................................................................................................... 73
2.1.3 Cycle time – AS/RS ......................................................................................................................................................................... 73
2.1.4 Fixed cycle time – AS/RS ................................................................................................................................................................ 73
2.1.5 Variable cycle time – AS/RS ........................................................................................................................................................... 73
COMPARISON BETWEEN SINGLE AND DUAL COMMAND CYCLES ....................................................................... 74
OPTIMIZATION OF THE DUAL COMMAND CYCLES ........................................................................................ 74
2.3.1 No Cost Zone .................................................................................................................................................................................. 74
CLASS BASED STORAGE ...................................................................................................................... 74
2.4.1 Optimization of the class based storage....................................................................................................................................... 75
2.4.2 Number of classes.......................................................................................................................................................................... 75
2.4.3 UL allocation in a class-based storage system based on the AI .................................................................................................. 75

3 APPLICATION FIELDS ................................................................................................................... 75

AREA OF CONVENIENCE OF THE AUTOMATED WAREHOUSES .......................................................................... 75
3.1.1 Main drawbacks ............................................................................................................................................................................. 75

4 DESIGN METHODOLOGIES............................................................................................................. 76
ASSUMPTIONS ................................................................................................................................. 76
VARIABLES AND OBJECTIVE ................................................................................................................. 76
AUTOMATED WAREHOUSE DESIGN ......................................................................................................... 76
1. DETERMINATION OF THE DESIGN PARAMETERS ..................................................................................... 76
2. LAYOUT DESIGN ........................................................................................................................... 76
4.5.1 Bay design ...................................................................................................................................................................................... 77
4.5.2 Maximum dimensions of the racks ............................................................................................................................................... 77
4.5.3 Minimum number of aisles (S/R machines) ................................................................................................................................. 77
4.5.4 Storage area ................................................................................................................................................................................... 77
3. THROUGHPUT CAPACITY ASSESSMENT ................................................................................................ 78
4.6.1 Relationship between TC and # aisles .......................................................................................................................................... 78
4.6.2 Throughput capacity assessment ................................................................................................................................................. 78
ASSESSMENT OF THE OPERATIVE CYCLE TIMES ......................................................................................... 79
4.7.1 FEM 9851 rule – Cycle time assessment ...................................................................................................................................... 79

5 NEW TRENDS .............................................................................................................................. 79

AUTONOMOUS VEHICLE STORAGE AND RETRIEVAL SYSTEM ......................................................................... 79
5.1.1 Why is this technology interesting? .............................................................................................................................................. 80
5.1.2 Comparison between AS/R and AVS/R systems in case of small unit loads e.g. totes) ............................................................. 80
AUTOMATED FORKLIFTS ..................................................................................................................... 80
6 CONCLUSIONS ............................................................................................................................ 80
SUPPLEMENTARY READINGS................................................................................................................. 80
CHALLENGE QUESTIONS ..................................................................................................................... 80
CRITICAL THINKING .......................................................................................................................... 81

07. Order Picking and Assembly: systems and design principles ....................................... 83
0 INTRODUCTION........................................................................................................................... 83
LEARNING OBJECTIVES....................................................................................................................... 83
AGENDA......................................................................................................................................... 83
1 PICKING PROCESS ....................................................................................................................... 83
PICKING ........................................................................................................................................ 84
1.1.1 Purchase order: an example ......................................................................................................................................................... 84
1.1.2 Picking List (or Warehouse Order) ................................................................................................................................................ 84
1.1.3 Picking and Order Assembly Process ............................................................................................................................................ 84
1.1.4 KPIs of the picking process ............................................................................................................................................................ 85
1.1.5 Design parameters ......................................................................................................................................................................... 85
1.1.6 Complexity of the picking process ................................................................................................................................................ 85

V
 Summary

1.1.7 Order cycle, order assembly, picking & sorting ........................................................................................................................... 86

1.1.8 Sorting ............................................................................................................................................................................................ 86
1.1.9 Packaging and transport order consolidation .............................................................................................................................. 86
1.1.10 Importance of order assembly ...................................................................................................................................................... 86
1.1.11 General trends in Warehouse Management................................................................................................................................ 86
CLASSIFICATION OF ORDER PICKING SYSTEMS ........................................................................................ 86
2 MANUAL PICKING SYSTEMS .......................................................................................................... 87
PICKER-TO-PARTS PICKING SYSTEMS..................................................................................................... 87
2.1.1 The basic idea................................................................................................................................................................................. 87
2.1.2 Ingredients ..................................................................................................................................................................................... 87
2.1.3 A classification framework ............................................................................................................................................................ 87
2.1.4 Examples of Picker-to-Parts .......................................................................................................................................................... 88
2.1.5 Configurations ................................................................................................................................................................................ 89
2.1.6 Design and management issues .................................................................................................................................................... 89
“PICK TO BOX” SYSTEM ...................................................................................................................... 89
2.2.1 The basic idea................................................................................................................................................................................. 89
2.2.2 Ingredients ..................................................................................................................................................................................... 89
2.2.3 Sorter .............................................................................................................................................................................................. 90
2.2.4 Configurations ................................................................................................................................................................................ 90
2.2.5 “Pick to belt/box” systems ............................................................................................................................................................ 90
2.2.6 Examples of Pick-to-box ................................................................................................................................................................ 90
“PICK AND SORT” SYSTEM .................................................................................................................. 91
2.3.1 The basic idea................................................................................................................................................................................. 91
2.3.2 Examples of Pick-and-sort ............................................................................................................................................................. 91
2.3.3 Design/management issues .......................................................................................................................................................... 91
PARTS-TO-PICKER PICKING SYSTEMS ..................................................................................................... 91
2.4.1 The basic idea................................................................................................................................................................................. 91
2.4.2 Configurations ................................................................................................................................................................................ 92
2.4.3 Design/management issues .......................................................................................................................................................... 92
2.4.4 Example – Sizing of the capacity of handling system in Parts-to-Pickers ................................................................................... 92
2.4.5 A classification framework ............................................................................................................................................................ 92
2.4.6 Examples of Parts-to-picker .......................................................................................................................................................... 92
2.4.7 Autostore........................................................................................................................................................................................ 93
2.4.8 KIVA robots..................................................................................................................................................................................... 93
“AUTOMATION” OF THE MANUAL PICKING SYSTEMS.................................................................................... 93
2.5.1 Examples of automation of the manual picking systems ............................................................................................................ 94

3 AUTOMATED PICKING SYSTEMS ..................................................................................................... 94

3.1.1 Configurations ................................................................................................................................................................................ 94
3.1.2 Design/Management issues .......................................................................................................................................................... 94
EXAMPLES OF AUTOMATED PICKING SYSTEMS ........................................................................................... 95
3.2.1 Singularization & Sorting ............................................................................................................................................................... 95
3.2.2 Identify and pick............................................................................................................................................................................. 95
SUMMARY OF PICKING SYSTEMS ........................................................................................................... 95
3.3.1 Five different KPIs .......................................................................................................................................................................... 96
3.3.2 Picker-to-parts ............................................................................................................................................................................... 96
3.3.3 Pick-to-box ..................................................................................................................................................................................... 96
3.3.4 Pick-to-sort ..................................................................................................................................................................................... 96
3.3.5 Parts-to-picker ............................................................................................................................................................................... 96
3.3.6 Automation of manual systems .................................................................................................................................................... 96
3.3.7 Automatic systems......................................................................................................................................................................... 96

4 PRINCIPLES OF PICKING SYSTEM DESIGN ....................................................................................... 97

MAIN DESIGN AND MANAGEMENT DECISIONS ........................................................................................... 97
4.1.1 1.1 - Segmentation of the picking problem in sub-problems (clusters of homogeneous items) .............................................. 97
4.1.2 Key Data: Order Profile .................................................................................................................................................................. 97
4.1.3 2.1 - Forward-Reserve Problem .................................................................................................................................................... 98
4.1.4 3 – Management policies .............................................................................................................................................................. 98
4.1.5 Key Performance Indicators .......................................................................................................................................................... 98
4.1.6 Selection among the available picking systems ........................................................................................................................... 99
4.1.7 Design parameters ......................................................................................................................................................................... 99
EXAMPLE – DESIGN PROCESS ............................................................................................................... 99
5 FORWARD RESERVE PROBLEM ....................................................................................................... 99
“PICKING STOCK” AND “GENERAL STOCK” ............................................................................................... 99
5.1.1 Picking from Forward stock ......................................................................................................................................................... 100
5.1.2 Examples ...................................................................................................................................................................................... 100

VI
 Summary

FORWARD RESERVE PROBLEM (FRP) .................................................................................................. 100

5.2.1 Hierarchy of decisions ................................................................................................................................................................. 100
5.2.2 Embedded problem: space allocation ........................................................................................................................................ 101
FRP POSSIBLE SOLUTIONS ............................................................................................................... 101
5.3.1 Equal space supply....................................................................................................................................................................... 101
5.3.2 Equal time supply......................................................................................................................................................................... 102
5.3.3 Replenishment frequency in the equal space or equal time supply ......................................................................................... 102
5.3.4 Optimal allocation policy ............................................................................................................................................................. 102
5.3.5 Efficiency index for the optimal solution .................................................................................................................................... 102

6 SEMINAR: ARTSANA GROUP ........................................................................................................ 103

PICKING SOLUTIONS ....................................................................................................................... 103
FORWARD RESERVE PROBLEM ........................................................................................................... 103
6.2.1 Module-selective racks ................................................................................................................................................................ 103

7 CONCLUSIONS .......................................................................................................................... 104

SUPPLEMENTARY READINGS .............................................................................................................. 104
CHALLENGE QUESTIONS ................................................................................................................... 104
CRITICAL THINKING ........................................................................................................................ 104

08. Order Picking and Assembly: Management policies ................................................. 107

0 INTRODUCTION......................................................................................................................... 107
LEARNING OBJECTIVES..................................................................................................................... 107
THE REFERENCE CASE (FOR THE MANAGEMENT POLICIES) .......................................................................... 107
AGENDA....................................................................................................................................... 107
1 BATCHING POLICIES .................................................................................................................. 107
BATCH PICKING PROS AND CONS......................................................................................................... 107
1.1.1 Pros1: Increase in the picking density ........................................................................................................................................ 108
1.1.2 Pros2: Overlapping of the order lines ......................................................................................................................................... 108
MAIN ISSUES................................................................................................................................. 108
1.2.1 Sorting .......................................................................................................................................................................................... 108
ZONE PICKING ............................................................................................................................... 109
1.3.1 Zone Picking with pre-defined areas .......................................................................................................................................... 109

2 ROUTING POLICIES.................................................................................................................... 109

MAIN ISSUES................................................................................................................................. 110
2.1.1 The problem ................................................................................................................................................................................. 110
2.1.2 The approaches............................................................................................................................................................................ 110
POSSIBLE SOLUTIONS ...................................................................................................................... 110
2.2.1 Traversal policy ............................................................................................................................................................................ 110
2.2.2 Return policy ................................................................................................................................................................................ 110
2.2.3 Mid-point return policy ............................................................................................................................................................... 110
2.2.4 Largest- gap return policy............................................................................................................................................................ 110
2.2.5 Optimal solution .......................................................................................................................................................................... 111
ASSESSMENT OF THE DIFFERENT ROUTING POLICIES ................................................................................ 111
2.3.1 First case....................................................................................................................................................................................... 111
2.3.2 Second case .................................................................................................................................................................................. 111

3 STORAGE POLICIES .................................................................................................................... 111

MAIN ISSUES................................................................................................................................. 112
ACCESS INDEX - BASED STORAGE ....................................................................................................... 112
3.2.1 Access index in the picking system ............................................................................................................................................. 112
3.2.2 ABC curve “Picking frequency – Space”...................................................................................................................................... 112
AI-BASED STORAGE - ROUTING POLICIES ............................................................................................. 113
3.3.1 AI-based storage with a traversal routing policy........................................................................................................................ 113
3.3.2 AI-based storage with a return routing policy............................................................................................................................ 113
CORRELATED STORAGE .................................................................................................................... 113
4 ASSESSMENT OF THE PICKING TIME .............................................................................................. 114
1. THE MAIN COMPONENTS OF THE PICKING TIME ................................................................................... 114
2. THE TRAVEL TIME IN A PICKING TOUR .............................................................................................. 114
4.2.1 The travel components (in the narrow-aisle configuration) .................................................................................................... 115
4.2.2 The within aisles travel distance ................................................................................................................................................. 115
4.2.3 The across aisles travel distance ................................................................................................................................................. 115
4.2.4 Travel distance (Example) ........................................................................................................................................................... 115

VII
 Summary

3. THE EXPECTED TRAVEL DISTANCE IN A PICKING TOUR ........................................................................... 115

4.3.1 Case A ........................................................................................................................................................................................... 116
4.3.2 Case B ........................................................................................................................................................................................... 117
4.3.3 Case C ........................................................................................................................................................................................... 117
4. DEFINITION OF A GENERAL FRAMEWORK ........................................................................................... 118
4.4.1 Customer order profile (P(N)) ..................................................................................................................................................... 118
4.4.2 The expected travel distance in a pick tour................................................................................................................................ 118
4.4.3 The expected travel distance in a picking tour .......................................................................................................................... 118

5 CONCLUSIONS .......................................................................................................................... 118

SUPPLEMENTARY READINGS .............................................................................................................. 118
CHALLENGE QUESTIONS ................................................................................................................... 118

09. Transportation: general overview............................................................................. 119

0 INTRODUCTION......................................................................................................................... 119
LEARNING OBJECTIVES..................................................................................................................... 119
AGENDA....................................................................................................................................... 119
1 TRANSPORTATION MODES ........................................................................................................... 119
THE LOGISTIC CHANNEL ................................................................................................................... 119
IMPORTANCE OF TRANSPORTATION IN SUPPLY CHAINS ............................................................................. 119
1.2.1 Transportation has an impact on service ................................................................................................................................... 120
1.2.2 Transportation has a major impact on costs .............................................................................................................................. 120
1.2.3 The impact of transportation on costs is growing ..................................................................................................................... 120
1.2.4 The importance of freight transportation is growing ................................................................................................................ 120
COMPLEXITY OF TRANSPORTATION ...................................................................................................... 120
DEFINITION .................................................................................................................................. 121
1.4.1 Units of measure .......................................................................................................................................................................... 121
VARIABLES DEFINING A TRANSPORTATION PROBLEM ................................................................................ 121
1.5.1 Variables related to the customers demand .............................................................................................................................. 121
1.5.2 Variables related to the logistics service market ....................................................................................................................... 122
MODES OF TRANSPORTATION ............................................................................................................ 122
1.6.1 Intermodal transportation .......................................................................................................................................................... 122
1.6.2 Comparison of transportation mode performances .................................................................................................................. 122
1.6.3 Adoption of transportation modes inside the EU ...................................................................................................................... 122

2 ROAD TRANSPORTATION............................................................................................................. 123

ROAD – FULL TRUCK LOAD ............................................................................................................... 123
ROAD – LESS THAN TRUCK LOAD ....................................................................................................... 124
2.2.1 Main phases of the LTL transportation....................................................................................................................................... 124
2.2.2 Typical logistics network of an express courier.......................................................................................................................... 125
2.2.3 Local distribution ......................................................................................................................................................................... 125
ROAD - SERVICE PROVIDERS ............................................................................................................. 125
ROAD - CHARGEABLE WEIGHT ........................................................................................................... 125
2.4.1 Notations ...................................................................................................................................................................................... 126
2.4.2 Example ........................................................................................................................................................................................ 126

3 RAIL TRANSPORTATION ............................................................................................................. 127

RAIL - TRADITIONAL SERVICE ............................................................................................................ 127
3.1.1 Main characteristics..................................................................................................................................................................... 127
3.1.2 Different types of cars ................................................................................................................................................................. 127
3.1.3 Main application fields ................................................................................................................................................................ 128
RAIL – SERVICE PROVIDERS.............................................................................................................. 128
MODES OF TRANSPORTATION RAIL – EFFICIENCY ................................................................................... 128
4 WATER TRANSPORTATION .......................................................................................................... 128
TYPES OF SHIPS ............................................................................................................................. 129
ISO CONTAINERS .......................................................................................................................... 129
TYPES OF WATER TRANSPORTATION .................................................................................................... 129
4.3.1 Coasting / Internal waterways .................................................................................................................................................... 129
4.3.2 Transoceanic shipments .............................................................................................................................................................. 129
WATER – SERVICE PROVIDERS .......................................................................................................... 130

VIII
 Summary

5 AIR TRANSPORTATION ............................................................................................................... 130

AIR – DIFFERENT AIRPLANE TYPES ...................................................................................................... 130
5.1.1 Airplane load: narrow body-wide body ...................................................................................................................................... 130
5.1.2 Airplane load: example ................................................................................................................................................................ 131
AIR TRANSPORT UNIT LOADS ............................................................................................................. 131
AIR – SERVICE PROVIDERS ............................................................................................................... 131
6 INTERMODAL ............................................................................................................................ 132
INLAND TERMINAL .......................................................................................................................... 132
INTERMODAL – PIGGYBACK ............................................................................................................... 132
6.2.1 Intermodal Transport Unit (I.T.U.) .............................................................................................................................................. 132
6.2.2 Loading ......................................................................................................................................................................................... 132
6.2.3 Handling equipment for containers ............................................................................................................................................ 132
INTERMODAL – ROAD VS INTERMODAL................................................................................................. 133
FISHYBACK ................................................................................................................................... 133
6.4.1 Intermodal Transport Unit (I.T.U.) .............................................................................................................................................. 133

7 FOCUS ON FTL ROAD TRANSPORTATION ....................................................................................... 134

INTRODUCTION ............................................................................................................................. 134
7.1.1 Learning objectives ...................................................................................................................................................................... 134
7.1.2 Agenda.......................................................................................................................................................................................... 134
COST STRUCTURE OF ROAD TRANSPORTATION........................................................................................ 134
7.2.1 Economies of scale ...................................................................................................................................................................... 134
7.2.2 Key variable .................................................................................................................................................................................. 134
7.2.3 Example ........................................................................................................................................................................................ 135
7.2.4 In brief .......................................................................................................................................................................................... 135
OPTIMIZATION OF ROAD TRANSPORTATION ........................................................................................... 135
7.3.1 Key variables ................................................................................................................................................................................ 135
WAYS TO OPTIMIZE ........................................................................................................................ 137
7.4.1 Solutions that affect the annual travel distance per truck ........................................................................................................ 137
7.4.2 Solutions that affect the truck capacity utilization .................................................................................................................... 138

8 CONCLUSIONS .......................................................................................................................... 139

SUPPLEMENTARY READINGS .............................................................................................................. 139
CHALLENGE QUESTIONS ................................................................................................................... 139

10. Logistics & B2C eCommerce ..................................................................................... 141

0 AGENDA .................................................................................................................................. 141
1 GENERAL OVERVIEW .................................................................................................................. 141
DEFINITION OF B2C E-COMMERCE ...................................................................................................... 141
B2C ECOMMERCE DIFFUSION ............................................................................................................ 141
1.2.1 Importance of the online channel .............................................................................................................................................. 141
1.2.2 Types of merchants ..................................................................................................................................................................... 142

2 KEY FEATURES AND PERFORMANCES.............................................................................................. 142

LOGISTICS AS A CRITICAL SUCCESS FACTOR ........................................................................................... 142
2.1.1 Factors .......................................................................................................................................................................................... 142
2.1.2 Logistic impact ............................................................................................................................................................................. 142
LOGISTICS PROCESS IN B2C E-COMMERCE ............................................................................................ 143
2.2.1 Order profile................................................................................................................................................................................. 143
2.2.2 Point of delivery ........................................................................................................................................................................... 144
2.2.3 Customer service ......................................................................................................................................................................... 144
2.2.4 Stock out management ............................................................................................................................................................... 145
2.2.5 Tracking e tracing......................................................................................................................................................................... 145

3 DESIGN OF LOGISTICS STRATEGIES............................................................................................... 146

KEY CHOICES ................................................................................................................................ 146
INVENTORY OWNERSHIP: MODELS....................................................................................................... 146
3.2.1 Merchant without inventory ....................................................................................................................................................... 147
3.2.2 Merchant with inventory............................................................................................................................................................. 147
DISTRIBUTION NETWORK DESIGN ....................................................................................................... 147
DELIVERY NETWORK ALTERNATIVES FOR TRADITIONAL RETAILERS ............................................................... 148
3.4.1 Distribution network design: alternatives .................................................................................................................................. 150
3.4.2 Picking system configuration ...................................................................................................................................................... 151

IX
 Summary

4 4. LAST MILE DELIVERY .............................................................................................................. 151

5 CHALLENGE QUESTIONS .............................................................................................................. 151

11. Distribution Network Design .................................................................................... 153

0 INTRODUCTION......................................................................................................................... 153
LEARNING OBJECTIVES..................................................................................................................... 153
AGENDA....................................................................................................................................... 153
1 DISTRIBUTION SYSTEM (RECAP) .................................................................................................. 153
DEFINITION .................................................................................................................................. 153
1.1.1 Trade channel and logistic channel: inter-relationships ............................................................................................................ 153
1.1.2 Logistic channel............................................................................................................................................................................ 154
1.1.3 Logistic channel: functions .......................................................................................................................................................... 154
PRINCIPLES OF DISTRIBUTION NETWORK DESIGN .................................................................................... 154
1. PROBLEM SETTING AND DECISION FRAMEWORK................................................................................... 155
1.3.1 Distribution design problem (in a nutshell) ................................................................................................................................ 155
1.3.2 Contextual variables [explained later] ........................................................................................................................................ 155
1.3.3 Design decisions........................................................................................................................................................................... 155
1.3.4 Unit of analysis & Sales and Marketing strategies + Sourcing strategies ................................................................................. 155
1.3.5 Product types/features ................................................................................................................................................................ 156
1.3.6 Objective function ....................................................................................................................................................................... 156
1.3.7 The decision framework .............................................................................................................................................................. 156
2. DISTRIBUTION PROBLEM PROFILE.................................................................................................... 156
3. TYPOLOGIES OF DISTRIBUTION NETWORK ......................................................................................... 157
1.5.1 Direct shipment ........................................................................................................................................................................... 157
4. DESIGN PRINCIPLES & 5. EXAMPLES ................................................................................................ 157
2 1-ECHELON DISTRIBUTION NETWORK ........................................................................................... 158
ROLE OF CENTRAL WAREHOUSES IN A 1-ECHELON NETWORK ...................................................................... 158
EMPIRICAL EVIDENCE - EXAMPLES....................................................................................................... 159
2.2.1 Example: Bassetti ......................................................................................................................................................................... 159
2.2.2 Example: 3M Europe ................................................................................................................................................................... 159
2.2.3 Example: Unilever Home & Personal Care ................................................................................................................................. 159
EXAMPLES SUMMARY ....................................................................................................................... 160
DESIGN PRINCIPLES ........................................................................................................................ 160
2.4.1 Number of central warehouses .................................................................................................................................................. 161

3 TWO-ECHELONS DISTRIBUTION NETWORKS.................................................................................... 163

ROLE OF REGIONAL WAREHOUSES (2ND TIER OF THE DISTRIBUTION NETWORK) .............................................. 163
EMPIRICAL EVIDENCE - EXAMPLES....................................................................................................... 164
3.2.1 Example: Ferrero Italia ................................................................................................................................................................ 164
3.2.2 Example: Iveco Europe Distribution network of spare parts ..................................................................................................... 164
3.2.3 Example: Nestlé Italia (bakeries) ................................................................................................................................................. 164
EXAMPLES SUMMARY ....................................................................................................................... 165
DESIGN PRINCIPLES ........................................................................................................................ 165
4 1-ECHELON + TRANSIT POINT DISTRIBUTION NETWORK .................................................................. 166
ROLE OF TRANSIT POINTS................................................................................................................. 166
EXAMPLES .................................................................................................................................... 166
4.2.1 Example: Chicco Artsana Italia .................................................................................................................................................... 166
4.2.2 Example: Nestlé Italia (dry food) ................................................................................................................................................. 166
4.2.3 Example: Unilever Italia - Food ................................................................................................................................................... 166

5 MIXED DISTRIBUTION NETWORKS (UP TO 2 ECHELONS) ................................................................... 167

“THE RATIONAL” FOR MIXED NETWORKS ............................................................................................... 167
5.1.1 Example - Barilla Italy................................................................................................................................................................... 167
5.1.2 Example: Nestlé Purina Pet Care Italia ....................................................................................................................................... 168
5.1.3 Example: Cola Italia...................................................................................................................................................................... 168

6 IDEAL BASE LINE OPTION NETWORK FOR A COMPANY ....................................................................... 169

STEPS ......................................................................................................................................... 169
EXAMPLE – CONSUMER ELECTRONICS PRODUCER (SONY, SAMSUNG, …)....................................................... 169
6.2.1 Distribution Problem(s) Profile ................................................................................................................................................... 169
6.2.2 Ideal Network structure .............................................................................................................................................................. 170

X
 Summary

7 3-ECHELON DISTRIBUTION NETWORK ........................................................................................... 170

7.1.1 Example: Galbani Italia ................................................................................................................................................................ 171

8 OPTIMIZATION FUNCTION AND DECISION VARIABLES....................................................................... 171

PROBLEM DEFINITION ...................................................................................................................... 171
8.1.1 Distribution costs ......................................................................................................................................................................... 171
INVENTORIES IN THE DISTRIBUTION SYSTEM ......................................................................................... 171
8.2.1 Distribution costs ......................................................................................................................................................................... 172
8.2.2 Customer service ......................................................................................................................................................................... 172
OPTIMISATION FUNCTION: HOW TO MERGE COST AND SERVICE OBJECTIVES ................................................... 172
8.3.1 The profit optimization approach ............................................................................................................................................... 172
8.3.2 The cost of poor service and the total cost optimization .......................................................................................................... 172
MAIN DECISION VARIABLES ............................................................................................................... 173
CONSTRAINTS ............................................................................................................................... 173
9 RECAP OF DISTRIBUTION NETWORK ............................................................................................. 173
10 AMAZON CASE........................................................................................................................... 174
DESCRIBE THE DISTRIBUTION PROBLEM................................................................................................ 174
10.1.1 POO ............................................................................................................................................................................................... 174
10.1.2 Products ....................................................................................................................................................................................... 174
10.1.3 POD ............................................................................................................................................................................................... 174
10.1.4 Service .......................................................................................................................................................................................... 174
DESIGN THE DISTRIBUTION NETWORK ................................................................................................. 175
10.2.1 Do I need a 1-echelon? ................................................................................................................................................................ 175
10.2.2 How many Warehouses do I need at L1? ................................................................................................................................... 175
10.2.3 Do I need a 2-echelon? ................................................................................................................................................................ 175
HOW WE CAN DEFINE THE PROBLEM .................................................................................................... 176
10.3.1 What are the main cost items we have to consider? ................................................................................................................ 176
10.3.2 What are the main performances of customer service? ........................................................................................................... 177
10.3.3 How is the methodology structured? ......................................................................................................................................... 177

11 CONCLUSION ............................................................................................................................ 177

SUPPLEMENTARY READINGS .............................................................................................................. 177
CHALLENGE QUESTIONS ................................................................................................................... 177

12. Distribution network design: design methodology and models – Part 1 ................... 179
0 INTRODUCTION......................................................................................................................... 179
LEARNING OBJECTIVES..................................................................................................................... 179
1 METHODOLOGY ......................................................................................................................... 179
DESIGN METHODOLOGY FRAMEWORK ................................................................................................... 179
DESIGN METHODOLOGY MAIN PHASES ................................................................................................. 179
DATA ANALYSIS ............................................................................................................................. 179
1.3.1 Models .......................................................................................................................................................................................... 180

2 NETWORK SELECTION MATRIXES .................................................................................................. 180

OBJECTIVE OF THE NETWORK SELECTION MATRIXES................................................................................. 180
NETWORK SELECTION MATRIXES......................................................................................................... 181
2.2.1 1. Data collection ......................................................................................................................................................................... 181
2.2.2 2. Distribution problem representation ..................................................................................................................................... 181
2.2.3 3. Positioning on the matrixes - Example ................................................................................................................................... 184
2.2.4 4. Macro decisions on the network structure ............................................................................................................................ 184

3 COST MODELS ........................................................................................................................... 185

3.1.1 Number of warehouses ............................................................................................................................................................... 185
THE COST ANALYSIS ........................................................................................................................ 185
4 COST ANALYSIS - IN DETAIL ........................................................................................................ 187
TRANSPORTATION COSTS ................................................................................................................. 187
4.1.1 Full truck load transportation costs ............................................................................................................................................ 187
4.1.2 Local distribution costs ................................................................................................................................................................ 187

XI
 Summary

INVENTORY COSTS.......................................................................................................................... 188

4.2.1 Cycle Stocks .................................................................................................................................................................................. 189
4.2.2 In-transit stock ............................................................................................................................................................................. 189
4.2.3 Safety stocks................................................................................................................................................................................. 189
4.2.4 Cost per unit ................................................................................................................................................................................. 190
HANDLING COSTS ........................................................................................................................... 190
TOTAL DISTRIBUTION COSTS ............................................................................................................. 190
5 CONCLUSION ............................................................................................................................ 190
SUPPLEMENTARY READINGS .............................................................................................................. 190
CHALLENGE QUESTIONS ................................................................................................................... 190

13. Distribution network design: design methodology and models – Part 2 .................. 191
0 INTRODUCTION......................................................................................................................... 191
LEARNING OBJECTIVES..................................................................................................................... 191
AGENDA....................................................................................................................................... 191
1 LOCATION AND ALLOCATION PROBLEM .......................................................................................... 191
ALLOCATION ................................................................................................................................. 191
LOCATION .................................................................................................................................... 192
1.2.1 Single vs multi-warehouse location ............................................................................................................................................ 192
MODEL CLASSIFICATION ................................................................................................................... 192
2 SINGLE WAREHOUSE LOCATION.................................................................................................... 193
CENTER OF GRAVITY ....................................................................................................................... 193
2.1.1 Steps ............................................................................................................................................................................................. 194
2.1.2 Distances computation ................................................................................................................................................................ 194
2.1.3 Key messages ............................................................................................................................................................................... 195

3 ALLOCATION ............................................................................................................................ 195

REFERENCE SCENARIO ..................................................................................................................... 195
3.1.1 Example (1)................................................................................................................................................................................... 196
3.1.2 Example (2)................................................................................................................................................................................... 196

4 MULTI-WAREHOUSE LOCATION .................................................................................................... 197

MODELS FOR MULTI-FACILITY ALLOCATION............................................................................................ 197
CLUSTERING METHODS .................................................................................................................... 197
5 PIRELLI PRACTICAL LESSON ........................................................................................................ 199
5.1.1 Their solution ............................................................................................................................................................................... 200
MIXED INTEGER PROGRAMMING ......................................................................................................... 201
CONCLUSIONS ............................................................................................................................... 202
6 CONCLUSIONS .......................................................................................................................... 202
SUPPLEMENTARY READINGS .............................................................................................................. 202
CHALLENGE QUESTIONS ................................................................................................................... 202

14. Logistics Outsourcing ............................................................................................... 203

0 INTRODUCTION......................................................................................................................... 203
LEARNING OBJECTIVES..................................................................................................................... 203
AGENDA....................................................................................................................................... 203
1 DEFINITIONS & STRATEGIES ....................................................................................................... 203
GENERAL OVERVIEW ....................................................................................................................... 203
LOGISTICS OUTSOURCING STRATEGIES ................................................................................................ 204
PROFILE OF LOGISTICS OUTSOURCING ACTIVITIES ................................................................................... 205
1.3.1 Outsourcing percentage per industry in Italy............................................................................................................................. 205
1.3.2 Innovative services....................................................................................................................................................................... 205
EXAMPLES .................................................................................................................................... 205
1.4.1 Real examples .............................................................................................................................................................................. 206
OTHER DEFINITIONS ....................................................................................................................... 207
1.5.1 General drivers............................................................................................................................................................................. 207
1.5.2 Core competencies ...................................................................................................................................................................... 207
1.5.3 Strategic fit ................................................................................................................................................................................... 207

XII
 Summary

2 TYPES OF PROVIDERS AND MARKET SIZE ........................................................................................ 207

SUPPLY CHAIN OF LOGISTIC SERVICES .................................................................................................. 207
EXAMPLES OF DIFFERENT PROVIDERS ................................................................................................... 208
2.2.1 Third Party Logistics Providers .................................................................................................................................................... 209
2.2.2 The Contract Logistics Industry ................................................................................................................................................... 209
2.2.3 Contract Logistics Net Revenue .................................................................................................................................................. 209
CLASSIFICATION ............................................................................................................................ 210
EXAMPLE OF LOGISTIC SERVICES SUPPLY CHAIN IN ITALY ......................................................................... 210
3 VALUE CREATION, RISK & COST AND PROVIDERS’ STRATEGIES ........................................................... 211
LOGISTICS COST STRUCTURE ............................................................................................................. 211
FACTORS FOR CREATE VALUE ............................................................................................................. 211
3.2.1 Value Tree .................................................................................................................................................................................... 211
3.2.2 Economies of scale ...................................................................................................................................................................... 212
3.2.3 Learning economies ..................................................................................................................................................................... 212
3.2.4 Ability to introduce innovation ................................................................................................................................................... 212
3.2.5 Flow management capabilities ................................................................................................................................................... 212
RISKS AND COSTS........................................................................................................................... 213
PROVIDERS’ STRATEGIC APPROACHES .................................................................................................. 213
4 SETTING OUTSOURCING RELATIONSHIPS ....................................................................................... 214
MAIN IMPLICATIONS ....................................................................................................................... 214
4.1.1 ‘‘Baby sitter’’ syndrome ............................................................................................................................................................... 214
OUTSOURCING PROCESS FRAMEWORK .................................................................................................. 214
4.2.1 Step 1: AS-IS situation assessment ............................................................................................................................................. 214
4.2.2 Step 2: Competitive advantages related to logistics outsourcing ............................................................................................. 215
4.2.3 Step 3: LSP assessment and selection ........................................................................................................................................ 215

5 CONCLUSIONS .......................................................................................................................... 216

SUPPLEMENTARY READINGS............................................................................................................... 216
CHALLENGE QUESTIONS ................................................................................................................... 216

15. Advanced Inventory Planning ................................................................................... 217

1 WHAT IS ADVANCED INVENTORY PLANNING? ................................................................................. 217
1.1.1 What is the meaning of inventory planning? ............................................................................................................................. 217

2 RECAP ABOUT INVENTORY PLANNING ............................................................................................ 217

WHAT IS THE ROLE OF INVENTORIES WITHIN THE SUPPLY CHAIN? ............................................................... 217
WHAT ARE THE MAIN TYPES OF INVENTORY WITHIN THE SUPPLY CHAIN? ....................................................... 218
2.2.1 What are the main drivers to size these stocks? ....................................................................................................................... 218
WHAT ARE THE MAIN KPIS OF INVENTORY PLANNING? ............................................................................. 218
2.3.1 Efficiency ...................................................................................................................................................................................... 218
2.3.2 Effectiveness ................................................................................................................................................................................ 219
WHAT ARE THE MAIN INVENTORY PLANNING MODELS?.............................................................................. 219
2.4.1 Pull inventory model -> independent work ................................................................................................................................ 219
2.4.2 Push inventory model -> dependent work ................................................................................................................................. 219
WHAT ARE THE MAIN PULL INVENTORY MODELS? .................................................................................... 220
2.5.1 Re-order point model .................................................................................................................................................................. 220
2.5.2 Periodic review model ................................................................................................................................................................. 220
2.5.3 What are their main variables/parameters? .............................................................................................................................. 220
HOW CAN THESE VALUES BE OPTIMIZED? .............................................................................................. 221
2.6.1 Re-order point .............................................................................................................................................................................. 221
2.6.2 Periodic review model ................................................................................................................................................................. 221

3 SAFETY STOCKS......................................................................................................................... 221

HOW CAN SAFETY STOCKS BE DETERMINED? .......................................................................................... 222
3.1.1 Re-order point case - example .................................................................................................................................................... 222
3.1.2 Re-order point case - General case ............................................................................................................................................. 223
3.1.3 Periodic review model ................................................................................................................................................................. 224

4 DISTRIBUTION NETWORK DESIGN: INVENTORY ALLOCATION ............................................................. 225

LEARNING OBJECTIVES..................................................................................................................... 225
SAFETY STOCK IN THE DISTRIBUTION NETWORK ..................................................................................... 225
4.2.1 Safety stock in a 2-echelon distribution network ...................................................................................................................... 225

XIII
 Summary

VARIABILITY ................................................................................................................................. 225

4.3.1 Coupled system............................................................................................................................................................................ 225
4.3.2 Independent system .................................................................................................................................................................... 226
SAFETY STOCKS CALCULATION ........................................................................................................... 226
4.4.1 Assumptions ................................................................................................................................................................................. 226
COUPLED SYSTEM ........................................................................................................................... 226
4.5.1 Independent system .................................................................................................................................................................... 226
SAFETY STOCKS AND NUMBER OF WAREHOUSES ...................................................................................... 227
4.6.1 Coupled system............................................................................................................................................................................ 227
4.6.2 Independent system .................................................................................................................................................................... 227

5 CONCLUSIONS .......................................................................................................................... 227

SUPPLEMENTARY READINGS .............................................................................................................. 227
CRITICAL THINKING ........................................................................................................................ 227
5.2.1 What are the main levers to reduce safety stocks? ................................................................................................................... 227
5.2.2 What are the levers to reduce OCT? .......................................................................................................................................... 228
5.2.3 Centralization/Decentralization of safety stocks in the network (allocation problem) ........................................................... 228

XIV
 01. Objectives, Teaching methods, Assessment

01. Objectives, Teaching methods, Assessment

0 AGENDA
- Objectives, teaching methods, assessment
- Programme and contents
- Syllabus and Programme Details

1 OBJECTIVES, TEACHING METHODS, ASSESSMENT

The Logistics Management course aims to provide the students with the main theories, criteria, methodologies,
analytical models and tools (what we have called principles) that will allow them to develop (during the course
and later):
- a good knowledge of logistics systems and processes
- an analytical mind-set and analytical skills when facing both the design and the management of logistics
systems and processes
- the key criteria to describe logistics problems and assess/address solutions

ILOS (INTENDED LEARNING OUTCOMES)

Three main ILO:
ILO1. Know and Understand, the basis of the ILO -> Understand and describe:
- the main logistics strategies and their consistency with the general company strategy
- the main logistic solutions in the fields of distribution, warehousing (storage and picking), transportation,
inventory management, the core blocks of logistics
- the main models/methodologies that are intended to convey the key principles and criteria to design and
manage logistics systems
These objectives are within the main objective Know and Understand

ILO2. Apply, Make Judgements

Apply the theories, models, principles and criteria to solve complex “likely logistics problems (similar to the real
world)” and to understand and address “real-world logistics problems”

ILO3. Create, Communicate, Learn how to learn

Design and illustrate innovative logistics solutions in new fields (e.g. B2c eCommerce)

These two final objectives are more related to the ability to apply and to learn, to develop something new

TEACHING METHODOLOGIES
- Theoretical lessons: “traditional” lessons to provide the students with both a deep knowledge on logistics
systems/processes and criteria, methodologies, models, tools to design and manage logistics systems
- Practical lessons: lessons going through structured exercises to let the students apply what they learnt
during the theoretical lessons
- Seminars: lectures given by external guests (top managers from multinational companies) to provide the
students with strong connections with the real applications of what has been learnt during the lessons
- Virtual visits: virtual (with videos) visits to show the students the logistics systems and processes of
multinational companies
- Contest: project with a company where the students are required to design and present a solution to
innovate in a specific logistic area

1
 02. Objectives, Teaching methods, Assessment

TEACHING METHODOLOGIES AND ILOS

The main idea is that you get Concepts +
Models/Methods + Real-world Insight in order to
achieve all the Intended Learning Outcomes

ASSESSMENT

Final Assessment (by professors): Written (compulsory) and Oral (elective) exams + Contest (elective)
- Written Exam
o 3 Quantitative Exercises: practical exercises with numbers where students are required to design
a logistics system (e.g. a warehouse, a picking system, …) applying the models learnt during the
course
o 2 Theoretical Questions: open questions where the students have to prove to be able to both
present/describe logistics strategies/solutions and apply them to specific contexts/problem
- Oral Assessment
2 questions that might span all the topics of the course and can modify the overall score from “minus
infinite” to around “plus 2 points” in the 0-30 scale. Access to the oral examination is allowed only if the
written exam is sufficient. Oral exam is mandatory for students who got a mark equal to or higher than
28. It is elective for all the other students.
- Contest
Esselunga Logistics Challenge Report (elective): short document where the students have to present their
idea with reference to a specific objective given in the contest

Self Assessment
- Exam simulation
Written exam simulation to get familiar with the exam structure and test your ability to properly answer
both quantitative exercises and theoretical questions. The exam simulation is planned in the middle of
the course
- Critical Thinking
Challenge questions at the end of each topic. You will be asked to answer, using the “Think-Pair-Share”
approach: provide your answer first, then discuss it with one colleague and eventually change it. Answers
will be then discussed together with the professors

1.4.1 Critical Thinking

- The tool adopted to support the Critical Thinking activity is Poll Everywhere
- We will ask you to answer questions in real time (according to the model “Think-Pair-Share)
- You can participate using your Pc, Tablet or Smartphone
- The procedure is the following:
o Type PollEv.com/logman on your laptop/smartphone/tablet
o Select/Write the answer

2
 01. Objectives, Teaching methods, Assessment

ASSESSMENT AND ILOS

2 PROGRAMME AND CONTENTS

ASSUMPTIONS
The course of Logistics Management is based on the following assumptions:
- The students might have some basic knowledge of logistics
- The course aims to:
o consolidate the basic knowledge on logistics execution and planning
o develop, on top of that, advanced knowledge on logistics systems design and planning
- NB: The basics will be recapped and catch-up material will be referenced/provided

BASICS OF LOGISTICS MANAGEMENT

The students could have some basic knowledge of logistics, mainly at the Execution, Planning and Performance
Measurement System Levels

2.2.1 Logistics execution

- The Warehousing (storage and handling) Systems -> General Overview
- The main Transportation modes -> General Overview
- The functional and technical features of the packaging

- ? Automated warehouses
- ? Picking/Order Assembly/Sorting Systems

➔ A recap of the basics related to Logistics Execution will be made during the course

2.2.2 Logistics planning

- Demand forecasting We strongly suggest you to attend the MOOC “Fundamentals of Operations”
o the demand forecasting process
o the basic forecasting techniques
o the accuracy metrics

3
 02. Objectives, Teaching methods, Assessment

- Inventory management
o the types of inventory
o the main inventory management techniques
o the safety stock formula
- Production planning
o the models for the aggregate production planning
- Replenishment planning
o the MRP techniques

We strongly suggest you to attend the MOOC “Fundamentals of Operations” on POLIMI OPEN KNOWLEDGE
www.pok.polimi.it
Supplementary material on -Demand Forecasting -Inventory Management is also available on
Beep.metid.polimi.it

2.2.3 Performance measures

- Customer service
o the main customer service factors
o the main customer service metrics
- Logistics costs
o transportation costs
o handling costs
o inventory costs
o stock out costs
➔ A recap of the basics related to Performance Measures will be made during the course

THE PROGRAMME OF THE LOGISTICS MANAGEMENT COURSE

LOGISTIC MANAGEMENT
2.4.1 Logistic System Design and Strategy
- Logistics management as a competitive advantage
- How companies create value through logistics
- Types of logistic/distribution networks
- Methodologies for logistic/distribution network design
- Models of cost analysis
- Optimization models for distribution network design (centre of gravity, linear programming, …)
- Optimization models for network management (stock allocation, local distribution,…)
➔ Lessons + practical exercises + lectures + basic and supplementary material

2.4.2 Logistics Planning (advanced)

- Recap of basics (Demand and Inventory Management)
- Advanced Inventory Management
o Multi-item Inventory Management
o Multi-warehouse Inventory allocation
➔ Lessons + practical exercises + lectures + basic and supplementary material

4
 01. Objectives, Teaching methods, Assessment

2.4.3 Logistics Execution (recap of basics + advanced topics)

- Warehouse Management: the main design and management
criteria
- Materials handling automation:
o Automated warehouses: the typologies and the main design and management criteria
o Picking : the system typologies and the main design and management criteria
- Transportation Management
➔ Lessons + practical exercises + basic and supplementary material + videos + virtual visits

SUPPLY CHAIN COLLABORATION

- Models/processes of collaborative planning
o Vendor Managed Inventory
o Collaborative Planning Forecasting e Replenishment
- eCommerce
➔ Lessons + practical exercises + lectures + basic and supplementary material

3 SCHEDULE OF THE COURSE

Lessons are of two types.
1. “In-classroom” lessons, carried out during the academic timetable. We expect that attendance will be
real-time and interactive. They will for sure be digitally delivered in the first part of the course, and – if
and when possible – we will switch to a face-to-face mode. Lessons given on Teams will be recorded and
you will be able to watch them also later at any time.
2. “Out-of-classroom” lessons/activities, carried out in time slots not included in the academic schedule.
They will be digitally delivered and recorded as well, and you are free to access them whenever you like.
Nonetheless, dates and timing will be communicated in advance, to facilitate real-time attendance. They
include tutorship sessions, seminars and virtual visits to distribution centers. Two virtual visits (Comifar
and Artsana) will be part of the program, the other ones – even if recommended – are elective.

OUT-OF-CLASSROOM LESSONS
3.1.1 Tutorship sessions
Topic Date
- Tutorship sessions will start at 5:30 p.m. Logistics&Strategy W 25/3
- They will be mainly devoted to the last topics tackled in the Warehouse Management W 15/4
course, but they must be intended as open sessions where
Picking/Order Assembly Tu 5/5
you can make all the questions you want
Correction of exam simulation Th 7/5
- In order to be more effective, questions will be collected in
Distribution Network Design (1) Tu 26/5
advance
- Instructions to access the virtual room will be Distribution Network Design (2), W 10/6
communicated in advance Transportation and Outsourcing
Advanced Inventory PlanningM 15/6
3.1.2 Virtual visits to Distribution Centers
We will arrange virtual visits to some distribution centers:
- Comifar: distribution center in Novate Milanese
- Artsana: distribution center in Grandate
- Amazon: the new distribution center in Torino
- Zara: the Italian Distribution Center for the B2c eCommerce and Click&Collect channels
- Esselunga: the warehouse dedicated to the B2c eCommerce initiative in Sesto Ulteriano. This virtual visit
and the presentation of the elective project will be carried out in the same session.

Rules of the game:

- Comifar and Artsana visits will be part of the program
- Amazon, Zara and Esselunga visits are not mandatory
Details about the dates for the virtual visits will be available soon. All the visits will be recorded (and thus
accessible whenever you want).

5
 02. Objectives, Teaching methods, Assessment

3.1.3 The Esselunga Logistics Challenge

- Task: You will be asked to work on a specific task in the logistics & distribution field in B2c eCommerce
area that requires both the application of your knowledge and an innovative/original idea to be tackled
- Output: Short report
summarizing the idea
- Teams: Teams of 4/5
students (up to you!)
- Evaluation criteria
o Projects will be evaluated by both Polimi and Esselunga:
▪ Polimi will assess the methodological approach
▪ Esselunga will evaluate the project in terms of customer impact, feasibility, scalability
o As for Polimi’s evaluations:
▪ The two best teams will get 3 additional points
▪ The second four best teams will get 2 additional points
▪ The other teams will get an evaluation from 0 to 2 additional points
N.B.: the results of Polimi and Esselunga’s evaluations could lead to two different rankings

3.1.4 Talks on “Sustainable and Fair Logistics"

1. "Last mile logistics: technology and sustainability", 23rd April
2. "Environmental-conscious Logistics", 14th May
3. “Artificial Intelligence for Supply Chain Planning”, 21th May

- 1,5-hour Evening Seminars with external guests

- Seminars will be interactive. A large part of each seminar will be devoted to the questions you may have
for the speakers. We encourage you to prepare the questions in advance.

TEACHING MATERIAL
- Set of slides supporting the in-class lectures
- Practical lessons/exercises
- Exercises selected from past exams
- Supplementary bibliography (selection of papers and chapters of relevant books, etc.)
- All the material will be available before the related lesson
- All the digital material (slides, texts of the practical lessons, exercises) will be available on the website
http://beep.metid.polimi.it

3.2.1 Suggested Bibliography(*):

- Coyle, Langley, Gibson, Novack, Bardi. “Supply Chain Management: A Logistics Perspective. (from 8th
edition, 2008)”. South-Western College Publications (ISBN: 978-0324376920)
- Ballou, R.H. “Business Logistics Management (from 5th Edition, 2003)”. Prentice Hall. (ISBN: 978-
0130661845)
- Chopra S., Meindl P., “Supply Chain Management: strategic, planning and operations”. Prentice Hall, 2004
(ISBN 0131217453)
(*) These bibliographies are not mandatory to successfully pass the exam. It is only recommended to supplement
the basic material and as a reference for future personal in-depth study

3.2.2 How to find us

- Prof. Alessandro Perego
Politecnico di Milano, Dipartimento di Ingegneria Gestionale
alessandro.perego@polimi.it Tel: 02-2399-2705
- Prof. Riccardo Mangiaracina
Politecnico di Milano, Dipartimento di Ingegneria Gestionale
riccardo.mangiaracina@polimi.it Tel: 02-2399-4051
- Eng. Arianna Seghezzi
Politecnico di Milano, Dipartimento di Ingegneria Gestionale
arianna.seghezzi@polimi.it Tel: 02-2399-4051

6
 02. Introduction to Logistics and Supply Chain Management

02. Introduction to Logistics and Supply Chain

Management
1 INTRODUCTION
The insights are fundamental to understand the
principles of logistics management LM, which are part
of the principles of management engineering
For this course, is more important to give us the ability
to learn, not just the principles, because what we are
studying now could change in the future. We should
have the capability to learn

Ability to learn
- Curiosity -> be open to learn, personal attitude
- Modelling skills (or problem setting & solving)
-> ability to understand the main variables and
to model the relations between them, and to
set the main methods to solve the problem
- Critical thinking
o judgment oriented about the
experience
o informed by evidence, thinking that
starts from facts
o rational/logical
o systematical, it takes into
consideration all the elements about the problem

Teacher-students Agreement -> right attitude when attending this course

- dedicated
- curious -> ask questions
- interactive
- respectful
- transparent

Content of the course

What does the name of the course, Logistics Management, mean?

7
 01. Introduction to Logistics and Supply Chain Management

INTENDED LEARNING OBJECTIVES

At the end of the module the students will be able to knowingly discuss the following topics:
- “Logistics” and “Supply Chain Management”: key-words and definitions
- The evolution of Logistics from Physical Distribution System in the
’70s to Supply Chain Management in the ’90s
- A 3-stack model of Logistics/Supply Chain Management : from logistics execution, to supply chain
planning, to supply chain design and strategy
- The main activities of Logistics Management
- Why Logistics and Supply Chain Management are key business topics

AGENDA
1. Logistics and supply chain management
2. The 3 “stacks” of logistics management
3. The importance of Logistics

2 LOGISTICS MANAGEMENT
“Logistics Management is that part of supply chain management that plans, implements, and controls the
efficient, effective forward and reverse flow and storage of goods, services and related information between the
point of origin and the point of consumption in order to meet customers' requirements.”
CSCMP (Council of Supply Chain Management Professionals), nowadays
Definition of Logistic given by CSCMP, United States
We will have all the course all the elements of this definition, now we will just understand the scope of Logistic
management

Logistics characteristics
- Logistics is not a function, but a process, a systemic process, we are optimizing the process/system
-> set of interrelated activities, correlated into a process, aimed at meeting customer requirements
- Supply chain process -> it is a process that ideally spends all the supply chain, both internal and
internal, so the one which is within a company and the one which connects the customer and supplier
- Logistic is about some activities
o execution/implementation of the management of storage and flow of goods
o planning and control part of the process -> demand forecasting, production planning,
transportation control…
o design of the system -> it has to do with the overall design of the process

Main objectives of logistics

𝑜𝑢𝑡𝑝𝑢𝑡
- Efficiency -> increasing the ratio 𝑖𝑛𝑝𝑢𝑡 , doing more output with less input
o Cost reduction
o Productivity increase
- Effectiveness -> meeting customers’ expectations, so we need to know them
We will talk a lot about service level, because it able us to know the real customer expectation

2.1.1 Council of Supply Chain Management Professionals (www.cscmp.org)

Founded in 1963, it is the preeminent association for individuals involved in supply chain management
CSCMP was originally founded as the National Council of Physical Distribution Management (NCPDM) in January
1963. NCPDM was formed by a visionary group of educators, consultants, and managers who envisioned the
integration of transportation, warehousing, and inventory as the future of the discipline. At that time, physical
distribution was just beginning to edge its way into the corporate lexicon and make its considerable presence felt
in the business community.
In 1985, recognizing the growing field of logistics, the association's focus broadened as it changed its name to the
Council of Logistics Management (CLM). It stayed that way until 2004 when CLM's Executive Committee voted to
become CSCMP, the Council of Supply Chain Management Professionals (CSCMP)

8
 02. Introduction to Logistics and Supply Chain Management

THE EVOLUTION OF LOGISTICS

To better understand the scope of

Logistics we will pass through 50 years of
definitions of Logistics and analyse them
with a radar chart, with 4 different
dimensions -> product, activity,
performance, system

THE EVOLUTION OF LOGISTICS: DISTRIBUTION

“A term employed in manufacturing and commerce to describe the broad range of activities concerned with
efficient movement of finished products from the end of production line to consumer, and in some cases includes
the movement of raw materials from the source of supply to the beginning of the production line.”
NCPD (National Council of Physical Distribution Management), 1967
Old name of National Council of Supply Chain Management Professionals

With this definition we have a first vision of Logistic from the industrial point of view, not just the military one,
because we are after the second Worldwide War

- Efficient: emphasis on cost reduction

- From production line to consumer: emphasis on the distribution system
- Movement of finished products: emphasis on materials handling and transportation

Radar chart
Initial scope of logistic was very narrow, it was related to just few elements
Four main dimensions:
- Product -> finished products (just sometimes also raw materials)
- Activity -> execution/movement: transportation and handling
- Performance -> efficiency
- System -> distribution part of the supply chain, in terms of the connection between the consumer
and the producer

THE EVOLUTION OF LOGISTICS: INTEGRATED LOGISTICS

2.4.1 NCPD (National Council of Physical Distribution Management), 1976
“The integration of two or more activities for the purpose of planning, implementing, and controlling the
efficient flow of raw materials, in-process inventory and finished goods from point-of-origin to point-of-
consumption.”

- Integration: emphasis on the value of global optimization (as opposed to local optimization)
- Plan, implement and control: not only execution, emphasis on logistics as a business management
discipline
- From point of origin to point of consumption: first inclusion of the (internal) “supply chain perspective”

9
 01. Introduction to Logistics and Supply Chain Management

Radar chart
- Products -> raw material, in process inventory, finished goods
- Activity -> plan, implement, control
- Performance -> (same) efficiency
- System -> from point-of-origin to point-of-consumption
internal supply chain: procurement, production, distribution
The system is more extended, we have a wider scope of logistic

2.4.2 CLM (Council of Logistics Management), 1985

“The process of planning, implementing, and controlling the efficient, cost-effective flow and storage of raw
materials, in-process inventory, finished goods, and related information from point-of-origin to point-of-
consumption for the purpose of conforming to customer requirements.”

- Process: it is the first explicit recognition that Logistics is a process

- Related information: not only flow of goods but also flow of information
- Conforming to customer requirements: it is the ultimate objective of logistics, not only efficiency and
cost-effectiveness

Radar chart
- Product -> … + information elements
- Activity -> (same) plan, implement, control
- Performance -> efficiency + effectiveness in customer requirements
- System -> (same) internal supply chain

2.4.3 CLM (Council of Logistics Management), 1992

“The process of planning, implementing, and controlling the efficient, effective flow and storage of goods,
services, and related information from point-of-origin to point-of-consumption for the purpose of conforming to
customer requirements.”
- Effective: not cost-effective but effective tout court
- Services: logistics management applies also to the provision of services (healthcare, after-sales,
engineering&contracting, …)

Radar chart
- Product -> … + services, logistics does not apply on just product industry, but it is also very
important in industries of services (health, tourism…)
- Activity -> (same) plan, implement, control
- Performance -> (same) efficiency + effectiveness
- System -> (same) internal supply chain

THE EVOLUTION OF LOGISTICS: SUPPLY CHAIN MANAGEMENT

2.5.1 CSCMP (Council of Supply Chain Management Professionals), nowadays
“Logistics Management is that part of supply chain management that plans, implements, and controls the
efficient, effective forward and reverse flow and storage of goods, services and related information between the
point-of-origin and the point-of-consumption in order to meet customers' requirements.”

Radar chart
- Product -> (same)
- Activity -> reverse logistics of goods, services, information from the market to the levels at the
beginning of the chain, there is not only the forward logistics
- Performance -> efficiency + effectiveness, elimination of just “cost-effectiveness”
- System -> internal + external supply chain, we are talking about the overall supply chain

10
 02. Introduction to Logistics and Supply Chain Management

SUMMARY OF RADAR CHART

- Black -> 1967
- Red -> 1976
- Blue -> 1985
- Pink -> 1992
- Green -> nowadays

3 MAIN TRENDS
1. From the optimisation of single activities
to an integrating function (system)
2. From physical distribution management
to supply chain management (system)
3. From execution to planning
to design of logistics system (activity)
4. From cost reduction/optimization
to customer service and profit optimization
(combination of the two) (objective/performance)
5. From the management of direct flows of products
to the management of the flow of information, services, money, …
to reverse logistic of all these products, moving forward a circular paradigm (product)

1. FROM THE OPTIMISATION OF SINGLE ACTIVITIES TO AN INTEGRATING FUNCTION

ex: let’s consider a manufacturer of consumer electronics in China

Additional Data
- Value density: 10.000 €/m3 (medium high value density)
It’s the value compared in terms of volume or weight of the goods
- Annual flow: 1500 pallet loads = 1500m3/year (=50 standards containers, one is more or less 30 m3)
- By ship -> LT = 6 weeks
Ordering period T = 3 months (between an order and another one)
Cost = 1500 €/container
- By airplane -> LT = 1 week
Ordering period T = 2 weeks (if LT is shorter, you tend to reorder more frequently)
Cost = 15.000 €/Container
[1 months = 4 weeks 1 year = 12 months]

11
 01. Introduction to Logistics and Supply Chain Management

Solution
Objective function -> we are considering the cost, we want to know the least expensive solution, so we will not
consider the impact on the service level, it’s a simplification of the problem
The costs that are affected are the transportation cost but also
inventory cost, the cost of keeping inventorying within the logistic
system

Transportation costs = annual flow * unitary transportation cost

- 𝑏𝑦 𝑠ℎ𝑖𝑝 = 50 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟𝑠/𝑦𝑒𝑎𝑟 ∗ 1500 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 = 75.000 €/𝑦𝑒𝑎𝑟
- 𝑏𝑦 𝑎𝑖𝑟𝑝𝑙𝑎𝑛𝑒 = 50 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟𝑠/𝑦𝑒𝑎𝑟 ∗ 15.000 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 = 750.000 €/𝑦𝑒𝑎𝑟

Inventory costs = Average Inventory Level (AIL) [containers] * Unitary Inventory Cost (UIC) [€/containers year]
- Unitary Inventory Cost = cost of mantaining one unit of products within a system
= value of 1 container [€/container] * % inventory cost [%/year]
= [€/container per year]
o 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 1 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 = 10.000 €/𝑚3 ∗ 30 𝑚3 /𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 = 300.000 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟
o 𝑈𝐼𝐶 = 300.000 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 ∗ 30% = 90.000 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟
In the percentage there is an additional assumption, it considers the cost of maintaining the good and
also the cost of the risk of obsolecence in keeping the goods as inventory

- Average Inventory Level = average quantity of inventories we have within a system

= AIL in time [years] * Average flow [containers/year] = [containers]
To compute the AIL in terms of time, we have to consider both the
in-transit and the warehouse inventories
o The in transit invetories mainly depend on the lenght of the
transportation arc, which influences the LT
o The inventories in the warehouse mainly depends on the
average period they spend into the warehouse -> T/2

The measure of inventories in the warehouse is T/2 because if we

order a quantity T each period T, this quantity takes a period equal to
T for going from the highest level of inventories to zero, so the
evolution of the warehouse in time has this shape
On average, you have T/2 value of stocks within the warehouse

-> AIL in time = LT + T/2 = [year]

AIL in time =
6 𝑤𝑒𝑒𝑘𝑠 3 𝑚𝑜𝑛𝑡ℎ𝑠
- 𝑏𝑦 𝑠ℎ𝑖𝑝 = 4 𝑤𝑒𝑒𝑘𝑠/𝑚𝑜𝑛𝑡ℎ∗12 𝑚𝑜𝑛𝑡ℎ𝑠/𝑦𝑒𝑎𝑟 + 2∗12 𝑚𝑜𝑛𝑡ℎ𝑠/𝑦𝑒𝑎𝑟 =
3 𝑚𝑜𝑛𝑡ℎ𝑠
12 𝑚𝑜𝑛𝑡ℎ𝑠/𝑦𝑒𝑎𝑟
= 0,25 𝑦𝑒𝑎𝑟
1 2 0,5
- 𝑏𝑦 𝑎𝑖𝑟𝑝𝑙𝑎𝑛𝑒 = + = = 0,042 𝑦𝑒𝑎𝑟
4∗12 2∗4∗12 12

Unitary Inventory costs =

- 𝑏𝑦 𝑠ℎ𝑖𝑝 = 3/12 𝑦𝑒𝑎𝑟𝑠 ∗ 90.000 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 𝑝𝑒𝑟 𝑦𝑒𝑎𝑟 ∗ 50 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟𝑠/𝑦𝑒𝑎𝑟 =
= 12,5 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟𝑠 [𝐴𝐼𝐿 𝑖𝑛 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟𝑠] ∗ 30.000 €/𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑒𝑟 = 1.125.000 €/𝑦𝑒𝑎𝑟
- 𝑏𝑦 𝑎𝑖𝑟𝑝𝑙𝑎𝑛𝑒 = 0,5/12 ∗ 90.000 ∗ 50 = 2,08 ∗ 30.000 = 187.500 €/𝑦𝑒𝑎𝑟

Transportation costs Inventory costs Overall logistic costs

By ship 75.000 1.125.000 1.200.000
By airplane 750.000 187.500 937.500

In this case, the ship option has a lower transportation cost and the airplane option has a lower inventory cost
The lowest overall logistic cost is sustained by using the airplane transportation

12
 02. Introduction to Logistics and Supply Chain Management

This example is useful in explaining how the logistic man has to consider more and more activities and situations
along the same system in order to find the right solution for each logistic problem
If you optimize single activities, you arrive at results which do not reflect the real situation, which are not the right
solution; you need to follow a systemic perspective, so taking into account both transportation and inventory
activities/costs

Of course, if we change the assumptions made at the beginning, the result will obviously change
The most important variable here is the value density of the product, because if you consider less expensive
products, with lower value density, the inventories costs will change, but the transportartion costs do not change,
so in this case, maybe the ship could be a better solution compared to airplane
Other changes may happen if we change the LT or the period time T

3.1.1 Physical Distribution/Materials Management

Focus
- Distribution of the end products or
supply of raw materials/production
- Single company
- Reduction of the distribution or supply/production costs

➔ Integration of “parts” of the internal supply chain

Transportation + Warehousing + Inventory Management = Physical Distribution System

In an extended way, we can connect distribution to production

-> Production + Distribution = Outbound Supply chain

We can also try to connect the supply of raw material to the production line
-> Supply + Production = Inbound Supply Chain

- Focus: distribution of end products with a systemic view of transportation, storage, inventory
management, packaging and materials handling
- Methodologies: systemic approach, total cost analysis, trade-off analysis, operational research
techniques (inventory management models, demand forecasting algorithms, site location, etc.)
- Organization: “Distribution Manager”
- Awareness: the National Council of Physical Distribution Management is founded in 1963

3.1.2 Integrated Logistics

Focus
- Supply+production+distribution
- Single company
- Reduction of the overall logistics costs

➔ Integration of the internal supply chain

- Focus: focus on the overall logistics system with the integration of the different sub-processes of
Materials Management, Production Management, Physical Distribution
- Methodologies: systemic approach, total cost analysis, integrated planning principles, competitive value
of the customer service
- Organization: “Logistics Manager”
- Awareness: the National Council of Physical Distribution Management becomes the Council of Logistics
Management

Integrated function means to have a systemic approach, also called as total-cost analysis
We are moving from considering the “parts” of the internal supply chain to consider all the supply chain as a
single block

13
 01. Introduction to Logistics and Supply Chain Management

Supply + Production + Distribution = Integration of internal Supply Chain

-> inclusion of all the activities in the company, broadest logistic system you can have within the company

Suppliers + Internal Supply Chain + Distribution + Customers = Extended Supply Chain (real idea of SC)
-> inclusion of all the activities in and out the company, broadest logistic system you can have

2. FROM PHYSICAL DISTRIBUTION MANAGEMENT TO SUPPLY CHAIN MANAGEMENT

Second trend as the conclusion of the first trend, how to pass from the physical distribution to an extended
supply chain management

3.2.1 The Supply Chain

It includes the different members, infrastructures and
resources, processes and activities (and all the links
among them) that allow the replenishment of raw
materials, their transformation in components and
products and the distribution of products to end
customers

3.2.2 The Food Supply Chain 3.2.3 The Pharma Supply Chain

3.2.4 The evolution of Logistics: Supply Chain Management

“Logistics Management is that part of supply chain management that plans, implements, and controls the
efficient, effective forward and reverse flow and storage of goods, services and related information between the
point of origin and the point of consumption in order to meet customers' requirements.”
CSCPM (Council of Supply Chain Management Professionals), 2009

- Part of supply chain management: Logistics is part of a more comprehensive concept, Supply Chain
Management
- Reverse flow: Logistics is also about the reverse flow of goods and information from the point-of-
consumption backward

14
 02. Introduction to Logistics and Supply Chain Management

“Supply chain management is an integrating function with primary responsibility for linking major business
functions and business processes within and across companies into a cohesive and high-performing business
model. It includes all of the logistics management activities, as well as manufacturing operations, and it drives
coordination of processes and activities with and across marketing, sales, product design, finance, and
information technology”
CSCPM (Council of Supply Chain Management Professionals)

Supply Chain Management includes all the inter-company processes, i.e. the processes which extend outside the
company boundaries and span over the whole supply chain (logistics, new product development, marketing, etc.)

SCM = the integration (integrated management) of business process along the extended supply chain
Business processes can be logistic, marketing, new product development…
Along the extended supply chain means taking a supply chain perspective

Focus
- Entire supply chain
- Value of the customer service
- Overall logistics cost to serve the final customer
- Supply chain based competition

➔ Integration of the external supply chain

- Focus: focus on the Supply Chain as a whole; competition among supply chains; logistics as a
differentiation lever
- Methodologies: analysis of supply chain performance, supply chain integration, collaborative planning
models (VMI, CPFR)
- Organization: “Supply Chain Manager”
- Awareness: the Council of Logistics Management becomes the Council of Supply Chain Management
Professionals

3.2.5 Supply Chain Based Competition

When we talk about supply chain perspective, we mainly talk about supply chain-based competition, which is a
strategic approach where:
- Competitiveness of the company is strongly impacted by the actions of the other supply chain members
Competitiveness of one company depends on how the other players of the chain behave
- Competitiveness is, first of all, between different supply chains (and not between companies of the same
supply chain); a company compete with all the competitors, included all the players in the chain, in order
to be enough comprehensive

Strategic implications
- Competitiveness of the company can be improved focusing on the interface processes with the other
supply chain members, through integration and collaboration -> if you work on these two aspects, you
can improve your performances
- In many cases the action on the interface processes is more effective (in terms of impact on KPIs) than
that on internal processes -> in order to review your logistic costs, it’s better to work on supply chain and
customer costs, not just the internal ones

3.2.6 Performances of the FMCG Supply Chain

FMCG: Fast Moving Consumer Goods
ex: food industry, from the manufacturer to the retailer
- the retailer has to keep some inventories, near to the POS (point of sales)
- the manufacturer has to keep finished products inventoried in the warehouse
-> we have 2 different systems, and the two players can optimize them in an internal perspective

15
 01. Introduction to Logistics and Supply Chain Management

What is the amount of inventory (of finished products) in the warehouses of

manufacturers and retailers?
- 40 days of inventory (on average) in the warehouses of the
manufacturers (including both cycle and safety stock) -> very high
risk of obsolescence
- 20 days of inventory in the warehouses of the retailers (both
distribution centres and backrooms) (including safety, cycle and
speculative stock)
-> 60 days of overall inventories

What is the average value of the on-shelf availability in the stores of the retailers?
What level of availability do the customers see?
-> 93% as an average of all the different types of products. This means that 7 times out of 100, consumer does not
find what he is looking for (Source: GS1 Europe)

We can work to improve these two performance indicators following two different options:
- Internal supply chain, on both manufacturer and retailer, which means for example to synchronize the
production and the deliver
- Extended supply chain, integration and collaboration along the same processes, in order to reduce the
safety stocks
A big portion of the 20 days/inventories for the retailer are safety stocks, taken because the retailer does not
trust the manufacturer about reliability
A big portion of the 40 days/inventories for the manufacturer are safety stocks taken because it does not know
how frequently, and in which amount the retailer will ask finished goods
➔ These safety stocks can be reduced with some collaboration, with synchronization of the production and
the request

Working with a supply chain perspective means to look at the overall costs -> the effort is very high, but also the
benefits will be higher than the benefits you can have looking just at your internal supply chain, because you do
not consider enough variables

How much does the order-delivery-invoicing-payment cycle cost?

- Approximately 80 €/cycle (from order to payment), of which 50% due to the management of the activities
generated by errors…
- …equally split among manufacturer and retailer
(Source: Electronic Invoicing Observatory - Politecnico MI)

What is the average saturation of the means of transportation used by manufacturers to deliver their products to
retailers?
- Approximately 70% for the single trips (do hope that returns are not empty!)…
- … with a very limited adoption of “multi-drop”deliveries

16
 02. Introduction to Logistics and Supply Chain Management

3. FROM EXECUTION TO PLANNING TO DESIGN OF LOGISTICS SYSTEM

3.3.1 The 3 “stacks” of logistics management

Logistics consists of at least 3

layers of different activities
- Execution
- Planning
- Design

We will understand what each

stack considers

3.3.2 1. Logistics Execution

It has of course to do with movement, so with transportation
and handling (movement of goods within the warehouse),
with handling and storage (which together build the
warehousing activity)
It includes also the production and packaging
These are related to products, but we need to consider also
information management

Logistics Execution includes the design and management of

the processes that support the flow of materials (handling,
storage, physical transformation and transport activities)
and data/documents along the supply chain

- It is the most traditional “scope” of Logistics (but it is only part of its most comprehensive scope)
- It is the base upon which all the other stacks are built (therefore it represents the “basics” a good
logistician must know)

17
 01. Introduction to Logistics and Supply Chain Management

3.3.3 2. Logistics Planning

- Demand management (forecasting is only a component of
management)
- Inventory management, which can be related to
components, raw material and so on
- Distribution management, so the synchronization of
transportation to, from and within the warehouse
- Production planning
- Procurement planning
The idea is that we should be able to manage all these
activities, and moreover to link them within the logistic
planning stack and in particular with the execution stack

Logistics planning is the process that plans for the adjustment of

the operational capacity and for the execution of the operational
activities linking the demand and the supply sides of the supply
chain. It aims at “getting the supply chain aligned and tuned”
NB: it’s impossible to do that if you do not consider the operational
aspects/constraints of the process

- This is the stack that “plans and manages” the operational level of logistics
- It should be an integrated process (and not merely a collection of phases)
- It should consider both the structure of the distribution network and all the specific operational
constraints

3.3.4 3. Logistic System design and strategy

- Network or supply chain structure
-> consideration of both arcs and nodes
o Plants
o Warehouse
o Suppliers
- Make or Buy decision
- Target KPIs
o Service level -> ex: delivery time
o Logistic costs
- Information system, that you want to have in order to
support your logistic system

Logistic System design and strategy aims at defining:

1. the logistics network structure (supply network,
production network and distribution network)
2. the transportation modes (road, rail, ship, inter-
modal, etc.)
3. the “make or buy” policies

18
 02. Introduction to Logistics and Supply Chain Management

These decisions must be strongly aligned with the overall

company strategy (markets served, product range, trade
channels, etc.), a lot of decisions at logistic level has to do
with other kind of decision within the same company
These choices have to consider the specific
features/constraints of both the planning process and the
logistics execution process, so this level must be connected
with the other two levels, otherwise you can only have a
very theoretical solution, not a quite real one

3.3.5 The performance measurement system

Main choices:
- the metrics
- the measurement techniques
- the process: phases, responsibilities,
frequency, reports, etc.

[From tutorship session] NB: to talk about strategy we need to know the basics
We can identify 3 different levels, and we know that they are strongly interrelated
At the execution level it is very important to understand some performance drivers, in order to do the good
choice for the strategy used at the design level

ex: at the strategic level, we would like to reduce the transportation costs
In order to do this, we should how we can reduce the costs, which means to understand the cost drivers of
transportation at the execution level
Transportation can be of different types
- Primary transportation (point-to-point, long distances)
- Secondary transportation (multi-point, short distances)

We can assume to concentrate on primary transportation, we can consider as “direct length” the distance
between the point of origin to the point of destination
After we have moved the product from the POO to the POD, then we have to return the truck to the POO, but it
can be different from the first one -> in this case we consider the “return length”

In this case, the cost drivers of the direct length are:

- Origin-destination distance
- Load factor -> ex: what percentage of the truck load we are using for the transportation?
Ideally, we should work with FTL (full truck load) -> we use the 100% of truck load

While the cost drivers of the return length are just:

- Destination-next origin distance -> we need to make this distance as short as possible, so as close as
possible to the destination

Once we have defined the cost drivers of the primary transportation, we can find the right policy to reduce the
transportation costs
- We can work on transportation planning -> if we can plan the transportation in a good way, we can
increase the load factor, and also to reduce the return length
- We can work on the network design, because it determines the distance between the nodes
- We can go for transportation outsourcing, for a logistic provider which manage flows of different
customers, because it has more point of origin, and in this way the return length can be reduced
- We can go for a packaging redesign, in order to have a better load factor
- We can also go for collaborative planning, in order to synchronize the transportation, to be both more
efficient

19
 01. Introduction to Logistics and Supply Chain Management

4. FROM COST REDUCTION/OPTIMIZATION TO CUSTOMER SERVICE AND PROFIT OPTIMIZATION

3.4.1 (1) Cost reduction/optimization
Logistic costs =
- Transportation costs
- Handling costs, movement within the warehouse
or the plants
- Storage costs
-> Warehousing costs = handling + storage costs
- Packaging costs, both of finished goods but also
intermediary goods
- Info/documents management and administrative costs

𝐿𝑜𝑔𝑖𝑠𝑡𝑖𝑐 𝑐𝑜𝑠𝑡𝑠
𝑅𝑒𝑣𝑒𝑛𝑢𝑒𝑠
This ratio depends a lot on
- which kind of industry
- which type and role of the company/player in the network
- the kind of supply chain perspective we are considering (internal or external supply chain…)
90% of instances vary in the 5 – 15% range, but the company nearness to one of the two extremes depends on
these variables

Examples:
- pharma manufacturer is near to the 5%, because the value
density of these product is very high, so the incidence of
transportation costs is lower
- grocery low value products (ex: water), logistic costs are closer to
15%, because the incidence of transportation costs becomes
higher because the value density is lower
- fashion manufacturer is between the 5-15% range
- e-commerce supply chain of food product, the incidence of
logistics costs is much higher than 15%, because the
management of the e-commerce supply chain is much higher, so
the ratio is closer to 20-25%
NB: not all logistic costs have value density as driver, it is mainly driver of inventory costs,
but for example usually you have not the link with the transportation costs, it is not the
most important driver for it, we should prefer the density of the goods, not the value
density

3.4.2 Importance of Logistics and Supply Chain Management

In a nutshell the importance of logistics and supply chain management depends on the following reasons:
- L&SCM are non-discretionary activities that must be accomplished to deliver the right value to the right
customer
- These activities are costly: L&SCM related costs represent on average between 5 and 15% of revenues and
more if a supply chain perspective is taken
- They have a substantial impact on the assets of a company (inventories, manufacturing plants, transport
vehicles, …)
- ….

20
 02. Introduction to Logistics and Supply Chain Management

3.4.3 The incidence of Logistic costs: An analysis by industry

The incidence of the logistic costs on the revenues

Source: ELA-AT Kearney (early 2000s)
There are differences in terms of industries and components
of logistics costs

3.4.4 Cost of Logistics as % of sales

Source: European Logistics study 2008-
2009 (ELA – AT Kearney)

The incidence varies also over time

This picture is the average among
different industries

There is a U-curve, we got a reduction

of logistics costs in the two decades
before 2000, but then we can see an
increase of them

- There is an increase in transportation costs, because there is a higher attention in impact on the
environment
- There is an increase in inventory costs and so in the supply chain risks: in a global society the risk of
supply chain management is obviously higher

3.4.5 Globalization: Export grows more than GDP

Source: World Economic Outlook Database, October

2013 (International Monetary Fund)

3.4.6 The evolution of the market: trends in factors

of production in Europe

In the last years the logistics sector

has been facing an increase in the
costs of the factors of production
(most of all labour, energy and fuel)

21
 01. Introduction to Logistics and Supply Chain Management

3.4.7 Cost of Logistics as % of sales in different industries

Source: the Benchmarking Institute 2007

3.4.8 The incidence of Logistic costs: An analysis by Supply Chain – the textile/apparel

Consideration of all the cost of

logistics in an extended supply
chain perspective
In this way you have to consider all
the green areas, so the costs of all
the three players
-> 3,2 + 3,3 + 8 = 14,5

You are considering some costs

which are usually hidden
So, if you consider the overall
incidence of logistic costs, of course
this will increase

3.4.9 (2) Customer service and profit optimization

Service level =
- Time-related elements
o Order cycle time -> time between order and deliver,
time seen by the customer; it is
not the time spent along all the
supply chain
o Punctuality -> on-time delivery, being faster or
reliable in terms of delivery time
- Availability -> on-shelf availability
- Quality
o Products -> delivering the right products
o Packaging -> delivering the right products
with the right packaging
- …

3.4.10 Relation between logistic costs and service level

There are some aspects very important for the customer which depends mainly on the structure of the logistic
system, which has also impact on the logistic costs
➔ Logistics Management Objective function consider both logistic costs and service level

If we increase the service level from 1 to 2 weeks, you increase also the logistic costs needed to do that; the
impact is more than proportional, so we have a curve, not a linear proportionality

22
 02. Introduction to Logistics and Supply Chain Management

The connection between cost and service is called cost-to-serve relation

In a nutshell the importance of logistics and supply chain management (L&SCM) depends on the following
reasons:
- L&SCM are non-discretionary activities that must be accomplished to deliver the right value to the right
customer
- These activities are costly: L&SCM related costs represent on average between 5 and 15% of revenues
and more if a supply chain perspective is taken
- They have a substantial impact on the assets of a company (inventories, manufacturing plants, transport
vehicles, …)
- Last but not least L&SCM affects the top line of the profit & loss, i.e. revenues (customer service, lost
sales, …)

5. FROM DIRECT FLOWS OF GOODS TO DIRECT/INDIRECT (REVERSE) FLOWS OF INFORMATION, SERVICES,

MONEY AND REVERSE LOGISTICS OF GOODS

3.5.1 Information
- Transactional data, all the data which are
talking about a transaction
ex: orders, invoices, transport documents,
order confirmation
-> order cycle, better expressed as order-to-
payment cycle
The flow can be obviously direct or reverse
- Planning data, which have to do with the
planning process
Demand forecast, inventory status, production
plans, capacity (operational)
- Product and marketing data: examples
o product range, SKUs (Stock Keeping Units, considering all the products with high level of detail),
all the characteristics you should know about the product in order to manage them
o phase-in/phase-out
o promotions (artificial trigger, which can change the natural pattern of the demand)

23
 01. Introduction to Logistics and Supply Chain Management

3.5.2 Customer service

- healthcare industry
Drugs/medicines, supplies, food, medical
equipment, documents, people (patients and
doctors and nurses)
- Aftersales service
- Hospitality/tourism
- Food service
- Humanitarian

Many customer service factors are strongly affected by

the logistic process and should be considered as key
performance indicators of the process itself:
- the order-delivery cycle time
- the delivery accuracy
- the order fill rate
- tracking information from order to delivery

3.5.3 Reverse logistics

- unsold products, typical of many sectors (ex:
books sector, e-commerce -> returns management)
- End of life products
- Logistic of packaging (both final and all the others
useful/instrumental along the supply chain)

Differences between direct and reverse flow

The main problem with reverse logistic, compared to the
direct/forward logistic, is in terms of complexity

1. Divergent vs convergent flow

In many industries the direct/forward flow is usually divergent, you start from some sources and you
reach millions of delivery point
In the reverse logistic type of flow, the flow is convergent, you start from a lot of source points and arrive
a single warehouse in which you collect the products
2. With forward logistic you manage items which are identified from the very beginning, for instance
through bar codes, Radio Frequency Identification (RFID)…
With reverse logistics we have a lot of non-identified items, so it is more difficult to find them
3. Might-have transformation for the forward logistics vs must-have transformation for reverse logistics
Once we collected all the non-identified items, we have to transform them

24
 02. Introduction to Logistics and Supply Chain Management

ex: considering unsold edible food, so very useful to fulfil the need
of poor/hungry people
You need to collect this type of products all along the supply chain
(manufacturer, supplier, and maybe also from consumer, very
difficult task)
This process of collection is very difficult and complex, because you
need to collect a lot of different goods from different points of
origin to a single “food bank”, in order to after delivering them to
people in need

3.5.4 The “weight” of marketing mix variables

Already 30 years ago the importance of Customer Service was in line with the importance of more traditional
product mix variables such as price and product
“Importance of Marketing Variables by Industry” Council of Logistics Management 1988

3.5.5 Service level expected by the customers of the Logistic Service Providers (LSPs)

Source: European Logistics study 2008-2009 (ELA – AT Kearney)

• Service level expected by the customers of the LSPs
• Service level given by LSPs

25
 01. Introduction to Logistics and Supply Chain Management

4 CONCLUSIONS
4.1.1 Supplementary Readings
- Cooper, Lambert, Pagh, 1997, Supply chain management: more than a new name for logistics,
International journal of logistics management, volume 8, pp.1-14
- Coyle, Bardi, Langley, Supply Chain Management, The management of business logistics, a supply chain
perspective, pp. 1-73

4.1.2 Relevant websites

Logistics and Supply Chain Management association websites:
- www.cscmp.org
- www.elalog.org
- www.sole.org

Practitioner Websites and industry insights:

- www.logisticsworld.com
- www.logistics.about.com
- www.logisticstoday.com
- www.supply-chain.org
- www.logisticsmanager.com
- www.scmr.com
- www.supplymanagement.co.uk

4.1.3 Challenge questions

- Suppose you are asked to tell why Logistics is not only Warehousing and Transportation. What would be
your line of reasoning? Which could be the activities under the responsibility of a Logistics or Supply Chain
Executive?
- What is the main idea behind the foundation of Logistics as a discipline?
- What is the difference between Logistics and Supply Chain Management?
- Build a strong case in favour of the following statement: Logistics is crucial for a Retailer in the Fast
Moving Consumer Goods sector

26
 03. Logistics Management and Strategy

03. Logistics Management and Strategy

0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples
-> what is the impact of logistics? It is relevant to corporate strategy and how?
- Logistics management as a competitive advantage -> How can logistics affect corporate profitability?
o Efficiency and Effectiveness
o The impact of Logistics Management on ROA
o The main Logistics Strategies/Approaches
- The way companies create value through Logistics and Supply Chain Management
-> what are the main logistics strategies?

AGENDA
- We will see different frameworks and models in order to understand these topics
o ROA model
o Supply chain strategies model
- Then we will follow a couple of seminars (Barilla and AB InBev) and real case examples, useful to
understand the models
- Finally, we will use critical thinking
[Supplementary readings: 4th (ROA model) and 6th (SCS model)]

1. Logistics and the marketing mix

2. The impact on ROA
3. The main Logistic Strategies
4. Case studies and executives

1 LOGISTICS AND THE MARKETING MIX

LOGISTICS MANAGEMENT: OBJECTIVES

“Logistics Management is that part of supply chain management that plans, implements, and controls the
efficient, effective forward and reverse flow and storage of goods, services and related information between the
point of origin and the point of consumption in order to meet customers' requirements”
CSCPM (Council of Supply Chain Management Professionals)

- Effectiveness = to meet customers’ requirements

- Efficiency = to control costs

27
 03. Logistics Management and Strategy

LOGISTICS AND THE SOURCES OF VALUE

2 THE IMPACT ON ROA (RETURN ON ASSETS)

ROA (Return on Assets) is a KPI of

profitability

𝑃𝑟𝑜𝑓𝑖𝑡
𝑅𝑂𝐴 =
𝐴𝑠𝑠𝑒𝑡𝑠 𝑖𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑠
𝑅𝑒𝑣𝑒𝑛𝑢𝑒𝑠−𝐶𝑜𝑠𝑡𝑠
= 𝐹𝑖𝑥𝑒𝑑 𝐶𝑎𝑝𝑖𝑡𝑎𝑙+𝑊𝑜𝑟𝑘𝑖𝑛𝑔 𝐶𝑎𝑝𝑖𝑡𝑎𝑙

NB: this indicator is useful to understand

the impact of logistic costs on the
corporate profitability

There are at least 4 different ways through which logistic can

impact on ROA, but they are more
1. On costs -> we need to understand the type of
cost and the single impact
2. On revenues, through a direct impact on service level
3. On profit -> we need to understand the service-
cost trade-off -> what is the right trade-off which has
the right impact on profit?
4. On working capital
5. On fixed capital -> consideration of logistic-related assets
6. ROA trade-off -> consideration of the trade-off
between all these different impacts

NB: In this curse, we will address just to the impact that logistics has on the ROA,
so on a relative relation, we do not talk about it in absolute terms (we will not
assess a good/bad value of ROA, but just variations
This because ROA is affected by many other factors, for which is needed a
strategic planning: finance, product development, sales, marketing, risk, time
period…

28
 03. Logistics Management and Strategy

(1) REDUCING LOGISTICS COSTS

- By improving the efficiency/productivity of logistic
activities which generate logistic costs:
o transportation
o handling + storage -> warehousing
o use of packaging
o administrative procedures
- By increasing the efficiency/productivity of the entire
logistics process by integrating the individual activities
(looking at the overall logistic costs through a total cost
analysis, trade-off analysis, systemic analysis)
- By improving the quality of the logistic activities/process
and thus reducing the non-conformity management costs
[logistics costs can consider also the inventory-related costs -> storage, cost of capital, cost of obsolescence; but it
is a difficult element, because it considers components which are related both to costs and revenues]

ex: suppose you are a retailer in a supply chain composed also of a

supplier and a manufacturer
If 100 is the final value of the product, a part of this value (ex: 5%)
is logistics costs in your perspective as retailer; another part of
100 (65€) is the cost the retailer pays to the manufacturer to
acquire the product
In the perspective of manufacturer, the 65€ seen by the retailer
consider a part of logistic costs (ex: 6%) and a part (40€) to
acquire the raw material from the supplier
Even the supplier has its own logistic costs, which are part (5%) of the 40€ payed by the manufacturer
So, considering a single perspective you have lower value of logistics costs on the value of product, but if you take
the overall perspective, the value is of course higher = 5 + 6 + 5 = 16%

(2) INCREASING THE REVENUES -> SERVICE LEVEL

- By improving customer service
o getting a “premium price” for the better service
o increasing customer retention
o increasing the number of customers
o entering new markets
- By “active” outsourcing, that is by providing logistic services to other companies

2.2.1 Classes of indicators

- Product Range. It is related to both the product families (width) and the number of items per each family
(depth). The higher the product range, the higher the probability to fulfil customer needs but the greater
the supply chain complexity
- Availability-related: Fill Rate. Indicators that measure the capability of the supplier to satisfy the requests
of the customers through the inventory on hand in the warehouse
- Time-related. Indicators that have to do with speed or punctuality in relation to what the customer
expects or was promised
- Accuracy. It includes all the indicators that measures the capacity of the supplier to be accurate and
compliant with the requests of the customer
- Compound Indicators. Mix of more than one indicator belonging to the different categories listed above.
They provide a concise and usually more customer-related measure of customer service

2.2.2 Availability related - Fill Rate

𝑓𝑢𝑙𝑓𝑖𝑙𝑙𝑒𝑑 𝑑𝑒𝑚𝑎𝑛𝑑
𝐹𝑖𝑙𝑙 𝑟𝑎𝑡𝑒 = 𝑟𝑒𝑐𝑒𝑖𝑣𝑒𝑑 𝑑𝑒𝑚𝑎𝑛𝑑 , it is the most important indicator we use to identify the availability level
- Num: Demand fulfilled with the inventories you have
- Den: Received demand, considering also the potential demand

29
 03. Logistics Management and Strategy

Our units of analysis can be order level, item or SKU level

- Order Fill Rate. Percentage of customer orders satisfied from stock on hand. It depends on the Fill Rate
measured on both the order lines (Lines Fill Rate) and the single pieces (Item Fill Rate).
- Item Fill Rate. It’s the Fill Rate measured at the single Item/SKU Level. It can be calculated with reference
to different “units”
o Lines. Line (Item) Fill Rate
o Cases. Case (Item) Fill Rate
o Pieces. Piece/Unit (Item) Fill Rate
- Stockout. It’s the opposite of the Fill Rate (1 – IFR)
- Order Completeness. The delivery is complete when the goods requested (i.e. all the order lines and all the pieces
per line) are delivered outright. It depends on both the Item Fill Rate and the delay the customer is wiling to accept
in case of partial stockout

NB: OFR and IFR are obviously related

ex: consider 1 order consisting of 10 order lines, considering that 1 order line = 1 product/SKU
if IFR = 99%, which means that out of 100 of item requested, 99 can be fulfilled with the inventories, while 1 no
in this case, 𝑂𝐹𝑅 = 𝐼𝐹𝑅10, because each order considers 10 order lines -> 𝑂𝐹𝑅 = (99%)10 ≈ 90%

The connection has to do with the order profile; it is obvious that OFR is always lower than IFR
Both the indicators are useful, but we should prefer one of them according to the type of customer we have
IFR is preferable if the customer prefers to get some different products and doesn’t look too much if the order is
not complete; it is an indicator very important for the logistic manager, because you work at the level of the
single product unit, you work on the IFR in order to improve the OFR

2.2.3 Time-related indicators

There are at least two main time-related indicators
- Order Cycle Time. It is the time elapsed between the order issuing and the order delivery. It depends on
both the time required to carry out the activities and the Order Fill Rate (OFR)
It is the time to fulfil an order, and indicates how fast is the system, it is a measure of the network speed
- Punctuality/On Time. It is the ability of the supplier to deliver in compliance with the time window (ΔT)
arranged with the customer. In case there is no specific agreement on the time window, punctuality is the
ability to be compliant with the promised Order Cycle Time
How the company is reliable in terms of prompts it made to the customer
- Delivery Frequency. It is the number of deliveries planned in a reference time window (day, week, month,…). It is
related to the minimum dimension of the order. It is affected by the shelf-life of products, by the storage capabilities
of the customer and by the incidence of transportation costs on the product value. The higher the frequency the
shorter the Order Cycle Time perceived by the customer

NB: the first two indicators are very different and require two different kind of logistic chain
- Order-cycle time requires a chain very fast
- Time-reliability requires a chain very reliable
It is obviously very difficult to have a chain which is both fast and reliable

2.2.4 Quality/Accuracy related

Order compliance. It is the ability of the supplier to deliver goods that are compliant with the order
specifications, in terms of both items and quantities. it can be measured on different levels (orders, order
lines, units) Attention in delivering the right products, in terms of right SKU and right quantity
- Document compliance. It is the ability of the supplier to create and send documents consistent with
customer specifications and with the goods delivered
Attention in delivering the right documents, which means that these documents are consistent along the
order cycle and with the delivered goods
- Packaging compliance. It is the ability of the supplier to use packaging compliant with customer
expectations (i.e. consistent with the customer storage system, adequate protection of the goods, ….).
Attention in delivering the right packaging
ex: if the space in the warehouse can take into a box high 1.5 m, the packaging of your product can’t be
higher than this, otherwise it is a cost in managing the issues

30
 03. Logistics Management and Strategy

2.2.5 Compound indicators

DIFOT (Delivered In-Full, On-Time) or OTIF (On-Time In-Full). It is a combination of two indicators:
o Completeness in full
o Punctuality on time
o [Accuracy]
It measures if the Supply Chain is able to deliver the expected product (reference and quality) in the quantity
ordered by the customer at the place agreed by the customer at the time expected by the customer (in most of
the cases, with a tolerance defined with the customer)

NB: it has the advantage to measure the performance of the whole logistic organization to meet the customer
service expectations. To reach a good OTIF level, all the functions of the supply chain (e.g. orders taking,
procurement, suppliers, warehouses, transport...) have to work at their best level. The main contribution of OTIF
is to allow to see at a glance how the company delivers its customers.

How we can create value increasing the service level?

We can use the service level to have a positive impact on ROA, because you can increase it in order to
- Have a premium price which able you to provide a better service
- Increase the customer base, increase the volumes

(3) INCREASING THE PROFIT: THE COST-SERVICE TRADE-OFF

2.3.1 The “margin-service” relationship

The connection between service level and logistic

costs is called cost-to-serve curve

The connection between service level and revenues is

called revenue-service curve, and it is more difficult to
be drawn; in many industries it has this shape

When you increase service level, you usually increase

revenues, with a more than proportional relationship
After a certain point, we have extra-service, from
which the revenues increase less quickly, because
there are less customers available to pay more for the
extra service

What the logistic man aims to do is find out which is the point that brings the maximum profit, that is the one in
which revenues are very distant from the logistic costs curve
In order to find out this point, we need to have in mind that the best service level is not the highest one, but the
one which is able to optimize both the cost and revenues
NB: the service level optimization depends a lot on what the customer wants, so on the type and size of the
cluster you are considering; we will have that for some customer the service level optimization point is different
from others, because for some customer having too high SL is not a real value, and it becomes just a cost for you

(4) THE WORKING CAPITAL

2.4.1 Reducing the current assets
- By reducing inventories
o removing “dead” inventories
o right-sizing and deploying the safety stocks (centralization/decentralization)
o right-sizing the cycle stocks
o …
- By reducing/controlling the “cash-to-cash” cycle time

31
 03. Logistics Management and Strategy

2.4.2 The Cash to Cash cycle (C2C)

It is the time between when you use money to produce the good and the moment in which you collect the money
from the purchase of the good
If the C2C is too long, you have of course a financial problem, if it is negative, this is an advantage, because you
have money that you can invest in other tasks
The Cash-to-cash cycle (working capital) can be defined as ‘‘the average days required to turn a dollar invested in
raw material into a dollar collected from a customer’’ (Steward, 1995) and can calculated as:

𝐶2𝐶 = 𝐷𝑆𝑂 + 𝐷𝐼𝐻 − 𝐷𝑃𝑂 -> working capital in terms of time

Where:
- DSO, days of sales outstanding, are the
average days required to collect accounts
receivable from customers
- DIH are the average day of inventory
holding -> inventory period, lead time of
the supply chain
➔ The sum of these two elements is the time
between when you start the production
and the moment in which you are paid; the
longer is this period of time, the worse for
you in terms of financial impact

- DPO, days of payables outstanding, are the average days allowed by suppliers to settle accounts payable
-> it is an advantage, because it reduces the overall C2C cycle

Working capital = Inventories + Account receivable + Account payables

-> working capital in terms of money
[Inventories = raw material + WIP + finished products]

➔ Logistics affect working capital because it affects the inventories, which is one of the main components

2.4.3 Examples of C2C (Farris & Hutchison, 2002)

(5) REDUCING THE INVESTMENT IN FIXED ASSETS

Fixed assets:
- Equipment/infrastructures (ex: trucks, warehouses…)
- Information system
- [People: they can be considered as fixed asset in the ROA model if the contracts are for long periods, if
you have difficulties in managing them in terms of capacity]

32
 03. Logistics Management and Strategy

A company can create value working on fixed assets through:

- By better logistics/supply chain planning (less variability and less unpredictability)
Increase asset utilisation and productivity (ex: increase the usage of trucks or warehouses)
o reducing the manufacturing and warehousing capacity
o reducing the transport assets
o reducing personnel/staff
- By outsourcing logistic activities -> utilisation of the make-or-buy lever
The main benefit you have in this case is the reduction of denominator of ROA, not just to reduce logistic
costs or increase production level

(6) MANAGING THE TRADE-OFFS

Managing the trade-offs between the different sources of value (with a net positive effect on ROA)
When we took some decisions about one single source of value in order to reduce the ROA, also other source of
value can be modified -> what is the overall/real impact of the ROA?
- By improving productivity (meaning reducing logistic costs) through investments in fixed assets
(increasing the fixed capital) (e.g. information systems)
- By improving the efficiency of the logistics activities (so reducing the logistics costs) accepting an increase
in the inventory level (increasing the working capital)
- By improving customer service and revenues accepting a less than proportional increase in the logistics
costs
- By reducing the fixed capital through outsourcing accepting an increase in the logistics costs

All these elements are of course used through a qualitative approach, but they can be used also in a quantitative
approach

EXERCISE
Assume that a manufacturer in the FMCG industry has the following base-line performance:
- Annual revenues: € 100m
- Annual operating costs: € 90m
- Fixed assets: € 50m
- Current assets: € 50m

2.7.1 Questions
1. First, assess the base-line ROA
2. Second, assess the impact on ROA of the outsourcing of distribution to a 3PL (third party logistic service
provider), given the following assumptions:
o Additional annual operating costs: € +1m
o Reduction in Fixed assets: € -5m
3. Third, evaluate what the increase in revenues should be to justify a Customer Service improvement
project (i.e. reducing order cycle time through more inventories and better information systems),
assuming the following:
o Impact on Current assets (inventories): € +2m
o Increase in Fixed assets (warehouse and information systems): € +2m

2.7.2 Solution
𝑝𝑟𝑜𝑓𝑖𝑡 𝑟𝑒𝑣𝑒𝑛𝑢𝑒𝑠−𝑐𝑜𝑠𝑡𝑠 100−90 10
1. 𝑅𝑂𝐴(1) = = = = = 10%/𝑦𝑒𝑎𝑟
𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑑 𝑓𝑖𝑥𝑒𝑑 𝑎𝑠𝑠𝑒𝑡𝑠+𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑎𝑠𝑠𝑒𝑡𝑠 50+50 100
10−1 9
2. 𝑅𝑂𝐴(2) = 100−5 = 95 = 9,47% This project does not make sense in terms of impact on ROA
𝑅(3) −𝐶(3)
3. What are the impacts on revenues that make the ROA(3) = ROA(1) 𝑅𝑂𝐴(3) = 100+2+2 = 10 %/𝑦𝑒𝑎𝑟
First of all, we have to understand the effect that service level increase has on revenues

33
 03. Logistics Management and Strategy

Assumptions
o We use the increase in service level in order to obtain/apply a delta price -> premium price
In this case, we have not an impact on costs -> C(3)=C(1)
𝑅(3) − 90
𝑅𝑂𝐴(3) = → 𝑅(3) = 10% ∗ 104 + 90 = 100,4
104
-> Δ𝑅 = 0,4 %/𝑦𝑒𝑎𝑟 Δ𝑝𝑟𝑖𝑐𝑒 = 0,4%

o We use the increase in service level in order to increase the market share
In this case, we do not change the price, but we increase the volumes
Another assumption inside this is that we have the same cost structure -> C(3)=0.9*R(3)
𝑅(3) − 0.9 𝑅(3) 10% ∗ 104
𝑅𝑂𝐴(3) = → 𝑅(3) = = 104
104 1 − 0.9
-> Δ𝑅 = 4 %/𝑦𝑒𝑎𝑟 = +4%

Beyond the exercise, we get that the impact on revenues depends on how the customer service improvement is
used to create values -> there are different ways (ex: applying premium price or increasing volumes, so increasing
market share)
If we do not express how the customer service improved is used, we cannot understand the real impact on ROA

3 THE MAIN LOGISTIC STRATEGIES

The main logistics strategies can be identified and classified in different ways. Two examples are:
1. Lean Vs Agile Supply Chain Strategies
Lee, 2002, Aligning Supply Chain Strategies with Product Uncertainties, California
Management Review , volume 44 number 3, pp.105-119
2. Time-based, Asset-based, Technology-based Relationship-based Supply Chain Strategies
Coyle, Langley, Gibson, Novack, Bardi, 2009, Supply Chain Management
A Logistics Perspective, chapter 16

1ST SUPPLY CHAIN STRATEGY (BASED ON HAU L. LEE’S FRAMEWORK)

Main Idea: Supply Chain strategic approach and policies should be consistent with the “profile” of the Supply
Chain in which the company operates (from Supply Chain profile to Supply Chain strategy)

References:
- Lee, 2002, Aligning Supply Chain Strategies with Product Uncertainties,
California Management Review , volume 44 number 3, pp.105-119
- Lee, 2004, The Triple-A Supply Chain, Harward Business Review, October 2004

3.1.1 Define the Supply Chain strategy - Method

1. Know your(s) Supply Chain(s) -> SC profile, characteristics of SC and context in which you operate
2. Choose your Supply Chain strategic approach
3. Define your Supply Chain strategic policies (consistently) -> relation between profile and strategy
4. Review your strategy over time
This model is very interesting because it is also very simple to be understood

3.1.2 1. Know your Supply Chain -> supply chain profile

We can describe our SC profile with a matrix built on two dimensions Demand-side
- Demand-side uncertainty, how is uncertain your demand, uncertainty
downwards in the supply chain, in the market Low High
- Supply-side uncertainty, how much is complex and uncertain you SC Supply-side Low
uncertainty High
looking upwards, from your suppliers

34
 03. Logistics Management and Strategy

Demand-side uncertainty -> Main Factors: Supply-side uncertainty -> Main Factors:
- Product variety -> wider product range means more - Number, uniqueness, localization of suppliers
uncertainty - Stability of the supplier base
- Consumer demand volatility/unpredictability - Duration of supply contracts
- %Sales induced by promotions/discounts - Level of dependency on the supplier base
-> promotions make the demand more unstable how much you depend on the rate of your suppliers in
- Price elasticity order to reach the objective
- Price fluctuation - Number of echelons in the upstream supply chain
- Product life cycle (length and position in the curve) - Financial performance of suppliers
- Introduction of new products - Country risk of suppliers
- Obsolescence risk -> product with shorter time on the - Maturity level of processes and technologies
market, demand-side is more uncertain if they are old or very new, If you do not expect an
- Number and characteristics of the market segments evolution in terms of technology, the environment is
served more stable , and so the production is more reliable for
- Competition level of the market you and even for suppliers
- Number of echelons in the downstream supply chain In other case, the supply network can evolve following
- Importance of time-to-market the changing of the technology, maybe your provider can
- Penetration of new markets be different
- Number and variety of the sales channel -> the - Product complexity/Number of components
complexity increase if you have to manage a lot of Complexity of Bill of Materials: if we have a stable and
channels (ex: supermarkets and e-commerce) clear BOM, it is easier to build the production plan, for
- …. both you and your supplier
- Quality level required
There are a lot of different variables through which you can - Evolution of raw materials
describe your demand-side uncertainty and rank it Which type of components you need (ex: if you need a
lot of electronic components that could be a problem,
because the SC is more complex)
- ….

Even in this case you can describe the supply-side uncertainty

through a lot of different variables, and rate them as low,
medium and high impact on the uncertainty

The first step that Lean ask to you is to place your SC in this matrix

3.1.2.1 Supply Chain Profile in different sectors Demand-side uncertainty

The Lee’s model can be applied to different units of analysis: entire industries, sub-industries, individual
companies; what we try to do applying this model, we aim to conduct a relative positioning in the matrix, not an
absolute positioning, so it is better to compare different industries, in order to define if they are similar or not
We can even compare a unit of analysis over time

The fashion industry has The consumer eletronics industry has of Historically, the food industry is
- Low supply-side uncertainty course higher supply-side uncertainty, placed in the top-left quadrant of the
- High demand-side because the SC is much more complex matrix
and it is short of capacity and the
uncertainty
volatility of price is high; this is higher
than the demand-side uncertainty

35
 03. Logistics Management and Strategy

ex: comparison between Food and Electronics (with some simplifications)

Drivers of positioning Food Electronics
Production technologies Relatively stable Rapidly evolving
Components BOM Stable Rapidly evolving
SS uncertainty
Location of suppliers Local and global Global
… -> Low medium -> High
Product range Low High
Price volatility Low High
DS uncertainty Demand unpredictability Medium/Low High
Product lifecycle Long Short
… -> Low medium -> High

NB: the real value of this exercise is to understand the drivers of

positioning, both of supply side and demand side

It is important to position the company and the industry in this

matrix, but is much more important to see if this position is
moving along time, looking if there are some trends in this matrix,
in order to link this to the supply-chain strategy

3.1.3 2. Supply Chain Strategies (Strategic Approaches)

1. Lean (efficient) Supply Chain
2. Responsive Supply Chain
3. Risk hedging Supply Chain
4. Agile Supply Chain

We will concentrate just on the first and the last approaches, because they are at the extremes, and the other
two are a sort of connection between the extremes; this does not mean that are worse, it’s a matter of
consistency between profile and strategy, no one of these approaches is better than the others

1. “Lean” Supply Chain Strategy

- Objective: The “Lean” Supply Chain strategy aims to maximize the “expected” Return On Assets (ROA)
creating both cost and value (service-related) competitive advantages
It aims at optimizing the supply chain, which can be translated into 𝑚𝑎𝑥 ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
𝑅𝑂𝐴𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑
- Policies consistent with this approach (examples):
o Non value adding activities elimination
o Focus on economies of scale and scope, working on volume and usage of assets you have
o Stock control and centralized management, inventory management optimisation
o Optimization techniques aim at distribution and production capacity maximization
o Client-supplier info sharing automation (in handling, process…)
The context more consistent with this approach should be a stable context
You can optimize the system if it is stable over time, if it does not change every few months; you can always
optimize it, but of course is more difficult, because you have to adjust the calculation every time the context
changes -> the main assumption is the stability of the context

2. “Responsive” Supply Chain Strategy

- Objective: The “Responsive” Supply Chain strategy aims to mitigate risks (related to the changes and the
diverse needs of the customers) on the demand-side
- Policies (examples):
o Build-to-order and mass customization approach, suitable to satisfy market specific demand
o Postponement strategies
o Extra-capacity
o Short lead times
o Reduction of Time-to-market

36
 03. Logistics Management and Strategy

3. “Risk Hedging” Supply Chain Strategy

- Objective: The Risk Hedging” Supply Chain strategy aims to reduce risks in supply disruption
- Policies (examples):
o Risk management orientation (resilience)
o Inventories of raw materials/components
o Back up suppliers
o Resource sharing inside supply chain, in order to share the risk of supply interruption (e.g. co-
ownership of raw materials and components stocks with other firms) − ICT adoption
o Real time information on inventory and demand
o Dynamic allocation of inventory and demand between partners who share the same warehouse
inventories

4. “Agile” Supply Chain Strategy

- Objective: The “Agile” Supply Chain aims to reduce risks on both the supply and the demand side,
combining Risk Hedging and Responsive Supply Chain strategies
This method aims to
o mitigating the supply chain risk, which means manage and reduce uncertainty,
minimizing the standard deviation of ROA, not the ROA itself -> 𝑚𝑖𝑛 𝜎𝑅𝑂𝐴
o working on empowering flexibility
- Policies (examples):
o Inventory and other capacity resources shared between partners to face stock out and capacity
interruption
o Short lead times, a very long supply chain makes you not agile at all
o Short TTM, making easier to introduce new products in the market
o Postponement
o Supply Chain visibility of what happen along the chain, in order to better understand all the
possible risks along the chain and reduce them
o Supply Chain collaboration between players along the chain
o Alternative suppliers/sources, in order to be independent from them

NB: it is important to notice how this approach is very different from the Lean approach

3.1.4 3. Consistency between Supply Chain profile and Supply Chain strategy
- Lean approach can be used in a very stable environment
- Agile approach can be used in a very uncertain context
- Responsive approach uses all the policies of the agile approach
in order to manage the demand-side uncertainty, but has low
supply-side uncertainty
- Risk hedging approach uses all the policies of the agile approach
in order to manage the supply-side uncertainty, but has low
demand-side uncertainty

You have to check out if your strategic approach is consistent with your supply-chain profile, and you have to
check it over time, because the trend of you supply-chain profile can be different in the future

Distribution network optimization policy Supply chain visibility

ex: how many warehouses you need ex: relation with customers

37
 03. Logistics Management and Strategy

3.1.5 1st Supply chain Strategy: a recap

Comparison between Supply Chain Strategic Approaches
Lean Approach Agile approach
Main Objective Fulfill demand at the lowest cost Fulfill demand quickly
Product Design Maximize production at the minimum Create modules in order to postpone the
Strategy costs product differentiation
Price Strategy Lower margins: price is one of the most Higher margins: price is not an important
important driver for the customer driver for the customer
Production Cost reduction through high utilization Flexibility of capacity to have buffers to cope
Strategy with demand/supply uncertainty
Inventory Minimize inventory in order to reduce Keep inventory in order to face
Strategy costs demand/supply uncertainty
Lead Time Lead time reduction, with no impact on Lead time reduction, even though the costs
Strategy costs are high
Suppliers Suppliers selected on both costs and Selected on the basis of their lead times,
strategy quality flexibility, reliability, and quality

2ND SUPPLY CHAIN STRATEGY

3.2.1 1. Time based strategies

- Cycle time reduction
o Reducing logistic processes time
o Information and Communication Technology to have faster (and more accurate) information
o Faster decision making
- Time-reduction logistic initiatives
o Cross docking (no storage, products flow faster)
o Collaborative Planning (such as VMI = Vendor Managed Inventory) to reduce the order cycle
times
- Pull approach
o Fast manufacturing systems to reduce cycle time for ATO (assembly to order) productions
o Postponement

3.2.2 2. Asset productivity strategies

- Inventory reduction
o Collaborative Planning allows inventory reduction for both the trade partners
o Cross Docking
- Equipment utilization
o Reduction of logistic facilities
o Use of technology (such as handheld terminals) to improve the utilization rate of the assets to
move and store products
o Direct delivery, if possible
- Third party/Contract logistic Services
o Outsourcing of Transportation
o Outsourcing of the whole logistic process

38
 03. Logistics Management and Strategy

3.2.3 3. Technology based strategies

- eCommerce B2b (Business To Business)
o eProcurement
o eSupply chain
- Automatic Identification projects
o RFId (Radio Frequency Identification)-enabled processes
- Automation of Warehousing and Materials Handling
o Automated Warehouses
o Automated Picking and Sorting Systems
o Wireless Technology (Wi-Fi in the warehouses, Mobile Workspace, …)

3.2.4 4. Relationship based strategies

- Integration of Supply Chain processes
o Digitalisation and integration of the order-to-payment cycle (the process at the interface between
supplier and buyer)
o Standardisation of industrial labels and packaging
- Collaboration
o Exchange of supply chain planning information
o Collaborative planning (VMI, CRP, CPFR, …)
o Restructuring of the logistics network to streamline the flow of goods and reduce the number of
facilities

Listen carefully to the presentations and answer the following questions

- What is the role of the Logistics/Supply Chain Executive in the corporate organization chart?
- What are the activities under his responsibility?
- Why is Logistics/Supply Chain Management important in the company and how does logistic/supply chain
performance relate to the overall corporate performance?
- How is the Logistic/Supply Chain Executive aligning the Logistic/Supply Chain strategy with the overall
corporate strategy?

4 CONCLUSIONS

SUPPLEMENTARY READINGS
- Ballou, Defining a sales-service relationship, Business logistics Management, pp.92101
- Braithwaite, Samakh, 1998, The cost-to-serve method International journal of logistics management,
volume 9, pp.69-84
- Christopher, The corporate role of logistics, The strategy of distribution management, pp.14-33
- Christopher, Ryals, 1999, Supply chain strategy: its impact on shareholder value,
- International journal of logistics management, volume 10, pp.1-10
- Fisher, 1997, What is the Right Supply Chain for Your Product, Harward Business Review, March-April
1997
- Lee, 2002, Aligning Supply Chain Strategies with Product Uncertainties, California
- Management Review , volume 44 number 3, pp.105-119
- Lee, 2004, The Triple-A Supply Chain, Harward Business Review, October 2004
- Riedl, J., Farag, H. and Korenkiewicz, D., 2016, Transportation and logistics in a changing world: the
journey back to profitable growth, (BCG report)
- Slone, Mentzer, Dittmann, 2007, Are you the weakest link in your company’s supply chain?, Harvard
business review, September 2007, pp 116-127

39
 03. Logistics Management and Strategy

CHALLENGE QUESTIONS
- What’s the difference between effectiveness- and efficiency-related performance in Logistics? Give some
examples of effectiveness related performance?
- What is the meaning and the use of the ROA model?
- Provide some examples of different logistic strategies, their expected impact on ROA and their alignment
with the corporate strategy?
- What are the main types/models of supply chain strategy according to Lee’s Model? How can these
approaches be related to the main variables describing the external context? What evidence of these
approaches could you find in examples/seminars/visits seen in class or that you know?

CRITICAL THINKING

4.3.1 Rank in decreasing order of logistics costs (as % of revenues)

Main drivers of logistic costs as %revenues, in relative terms
- Value density of products logistic costs as %revenues decrease if value density increases, because
revenues increase direct proportionally as value density, while not all the logistic costs increase direct
proportionally as value density, just some of them
- Complexity it drives up logistics costs; given by, for example
o service level target
o granularity
1
𝑙𝑜𝑔𝑖𝑠𝑡𝑖𝑐 𝑐𝑜𝑠𝑡𝑠 ≈ ∗ 𝑐𝑜𝑚𝑝𝑙𝑒𝑥𝑖𝑡𝑦
𝑣𝑎𝑙𝑢𝑒 𝑑𝑒𝑛𝑠𝑖𝑡𝑦
We can build a matrix with two dimensions: value density and complexity; of
course, if we increase the complexity level, also the logistic costs increase;
we can build some iso-curves of logistic costs (as percentage of revenues)

ex:
- pharmaceutical sector -> medium complexity, high value density
- water bottle -> low value density, low complexity in terms of logistic
- e-commerce of fashion -> medium-high level of both complexity and value density
- e-commerce of food -> medium-low level of value density, very high level of complexity
-> we can see that e-commerce and water bottle are quite on the same iso-curve, so they are comparable in
terms of logistic costs, while the other two are not
[Granularity moves higher when you move small packages, when all starts with big pallets, but at the end of the
supply chain we work with single products]

4.3.2 What is the difference between Order Cycle Time and Supply Chain Lead Time?
They are two very different concepts, they are not synonymous at all
Supply chain lead time or logistic lead time or operation distribution lead time
We have many operational stages/phases of the supply chain we are considering,
which could be phase of transformation, transportation, storage… each of these
phases takes time
It measures how long is the supply chain in terms of time
➔ The supply chain LT is the sum of the LT of each single operational stage

Order cycle time or the LT seen by the customer

Usually it is not the overall supply chain LT, because along it we can have some decoupling points, from which you
divide your SC in two parts:
- the one on the left, you work by forecast
- the one on the right you work by order
Inside the decoupling point you have finished products, so you can fulfil the orders from this point on
➔ The OCT is the sum of
o Order LT, time to collect and process order from the customer
o Customer LT, which is the part from the decoupling point to the end of the SC

40
 03. Logistics Management and Strategy

As logistic man you have to position the decoupling point in the right place: if you are closer to the customer, you
can reduce the Customer LT, so also the OCT; on the other hand, you are also increasing the portion of
operational phases which have to work on forecast

4.3.3 What is the relation between Order Fill Rate (OFR) and Inventories?
- OFR directly proportional to inventories
- OFR inversely proportional to inventories
- OFR an exponential function of inventories
- Inventories an exponential function of OFR

𝑜𝑟𝑑𝑒𝑟 𝑓𝑢𝑙𝑓𝑖𝑙𝑙𝑒𝑑 𝑓𝑟𝑜𝑚 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑖𝑒𝑠

𝑂𝐹𝑅 =
𝑜𝑟𝑑𝑒𝑟 𝑟𝑒𝑐𝑒𝑖𝑣𝑒𝑑 (+ 𝑝𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙)

Of course, the OFR has a limit, which is 100%, which means that
you are able to fulfil all the demand, also the potential one, with
your inventories
Every time we want to increase the OFR, we have to increase also the inventories level, in an exponential way, in
order to fulfil all the orders

This curve is a very good example of cost-to-serve curve:

- Inventories in this case are a good example of logistic costs
- OFR is a good example for service level

[from tutorship session]

Why is the relationship between OFR and
Inventories exponential (and not linear)?

The first curve describes the demand

coming from the market in terms of
probability
We have an expected demand and other
possible values, both lower and higher; in
this case we have a Normal distribution

Safety stocks are the goods we maintain in order to satisfy the demand when it is higher than the expectation; in
the graph on the right, we have the relation between safety stocks and the service level
- If we are on the mean value, we need 0 safety stocks
- Moving from 50 to 75, which means to cut out the 25% of probability to go out of stocks -> in order to do
that, we need K1 quantity of safety stocks
- Moving from 75 to 87,5%, you cut out other 12,5% of probability to go stock out -> we need a K2 quantity
of safety stocks
The difference between K1 and K2 is higher than the difference between 0 and K1

41
 03. Logistics Management and Strategy

4.3.4 How are service level indicators chosen? And the target values
for these indicators?
We should use the connection between the service level with logistic
costs and revenues increase (∆revenues) -> revenue-service curve
- Logistics costs increase more than proportionally compared to
service level, in an exponential way
- Revenues increase with service level, but we have an inflection
point, in which we have saturation of customers, so from this
point the customer mess the increase in service level

The target service level should be chosen considering the impact on profit, not just on cost or on revenues
- The logistic manager has to understand the shape of the logistic costs curve
- The marketing manager has to understand the shape of revenues curve

The curves change according to In order to apply this idea in the real world, we need customer
- customer segmentation segmentation, which means to have
- technological innovation - Different service level
- consumer behaviour and - Different KPIs
expectations - Different targets for the different cluster of customers

-> these curves change over time, so we need to understand their pattern every time we can
ex: customer expectation in terms of OCT changed a lot in the lasts years, think about Amazon

4.3.5 What is the correct use of the ROA model regarding Logistics Strategy?
- Think about the possible impacts of logistic strategies on corporate profitability
- Calculate the impact of logistic strategies on corporate profitability
- Understand the trade-offs between possible impacts of logistic strategies on corporate profitability
- Assess the consistency between logistics strategy and corporate strategy

The most important use of ROA is not to calculate quantitatively the different impacts on corporate profitability,
but to understand qualitatively the trade-off between them

𝑅−𝐶
ex: impact of full logistic outsourcing strategy on corporate profitability 𝐹𝐶+𝑊𝐶
- Reduction of logistic costs, because the company moves to an experienced third-party logistic provider,
with capabilities to use economies of scale/scope
- Reduction of fixed capital, because you can reduce the assets which tied up in logistic within the internal
supply chain
- No impact on working capital, because you are not outsourcing also the inventories, the planning part
- The impact of revenues depends on the impact on service level, and we can expect an increase of it,
because it is managed by an experienced third party, which can work on on-time deliveries and on OCT

42
 03. Logistics Management and Strategy

4.3.6 Is a logistics outsourcing policy more consistent with a lean or an agile supply chain strategy?
It depends, we cannot say for sure

Let’s look at the different elements of the outsourcing process, of the relation between the company and the
third-party provider
- Control -> if the company which is outsourcing is still able to control the supply chain, the policy is
consistent with the agile approach
- Flexibility -> the network of the provider can be more or less flexible than yours; if it is more flexible,
the policy is consistent with the agile approach
- Resiliency and robustness -> if the provider has a network which is more resilient and robust than
the company one, the policy is consistent with the agile approach
➔ If we can say that we can keep control and the logistic provider is more flexible and more resilient/robust,
then the answer is YES, the logistic outsourcing is a good movement according to an agile strategy
If at least one of these elements are not accomplished, you are losing part of your agility, so the policy is not
consistent with an agile approach

It is difficult to understand the answer if you do not go in detail of your logistic process

43
03. Logistics Management and Strategy

44
 04. Warehouses: general overview

04. Warehouses: general overview

0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples
- The role of warehouses within the distribution networks
- The types of warehouses: depots/distribution centres vs transit points
- The main warehousing operations and functional areas
- The performance indicators of a storage system
- The main types of storage systems: performances and main application fields

AGENDA
1. Role of warehouses
2. Types of warehouses
3. Warehouse operations
4. Storage system

1 ROLE OF WAREHOUSES
What is the main strategic objective of the course? Looking at the time table, we will learn how to design a
logistic system, and more in detail we will learn how to design a distribution network, so the main principles
behind it and the method to structure it
We will not concentrate just on the distribution network, but also on the transportation system
These two together represent the logistic system

THE DISTRIBUTION SYSTEM

Distribution system: all (company and extra-company) resources and structures through which the goods and the
associated services are sold and transferred to the end customers

Distribution System:
- Logistic channel -> related to the physical movement of the goods
distribution network + transportation system
- Trade channel -> related to the movement of ownership of the products

THE LOGISTIC CHANNEL

The logistic channel includes:
- the distribution network, whose nodes are the
plants, the warehouses (central and regional
warehouses, transit points) and the points of sale
we have some points of origin and other of
destinations: the distribution network is the
infrastructures we need to move the goods
between these points, from origin to destination
- the transportation systems (connections) which connect the nodes of the network, which can be points
of origin, destination or other infrastructures in the middle

45
 04. Warehouses: general overview

BASIC WAREHOUSING DECISIONS

We can identify different dimensions we
have to take into consideration when we
are designing a distribution network; we
have to decide:
- Number of echelons, the number
of levels through which the
product has to pass to reach the
final point of destination
- Type of warehouse/infrastructure,
for each level
ex: with or without inventories
- Number of warehouses for each
echelon

In order to do that, we need to know some basics about the warehouse, and we need also to learn different
things in order to design the warehouse, which means to design each infrastructure area, in order to design the
different flows inside them and the operating activities, taking decisions about these two main characteristics:
- Requirements
o Type -> what is the warehouse we require
o Location -> where we should put our warehouse
o Performances requirement ex: number of loads, capacity, flows…
- Output
o Main functional areas that we have to consider in our infrastructure (storage, preparation,
sorting…); we will learn which ones we need and how to design them
o For each area, we have to determine
▪ Real capacity
▪ Technology -> traditional (activities carried out by operators)
or automated area (machines)
▪ Layout -> the structure
▪ Management policies ex: how to manage the operative cycle
-> batching, storage, routing policy
About the design of transportation
- Number of nodes
- Nodes features activities done into each node
- Application fields when/in which conditions we can use them

Importance of execution
If we have in mind numbers, performances, factors/drivers related to the performances of the system, and how
these factors impact on the system, we are able to make strategic choices
We will understand how with low amount of data we can manage the main strategic choices
ex: consider the European distribution network of an international company, that is made by only one warehouse
between the points of origin and destination
The question is if the company should add a warehouse, so it has also to decide the type of system inside it
We do not have a lot of data to manage this situation, but we have to provide quickly our position and justify it

SC = 60.000 PL
TCIN = 2.000 PL/day
TCOUT = 1.000 orders/day
1 order = 2 PL

46
 04. Warehouses: general overview

Storage area
𝑆𝐶 [𝑃𝐿] 60.000 𝑃𝐿
Traditional storage system 𝑠𝑡𝑜𝑟𝑎𝑔𝑒 𝑎𝑟𝑒𝑎 = = = 60.000 𝑚2
𝐴𝑈𝑅 [𝑃𝐿/𝑚2 ] 1 𝑃𝐿/𝑚2
AUR = area utilisation rate -> 1 PL/m2 this value depends on the type of system we choose

𝑆𝐶 [𝑃𝐿] 60.000 𝑃𝐿
Automated storage system 𝑠𝑡𝑜𝑟𝑎𝑔𝑒 𝑎𝑟𝑒𝑎 = 𝐴𝑈𝑅 [𝑃𝐿/𝑚2 ] = 3 𝑃𝐿/𝑚2
= 20.000 𝑚2
AUR = area utilisation rate -> 3 PL/m2 in this case it is much more, because we can reach higher shelves

People
Traditional storage system
2.000 𝑃𝐿/𝑑𝑎𝑦 2.000 100 ℎ/𝑑𝑎𝑦
- Inbound 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 [𝑃𝐿/𝑑𝑎𝑦]
= 20 = 100 ℎ𝑜𝑢𝑟𝑠/𝑑𝑎𝑦 → 8 ℎ/𝑑𝑎𝑦+𝑝𝑒𝑟𝑠𝑜𝑛 ≈ 15 𝑝𝑒𝑜𝑝𝑙𝑒
60 𝑚𝑖𝑛/ℎ
𝑤𝑖𝑡ℎ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 = 3 𝑚𝑖𝑛/𝑐𝑦𝑐𝑙𝑒
= 20 𝑃𝐿/ℎ we are assuming that the time for a PL is 3 minutes

1.000 𝑜𝑟𝑑𝑒𝑟𝑠/𝑑𝑎𝑦 1.000 334 ℎ/𝑑𝑎𝑦

- Outbound = = 334 ℎ/𝑑𝑎𝑦 → ≈ 50 𝑝𝑒𝑜𝑝𝑙𝑒
𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 [𝑜𝑟𝑑𝑒𝑟/ℎ] 3 8 ℎ/𝑑𝑎𝑦+𝑝𝑒𝑟𝑠𝑜𝑛
60 𝑚𝑖𝑛/ℎ
𝑤𝑖𝑡ℎ 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 = 20 𝑚𝑖𝑛/𝑐𝑦𝑐𝑙𝑒
= 3 𝑜𝑟𝑑𝑒𝑟𝑠/ℎ we are assuming that the time for a PL is 20 minutes
➔ Total needed people = 15 + 50 = 65 people

NB: we consider more than the exact result because we should approximately consider that people do not work
always at 100%

Automated storage system -> 10 people, we need just some people to control the activities

Area People + other ex: Flexibility -> Layout and handling system
elements independent
not easy to
Traditional 60.000 m2 65 be taken
into
account
concurrent, if we change the capacity of one, we have to change
Automated 20.000 m2 10 𝑆𝐶
the capacity of the other one -> = 𝑓𝑖𝑥𝑒𝑑 𝑣𝑎𝑙𝑢𝑒
𝑇𝐶
-> investments -> operative costs
+ investments in handling system and
storage system

Knowing just two data, we came out with big results, and we can make choices with some little computations; of
course, we cannot finish here
NB: it is obvious that in order to make the best choice we need other data, such as
- Cost of the land
- Investment in handling and storage system
- Features of the demand (ex: growth needs flexibility)

WAREHOUSES: THE BASIC FUNCTIONS

1. Storage: Keep the inventories
- guarantee a determined safety stock coverage -> ex: needed to cover an unexpected demand
- decouple asynchronous processes -> there could be upstream and
downstream processes, and they can be asynchronous
ex: the production plant (upstream process) is able to produce big lots
in short time, but the point of sales (downstream process) needs
smaller quantites every day
- keep the goods safe

47
 04. Warehouses: general overview

- ….. (the main roles of the inventories)

2. Flow management/materials handling: Transform the flows

- From full PL to customer orders (mixed PL, different product)
- From unpacked products to packed products
- From untailored products to tailored products

2 TYPES OF WAREHOUSES
- With inventories -> warehouses,
distribution centres, depots
- Without inventories -> transit point

Depending on their function:

- Warehouses -> flow management + storage
- Transit point -> flow management

Warehouse
Orders from downstream are fulfilled by using inventories.
Upstream orders aim at replenishing the inventories in the
warehouse

Transit point
Orders from downstream are sent straight to upstream.
Orders are fulfilled upstream, but they are delivered and
sorted through the transit point.

The impact of these differences is obviously on performances and costs

- Decoupling point = warehouse: an order occurs from the POS and is fulfilled from the warehouse,
assuming that the goods needed for the order are available in the warehouse
- Decoupling point = transit point: the order occurs in the production plant, then the goods go into the
transit point and then to the POS
- No decoupling point, direct shipment: the order occurs in the production plant and go directly to the POS

We have different OCT (Order Cycle Time) -> with the transit point we
have longer cycle time compared both to warehouse and direct shipment
With the warehouse we have the shortest cycle time possible, but
inventories mean costs, so the problem is to optimize both CT and costs

Warehouse Transit point

- It needs materials planning and materials - It does not need any activity of materials
management planning and management
- It contains safety and cycle stocks - It generates only “in transit” stocks
- It accomplishes all the activities/tasks typical - It accomplishes only sorting activities:
of a depot (see next paragraphs) o re-packing
o cross docking
o delivery

48
 04. Warehouses: general overview

TRANSIT POINTS
Cross docking activity in a transit point
The transit point is very important because it can manage
the flow, it can transform the flow in order to reduce the
transportation costs, because with it we can be closer to
our final destination; in this way it is possible to use smaller
trucks to deliver the goods

2.1.1 The role of transit point

It aims to optimize transports to the end customers (local distribution)…
… by accepting a longer cycle time (+ 1 or 2 days on average) than the 1-echelon distribution network
Transit points can adopt at least two different operative models:
- only sorting of Transport Units which have been completely picked upstream, we accomplish only the
sorting activities (warehouse/production plant)
- receipt of a massive delivery (=aggregation of different customer orders), picking and consolidation of
Transport Units for each customer; the company collects, in a time frame, all the products to fulfil the
order, and then deliver them to the transit point
Receiving Inbound flow
3 WAREHOUSE OPERATIONS Put-away -> inbound logistic
Storage
THE MAIN “FLOW MANAGEMENT” ACTIVITIES Picking Outbound flow
Shipping -> outbound logistic

3.1.1 The inbound flow

Receiving - The carrier is scheduled to deliver the goods at a specific time in
- Schedule carrier order to improve the warehouse labour productivity -> very
- Unload vehicle important activity
- Inspect for damage - The goods are unloaded from the vehicle and moved to the
- Compare to P/O receiving dock (usually from trucks)
- Once there they are inspected for damage and any damage is
noted on the carrier receipt
- The received goods are compared to the purchase order to check
that they are those ordered (there could be some mistakes)
- The put-away operation moves the goods from the docks to the
storage area

Put-away - The product is identified (e.g. by scanning the bar code)

- Identify product - The location to store the product is identified as well
- Identify storage location - The product is moved to the location
- Move products - The warehouse inventory records are updated to reflect receipt of
- Update records the item and its location

3.1.2 The outbound flow

Picking - Based on their size, orders can be fulfilled in different ways: full
- Manage order info pallet retrieval or picking activity
- Pick goods - When picking is needed, the order information is provided to the
- Move goods personnel on a picking list
- Label package - The items on the picking list are arranged so as to minimize the
distance the picker has to walk through the aisles

49
 04. Warehouses: general overview

- Once they arrive to the shipping preparation area, the items are
placed in a shipping package (or a pallet that could be wrapped)
- The shipping label (i.e. a label that indicates the customer address)
is attached to the shipping package

Shipping - The carrier is scheduled to pick up the goods at a specific time

- Schedule carrier - The goods are moved from the staging area to loading dock
- Load vehicle - The goods are loaded to the carrier vehicle
- Bill of lading - The carrier signs a bill of lading
- Record update - The warehouse management system is updated to reflect removal
of the products from the warehouse

3.1.3 The “new” value-adding activities

- Customization ->
o by market area
o by sales channel ex: labelling of some products when there are some promotions
o on a customer base ex: different feature for different customer
o …
- Technical support
o repairing activities
o replacing activities
o …
- Return management -> very important nowadays, especially in e-commerce sector
o collection of the return
o products/packaging
o collection of the expired products
o …

THE MAIN FUNCTIONAL AREAS OF THE WAREHOUSE

- Storage
- Picking/Order Assembly
- Receiving/Shipping
The two most important areas are storage and picking systems
NB: these are not the only ones, there are a lot of different areas

3.2.1 The chart of the main flows

All the activities which are made inside

a warehouse are a lot and very
complex
The flows are many, and there are flow
more important than other

The overall picture of the system is

very complex, and of course we need
an informatic system, which can be
summarized as warehouse
management system WMS

50
 04. Warehouses: general overview

3.2.2 Warehouse Management System

The WMS enables the accurate management of the receiving, put
away, packing, shipping, storage location, work planning, warehouse
layout and analysis activities
The WMS increases efficiency of work planning

Starting at the receiving dock the WMS:

- Verifies the item against the purchase order
- Adds the item to inventory
- Determines the stock location in the warehouse
- Keeps track of the item location
- Determinates the item picking arrangement on the picking list
- Determines and keeps track of the staging location of the complete orders.

More advanced WMS are integrated to the materials handling equipment, picking and sorting systems

4 STORAGE SYSTEMS

PARAMETERS
4.1.1 Design parameters
- Storage capacity (SC) Number of Unit load locations (with dimensions: a x b x h and a determined
weight capacity) -> they can be very different, in type (pallet load or other
measure), size, dimension [UL = unit load]
- Throughput capacity (TC) Flow of Unit Loads [UL/h] -> number of UL that can be handled by
↓ the system in a specific time frame/window
o Input capacity (only storage activities)
o Output capacity (only retrieval activities)
o Throughput capacity
We can have two different values for inbound and outbound capacity if these activities do not share the
same handling system or unit loads: in this case is better to manage them separately, however, we can
consider just the throughput capacity
ex: in input we can have panel loads, in output we can have smaller loads
These two parameters are a sort of constraint, we need to achieve them

4.1.2 Descriptive parameters

- Area utilization rate (AUR) SC / Surface of the storage area [UL/m2]
Number of UL possible to store in a certain area, it is a density measure
The higher is AUR, the lower is the needed area
- Cube utilization rate (CUR)
o SC / volume of the storage area [UL/m3]
o Average volume of the UL / Average volume of the Unit load location [< 1]
- Selectivity capacity to access a specific unit location in the warehouse; it is possible to have:
o Directly accessible UL locations / SC [≤ 1]
o Number of directly accessible items / total number of items stored [≤ 1] -> not direct access
ex: I can have something in front of me that I have to move in order to reach what I really want
Selectivity is very important because if you want to access all the unit location, so to have a full selectivity,
you need a lot of space and optimization of it
- Saturation coefficient of the SC (Average or Maximum) Number of UL stored / SC [≤ 1]
Percentage of “used” storing capacity

4.1.3 Cost indicators

- Storage cost / (UL location) [€/(UL*year)]
o Annual cost of:
▪ Building and facility services -> this can be considered also as an investment
▪ Racks

51
 04. Warehouses: general overview

o Part of the annual cost related to general facility services -> these costs depend on how we decide
to structure our facilities
- Handling cost / (UL handled) [€/UL] -> Annual cost of:
o Materials handling system
o Labour (handling operators), in particular for a traditional system
o Energy consumption related to handling, in particular if we have an automated system
o Handling equipment maintenance, which can be really significant in terms of automated system

4.1.4 Impact of storage/handling systems on performances

The main impacts of handling system are on
- Throughput Capacity, impacted by truck speed
- Costs of investments
- Storage system -> layout of the system -> AUR [ULs/m2]
o #levels -> max height achievable
o Aisle width -> length of trucks (with the UL on
board)

PALLET HANDLING TRUCKS

Three main types of pallet handling trucks:
- Walkie stackers Mainly used to move horizontally the pallet loads and to load/unload the trucks
o operator moves manually
o short distances
o cheap solution
- Front loading forklift trucks Used to move the pallets both horizontally and vertically
Two main typologies:
o counterbalance forklift trucks (Max forks height 6 m, Min aisle width 3 m)
o straddle reach trucks (Max forks height 10 m, Min aisle width 2,5 m)
- Side loading forklift trucks Used to move the pallets both horizontally and vertically, usually in the
aisles, more expensive than front loading forklift trucks
Two main typologies:
o side loaders (Max forks height 14 m, Min aisle width 1,5 m)
o turret trucks (Max forks height 14 m, Min aisle width 1,5 m)

NB: the choice of the handling system is very important because has an impact on the AUR, the handling system
influences the area and so the strategic decision we have to take
- the higher the height, the higher the AUR, the smaller the needed area
- the higher the weight, the lower the AUR, the bigger the needed area
In general, we have to understand which is the thing we need to focus on, if we want to be fast of if we do not
have space

4.2.1 Walkie stackers

- Pallet jack: https://www.dropbox.com/s/h0zbm0v5p6498u4/transpallet_cutted.mp4?dl=0
- Counterbalance: https://www.youtube.com/watch?v=TJNhIyvPILk

4.2.2 Front loading forklift trucks

- Counterbalance forklift truck: https://www.youtube.com/watch?v=ZcATrAwOG64
o Most diffused one; it has weight in the back to balance the pallets’ weight
o Maximum length = 6/7m
o Cheaper than other trucks (30k)
o In order to use this truck, we need wide aisles
- Straddle reach truck: https://www.youtube.com/watch?v=dl8w3UIdDp8
o It reaches higher shelves (8-9m)

4.2.3 Side loading Forklift trucks

- Side loader: https://www.dropbox.com/s/62hc4jqxkxgew6z/Side%20loader%20forklift%20truck.mp4?dl=0

52
 04. Warehouses: general overview

[from min. 1:00 to 1:40]

Height 10-13m + less space -> AUR increases -> area decreases but it is slow -> TC decreases
- Swing fork turret truck: https://www.youtube.com/watch?v=HU80HnmSX9E

4.2.4 AGV (Automated Guided Vehicles)

https://www.youtube.com/watch?v=x6kKLJTQ4oU
AGV Quad https://www.youtube.com/watch?v=EoQKemA8p_I
It is not so easy to use them in a complex system

→ It is possible to use a mix of these handling system in the same warehouse, and it is very useful

TYPOLOGIES OF STORAGE SYSTEMS

Storage systems for big size UL (typically pallet loads):
1. Block stacking
2. Rail racks (drive-in, drive-through, …)
3. Selective pallet racks (served by counterweight forklift trucks,
straddle reach trucks, turret trucks…)
4. Flow racks (gravity, push-back, …)

Storage systems for small size UL: <-> Picking systems

- Miniload
- Carousels (vertical, horizontal)
- Vertical AS/RS systems

4.3.1 1. Block stacking

- Storage systems that do not require any equipment (it is only needed to paint the lanes on the floor)
This is the easiest system we can have: it requires just a level floor on which we can stack/pile the
different products, one over the others
- The stock piling is limited by the characteristics of the goods and by stability requirements (max 3/4
levels) -> we cannot pile a great amount of product, because the maximum level is limited
Having a limited number of aisles limits you in having a great AUR
- They provide small selectivity of the UL, because we cannot access to all the different pallet loads
ex: in this case the products have different colour, because the colour indicates that they are the same
product, in order to “increase” the selectivity
At least one lane per item is required in order to obtain item selectivity equal to one
- If the lanes can be accessed only by one aisle it is necessary to operate with a LIFO rule within the lane,
but with more than one lane per article it is possible to operate with a lane FIFO rule
- Systems useful for items with high inventory level; this storage system is not suitable for goods with short
expiration date, but for goods that can have a high level of inventories
- Lanes need to be large enough to let the forklift trucks in
Block stacking: https://www.youtube.com/watch?v=7tzRHeVtV6g

4.3.2 2. Drive-in or drive-through systems (Rail tracks)

- Storage systems similar to block stacking, but provided with racks to carry the unit loads, it has a
structure; its advantage is that we have less problem of weight of products
- One item per lane
- They provide a higher exploitation in height, (particularly relevant if UL cannot be stacked)
- Special trucks are required to enter the lanes (width = 1 meter)
- In Drive-in systems access to the lanes is only from one side, thus implying a LIFO stock management for
the lane, while in drive-through systems it is from both sides thus permitting a FIFO stock management
for the lane

53
 04. Warehouses: general overview

- Like block stacking, these systems are useful for articles with high inventory level

Drive- in racks Drive-through racks

https://www.youtube.com/
watch?v=lVMFlfL-UQo

The sizes of the pallet load are standardized at European level -> 0,8*1,2m
In this system, is useful to understand in which direction we should put the pallet loads, considering that the
maximum width of the forks is 1 m, so they can enter for only 1 m in the pallet location

4.3.3 3. Selective pallet rack systems

- Storage systems typically made by couples of racks
separated by working aisles
- Racks are made by joining uprights (vertical
elements) and beams (horizontal elements)
- Selectivity is high. It is equal to 1 in case of single
deep storage racks (highest selectivity we can
reach); we can stock all the different kinds of
products and have access to all of them
- Each bay can hold more than one UL (depending on
the characteristics of the UL), i.e. a bay can have
more than one pallet location (ex: in this case, we
have 3 PL in one bay)
- They can be used simultaneously to stock full pallet
loads and as picking stock
- Their cost is low, around 20-30 euro per UL location
https://www.youtube.com/watch?v=8DZFdy23gvM

Cantilever Racks https://www.youtube.com/watch?v=vgYbW60OaoA

Most diffused system, the 90% of warehouses has this one

There are different aisles, and the distances depend on how the handling system is, since it determines the height
of the infrastructure (number of levels) we can reach

This kind of system can be used also for the picking activity, and it is not a negligible advantage
ex: the small track in the picture is conducting a picking activity, it is not part of the handling system: he picks
single products from the ground level, not entire pallets from the other levels
In this way we can share the space for picking and retrieving activity: some trucks which can reach higher level
move the needed products from the 2/3 level to the ground level, so smaller trucks can take them easily
This system is the most diffused for this reason, because it increases a lot the AUR

54
 04. Warehouses: general overview

4.3.4 4. Flow rack systems

- Adapt to robust UL with uniform weight
- Single side loading/unloading (LIFO rule)
- 3÷4 UL per channel

https://www.youtube.com/watch?v=L-Wtaw9fbL8

It is like the drive in, but there is a flow rack system which creates a
certain slope, and the pallet loads flow in the conveyor thank to the
gravity
In this way, we have the same features of the drive in -> same
levels, same selectivity and same rule (LIFO)
The main advantage is that the truck has not to enter in the
system, so the retrieving is quite more comfortable and quicker
The main problem is that this system is very expensive

Pallet/case flow rack systems

- Storage systems made by uprights sustaining slightly inclined roll conveyors, to enable gravity-based
movement
- Low selectivity (only the UL facing the aisle is directly accessible)
- Each channel is dedicated to one article
- Low retrieval times (high facing density)
- Useful for articles with a medium-high stock level
- Cost is high (150-200 euros per UL location)

COMPARISON BETWEEN TRUCKS SERVING SINGLE DEEP SELECTIVE RACKS

Assumption: selective pallet racks NB: reference values
HMAX Aisle widthmin AUR Storage Cost
m m UL/m2 € /UL
Counterbalance forklift truck 6 3.0 1-2 20-30
Straddle reach truck 10 2.5 - 2.8 1-3 20-30
Turret truck 14 1.5 - 1.6 3-5 30-40
S/R machine 35 1.0 - 1.4 6 - 12 >50

5 CONCLUSIONS

SUPPLEMENTARY READINGS
- Coyle, Bardi, Langley, Warehousing decisions, The management of business logistics, a supply chain
perspective, pp. 282-314
- Frazelle, World-class warehousing, pp. 198-229
- Eldemir, Graves, Malmborg, 2004, New cycle time and space estimation models for automated storage
and retrieval system conceptualization, International journal of production research, volume 42, pp.
4767-4783
- Dallari, Marchet, Melacini, Automated material handling systems: a way to increase efficiency of
distribution networks?, 18th International conference on production research

CHALLENGE QUESTIONS
- What is the role of warehouses in a distribution network? What are the main processes and activities in a
warehouse? And in a transit point?
- What are the main performance indicators to control the warehousing process?
- What are the main performance differences between a single-deep pallet rack and a drive-in system?
- What could be the relative “weight” of the functional areas within the warehouse of an electrical
equipment manufacturer? What in a pharmaceuticals wholesaler distribution center?

55
04. Warehouses: general overview

- Assume you are a cement manufacturer. Which storage system would you take into account for the
temporary storage of the cement bags on pallet loads that come out of the plant?

56
 05. Management Policies and Design Methodology

05. Management Policies and Design

Methodology
0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples:
- The main policies to manage the storage systems: store and retrieval cycle management and space
allocation
- The principles and the main methodology to design the storage system in a warehouse

Apply theories, models, principles and criteria to:

- design a “conventional” manual storage system (selective racks + forklifts)

1 MANAGEMENT POLICIES

STORAGE SYSTEM MANAGEMENT POLICIES

Two main policies:
1. Management of the operative cycles (storing/retrieving)
Operative cycles types
o Single command cycles to store or retrieve
o Dual command cycles to store and retrieve
2. Criteria to allocate the UL to the locations of the storage system
o Criteria to decide where to place the UL of an article
o Space subdivision criteria

OPERATIVE CYCLE TYPES – CONVENTIONAL TRUCKS

Variable times Fixed times
- Transfer with the load - Waiting for data transfer
- Transfer without the load - Maneuvers and curves
- Lifting with/without load - Positioning
- Descent with/without load - Loading/unloading the UL

57
 05. Management Policies and Design Methodology

1.2.1 Comparison between single and dual command cycles

The dual command cycles are in general more efficient than single command cycles because:
- It is possible to move 2 Unit Loads per cycle
- The average time to complete the dual command cycle is shorter than two times the average time for a
single command cycle
➔ BUT it is required that the storage and retrieving activities are concurrent

Dual cycle is more efficient, but it can not always be done, we need to have the “need” of storing and picking at
the same time. Sometimes we may store everything when the supplier arrives, and retrieve stuff later, so we
cannot use double cycle. If we manage to structure it as dual is definetly better.

2 RETRIEVING AND ACCESS INDEX

For a generic item i:
Retrieving index 𝑅𝐼𝑖,𝑇 = 𝑁° 𝑜𝑓 𝑈𝐿§ 𝑟𝑒𝑡𝑟𝑖𝑒𝑣𝑒𝑑 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑠𝑡𝑜𝑟𝑎𝑔𝑒 𝑠𝑦𝑠𝑡𝑒𝑚 𝑑𝑢𝑟𝑖𝑛𝑔 𝑇 (§) Full or cut UL
This index is referring to a certain item, and it refers to a time window T (week, month…)
This number tells something about the volume/demand of the item

𝑅𝐼𝑖,𝑇
Access index 𝐴𝐼𝑖,𝑇 =
𝑁° 𝐿𝑜𝑐𝑎𝑡𝑖𝑜𝑛𝑠 𝑎𝑠𝑠𝑖𝑔𝑛𝑒𝑑 𝑑𝑢𝑟𝑖𝑛𝑔 𝑇
This index is referring to a certain item, and it refers to a time window T (week, month…)
This indicator measures the number of UL retrieved for each location

ACCESS INDEX EXAMPLES

2.1.1 Example 1
[T = 1 week] A B
# ULs retrieved 100 200
# Assigned location 20 50
RI = 100 = 200
AI = 100/20 = 5 = 200/50 = 4
In this example, we have a higher RI for B and higher AI for A

We must decide which one of the two items is better to put closer to the I/O point
Basing on the volume, since B is the most require item, we can put it closer to the I/O point
BUT the correct answer is A, because we should take into account not the overall number of UL retrieved, but the
density to accesses of each unit locations, we will put those unit locations most visited, not the overall product

If we have two different points for I and O, the allocation depends on which is the main activity for the system
between storing and retrieving (when we need picking activity)

2.1.2 Example 2 2.1.3 Example 3

Assumptions: Assumptions:
- Item: soft drink 1500 ml (PET package) - Item: peeled tomatoes in cans weighting 500 gr
- Retrieval of full UL - Retrieval both Full UL and cut UL (picking)

Data: Data:
- Number of UL sent in a month: 80 - Number of UL sent in a month: 80
- Average stock: 10 - Average stock: 10
- Number of assigned locations: 20 - Number of retrieved UL: 200
- Number of assigned locations: 20
➔ RI = 80 retrievals/month
➔ AI = 80/20 = 4 accesses/(location*month) ➔ RI = 200 retrievals/month
➔ AI = 200/20 = 10 accesses/(location*month)

58
 05. Management Policies and Design Methodology

2.1.4 Allocation of the UL depending on the AI

- In order to minimize the operative cycle times it is possible to store the articles with higher AI in the
locations where they are most easily retrievable
- This entails an access concentration (operative cycles) in the locations closest to the I/O, and brings as a
consequence the minimization of the average operative cycle times
- The locations which are most easily accessible depend on the storage systems and on the handling system
- To apply the criterion a dedicated storage is required but a class based storage is in general working

ALLOCATION POLICIES
- Random Storage, the unit load can be stored in each pallet location, if available -> no dedicated locations
- Class Based Storage, the unit load of a specific family of items must be stored in a specific set of pallet
locations (usually on the basis of the AI)
- Dedicated Storage, the unit load of a specific item must be stored in a specific set of pallet locations
(usually on the basis of the AI, i.e. the items with the highest access index have to be stored in the pallet
locations closest to the input/output)

2.2.1 Dedicated storage

- Each item has dedicated locations in the storage system
- Usually items/locations are allocated according to the access index

2.2.2 Class Based storage

- Items are clustered in classes (2-4) depending on their access
index and classes are allocated to a fixed storage zone according
to the AI of the class
- In each storage zone there is a randomized storage

This solution is a kind of trade off between the other two solutions,
which are at the extremes

The dedicated storage policy reduces the operative cycle time the most,
because it concetrated the most visited locations near to the I/O point;
on the other hand, it is the most complex in terms of management

Characteristics of the classes

- How do we form the classes?
- How many classes should we consider?
- Which shape should have the classes?

ALLOCATION OF THE ITEMS IN A CLASS BASED STORAGE SYSTEM FOLLOWING AI

Ordering the articles depending on their access index
- for each item we can compute the AI
- then we can rank these AI from the highest to the lowest
- then we compute the average value of AI
- those products with an AI higher than the average value are put in the class A
- then we recompute the average value, considering all the AI minus those I put in the class A
- those products with an AI higher than the new average one are put in the class B, the others in the class C

59
 05. Management Policies and Design Methodology

2.3.1 ABC curve “Accesses-UL locations”

NB: articles are ordered with decreasing AI

CRITERIA TO FORM THE ZONES

- Shape of the zones:
o The boundary curves between a zone and the following are equal time curves starting from the
I/O point of the storage system
- Number of classes:
o The marginal advantage in reducing the average operative cycle times, coming from the
increasing of the number of classes (and therefore the number of zones) is decreasing
(significantly decreasing with more than 3 zones)
o In the meanwhile the marginal disadvantage in terms of storage capacity (and complexity),
coming from the increasing of the number of classes (and therefore the number of zones) is
increasing

2.4.1 Shape of the zones

Shape and disposition of the zones in a selective storage
system served by counterbalance forklift trucks

We should design the

different iso-time curve
-> all the points in this
curve need the same
time to be picked up

Considering the top view:

- In the example on the left, the I/O point is in the centre, so the zones have a triangular iso-time curve
- In the example on the right, the I/O point is all along one side, so the zones have a flat iso-time curve

2.4.2 Number of classes

ABC curve: Accesses / occupied space -> demand curve
if we take into account the first curve, we have that the 20% of UL
locations are responsible of the 40% of the accesses

Reduction of the operative cycle variable times with respect to the

random location storage systems
(Example: selective racks storage system, served by straddle reach
trucks; I/O in the middle of the front; triangular shape zones)

60
 05. Management Policies and Design Methodology

ex: if we decide to have 2 classes/zones and the curve of the demand is the 60/20 (so it says that the 20% of the
UL locations are responsible of the 60% of the access), we can reduce the variable time of single command cycle
time by 30%; if we go for 3 classes, we can reduce the variable time about 35%
-> the difference is not so high Δ80/20 > Δ60/20
NB: every time we increase the number of classes, the managerial complexity increases as well

PROS/CONS OF THE ALLOCATION CRITERIA

Pro and cons of the subdivision of the space:
- Random storage
o PROs: better use of the storage capacity that can be sized in order to store the maximum overall
stock (compared to the dedicated storage)
o CONs: limitations in the applicability of the UL allocation criteria in the locations
-> no allocation criteria, high operative cycle time
- Dedicated storage
o PROs: it is possible to use the UL allocation criteria on the basis of the access index
-> we can reduce the operative cycle time
o CONs: higher storage capacity required, it is required to have storage capacity equal to the sum of
the maximum storage capacity required for each item
- Class based storage
o It is a good trade-off as it is possible to allocate the items to the locations following the allocation
criteria, but with a good utilization of the storage capacity

2.5.1 Example of SC required

Jan Feb Mar Storage capacity:
P1 20 40 30 - Random -> maximum value that we have considering the sums of all
P2 50 20 10 the values of each month -> 100
P3 30 20 40 - Dedicated -> sum of the maximum values that we have for each month
sum = 100 80 80 -> 50 + 40 + 40 = 130

Considering these two options, the storage capacity needed for a dedicated allocation can be equal or higher than
the one required by a random storage system; it can be equal in the case in which all the products reach the
maximum stock value in the same period T

3 STORAGE SYSTEM DESIGN PRINCIPLES

1. Cluster the items, according to
o their features (size, weight, height)
o their environmetal requirements
o their request/demand
2. Determining the main design parameters -> SC- TC
because they mainly depends on the items we have to
store
3. Consideration of the constraints
o Urban -> which building we can use
o Technical -> conventional or automated system
o Economical -> main constraints
o Safety -> respect of the law

4. Selection of the storage system, which depends on all

the previous steps
5. Storage area pre-design -> layout
6. Modeling & simulation, necessary in order to conclude
the project
7. Technical-economical checks and assessments
➔ Result = design

61
 05. Management Policies and Design Methodology

Criteria to cluster the items

- Technical features Dimensions, weight, fragility, … Key elements for the storage system selection
- Enviromental requirements Temperature, humidity, ...
- Safety requirements Fire prevention, hazardous goods,
non-compatibility between different items, …

Design constraints
- Urban: Maximum heights, distances, ...
- Technical: Surface and shape of the ground, integration
with other activities, ...
- Safety: Escape routes, subdivision, fire loads, ...
- Economical: Available resources (maximum investment,
min IRR, max PayBack, …)

4 STORAGE SYSTEM DESIGN METHOD

ASSUMPTIONS
We will illustrate the method for the design of a “conventional” manual storage system (selective racks + forklifts)
- The system does already exist and historical data are available -> crucial in determining the TC
- The storage area is made of single deep selective pallet racks -> this is not the only alternative, but we will
consider only this one
- Storage of full pallet loads, no mixed pallet loads, all of the same size
- Handling systems are counterweight forklift trucks or straddle reach trucks
- The number of trucks is not connected to the number of aisles -> made for simplification, because in
reality after a certain number of trucks we can have a sort of congestation
- There are no picking activities in the storage area, no picking area on the ground level

THE DESIGN FRAMEWORK

There are 3 main steps in order to design a traditional storage system
1. Identification of the Design parameters Storage Capacity and Throughput Capacity
2. Layout design to have a given Storage Capacity
3. Throughput capacity assessment to reach a given Throughput Capacity
-> determination of the number of trucks

NB: there could be few connections between phases 2 and 3 (e.g.: the choice of the layout that minimize the
expected path)

1. DETERMINATION OF THE DESIGN PARAMETERS

Analyses to determine the storage and the throughput capacity
- Analysis of the physical flows -> throughput capacity
- Inventory analysis -> storage capacity

4.3.1 TC - Analysis of the physical flows

Fext = inbound flow from the external arrivals
(full unit load)
FINPUT Fback = return flow from the picking area
(cut unit load)
Fpick = flow to the picking area
Starting from the assumption of having historical (full or cut unit load)
data, we will consider them with also the new Fship = flow to the shipping area
requirements we have FOUTPUT (full unit load)
The flows can be both “from” and “to” our storage
area -> we have to conduct an analysis for each Fend = empty unit load
flow

62
 05. Management Policies and Design Methodology

It takes into account the values of the flows (per hour/per

day) got from the historical data related to a representative
period of time

Inbound flow forecasting Outbound flow forecasting

Throughput Capacity assessment

Depending on the fact that the inbound and outbound flow
share the same handling system,
- if they share it, we have to consider 𝑇𝐶 = 𝐹𝐼𝑁 + 𝐹𝑂𝑈𝑇
- otherwise we should consider them separately

4.3.2 SC - Inventory analysis

As we said for the TC assessment, we will use both the

historical data and the new requirements

Moreover, the SC depends on the type of allocation policy we

chose, which can be
- Random storage
- Dedicated storage
- Class based storage

SC assessment - Random storage

Input data: historical data on the number of j: index of the Tj period on which
𝑆𝐶 = 𝑚𝑎𝑥𝑗 {𝑁𝑆𝑗 } ∗ 𝐹𝐹
locations occupied in a representative period of the representative period is split
max value between months
time (NSj) FF: forecasting factor

SC assessment - Dedicated storage

Input data: historical data on the number of 𝑆𝐶 = ∑(𝑚𝑎𝑥𝑗 {𝑁𝑆𝑖,𝑗 } ∗ 𝐹𝐹𝑖 ) i: index of the item i
locations that are occupied by the item i during 𝑖
FFi: forecasting factor of the
a representative period of time sum of max values between months item i

SC assessment - Class based storage

k: index of the cluster of
Input data: historical data on the number of
𝑆𝐶𝑘 = ∑(𝑚𝑎𝑥𝑗 {𝑁𝑆𝑘,𝑗 } ∗ 𝐹𝐹𝑘 ) items (A, B, C, …)
locations that are occupied by the cluster of
𝑘
FFk: forecasting factor of the
items k in a representative period of time (NSk,j)
cluster of items k

63
 05. Management Policies and Design Methodology

4.3.3 SC assessment - The forecasting factor

- The forecasting factor takes into account:
- The expected evolution of the product range
- The expected evolution of the average stock of the items (or cluster of items)
- The expected evolution of the rack saturation (% of the occupied bays)
- The expected evolution of the correlation between the average stocks of the different items (or clusters
of items)
NB: with reference to a short-medium time horizon

2. LAYOUT DESIGN
Main Phases Constraints in the layout design
1. The choice of both the layout typology and the - Imposed by the handling equipment
location of the input/output points o Maximum height that can be reached by
2. Determination of the storage area the forks
o The bay design o Minimum width of the aisles
o The module design o ....
o The number of levels - Imposed by the building structure
o The determination of the required area o Net height of the building
3. Determination of the optimal shape o Maximum storage area
4. Storage area design o Position of the pillars
o Number of aisles o Maximum load admitted (N/m2)
o Number of bay columns o ....
o Real storage capacity - Imposed by the law (safety)

4.4.1 1. Layout typologies

In respect to the I/O point, we can have a
longitudinal or trasversal storage area
- With the longitudinal layout, we have
we have bigger access aisles and I can
use them in a better way, with less
interferences
- With the transversal layout the access
aisles occupy a smaller area, but we can
have more intererferences between the
machines
Depending on the throughput capacity we
would like to achieve, we can go for a layout or
the other, because it has an impact on the
number of trucks
If the TC is not so big, we can go for the trasversal solution, otherwise a longitudinal solution is better

Layout typologies - I/O Location

The I/O point can be also a constraints, that we already have,
because we are dsign the layout of the storage area, not of the
building; it can be in three different positions

Case 1. The entrance and the exit from the storage area are in the same point -> it can be in the middle of a side,
or in a corner (Example: transit through the same gate/aisle, interface with a fixed handling system (e.g. lift))

Case 2. The input/output points are distributed all along one side of the storage area
Example: receiving/shipping area with many accesses

64
 05. Management Policies and Design Methodology

4.4.2 2. Determination of the storage area

1. Bay design

Example: single deep racks / 2 europallets per bay

(d) : lateral distance (75 - 150mm) Main dimensions of the bay:
(c) : clearance between top of pallet load and beam above (100 - 150mm) - L (length) = l*2 + d*3
(e) : distance between two opposite UL in the two-faced racks (200 - 300 mm) - H (height) = h + c
(l) : length of the pallet load = 800mm - D (depth) = d + e
(h) : height of the pallet load
(d) : depth of the pallet load = 1200mm

Arrangement of the UL in the bays

Configuration A Configuration B
Example: warehouse
with single deep
selective racks
served by straddle
reach trucks
NB: the aisle width
depends on the
configuration
Better area uilization rate AUR Better ergonomicity to pick the cases (picking)
Module area: 11,2 m2 Module area: 12,32 m2

We can store the pallet load in two different ways -> the better way to store the bays is configuration A; we can
see it also with a real computation, which shows how the module area is lower (so AUR is higher) in the
configuration A

NB: considering a drive in, we should remember that the forks of the special truck that it needs have a maximum
width of 1 m, so we should put the PL in the opposite way compared to the common one, with a lower AUR

2. Module design
Module = if we replicate the
module we get the entire
storage area

Example: assumptions: 2 pallet loads per bay (europallet) served by straddle

reach trucks -> Module area: 1,9 m x (1,35 + 2,7 + 1,35) m ® 10,2 m2

65
 05. Management Policies and Design Methodology

3. Number of storage levels

The maximum number of storage levels depends on the more relevant constraints (strictest
constraint) between the maximum height that can be reached by the forks and the net height of
the building
𝑚𝑎𝑥 ℎ𝑒𝑖𝑔ℎ𝑡 𝑟𝑒𝑎𝑐ℎ𝑎𝑏𝑙𝑒 𝑏𝑢𝑖𝑙𝑑𝑖𝑛𝑔 ℎ𝑒𝑖𝑔ℎ𝑡
𝑁𝐿 = 𝑚𝑖𝑛 {⌈ ⌉;⌊ ⌋}
𝑏𝑎𝑦 ℎ𝑒𝑖𝑔ℎ𝑡 𝐻 𝑏𝑎𝑦 ℎ𝑒𝑖𝑔ℎ𝑡

NB: we round up the first ratio (between the maximum value of height and the height of the pallet bay) because
the machine has to reach the bottom of the pallet load stored at the last level, not its top
On the other hand, we round down the second ratio (between the height of the building and the height of the
bay) because the machine cannot go over the roof

4. Required area
Area Utilization Rate (AUR): to calculate the AUR it is sufficient to calculate it only for a 2 ∗ 𝑁𝑃𝐵 ∗ 𝑁𝐿
𝐴𝑈𝑅 =
module as the module is representative of the whole storage area 𝐴𝑚𝑜𝑑𝑢𝑙𝑒
𝑆𝐶 #𝑠𝑡𝑜𝑟𝑒𝑑 𝑝𝑎𝑙𝑙𝑒𝑡 𝑙𝑜𝑎𝑑𝑠 [𝑈𝐿/𝑚2 ] 𝑆𝐶
𝐴𝑈𝑅 = = 𝐴𝑈𝑅𝑚𝑜𝑑𝑢𝑙𝑒 = →𝐴=
𝐴 𝑀𝑜𝑑𝑢𝑙𝑒 𝑎𝑟𝑒𝑎 𝐴𝑈𝑅𝑚𝑜𝑑𝑢𝑙𝑒

Example: warehouse with single deep selective racks served by straddle reach trucks

Module area: 10,2 m2

Aisle width: 2,5 m
Number of levels: (7 / 1,5) = 4,6= 5 (*)
UL dimensions: 0,8x1,2x1,2 (h) m
Number of UL per module: 4 x 5 = 20
Bay dimensions (2 UL): 2x1,3x1,5 (h) m
AUR: 20 / 10,2 = 1,95 UL/m2
Max reachable height by forks: 7 m
(*) Rounding to whole upper

4.4.3 3. Optimal shape of the storage area

Assuming that:
- There are only single command cycles
- Each location has the same probability to be visited
- Rectilinear paths
- Racks as a continuous area

We can get the optimal rate between the width (U) and the depth (V) of the storage area through the
minimization of the expected path

The expected lenght of the path (to the average location and back) depends on the position of the I/O point

According to the I/O point position the path is more or less long; we can have three different alternatives, and the
following formulas determine the average length
𝑈 𝑉
1. I/O in the middle of the storage front 𝑃 = 2 ∗ (4 + 2)
-> we can consider the area as the sum of two equal subareas
𝑈 𝑉
1. I/O on the vertex of the storage front 𝑃 = 2 ∗ (2 + 2)
𝑈 𝑉
2. I/O distributed along the storage front 𝑃 =2∗( + )
3 2
[the demonstration of the 3 at the denominator is too long]

66
 05. Management Policies and Design Methodology

The minimum length of the path (to the average location and back) from the I/O point can be got by deriving:
𝑈 𝐴
𝑈 𝑉 𝑑𝑃 𝑑( + ) 1 𝐴 𝑎
Expected path: 𝑃 = 2 ∗ ( 𝑎 + 2 ) 𝑑𝑈
= 0 → 𝑎𝑑𝑈2𝑈 = 0 → 𝑎 − 2𝑈2 = 0 → 𝑈𝑜𝑝𝑡 = 2 ∗ 𝑉𝑜𝑝𝑡
We are looking for the minimum, so we can derive the general formula and put it equal to zero

The optimal rate between the two sides of the storage area (U e V)
- I/O point in the middle of the front: Uopt = 2 Vopt
- I/O point in a vertex of the front: Uopt = Vopt
- I/O point distributed along the front: Uopt = 1,5 Vopt

4.4.4 4. Storage area design

𝑈𝑜𝑝𝑡
- Number of aisles 𝑁𝐴 = → 𝑈 = 𝑁𝐴 ∗ (𝐴𝑊 + 2𝐷) [𝑚]
𝐴𝑊+2𝐷
𝑆𝐶
- Number of bay columns 𝑁𝐶 = ⌈2∗𝑁𝐴∗𝑁𝑃𝐵∗𝑁𝐿⌉ → 𝑉 = 𝑁𝐶 ∗ 𝑊 [𝑚]

➔ Real storage capacity: 2 ∗ 𝑁𝐴 ∗ 𝑁𝐶 ∗ 𝑁𝑃𝐵 ∗ 𝑁𝐿 [NPB = number of pallet locations in a bay]

[from tutorship session]

At the end of the computations, maybe we need to
round the NA, but we do not know if we should round
up or down it: in order to decide it in the best way, we
should compute the expected path considering both the
solutions, and then pick the best one, so the one which
minimize the expected path, which has the lowest SCCT
→ In general, we can round NA to the closest integer

3. THROUGHPUT CAPACITY ASSESSMENT

Determination of the number of trucks

1. Determination of the average operative cycle time and of the single truck throughput capacity
2. Determination of the number of trucks to satisfy the planned throughput capacity
3. Determination of the real throughput capacity

𝑇𝑆𝐶 = 𝑉𝑇𝑆𝐶 + 𝐹𝑇𝑆𝐶 [𝑠] - P = length of the path [m]

𝑃 𝑆
𝑉𝑇𝑆𝐶 = + - S = length of the vertical path (forks
𝑠𝐻 𝑠𝑉 going up and down) [m]
𝑈 𝑉
𝑃 =2∗( + ) - sH = horizontal speed of the truck [m/s]
𝑎 2
{
𝑁𝐿 − 1 - sV = vertical speed of the forks [m/s]
𝑆 =2∗𝐻∗( ) if the up and down speeds are the same
2
3600 𝑠/ℎ
Throughput capacity of a single truck 𝑇𝐶𝑡𝑟𝑢𝑐𝑘 = 𝑈𝐹 ∗ [𝑠𝑐/ℎ]
𝑇𝑆𝐶
UF = Truck utilization factor (operator breaks, truck availability, …)
𝑇𝐶𝑝𝑙𝑎𝑛𝑛𝑒𝑑
-> #𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑡𝑟𝑢𝑐𝑘𝑠 = ⌈ 𝑇𝐶 ⌉
𝑡𝑟𝑢𝑐𝑘𝑠
Real throughput capacity: 𝑇𝐶𝑛𝑒𝑒𝑑𝑒𝑑 = #𝑡𝑟𝑢𝑐𝑘𝑠 ∗ 𝑇𝐶𝑡𝑟𝑢𝑐𝑘

[from tutorship session]

NB: adding an aisle in order to achieve the TC, maintaining the same AUR, we
will maintain the same overall storage area because we are going to reduce
the length of the aisle (we can cut at the height of the horizontal red line)

67
05. Management Policies and Design Methodology

5 CONCLUSIONS

SUPPLEMENTARY READINGS
- Coyle, Bardi, Langley, Warehousing decisions, The management of business logistics, a supply chain perspective, pp.
282-314
- Frazelle, World-class warehousing, pp. 198-229
- Eldemir, Graves, Malmborg, 2004, New cycle time and space estimation models for automated storage and retrieval
system conceptualization, International journal of production research, volume 42, pp. 4767-4783
- Dallari, Marchet, Melacini, Automated material handling systems: a way to increase efficiency of distribution
networks?, 18th International conference on production research

CHALLENGE QUESTIONS
- What are the main warehouse management policies?
- What is the Access Index (AI)? How can it be calculated? What is its purpose?
- In which situations would you suggest a class based storage policy? Make an example
- Detail the main phases of the design of the storage system within a warehouse
- How the decoupling of the storage and retrieving activities during a day can influence the needs of trucks capacity?

68
 06. Automated Warehouses

06. Automated Warehouses

0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples:
- The functional and technical features of the automated warehouses
- The main management policies: the dual command storage-and retrieval cycle management and the
access-index-based item and space allocation
- The application fields of the automated warehouses

Apply theories, models, principles and criteria to:

- design an automated storage system (selective racks + AS/RS)

AGENDA
1. Technical features
2. Management policies
3. Application fields
4. Design methodology
5. New Trends

1 TECHNICAL FEATURES

AUTOMATED WAREHOUSES

NB: input and output can be on the same but

also on different levels; they are always
connected to an automated system

1.1.1 Aisle width

1.1.2 Area Utilization Rate

Example: AS/RS with single deep storage racks
- Aisle width: 1,4 m
- UL dimensions: 0,8 x 1,2 x 1,2 (h) m
- Bay dimensions (2 UL): 2 x 1,3 x 1,5 (h) m
- Available building height: 25 m
- Module area: 8 m2
- Number of levels: int inf (25 / 1,5) = 16
- Number of UL per module: 4 x 16 = 64
- Area utilization rate: 64 / 8 = 8 UL / m2

69
 06. Automated Warehouses

1.1.3 S/R machine features

The aisle is very narrow, the machines are very
quick both vertically and horizontally

1.1.4 S/R machine kinematics

The S/R machine can move the pallet load
horizontally and vertically at the same time, with a
positive impact on the TC; this is something that
the other handling systems cannot do: the forklift
trucks cannot do it mainly for safety reasons: the
pallet load can be moved either horizontally or
vertically at one time

MAIN DIFFERENCES WITH THE TRADITIONAL STORAGE SYSTEMS

- Racks are higher (higher number of levels) and longer
- Aisles are narrower (1 – 1,4 m)
-> these two have a high impact on AUR
- Higher requirements in terms of construction accuracy
The main issue is the cost, this system is a huge investment, both for the system and for the trucks,
because they have to reach high levels, so it is needed a great accuracy in building them
- An automatic input/output system is required, because we need to feed the automated storage system,
we should not have human interaction even between them
- No human intervention is required (except for maintenance)
impact on the kind of products we can storage (ex: goods which have to be in a specific environment)
- The number of machines is fixed, and it is connected to the number of aisles
-> impact on the flexibility, because in order to increase the TC, we should change the layout of the entire
system, we cannot just change the number of machines

1.2.1 Comparison among pallet racks storage systems

Assumption: selective pallet racks
HMAX Aisle widthmin AUR Storage Cost
m m UL/m2 € /UL
Counterbalance forklift truck 6 3.0 1-2 20-30
Straddle reach truck 10 2.5 - 2.8 1-3 20-30
Turret truck 14 1.5 - 1.6 3-5 30-40
S/R machine 35 1.0 - 1.4 6 - 12 >50
NB: reference values

70
 06. Automated Warehouses

RACK REPRESENTATION

- RL = Rack Length
- SL = Stroke Length, i.e. maximum horizontal path
- RH = Rack Height
- SH = Stroke Height, i.e. maximum vertical path
N.B.: this figure represents the 2 selective racks of the aisle
stroke because we are talking about the rails of the machines

Rack in spatial coordinates Rack in time coordinates

1.3.1 From “spatial” to “time” coordinates

𝑆𝐿 𝑆𝐻
𝑇𝐻 = 𝑆 𝑇𝑉 = 𝑆 we can measure the distances both in meter and seconds
𝐻 𝑉
- SL = stroke length, i.e. maximum horizontal path [m]
- SH = stroke height, i.e. maximum vertical path [m]
- SH = Horizontal speed [m/s]
- SV = Vertical speed [m/s]
We need this measure because if we would like to obtain the overall time to get an item, this will be the
maximum between the vertical time and horizontal time, and not the sum of the two as in the traditional
warehouse, because my machine can move up vertically and horizontally at the same time, concurrently

1.3.2 Example: from spatial to time coordinates

(𝑥𝐴 ; 𝑦𝐴 ) 𝑥 𝑦
𝐴={ → 𝑇𝐻,𝐴 = 𝑆 𝐴 ; 𝑇𝑉,𝐴 = 𝑆𝐴
(𝑇𝐻,𝐴 ; 𝑇𝑉,𝐴 ) 𝐻 𝑉

In the time representation, the slope of the line is

always 45°, till the vertical and horizontal time are
equal, because the machines moves vertically and
horizontally concurrently; then the representation will
move up if the vertical time is higher than the
horizontal time, and vice versa

The machine moves itself on a line inclined of 45° in the spatial representation till the vertical
and horizontal distances between the machine and the item are equal
NB: the speeds are never the same

1.3.3 Rack shape factor

𝑆𝐻 𝑆 𝑆𝐻 𝑆𝐿 𝑇
The “shape factor” can be defined as follows: 𝑆𝐹 = 𝑆𝐿
∗ 𝑆𝐻 → 𝑆𝐹 = /
𝑆𝑉 𝑆𝐻
= 𝑇𝑉
𝑉 𝐻

The more my area is similar to a square (SF=1), the more the vertical and horizontal speeds are similar

71
 06. Automated Warehouses

SYSTEM CONFIGURATIONS
- Structure
>1 immagine
o Standard building
cellulare 5-10 o Sunk building -> the system is put in a hole in the ground ---->
the advantage is
that the service o Rack supported building -> the rack basically constructs the
can continue if a
machine is out for
entire building (most diffused one) ---------------------------------------------------------->
problems, 𝑆/𝑅 𝑚𝑎𝑐ℎ𝑖𝑛𝑒𝑠
because the - #𝑎𝑖𝑠𝑙𝑒𝑠
number of machines for each aisle
machine on the
other side works o= 1 -> the #machines is fixed and is connected to #aisles
in the two parts. o< 1 -> one machine is shared between different aisles
this is really
useful in the case o> 1 -> there can be also the case in which you put more than one machine per
of short lead time
aisle, in order to increase the TC capacity
- Rack depth
o Single-deep
o Double-deep -> higher AUR, lower selectivity
- Telescopic forks per each S/R machine
o 1
o 2 -> it means an increase in TC

𝑆/𝑅 𝑚𝑎𝑐ℎ𝑖𝑛𝑒𝑠
#𝑎𝑖𝑠𝑙𝑒𝑠
< 1-> less than one machine per aisle

The machine shared between different aisles needs a

system to move around the aisles
- Rail + swaps
- Trans-shipments

In this case, the effectiveness decreases, because you

lose time in moving from an aisle to the other

1.4.1 Investment costs

E S/R machine Cs/r € 200.000 #s/r Equipment cost+
Building cost+
E Racks CR 85 €/SL SL Fire protection system+
Software cost
F Sprinkler CSP 110 SL/5 𝑆𝐿
€/sprinkler 𝐼𝑛𝑣 = 𝐶𝑠/𝑟 ∗ #𝑠/𝑟 + 𝐶𝑅 ∗ 𝑆𝐿 + 𝐶𝑆𝑃 ∗ + 𝐶𝑃 ∗ 𝑚2
5
F Sprinkler CP 50 €/m2 m2 +𝐶𝐼𝑂𝑆 + 𝐶𝐿 ∗ 𝑚2 + 𝐶𝐼𝑆
(pumping)
E I/O station CIOS € 120.000 #s/r = number of s/r machines, SL= storage locations
B Floor levelling CL 50 €/m2 m 2 NB: rough values, the cost of the land is not included
Information CIS € 100.000
C system
the software used to the moving of the storage machine

2 MANAGEMENT POLICIES
Two main policies:
1. Management of the operative cycles (storing/retrieving)
- Operative cycles types
o Single command cycles to store or retrieve
o Dual command cycles to store and retrieve
- Optimization of the dual command cycles:
o NCZ (No Cost Zone)
o MTB (Minimum Travel Between)
2. Class based storage (we will focus a bit more on the management of the operative cycle)

72
 06. Automated Warehouses

OPERATIVE CYCLES TYPES

2.1.1 Single command cycles – AS/RS
Storage cycle Retrieving cycle

PL: pallet location

Fork cycle (loading)

Acceleration
1.Travel to the PL (with load) Transferring variable time
Deceleration
Considering the times, we see that just
Positioning (at the PL)
the transferring time is variable, the
Fork cycle (inserting UL -> PL)
others are fixed
Acceleration
2. Return (without load) Transferring variable time
Deceleration
Positioning (at the I/O point)

2.1.2 Dual command cycles – AS/RS

If we have a dual command cycle,
we have 3 phases (not more 4)
-> it is more efficient than a single
command cycle, as in a
conventional storage system

2.1.3 Cycle time – AS/RS

- Variable times (VT), that depend on the pallet location to be reached
(transferring time) The cycle time is the sum of:
- Fixed times (FT), that do not depend on the pallet location to be reached 𝑪𝒚𝒄𝒍𝒆 𝑻𝒊𝒎𝒆 = 𝑽𝑻 + 𝑭𝑻
(positioning, fork cycle, acceleration, deceleration)

2.1.4 Fixed cycle time – AS/RS

SC DC In the double command cycle, we
have 3 positioning places and 3
Positioning 2-4s x2 x3
acceleration + deceleration phases
It has advantages in reducing the
Fork cycles 10-15s x2 x4
variable time, and so the overall
Acc+deceleration 6-8s x2 x3 SC = Single Command cycle cycle time
DC = Dual Command cycle

2.1.5 Variable cycle time – AS/RS

The variable time

depends on the position
of the arriving point in
respect to the 45°
diagonal

𝑇(𝐼/𝑂 → 𝐴) = 𝑚𝑎𝑥(𝑇𝐻,𝐴 ; 𝑇𝑉,𝐴 ) → 𝑇𝑉,𝐴 𝑇(𝐼/𝑂 → 𝐵) = 𝑚𝑎𝑥(𝑇𝐻,𝐵 ; 𝑇𝑉,𝐵 ) → 𝑇𝐻,𝐵

73
 06. Automated Warehouses

COMPARISON BETWEEN SINGLE AND DUAL COMMAND CYCLES

The dual command cycles are in general more efficient than single command cycles because:
- It is possible to move 2 Unit Loads per cycle
- The average time to complete the dual command cycle is shorter than two times the average time for a
single command cycle
NB: there is a prerequisite: the storage and retrieving activities have to be concurrent

OPTIMIZATION OF THE DUAL COMMAND CYCLES

Dual command cycles can be designed in order to maximize the throughput capacity, by choosing the best place
to store and retrieve the Unit Loads; there are different criteria, we will see two examples

Criteria to optimize the dual command cycles:

- MTB (Minimum Travel Between)
The idea is to couple a storage/retrieval in a specific pallet location with a retrieval/storage in the closest
pallet location (i.e. minimize the interleave without load)
- NCZ (No Cost Zone)
The idea is to couple a storage/retrieval in a specific pallet location with a retrieval/storage without
increasing the variable time of the cycle (i.e. the dual command cycle must have the same variable time of
the single command cycle, without increasing the costs)

2.3.1 No Cost Zone

The No Cost Zone is the area made by all the points Q that satisfy the following relationship:
𝑉𝑇(𝐼 → 𝑃) = 𝑉𝑇(𝐼 → 𝑄) + 𝑉𝑇(𝑄 → 𝑃)

We would like to avoid the increase of the variable time, so we will

choose the position for which the variable time will not increase
𝑇𝑃 = 𝐹𝑇𝑃 + 𝑉𝑇𝑃 → 𝑉𝑇𝑜𝑣𝑒𝑟𝑎𝑙𝑙 = 𝑉𝑇𝑃

This method is the best one, we cannot do better than maintaining the
variable time as the one we would have with a single command cycle, we
cannot reduce it, so we try to not increase it

ex: we can choose only one of the three points A, B, C we will

reach after we have stored an item in P
- MTB: we will minimize the travel between the point P
and the next point -> we would choose A
- NCZ: all the points belonging to the green area (no cost
zone, it is drawn by the machine while it reaches P) are
the points which do not modify the overall variable time
-> we would choose C
NB: in the prof representation we use the spatial coordinates, while on the slides are the time coordinates

CLASS BASED STORAGE

Retrieving Index 𝑅𝐼𝑖,𝑇 = 𝑁° 𝑜𝑓 𝑈𝐿 𝑟𝑒𝑡𝑖𝑒𝑣𝑒𝑑 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑠𝑡𝑜𝑟𝑎𝑔𝑒 𝑠𝑦𝑠𝑡𝑒𝑚 𝑑𝑢𝑟𝑖𝑛𝑔 𝑇

Access Index 𝐴𝐼𝑖,𝑇 =

𝑅𝐼𝑖,𝑇
- Dedicated storage #𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛𝑠 𝑎𝑠𝑠𝑖𝑔𝑛𝑒𝑑 𝑑𝑢𝑟𝑖𝑛𝑔 𝑇
𝑅𝐼𝑖,𝑇
- Class-based storage 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 # 𝑜𝑓 𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛𝑠 𝑑𝑢𝑟𝑖𝑛𝑔 𝑇

74
 06. Automated Warehouses

2.4.1 Optimization of the class based storage

What mainly changes from the conventional systems is the shape of the area
-> we have to consider the iso-time curves

The other things are the same

ex: considering the A zone, the time needed to
reach the three corner is the same

NB: in most of the cases, the horizontal dimension of the different

areas is higher than the vertical one because the horizontal speed is
lower, so the needed time is higher different classes, and for each class
we have different items

Shape and disposition of the zones in an AS/RS

If the I/O is not at the ground level,
the shape will be different,
because we have different iso-time curves

2.4.2 Number of classes 2.4.3 UL allocation in a class-based storage system based on the AI
ABC curve: Accesses / occupied space Reduction of the
operative cycle
variable times with
respect to shared
storage systems
Example: AS/RS,
SF=1 ; yI/O=0
This is the same as
conventional
systems

3 APPLICATION FIELDS
The automated warehouses are convenient if:
- A high storage capacity is required (roughly > 5.000 UL)
- There is low availability (or high cost) of the ground
-> it is more efficient in reducing the square meters and increasing the AUR
- A high throughput capacity is required (roughly: 25 - 35 UL/h per S/R machine)
- Need to protect goods (from damage, theft, fire) or to protect humans
- Need to control environmental conditions (e.g. frozen or pharmaceutical products)
-> the last two moves for having no human intervention

AREA OF CONVENIENCE OF THE AUTOMATED WAREHOUSES

NB: only from an economic point of view and in a deterministic
environment

3.1.1 Main drawbacks

- Low flexibility to changes of:
o the throughput capacity
o the storage capacity
o the Unit Load sizes
- High investments required
- High dependability on the system reliability: if the R/S
machines or the information system had a problem, the
activities of the warehouse would be partly or totally shut down

75
 06. Automated Warehouses

4 DESIGN METHODOLOGIES

ASSUMPTIONS
Basic case ➔ Design Principles
- 1 S/R machine per each aisle - Storage Capacity and Throughput Capacity (given the
- Single deep racks management policies) are determined by the warehouse
- 1 couple of forks per each S/R configuration
machine - The cost minimization requires the minimization of the number
of aisles, because we have 1 machine per aisle, so in this way it
is equal to minimize the number of machines

VARIABLES AND OBJECTIVE

Storage capacity (SC) (could be a variable)
Throughput capacity (TC) (could be a variable)
-> these two are the design parameters, and they can be both given or variable
Input data Maximum rack height (RHMAX)
Maximum rack length (RLMAX)
-> they depend on the racks and also on the type of machine we chose
They represent a constraint in our problem
Number of aisles (NA)
Number of storage levels (NL)
Variables
Number of columns of bays (NC)
Height of the input/output point (yI/O) -> in many cases it can be considered an input
Objective Minimize the overall costs of the system

AUTOMATED WAREHOUSE DESIGN

1. Identification of the Design parameters -> same as the conventional storage system
Storage Capacity and Throughput Capacity
2. Layout design
to have a given Storage Capacity (minimizing the number of aisles and, as a consequence, the number of
S/R machines)
3. Throughput capacity assessment
to check that the Throughput Capacity is higher than required

1. DETERMINATION OF THE DESIGN PARAMETERS

The method to determine the design parameters is the same as the one seen in the previous lesson (traditional
warehouse design)
The analyses to determine the storage and the throughput capacity are:
- Analysis of the physical flows -> throughput capacity
- Inventory analysis -> storage capacity

2. LAYOUT DESIGN
First, we design the bay (as for the
conventional system), then we will
determine the maximum number
of level and columns of bay,
because we have two constraints
given by the machines

76
 06. Automated Warehouses

4.5.1 Bay design

Example: single deep racks / 2 europallets per bay

(d) : lateral distance (75 - 150mm) Main dimensions of the bay:
(c) : clearance between top of pallet load and beam above (100 - 150mm) - L (length) = l*2 + d*3
(e) : distance between two opposite UL in the two-faced racks (200 - 300 mm) - H (height) = h + c
(l) : length of the pallet load = 800mm - D (depth) = d + e
(h) : height of the pallet load
(d) : depth of the pallet load = 1200mm

4.5.2 Maximum dimensions of the racks

𝑅𝐻𝑚𝑎𝑥
NL: maximum number of storage levels 𝑁𝐿𝑚𝑎𝑥 = ⌊ ⌋
𝐻
𝑅𝐿
NC: maximum number of bay columns 𝑁𝐶𝑚𝑎𝑥 = ⌊ 𝑚𝑎𝑥
𝐿
⌋
-> we round down because we cannot overcome these constraints, the maximum length and height
- RHMAX = maximum heights of the racks
- RLMAX = maximum length of the racks
- H = bay height
- L = bay length

4.5.3 Minimum number of aisles (S/R machines)

𝑆𝐶
NAMIN: minimum number of aisles 𝑁𝐴𝑚𝑖𝑛 = ⌈2∗𝑁𝐿∗𝑁𝐶 ⌉
𝑚𝑎𝑥 ∗𝑁𝑃𝐵
𝑆𝐶
NCMIN: minimum number of columns 𝑁𝐶𝑟𝑒𝑎𝑙 = ⌈2∗𝑁𝐴∗𝑁𝑃𝐵∗𝑁𝐿⌉
[we divide the number by 2 because we have 2 columns of bay in one module]
- NPB = number of pallet locations in a bay
- SC = storage capacity

[from tutorship session]

After computing the number of aisles, we re-compute
the number of columns and not the number of levels
because we do not want to affect the AUR, so we
decide only the number of columns in order to have
shorter racks and to not touch their heights

If we want to compute the real SC, since we have rounded up NCmin and NCreal, for sure we will have a SC which
is higher (and different) than the target SC, the one we have to achieve
NB: the number of columns we re-compute CAN’T BE higher than the NCmax, because we are minimising the
number of aisles

4.5.4 Storage area

- RL: rack length 𝑅𝐿 = 𝑁𝐶 ∗ 𝐿
- SCr: real storage capacity
- RH: rack height 𝑅𝐻 = 𝑁𝐿 ∗ 𝐻
𝑆𝐶𝑟𝑒𝑎𝑙 = 𝑁𝐴 ∗ (2 ∗ 𝑁𝐿 ∗ 𝑁𝐶 ∗ 𝑁𝑃𝐵)
- RFW: rack front side width 𝑅𝐹𝑊 = 𝑁𝐴 ∗ (2 ∗ 𝐷 + 𝐴𝑊)
NB: we will consider always the RL=SL (rack length is equal to stroke length); maybe they are quite different, but
the difference is so little that we do not care about it

77
 06. Automated Warehouses

3. THROUGHPUT CAPACITY ASSESSMENT

- If the TC is enough, we are fine
- If the TC is lower than the requested one, we have
to go back to the design of the layout

-> add a new aisle

-> add a machine
-> eliminate a new column of bay

[we assume that we keep fixed the number of levels,

because we cannot change the maximum height reachable
by the machine]

4.6.1 Relationship between TC and # aisles

In this graph we have the relationship between the

#aisles and the TC

Assuming that our TC is lower than the required one,

so we should add one aisle, we have an impact on the
other choices
First, we have to reduce the number of columns,
maintaining the same number of levels
-> in this way we reduce the size of one rack, and this
means that on average we need less time to reach the
pallet locations, so we reduce the time of operative
cycle, which means an increase in TC for one machine

Summarizing, I have a double impact:

- Increase of the overall TC, by adding one machine
- Increase of the TC per machine, by adding one aisle
If we increase the number of aisle and reduce the number of columns, we will increase the TC just for this reason,
plus the fact that we have one more machine

4.6.2 Throughput capacity assessment

In AS/RS throughput capacity refers to the number of unit loads going through the
system in the unit of time (i.e. pallets/hour) -> it goes into and it goes out
This means:
- To perform 2 cycles (storage cycle + retrieval cycle) in the case of single command cycles
- To perform 1 cycle in the case of dual command cycles

Per each S/R machine

3600
- TCSC: Throughput capacity with single command cycles 𝑇𝐶𝑆𝐶 = 𝟐∗𝐴𝑇𝑆𝐶 [𝑃𝑎𝑙𝑙𝑒𝑡 𝑙𝑜𝑎𝑑/ℎ]
ATSC = average time of a single command cycle = AVTSC + FTSC

3600
- TCDC: Throughput capacity with dual command cycles 𝑇𝐶𝐷𝐶 = [𝑃𝑎𝑙𝑙𝑒𝑡 𝑙𝑜𝑎𝑑𝑠/ℎ]
𝐴𝑇𝐷𝐶
ATDC = average time of a dual command cycle = AVTDC + FTDC

𝑇𝐶𝑆𝐶 (𝑜𝑛𝑙𝑦 𝑠𝑖𝑛𝑔𝑙𝑒 𝑐𝑜𝑚𝑚𝑎𝑛𝑑 𝑐𝑦𝑐𝑙𝑒𝑠)

TC: throughput capacity of the system 𝑇𝐶𝐷𝐶 (𝑜𝑛𝑙𝑦 𝑑𝑢𝑎𝑙 𝑐𝑜𝑚𝑚𝑎𝑛𝑑 𝑐𝑦𝑐𝑙𝑒𝑠)
%SC = % of single command cycles 𝑇𝐶 = 𝑁𝐴 ∗ %𝑆𝐶
3600 ∗ ( + %𝐷𝐶)
%CC = % of dual command cycles 𝟐
{ %𝑆𝐶 ∗ 𝐴𝑇𝐶𝑆 + %𝐷𝐶 ∗ 𝐴𝑇𝐷𝐶

78
 06. Automated Warehouses

ASSESSMENT OF THE OPERATIVE CYCLE TIMES

Determination of the average variable cycle time according to the FEM 9851 (1978) rule
Assumptions:
- Each bay has the same probability to be visited
-> we are not taking into account the access index based storage
- Racks described as a continuous domain
- Shape factor » 1 (in general 0,5 ≤ 𝑆𝐹 ≤ 2)
F.E.M. = Fédération Européenne de la Manutention

4.7.1 FEM 9851 rule – Cycle time assessment

X Y
The average variable cycle time is determined 1 2 1
on the basis of two reference points (P1, P2) P1 𝑆𝐿 𝑆𝐻 + 𝑦𝐼/𝑂
5 3 3
These two points can be used to describe the
rack’s variable times 2 1 1
P2 𝑆𝐿 𝑆𝐻 + 𝑦𝐼/𝑂
3 5 3
1. Average Variable time of the single command cycle (AVTSC)
- 𝑇(𝐼 → 𝑃1 ) = 𝑇(𝑃1 → 𝑂)
- 𝑇(𝐼 → 𝑃2 ) = 𝑇(𝑃2 → 𝑂)

𝐴𝑉𝑇𝑆𝐶 = ½ ∗ [𝑇(𝐼 → 𝑃1 ) + 𝑇(𝑃1 → 𝑂) + 𝑇(𝐼 → 𝑃2 ) + 𝑇(𝑃2 → 𝑂)]

2. Average variable time of the dual command cycle (AVTDC)

- 𝑇(𝑃1 → 𝑃2 ) = 𝑇(𝑃1 → 𝑃2 )

𝐴𝑉𝑇𝐷𝐶 = [𝑇(𝐼 → 𝑃1 ) + 𝑇(𝑃1 → 𝑃2 ) + 𝑇(𝑃2 → 𝑂)]

Example
Assumptions:
- I/O in the corner of the rack (yI/O =0)
- Squared rack (according to the time dimension) (SF=1; Tv=TH=T)
4 9 2∗𝐴𝑉𝑇𝑆𝐶−𝐴𝑉𝑇𝐷𝐶
𝐴𝑉𝑇𝑆𝐶 = 3 𝑇 𝐴𝑉𝑇𝐷𝐶 = 5 𝑇 -> Δ% = ( 𝐴𝑉𝑇𝐷𝐶
)∗ 100 ≈ 48%

5 NEW TRENDS

AUTONOMOUS VEHICLE STORAGE AND RETRIEVAL SYSTEM

- New technology based on the decoupling of the vertical
and horizontal movements
- Lifts to perform the vertical movements
- Autonomous vehicles to perform the horizontal
movements

Different configurations may be identified; for instance, based on:

- number of lifts
- number of vehicles
- type of vehicles
o tier-captive vehicles, i.e. vehicles can access only a specific tier
o tier-to-tier vehicles, i.e. vehicles can access any tier
- type of unit loads:
o palletised unit load
o small unit load (e.g. tote)

79
 06. Automated Warehouses

5.1.1 Why is this technology interesting?

Service level improvement
- Scalability
- Very high reliability
- Lower energy consumption

5.1.2 Comparison between AS/R and AVS/R systems in case of small unit loads e.g. totes)
- Time to perform 1.000 cycles: - Energy consumption considering 1.000 cycles: Source :
o AS/RS – about 8 hours o AS/RS – about 60kWh Material-
o AVS/RS – about 1,5 hours o AVS/RS – about 3kWh handling
providers

AUTOMATED FORKLIFTS
Automated forklifts have been developed: they do not only
transfer unit loads among different warehousing areas but also
perform storage and retrieval cycles

http://www.youtube.com/watch?v=jkoiiS81goU

6 CONCLUSIONS

SUPPLEMENTARY READINGS
- Frazelle, World-class warehousing, pp. 198-229
- Eldemir, Graves, Malmborg, 2004, New cycle time and space estimation models for automated storage
and retrieval system conceptualization, International journal of production research, volume 42, pp.
4767-4783
- Dallari, Marchet, Melacini, Automated material handling systems: a way to increase efficiency of
distribution networks?, 18th International conference on production research
- Marchet G., Melacini M., Perotti C, Tappia E., (2012) “Analytical model to estimate performances of
autonomous vehicle storage and retrieval systems for product totes”, International Journal of Production
Research, Vol. 50, No. 24, pp. 7134-7148
- Marchet G., Melacini M., Perotti C, Tappia E. (2013), “ Development of a framework for the design of
autonomous vehicle storage and retrieval systems”, International Journal of Production Research, Vol. 51,
No 14, pp. 4365-4387

CHALLENGE QUESTIONS
- Which are the main differences - in terms of technical features between the traditional and the
automated warehouses?
- Why do we translate the coordinates of the racks from spatial to temporal?
- Which are the main advantages of the dual command cycles with respect to the single command cycles?
How would you improve the dual command cycle performances?
- Try to explain why the design of the storage area in an automated warehouse affects both the Storage
and the Throughput capacity

80
 06. Automated Warehouses

CRITICAL THINKING
What is the relationship between the handling system and the AUR?
- There is no relationship between them
- The handling system affects the maximum height of the racks and the aisle width
- The handling system affects the length of the racks

Affecting the racks maximum height and affecting the aisle width -> impact on AUR
NB: we have of course an impact on AUR, but we have to consider also impact on other two dimensions, and not
always in the same way,
- TC performances (ex: according to different types of tracks, since they have different speeds)
- Costs
-> when we have to take decisions, we need to take into account at least these three factors (AUR, TC, costs),
considering the proper trade-offs according to our requirements

For a given product range and given inventory levels, what is the expected impact of Random Storage vs Class
Based Storage on the overall Storage Capacity?
- Random storage requires higher storage capacity
- Class based storage requires higher storage capacity
- The storage policy is irrelevant

If we have the class-based storage, we will pick up the sum of the maximum values for each month, with the
random storage we will pick up the maximum between the different sums of each month
- with a random storage we can explore better the SC (better use of the space), and it is lower
- With a class-based storage we need a higher SC, but we can increase the throughput capacity

What is the connection between Storage Capacity and Throughput Capacity in a Storage System?
- Independent decisions
- Loosely connected decisions
- Strongly connected decisions

It depends on the type of storage system: automated or conventional

- Automated storage system -> SC/TC = fixed -> they are strongly related
- Conventional storage capacity -> they are connected, but not necessarily strongly
it depends on different things
o the main factor is the level of congestion, because higher it is, more we cannot change the
storage capacity
NB: if we have a class-based storage, the congestion can be achieved sooner, because we are
concentrating more the items → higher relation with TC
o considering the optimal layout design, we consider the relation between TC and SC that aims to
achieve the objective: minimize the expected cycle time, so minimizing the number of trucks

A food multi-national manufacturer opens a - Block staging

new warehouse in China. Which storage - Single deep selective racks with counterbalance forklift trucks
systems would you suggest? - Single deep selective racks with AS/RS

What is the impact of decoupling the - No impact

management of the in-flows from the - Throughout capacity must be increased
management of the outflows in a warehouse? - Throughout capacity can be decreased

- Unit loads per hour

How would you express the Throughput - Unit loads per shift
Capacity of a Storage System? - Unit loads per day
- Unit loads per week/month

81
06. Automated Warehouses

82
 07. Order Picking and Assembly: systems and design principles

07. Order Picking and Assembly: systems and

design principles
0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples:
- The main phases of the order assembly process
- How the picking process affects both logistics costs and customer service
- The main categories of picking systems, their functional features and their application areas
- The main choice in the picking process design: the forward-reserve problem
- The main management policies: routing, storage and batching

AGENDA
1. Picking process what is picking and why it is important in logistics
2. Picking systems
what are the main typologies of picking system (architectures) and relative pros
o Manual
and cons (how we can compare them)
o Automatic
3. Principles of Picking System Design → design and management of picking system
what are the main steps/phases of the design process of the picking system and the most
important/relevant management policies
4. Forward Reserve Problem

1 PICKING PROCESS
In this architecture, inflow unit loads are the same
of outflows unit loads, there is not transformation,
so this case may have no picking activity, the
storage function is enough in a warehouse
ex: single item full pallet loads

Normally, in the real warehouses, the architecture

changes, and inside it we have first the storage
activity and then the picking one
This activity is usually necessary when the inflow unit loads are not equal to the outflow unit loads, because the
main trigger of this configuration is the customer order, because according to it, the outflow unit loads may be
different to what you have stored in the warehouse, which needs be “transformed”

Examples:
From To Picking Activity
Full pallet loads -> layers of cartons Layer picking
-> cartons/boxes Carton picking
Cartons -> pieces/item Item picking

83
 07. Order Picking and Assembly: systems and design principles

PICKING
Picking is retrieving lower-level unit loads from higher level unit loads (items from cartons, cartons from pallet
loads) to fulfil customer orders
NB: at the end of the process of picking, you have more unit loads to be managed, so the complexity increases

The selective retrieval of Unit Loads from high-level Unit Loads or single pieces/cases from racks or plastic crates
(where they were previously inserted) in order to fulfil customer purchase orders

- Retrieval of cases from pallet Unit Loads

- Retrieval of pieces from cases/bundles
- Retrieval of cases from racks
- ….

1.1.1 Purchase order: an example

As far as picking is concerned a Customer Order is a
collection of several Order Lines, each one requesting a
defined quantity of a specific SKU/item

We will see that there is a connection between the

customer order and the picking list

1.1.2 Picking List (or Warehouse Order)

Picking list is a sequence of products to be picked in a picking mission
It is a list of pick locations, items to be picked and quantities to be picked

It may correspond to one customer order (or a part of a customer order) or many customer orders
We can have different types of connection between it and the customer order
- Picking list <--> 1 customer order
- Picking list <--> part of 1 customer order -> we need consolidation activity
- Picking list <--> many customer orders -> batch picking
In this case we will have another activity, the sorting

The sequence on the picking list is chosen in order to optimize the picking activities (in terms of efficiency and/or
effectiveness)
The picking list may be on-paper or electronic (transmitted to a mobile device or via headphone)

1.1.3 Picking and Order Assembly Process

Picking is the “core” activity of the Order Assembly Process
1. Order processing -> It starts when the company receives the order from the customer, and it has to
do with the processing of the order information
2. Picking component -> physical component of the process, in which you physically prepare the order
3. Shipment of the goods

➔ Order Cycle Time = time of the Order Assembly Process

It is the sum of the times for these three different activities, we
have to consider all of them, because we have to consider the
perspective of the customer, so from the moment he gives the
order to the company to when he receives the fulfilled order

84
 07. Order Picking and Assembly: systems and design principles

1.1.4 KPIs of the picking process

- Service level:
o Picking lead time is not only the time to pick a product, but the time between the arrival of a
picking list and the moment in which the customer receives the picked products -> speed
it is important because it is part of the order cycle time
o Punctuality -> in the moment in which the customer asks for an order, he receives from the
company a delivery date, and the company should respect this date
o Accuracy -> right products in the right quantity (accuracy in terms of type and quantity)
NB: picking has nothing to do with the fact that you have not the product in the warehouse
- Costs: the picking costs are a component of the warehousing costs, in particular they are mainly related
to the handling costs, because they are the major part of it, it is the most expensive activity
-> what are the main resources we have in a picking process?
o People -> there is also some automation, but the majority of the picking activity is still
done by humans
o Space
o Equipment
o Informatic system (ICT) mainly for order processing

1.1.5 Design parameters

Main designer requirements you need to fulfil in designing a picking system -> storage and throughput capacity
NB: they are different from the warehouse storage and throughput capacity

- Picking storage capacity [pallet loads-> how many products we can manage]
The storage capacity has to do with the inventories policy, while the picking storage capacity has nothing
to do with it: it is mainly related to the forward reserve problem
This problem is how to split the overall stock, which is related to the
inventory policy, between the reserve area and the forward area, which
is mainly linked to the picking activity

- Picking throughput capacity -> it has more to do with the orders, so it is

measured in terms of [orders/time] or [orders line/time], because we are following the customer
perspective, and not in [pallet loads/time] as in the storage activity, which is more related to internal
perspective

1.1.6 Complexity of the picking process

ex: inflow = 1000 single-item full PL per day (full truck loads)
we assume that the customers order a multi-item PL (single-item cartons)
-> inflow is different from the outflow, so we need for sure the picking activity

We also assume that there is a packaging hierarchy, which means that from a pallet load you can pick 40 cartons
-> the right unit of measure is 40.000 cartons/day

We add another type of customer, which requires

multi-item cartons, so it requires single-item pieces
-> the packaging hierarchy has another level, which
is 12 pieces/carton
-> to describe this outflow, we should consider
480.000 pieces/day (= 12 * 40.000)

When we describe the picking process/outflow, we

cannot use an overall volume, but the measure
which has mainly to do with the customer order
-> the factor of complexity, in this case, is much
higher than the complexity of the inflow
- around 40 times for the first outflow
- around 500 times (= 12 * 40) for the second outflow

85
 07. Order Picking and Assembly: systems and design principles

1.1.7 Order cycle, order assembly, picking & sorting

1.1.8 Sorting
It is a process through which the goods are divided according to their destination (customer, geographical area,
etc). It is required only if the goods have been picked in “batch”, i.e. collecting together the goods requested in
more than one order. It can be manual or automated (sorting machine).

1.1.9 Packaging and transport order consolidation

Packaging is the process by which single pieces are assembled in boxes/cases (secondary packaging) and the
boxes/cases are weighted and labelled and/or boxes/cases are assembled into Unit Loads (tertiary packaging)

Transport Oder Consolidation is the process through which the Unit Loads (for single customers or single
destination) are assembled into a Transport Load, including also the final check and the matching with the
shipping note

1.1.10 Importance of order assembly

- Impact on Warehouse Costs
Picking/Order Assembly usually accounts for more than 50% of the Warehousing Costs
o High complexity: handling of fragmented unit loads
o High labour intensity: picking is difficult to automate
- Impact on Customer Service
o Order Assembly Time is a relevant part of the Order Cycle Time
o Order Picking Accuracy is substantial to achieve Delivery Accuracy
o Order Picking Flexibility is paramount to achieve Delivery Flexibility

1.1.11 General trends in Warehouse Management

Enlargement of the product assortment
Increase in the delivery frequency
↓
Average quantity per order line decreases (at least in mature markets)
↓
Increase in the complexity (and costs) of picking and order assembly activities

CLASSIFICATION OF ORDER PICKING SYSTEMS

The main classification between picking systems is

from manual to automated picking system
NB: manual solutions have a high level of automation,
while the automated solutions are fully automated
We will see that the majority of the picking systems
are manual systems
We will see also many under-categories of the
manual workers

86
 07. Order Picking and Assembly: systems and design principles

2 MANUAL PICKING SYSTEMS

- Picker-to-parts -> the pickers move within
the system in order to collect the parts
- Pick to box
- Pick and sort -> similar to the previous one
These two categories have in common an
optimisation in terms of space and time, in
order to reduce complexity and increase
productivity
- Parts-to-picker -> opposite of the first
system, in this case, the parts move within the
system going to the picker who collects them
- Automation of manual picking

PICKER-TO-PARTS PICKING SYSTEMS

2.1.1 The basic idea
The picker carries out a “picking mission” within the “picking area”, visiting in sequence all the locations which
are detailed in the “picking list”

2.1.2 Ingredients
- Picking area
- Pickers, who moves within the picking area
- Picking mission -> multiple command
Multiple in order to distinguish it from the single and dual
command we have seen in the storage system, because all these
points must be connected in order to complete the picking
mission

𝑀𝑢𝑙𝑡𝑖𝑝𝑙𝑒 𝑐𝑜𝑚𝑚𝑎𝑛𝑑 𝑡𝑖𝑚𝑒 = 𝑇𝑟𝑎𝑣𝑒𝑙 𝑡𝑖𝑚𝑒 (𝑇𝑇) + 𝑅𝑒𝑡𝑟𝑖𝑒𝑣𝑎𝑙 𝑡𝑖𝑚𝑒 (𝑅𝑇) = 𝑃𝑖𝑐𝑘𝑖𝑛𝑔 𝑡𝑖𝑚𝑒
- Travel time -> the time the picker spends to move from a pick point to the following one
- Retrieval time -> extraction time, the time the picker needs in order to physically pick the item from the
storage position
NB: both components depend on how many different picking
points the picker has to visit to complete the picking mission
NB: this picking time is not in terms of LT, but the needed time to
pick all the products of picking mission -> resource time

ex: in this picture, we just have a truck that help the picker to move
and collect the order, it has no forklifts as in the storage activity

2.1.3 A classification framework

We can have different types of storage system, and of course we can have different combinations with the
different types of handling system we have seen
Storage systems: Handling equipment:
- Large size Unit Loads: - Manual-driven order picker trucks
o Bay racks (pallet loads, ….) - Motor-driven order picker trucks
o Flow racks (gravity) - Conveyors
- Small size Unit Loads:
o Shelves/bins
o Flaw racks (gravity)
o Drawers

87
 07. Order Picking and Assembly: systems and design principles

2.1.4 Examples of Picker-to-Parts

Manual picking on an order picker truck
The peculiarity in the first picture is that
the ground level is reserved to the forward
area, so to the picking activity, while the
other higher levels are for the reserve area,
for the storage activity

In this case, the picker does not need a

liftable platform, it is needed if also the
first level is reserved for the picking activity
(picture on the right)

Picker drives an Order Picking Truck (low

level) or a manual Picking Cart

Amazon− Castel S. Giovanni (PC)

ex: Amazon warehouse

Manual picking on Narrow-aisle turret trucks

Instead, in this case the picker needs to reach all

the levels, because they are reserved both for
storage and picking activity

This solution is more complex and expensive than

the previous one, because you need a truck
which can move both horizontally and vertically

Picking on Automatic S/R (Storage/Retrieval) machines

This is the extreme solution, in
which there is an automatic storage
system, and the picker is on the
storage system

This is the most expensive one,

because of the highest level of
automation
-> It is not so used

88
 07. Order Picking and Assembly: systems and design principles

2.1.5 Configurations
- Low level (2d) -> picking missions are related only at the ground level, maximum at the first one
it is the most used because it has high level of productivity, because you have not
to reach the higher levels
- High level (2d, 3d) -> better use of space, but on the other hand you have a reduction of productivity,
because you are lifting people, not just the forklift (heavier)
[2d = two dimensions, the picker can move horizontally OR vertically]
[3d = three dimensions, the picker can move horizontally AND vertically, using some liftable trucks]

These configurations have to be combined with

- Storage equipment - Handling equipment
o Racks o Manual carts
o Shelves o Order picking trucks
o Flow racks o High-level trucks

2.1.6 Design and management issues

Reduction of the picking time, through a reduction of
- the travel time
o finding the best routing policy, which minimizes the path and movements between the different
picking points of a single picking mission
o finding the best storage policy, such as the access index policy for the storage system; the
difference is that in this case, the complexity is higher compared to storing activity, because we
have a multiple command, not just a dual or single command cycle
o using a batching picking, which means to collect together different customer
orders with products that are close
This modality aims you to reduce the number of travels, reducing the travel time
on the overall picking time, not on the single one; on the other hand, you need a higher picking
capacity of the picking system
o Reducing interferences, congestions of pickers between picking areas, or you can have a
slowdown of single picking activities

- the retrieval time, working on

o storage activity, storing what needs to be picked in easier positions
o ICT solutions, going for higher automation level
o batching policy, if we have overlapping

“PICK TO BOX” SYSTEM

2.2.1 The basic idea
The pick-to-box system divides the picking area into picking
stations (also called ‘picking zones’), each of which normally
dedicated to one or more pickers. The picking stations are
connected by a conveyor on which bins are placed, that are filled
with picked items. Each bin corresponds (partially or completely)
to a customer order (“order picking” policy)

2.2.2 Ingredients
- Picking area
- Zones -> compared to the previous solution, this picking area is divided into compartments
- Conveyors -> higher automation level, it is the connecting component
- Zone-dedicated pickers -> you try to increase the specialization and reducing the pickers movements
- (sorters)

89
 07. Order Picking and Assembly: systems and design principles

2.2.3 Sorter
- If at the end of the picking activity you want to connect the picking area directly with the packaging area,
you need to collect all the products of a single order, you have to prepare the order
-> pick-to-box approach, where the “box” is a place in which you can aggregate all the items picked up
- Otherwise, you have to move all the picked products in a sorting area, in which the items are aggregated
to create the single orders, which after go to the packaging area
-> pick-and-sort approach, you are adding a decoupling point
This solution is mostly used concurrently with a batching policy

Pros of zone-dedicated pickers:

- travel time reduction (compared to the pickers-to-parts), because the pickers move within smaller areas
- labour specialization, so you are trying to make each picker more confident with the products they have
to pick within their picking area

2.2.4 Configurations
- Pick by order -> pick to box
- Pick by item -> pick and sort / pick to conveyor

2.2.5 “Pick to belt/box” systems

In this case we have a picking system that merges the different

products and brings them to a sorting area

2.2.6 Examples of Pick-to-box

“Paperless” pick to box (pick to light)

Higher level of automation, in order to improve the
productivity and the accuracy: the pickers see the light
on the product they have to pick up, and the right
quantities

In this case we have a pick-to-box system, that uses

the blue boxes to collect the items for a single order

90
 07. Order Picking and Assembly: systems and design principles

“PICK AND SORT” SYSTEM

2.3.1 The basic idea

Pickers retrieve the quantity of each single item resulting from the batching of multiple orders (wave) and place
these onto a takeaway conveyor connecting the picking area to the sorting system. In some cases, takeaway
conveyor is not implemented
Usually you have a ring with some input points and
different outputs connected to different customer
orders
There are very high
automated systems able to
identify the right product
for the right output

2.3.2 Examples of Pick-and-sort

Arrow − Italy

Several, different technologies are available for automated sorting

systems. The choice between different discharge mechanisms is based
on the amount of goods to be sorted and the physical dimensions of the
items to be sorted and is influenced by factors such as weight, size, and
fragility of the items

2.3.3 Design/management issues

- Load balancing between zones -> it is very similar to an assembly line; all the zones has to be optimized
- Storage between zones, in order to reduce the travel time of the conveyors
- Batching management, both for picking and sorting activities
PARTS-TO-PICKER PICKING SYSTEMS
2.4.1 The basic idea
The pickers work in one or more picking stations
The pallet loads of the items detailed in the picking
list are retrieved from the storage area and carried in
sequence to the pickers who pick only the quantity required in the picking list. Then the pallet loads, unless
finished, are put to store again

Ingredients
- Storage area with high AUR (high density per m2)
- Picking bays with stationary pickers, each picker works in his area, he does not move
- Handling system, which moves the products from storage area to picking stations and back if they are not
picked up
This system is used to reduce the Travel Time to 0, because it is shifted from people to handling system, which
must be designed in order to manage these products movements for picking activity

91
 07. Order Picking and Assembly: systems and design principles

2.4.2 Configurations
- Traditional storage system (ex: pallet racks) and conveyors -> a bit unusual
- AS/RS and conveyors (ex: AVG) -> usually used
- Miniload AS/RS, automated warehouse for small unit loads (not pallet loads, but smaller)
-> ex: carousels, vertical storage systems… which links directly the storage area to the picking bays
- Racks (pods) and autonomous robots (ex: KIVA robots)

2.4.3 Design/management issues

- Capacity of the handling system -> the handling system is the connector/link between the 2 area, and it is
very important to design it in order to have the picking bays always feed, because they are the most
expensive areas of the system
- Batching policy -> it is very linked to the right size of the capacity of the handling system, so it is quite
common to use it; we will explain it with an example
- Ergonomics of picking stations, because pickers work all day in front of picking stations

2.4.4 Example – Sizing of the capacity of handling system in Parts-to-Pickers

Typical situation for an automated storage system AS/RS
inflow: 1000 pallets/day -> outflow: 1000 pallets/day

If we consider a picking system, we have differences, because within the

picking activity you are not picking all the items in the storage area
Assumptions:
- 1 pallet = 40 cases
- 1 order line = 4 cases
-> 1 pallet <-> 10 order lines
In this example, you need to manage 10.000 pallets/day (cycles/day) from the storage system to the picking bays,
and other 9000 pallets/day back; of course, the overall outflow is the same as the previous example, but you have
to manage 10 times the previous capacity, 10.000 in (=1000+9000) and 10.000 out

Using a batching policy, we can use other assumptions

- Batching policy = 5 orders
- Batch order line = 20 cases
-> in this way, you can manage less pallets/day
o 2000 pallets/day from AS/RS to picking bays
o 1000 pallets/day back (=2000 – 1000)

2.4.5 A classification framework

Large size Unit Loads: Small size Unit Loads:
- Picking from automated - miniload AS/RS (automated storage and retrieval system)
warehouses to picking - Vertical carousel (horizontal axis)
bays (and sorting systems) - Horizontal carousel (vertical axis)
- Automatic vertical storing system

2.4.6 Examples of Parts-to-picker

Automated warehouse + picking (and sorting) bays

Typical example, we have a

conveyor which connects the
automated warehouse to the
picking bays

92
 07. Order Picking and Assembly: systems and design principles

Miniloads AS/RS (Automated S/R Systems)

You are retrieving and storing
small unit loads
On the right, we have the
handling system in the middle
between the automated storage
AS/RS and the picking bays, which
is used as a sort of decoupling
point

Vertical and horizontal carousels

Throughput capacity: a few dozens of cycles/hour
depending on the use of the system
Automated
vertical storage
systems
No rotations,
just the lift of
the products

2.4.7 Autostore
Very intensive storage area, and a brilliant picking system, consisting in
robots which are able to pick the right box; the selectivity is lower than
one, so they have to move a lot of boxes in order to reach the right one,
so the picking activity is very low, but is also very accurate, because it is
highly automated

2.4.8 KIVA robots

Autonomous robots which retrieve the products
It is more flexible than all the other solution we have seen
What is brought to pickers, are singularly picked and then brought to a packaging
area

“AUTOMATION” OF THE MANUAL PICKING SYSTEMS

When we talk about manual solution, we have seen that all the classes are or can be very strongly automated,
using some devices/solutions
- Automatic identification of the items
o Bar Code Scanning
o Radio Frequency Identification (forthcoming)
- Radio Frequency Communication between the Warehouse Management System and the pickers
o Mobile devices aboard the picker trucks
o Hand-held mobile devices (terminals, headphones, finger-wearable, …)
- Display and lights on the picking locations
-> pick-to-light, in order to tell the picker where and in which quantity he has to pick
- Scale to check the number of items (on the picker truck, on the picking station)
- Computer visions, to see if you have picked the right product

93
 07. Order Picking and Assembly: systems and design principles

2.5.1 Examples of automation of the manual picking systems

Handheld Radio Data Terminals
Very typical, which scan the product in order to
increase the accuracy of the manual process
Amazon DC – Castel S. Giovanni (PC)

Handsfree (Wrist Mounted) Radio Data Terminal

It is also called finger-wearable, used to leave the hands free to pick the items

Light Technologies (i.e. Pick-to-light, Put-to-light)

For each location, you have a light and a display

Voice Headsets and Visual Picking Technology

Voice picking, the operators receive instruction through
voice, so they are totally hand-free

Radio Frequency Identification (RFId)

Used instead of the barcodes, used in order to identify the unit loads (pallet,
cartons, single items...); it makes easier the identification of the items

3 AUTOMATED PICKING SYSTEMS

As we said before, the 99% of the picking solutions are manual,
because the automated systems are those which are completely
automated

Because the level of automation is so high, there are a lot of

solutions which are still in R&D departments of the biggest
companies, they are still in development

The main idea talking about automated system, is to - Robots-to-parts

have machines instead of pickers, and then we can apply - Dispensers-to-conveyor
what we have seen before - Parts-to-robots

3.1.1 Configurations
We will deal with these configurations It is a lot easier to work
- Singularize and sorting -> it is easier to identify the items, since they are on the first idea, so
singularized singularize and sorting,
o Dispenser while the second one is
o (de)palletizer -> applicable both for picking layers and picking more complex through
cartons/cases automation, it is a
- (automatically) identify and pick -> more similar to what human picker does, solution still in the R&D
o Robots-to-part department
o Parts-to-robot

3.1.2 Design/Management issues

- Automatic identification -> most difficult part
- Computer vision -> machine capable to see and identify an item within a group of different items
- Automatic manipulation -> problem especially if you want to manage very fragile products, or product
with very different shapes
- Reliability, what happens when a machine breaks down, because it has a very predictable productivity,
but you do not know when and how it will stop for any kind of problem, and which will be its entity

94
 07. Order Picking and Assembly: systems and design principles

EXAMPLES OF AUTOMATED PICKING SYSTEMS

3.2.1 Singularization & Sorting
A-Frame Dispenser
- Small, regular, non-fragile, high-demand items
- High Retrieving capacity: up to » 3000 pieces/hour
Typical structure of a dispenser configuration
There are many channels, and each channel for one type of product
NB: each channel must be feed manually

V-Frame Dispenser
Lower productivity but higher space capacity than the A-Frame
configuration

Automated layer picking

It is able to pick entire layers of cartons and prepares a pallet
load which consist of different layers of items

Automated case picking

In this case, the robot is able to move single cartons,
so it has to know in advance the exact structure
of the single pallet load
-> more complex than the layer picking
3.2.2 Identify and pick
Automated item picking - Parts to robot
Inside the different boxes you have many types of products: there is a problem
of computer vision, because the robot has to identify exactly which products it
needs and then it can put them into the picking box

Robots-to-part
In this case we can see a robot which
moves through the aisles
Most complex architecture you can consider with automation of picking
It is just a testing solution, it is not real yet

SUMMARY OF PICKING SYSTEMS

What is important is not just to describe these solutions, we have to understand the logic of the solution, we have
to understand the main configurations. For this reason, we entered into the main design and management issues,
most of all, we have to be able to compare these solutions and see what their pros and cons are

Picker-to-parts Pick-to-box Pick-and-sort Parts-to-pickers Automated

Labour Productivity - - - (*) / + (*) ++ +++
Space Productivity - - - (*) -- --- ++ +++
Investments Costs +++ + - (*) -- ---
(++ for KIVA)
Flexibility +++ + - -- ---
Accuracy and Speed --- - + ++ +++
+ point of strength - point of weakness (*) trade off

95
 07. Order Picking and Assembly: systems and design principles

3.3.1 Five different KPIs

- Labour productivity -> how many picks per workers per unit of time
- Space productivity -> equivalent of the area utilization rate for the storage activity
- Investments costs -> different investment needs for each picking system
- Flexibility -> capability to modify the system if the demand changes from the market
- Accuracy and speed -> two main KPIs of the picking systems

3.3.2 Picker-to-parts
Points of strength are the investment costs, because it is a very manual option, you can have a lot of automation
in terms of information management, but not about the picking activity itself; for the same reason, it has high
level of flexibility, the design can be easily changed
On the other hand, always because you have a lot of human work, the labour productivity is quite low; we have
also a low space utilization, because usually you use a low-level picking activity
NB: there is a trade off between the labour productivity and the space productivity, because to increase the space
productivity in this picking modality, we should use a high level picking, which means to have specific picking
trucks, which reduce the labour productivity since the picker cannot just walk

3.3.3 Pick-to-box
The labour productivity, compared to the other solutions, is neither a strength neither a weakness
The overall balance of space productivity and investments costs are quite similar to pickers-to-part
You are reducing a bit your flexibility, you have some more constraints
Even the accuracy and speed are similar to pickers-to-parts; the main difference is that the pickers know very well
the working area

3.3.4 Pick-to-sort
The labour productivity increases and becomes a point of strength, because using the batching approach, you are
reducing the distance between picking points, and so there is a reduction of the overall transfer time
The investments costs are higher, because you need a conveyor and a sorting area
NB: there is a trade-off between the labour productivity and investments costs, since we need a sorting activity
Accuracy and speed are a big point of strength

3.3.5 Parts-to-picker
Labour productivity as the main strength; it tries to eliminate the travel time
It needs a lot of investments, because the level of automation is high
Accuracy and speed are high because you usually specialize the pickers to do one single task
NB: the low score of flexibility is true for the majority, but not for the KIVA and other autonomous robots’
solution, because the combination of independent racks and autonomous robots is a great example of high
flexibility, because you can easily change it

3.3.6 Automation of manual systems

The automation of manual system is used in order to increase the points of
weaknesses of pickers-to-parts and pickers-to-box, which are the labour productivity
and accuracy and speed
Although, there is a great increase in investment costs, which become a weakness
NB: the automation can be used also to reduce the level of alienation for workers

3.3.7 Automatic systems

The labour productivity is very high, because these systems aims to minimize the
people work, but it is no still equal to zero, some people are still needed to feed the machines
The main advantage is also the accuracy and speed
These solutions are not been used because they cost a lot and have not flexibility, because if the requirements
change, it is not so easy to change the design of an automated system
There is still a lot to be done in order to make the automated solution the best one

96
 07. Order Picking and Assembly: systems and design principles

4 PRINCIPLES OF PICKING SYSTEM DESIGN

MAIN DESIGN AND MANAGEMENT DECISIONS

NB: this is not the design in detail of the picking systems, but just the main principles
These are the main phases of the design process for the picking system
1. Selection of the picking system
1. Segmentation of the picking problem in sub-problems (clusters of homogeneous items)
2. Identification of the eligible picking solutions
We have to describe the different problems we are facing, in order to cluster them and after find the
right possible picking solution for each of them
2. Picking system capacity and design -> this phase has to be done for each eligible solution; for this reason,
phase one is necessary, in order to have the lowest number of eligible solutions as possible, because in
this way time and efforts are less
1. Picking System Storage Capacity (Solution of the Forward-Reserve Problem, FRP)
NB: the storage capacity of the warehouse is given by the market, while the storage capacity of
the picking system is an internal decision
2. Picking System configuration (layout, height, …) -> at the end, we have our physical system
It is a very technical phase, and we will not address it in detail
3. Management Policies Space allocation for each item (by-product of the FRP)
The application of the right management policies can dramatically change the productivity of the system
o Batching Policies
o Routing Policies (Sequencing)
o Location or Storage Policies
Only at the end of the process you can compare the different systems you could choose, and maybe go back to
the first assumptions you made in order to find a better solution -> usually, it is a iterative process

4.1.1 1.1 - Segmentation of the picking problem in sub-problems (clusters of homogeneous items)
Main variables to describe each picking problem
- Order profile: how the order is structured; it is describable through
o order lines (ex: Amazon has low level of order lines, for food industry is higher)
o volumes -> how many units loads
o unit loads -> which kind of unit load
- Item/SKU profile: description of the goods, not of the overall order
o Order frequency per item [order lines/tons]
o Physical elements: volume/weight/fragility -> different characteristics, different solutions
- Throughput capacity of the overall capacity [orders/time or order lines/time]
- Target service level (accuracy, speed, lead time)
- Level of flexibility -> it is linked to how much the problem you are describing is reliable in terms of time
-> is the description valid for a long period or it will change soon?

4.1.2 Key Data: Order Profile

Order profile based on volumes If we have this type of distribution, of Order profile based on volumes
(lines/order) course we are facing a big problem, (volume/order)
because we have the 32% of orders
which have 1 line/order, but also 20%
that needs more than 20 lines/order
-> two very different picking problems
need two very different picking
solutions
Order profile based on the picking unit In this case, for only 7% or the orders Order profile based on volumes
you do not even need picking, while ( % palletised ULs)
for 50% you need cartons and for
another 30% you need pieces-only
picking, which is much more
complicated

97
 07. Order Picking and Assembly: systems and design principles

4.1.3 2.1 - Forward-Reserve Problem

The Forward-Reserve decision
- How to split the overall inventory within the warehouse into a reserve stock and a forward stock
(= inventories dedicated to picking) -> the sum is exactly the same, but logically you have split it
The idea is to carry out the picking activity on a portion of the overall area, but of course there is the
problem of the replenishment activity
- The output of the FRP is first of all the “storage capacity” of the picking system, which minimize the costs
- A sub-problem, which is embedded in the main problem, is the allocation of the capacity to the item
classes or single items
- The basic trade-off is between the cost of space and handling costs, i.e. the replenishment costs on one
side and the picking costs - and effectiveness - on the other side

4.1.4 3 – Management policies

Routing policies
- The basic decision is about the sequencing of picks in the warehouse order (or picking list)
- It is a sort of TSP (Traveling Salesman Problem), with many variants in function of the travelling
constraints (which depends on the chosen solutions), so that in some instances the problem is
o mono-dimensional (carousels)
o bidimensional (low-level picker-to-parts systems, miniload AS/RS)
o tri-dimensional (high-level picker-to-parts system)
- The main trade-off is between
o picking costs (travelling time in picker-to-parts systems, idle time in parts-to-picker systems)
it is the main objective function
o software costs and operational complexity (nontrivial routes)

Location or storage policies

- The basic decision is about the allocation of single items or classes of items (class-based storage) to the
available picking locations; it is more complex since we have a multiple command
- The idea is to
o dedicate the most easily accessible locations to the most frequently requested items (actually to
the items having the highest access index as we will see), mainly in the picker-to-parts systems
o locate in proximity items that are frequently requested in the same customer order (correlated
storage), mainly in parts-to-picker systems
- The main trade-off is between
o picking costs
o software costs and operational complexity (congestion, periodic re-allocation)

Batching policies
- The basic decision is about the relationship between
o “warehouse orders” (= picking mission)
o customer orders (= the final output of the process)
- The main options are:
o Single Order Picking: one order at a time (or one portion of the order at a time)
o Batch Order Picking + Sorting: several orders in a warehouse order that before the end of the
process must be sorted into single customer orders
- The main trade-off is between
o picking costs, which decrease with the increase in picking density
o sorting costs, which are needed in the batching option, and increase with the size of the batch

4.1.5 Key Performance Indicators

Picking-related costs Customer service
- Picking costs (people, equipment, software and space) - Accuracy
- (Sorting costs) (people, equipment, software and space) - Order picking cycle time
- Replenishment costs (people, equipment, software and space) - Flexibility

98
 07. Order Picking and Assembly: systems and design principles

4.1.6 Selection among the available picking systems

4.1.7 Design parameters

Storage capacity (SC) Throughput capacity (TC)
- Number of Unit load locations - Number of Orders per hour/day
(with dimensions: a x b x h and a - Number of Line items per hour/day
determined weight capacity) - Number of boxes/cases/pieces per hour/day
- Volume Flow per hour/day

EXAMPLE – DESIGN PROCESS

1.1 segmentation in sub-problems
Small orders (10-20 orders lines), usually picked in cartons
Big orders (>20 order lines), usually picked in layers or full pallet loads
-> in this way, we are splitting the design process in 2 different paths
1.2 considering only the small orders problem, we consider different eligible solutions:
Pickers-to-parts Pick-to-conveyor
pallet racks, low level, order-picker trucks flow racks, zoning, conveyor, carton sorter
2.1 for each one of these eligible solutions, we need to solve the FRP
we need to define both the picking area and the
we need to define the picking area sorting area, considering also the design of the
numbers of channels needed to link them
2.2 Layout definition
in terms of layout and number of the zones of the
in terms of number and length of the aisles
picking area
3 Management policies
all routing, storage and batching policies can be a good the batching solution can be the best one, considering
solution also the load balancing

When we get to the end of the process, we can assess the performance of the two solutions, and also in
quantitative parameters, in order to define which is the best one

5 FORWARD RESERVE PROBLEM

“PICKING STOCK” AND “GENERAL STOCK”

99
 07. Order Picking and Assembly: systems and design principles

There are two main options in terms of integration between the storage and the picking functions:
- The two functions share the same stock, i.e. items are picked from the general stock
-> we can pick both full unit loads or smaller Uls from the same area
- The two functions have different stocks, i.e. there is both a general stock (reserve stock) and a picking
stock (forward stock):
o The picking stock is allocated to different storage systems with respect to the general stock
-> physical distinctions
o The picking and storage stocks are in different sections of the same storage system (e.g. the
picking stock is in the lower levels and the general stock in the upper levels of the same racks)
In this case you can pick both full UL when you need them, but in order to pick smaller unit loads, you
need a picking activity; in this case, we of course need also a replenishment activity

5.1.1 Picking from Forward stock

- Pros (< costs)
o Smaller working areas (only picking stock, size to be chosen, and we can find it through the FRP)
and therefore fewer picking costs
o Better working conditions, specific to ease the picking activities (lighting, ergonomics, …) and
therefore higher effectiveness (ex: higher service level and accuracy)
- Cons (> costs) -> additional costs
o Replenishment of the picking stock from the general stock is required

5.1.2 Examples
We can consider these two configurations used for different options of picking systems
RS = PS RS -> PS
Pickers-to-parts Storage area with pallet racks and high-level Storage area with pallet racks with low-level
picking picking, using only the first level for picking
stock and the other levels for reserve stock

Shelves and “pods” with all the stocks (ex: High-density storage area + flow
Amazon) racks/shelves for picking
Parts-to-pickers Automated storage system + picking bays, AS/RS + mini load/carousels + picking bays
as operational bays in which you carry on
picking activity, but without storage goods
Pick-to-conveyor Flow-racks (with all the stock) + conveyor High-density storage area + flow-racks +
conveyor

FORWARD RESERVE PROBLEM (FRP)

5.2.1 Hierarchy of decisions
Determine the value of the picking stock (how much storage capacity is dedicated to the picking area) in order to
reduce the sum of the picking costs and the picking area replenishment costs
The FRP decision requires to
- Determine the storage capacity of the picking area:
o Forward area: yes or no?
o Identify the items to be located in the picking area: all or only a part?
o Choose the replenishment frequency if there is a forward area separated from the reserve area
- Embedded problem: allocate the picking storage capacity to the items

1. Decision between two options: RS=PS (united) and RS -> PS (separated)

If we select the first option, the total cost is just the cost for picking, while for the second option, the total
costs is the sum of picking and replenishment activities; in order to define these costs, you have to know
the size of the areas

100
 07. Order Picking and Assembly: systems and design principles

1. Determination of the picking stock capacity (S) -> TC = f(S)

It is important because it determines the Total Cost of the option
Smaller the picking area, the better in terms of picking costs, but the
opposite happens for the replenishment costs, because the more
frequently you have to refill it
We need to find the optimal capacity (Sopt), so the size which can
minimize both the picking and replenishment costs, and then we can
compare this level of costs with the cost of only picking activity (with RS = PS option)

2. Determination of the cost of replenishment: we should consider that the costs for replenishment are a
function of the overall size, but also of how you allocate each item in the picking area
-> 𝐶𝑟𝑒𝑝𝑙𝑒𝑛𝑖𝑠ℎ𝑚𝑒𝑛𝑡 = 𝑓(𝑆) + 𝑓(𝑆𝑖 )
𝑆𝑖 = (𝑝𝑖𝑐𝑘𝑖𝑛𝑔 𝑠𝑡𝑜𝑐𝑘)𝑖 [𝑚3 ] where [i = 1, …, N] → 𝑆 = ∑𝑖 𝑆𝑖
picking stock dedicated to each single item stocked in the overall picking area

ex: assume that we have a RS -> PS system with overall capacity S

𝐹𝑖 = (𝑎𝑛𝑛𝑢𝑎𝑙 𝑓𝑙𝑜𝑤)𝑖 [𝑚3 /𝑦] -> overall flow for each item i during the year
𝑅𝑖 = (𝑎𝑛𝑛𝑢𝑎𝑙 𝑟𝑒𝑝𝑙𝑒𝑛𝑖𝑠ℎ𝑚𝑒𝑛𝑡)𝑖 [𝑛𝑢𝑚𝑏𝑒𝑟] -> number of replenishments for each item i
𝐹 𝐹
-> connection between S, F and R 𝑅𝑖 = 𝑆 𝑖 𝑅 = ∑𝑖 𝑅𝑖 = ∑𝑖 𝑆 𝑖
𝑖 𝑖
NB: we are assuming that we replenish the stock when it is completely empty, so the replenishment lot is
equal to the picking stock area for each item

𝐹
-> 𝐶𝑟𝑒𝑝𝑙𝑒𝑛𝑖𝑠ℎ𝑚𝑒𝑛𝑡 = 𝑓(𝑅) = 𝑓 (∑𝑖 𝑆𝑖) it is function of both S and Si
𝑖

We have a hierarchy of decisions because in order to made the first and most important decision we need to
enter in the second decision, and for take the second decision, we have to go to the third level

5.2.2 Embedded problem: space allocation

The Space Allocation problem is Variables:
about the repartition of the overall Fi Volume Flows for the i-th item (m3 in a time bucket)
space (Smax) to single items (Si) in Si Picking Storage Capacity for the i-th item (m3)
order to reduce the number of ri Replenishment Frequency for the i-th item (cycles in a time bucket)
total replenishments (Rtot) from Smax Total Picking Storage Capacity (m3)
the reserve area to the forward Rtot Total Replenishment Frequency (cycles in a time bucket)
area n Number of items

FRP POSSIBLE SOLUTIONS

We have 3 main types of solutions; the first 2 are quite - Equal space
good, the third one is the optimal one, but also the most - Equal time (frequency)
complex - Minimization of Replenishment Frequency R

5.3.1 Equal space supply

𝑆 𝐹 𝐹
Each item is assigned the same volume in the picking area 𝑆𝑖 = 𝑐𝑜𝑠𝑡𝑎𝑛𝑡 = 𝑁 𝑅𝑖 = 𝑆 𝑖 = 𝑆/𝑁
𝑖
𝑖
-> same space, but different replenishment frequency
The overall storage capacity of the picking area is the volume assigned to each item times the number of items
The replenishment frequency is different for each item and specifically it is the ratio between the item volume
flow and the volume assigned

Pros: High space saturation in the picking area since the same location can be used to keep different items
-> easy to allocate space in the design of layout
Cons: Replenishment is more complex and is concurrent with the picking activities -> very complex to manage it

101
 07. Order Picking and Assembly: systems and design principles

5.3.2 Equal time supply

𝐹 𝐹 𝐹𝐼
All the items are replenished at the same frequency 𝑅𝑖 = 𝑐𝑜𝑠𝑡𝑎𝑛𝑡 = 𝑆 𝑖 = 𝑆
𝑆𝑖 = 𝐹
∗𝑆
𝑖
The allocation of the overall space S is in proportion of the ratio between the flow of each item and the overall
flow, which means that the allocation coefficient of the space is in proportion to the flows
-> same replenishment frequency, but different space
The overall storage capacity depends on the chosen replenishment frequency (the less frequent the
replenishments the larger the storage capacity of the picking area)
-> The storage capacity is the ratio between the Daily Volume Flow and the Replenishment Frequency (times per day)
The overall picking capacity is allocated to the items according to their volume flow

Pros: Easy replenishment planning, i.e. the replenishment activities take place in a defined time window
-> we can organize the replenishment of all the goods at the same time, at the end of our time window
(end of the day, end of the week)
Cons:
- The replenishment quantities vary, and they are often different from a UL multiple
-> higher complexity in design the layout of the spaces
- If the locations are the same size, the space saturation could be low, or the locations must be divided

5.3.3 Replenishment frequency in the equal space or equal time supply

𝑛
With both policies (equal time and equal space supply) the total replenishment policy is 𝑅𝑡𝑜𝑡 = 𝑆 ∗ ∑𝑖 𝐹𝑖
𝑚𝑎𝑥
𝐹 ∑ 𝑖 𝐹𝑖 𝑛
For instance, in the equal space supply case: 𝑅𝑡𝑜𝑡 = ∑𝑖 𝑟𝑖 = ∑𝑖 𝑖 =𝑆 =𝑆 ∗ ∑𝑖 𝐹𝑖
𝑆𝑖 𝑚𝑎𝑥 /𝑛 𝑚𝑎𝑥
𝐹 𝐹 𝐹
In the equal time supply: 𝑅𝑡𝑜𝑡 = ∑𝑖 𝑅𝑖 = ∑𝑖 𝑖 = ∑𝑖 𝑖 =
𝑆𝑖 𝐹𝑖 /𝐹∗𝑆 𝑆/𝑁

-> the overall number of replenishments is the overall flow divided by the ratio between the overall capacity and
the number of items
These two solutions have the same computation of number of replenishments, but they are very different in
practical terms, because they have different pros and cons

5.3.4 Optimal allocation policy

The optimal solution is the one that minimizes R [not demonstrated]

The optimal policy is to allocate the total capacity (allocation coefficient) in proportion √𝐹𝑖
𝑆𝑖 = ∗ 𝑆𝑚𝑎𝑥
to the square root of the volume flows ∑𝑛𝑖=1 √𝐹𝑖
2
(∑𝑖 √𝐹𝑖 )
The resulting total replenishment frequency is as follows 𝑅𝑡𝑜𝑡 = ∑ 𝑟𝑖 =
𝑆𝑚𝑎𝑥
𝑖

The problem here is that we have both the disadvantages mentioned before
-> we have both replenishment frequency and space for different items, so we have complexity in both the design
of the layout and replenishment process
This option is unpractical: the solution could be to use it considering class of items, and not all the different items

An estimation of the benefit gained thanks to the application of FRP instead of heuristics can be obtained through
𝑅 (𝑜𝑝𝑡)
the calculation of an index of efficiency (IE): 𝐼𝐸 = (1 − 𝑅 𝑡𝑜𝑡(ℎ𝑒𝑢)) ∗ 100
𝑡𝑜𝑡
- Rtot(opt) Total number of replenishments in a defined time window (replenishments/unit
of time) for the optimal solution
- Rtot(heu) Total number of replenishments in a defined time window

5.3.5 Efficiency index for the optimal solution

102
 07. Order Picking and Assembly: systems and design principles

6 SEMINAR: ARTSANA GROUP

PICKING SOLUTIONS
We should start considering which kind
of unit loads they have to prepare
according to the customers’ order
- Multi-item cartons
- Single-item cartons
- Layers of cartons/special orders
- FPL

All these have different picking problems

and so they need different type of
picking solutions
NB: we are considering the AI of items
for a specific period of the year

200 items with very high

Pick-to-belt + Batching policy
Multi-item cartons Access Index, very frequently
sorting considering that after there is
They are prepared at requested in the orders
the sorting and packaging
the end of the sorting Low-level picker-
area; this allow to move 200-
area to-parts solution + 1800 items with lower AI
300 orders per time
sorting
Single-item cartons Single order policy
because they don’t need to
Low-level picker-to-parts
Layers of pass through the sorting
+ operational stage
cartons/special orders activity, they are already
packed
FPL No picking is required

FORWARD RESERVE PROBLEM

They have a very big reserve area
- First area with a heavy usage of AS/RS -> 46.000 PL
- Second area with selective racks (>= 2nd level)

Picking stock area

- Selective racks for the 1st level
- Pick-to-belt, which is replenished directly from the
reserve area -> 200 PL
From AS/RS and pick-to-belt there is a strong automated connection, because there are a lot of requests

6.2.1 Module-selective racks

4 𝑝𝑎𝑙𝑙𝑒𝑡 𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛𝑠
𝐴𝑈𝑅𝑝𝑖𝑐𝑘𝑖𝑛𝑔 = 5∗2,5 𝑚2
= 0,32 𝑃𝐿/𝑚2 ≈ 1/3
4∗3 𝑙𝑒𝑣𝑒𝑙𝑠
𝐴𝑈𝑅𝑟𝑒𝑠𝑒𝑟𝑣𝑒 = 12,5 𝑚2
= 0,96 𝑃𝐿/𝑚2 ≈ 1

Selective Racks Area = 13.000 m2

- = 13.000 ∗ 0,32 = 4.160 ≈ 4.100 pallet loads for picking
- = 13.000 ∗ 0,96 = 12.480 ≈ 12.000 pallet loads for reserve stock

103
 07. Order Picking and Assembly: systems and design principles

7 CONCLUSIONS

SUPPLEMENTARY READINGS
- Perego, 1997, Pick quick papers, Logistics Europe, pp. 54-60;
- Caron, Marchet, Perego, 1998, Routing policies and COI based storage policies in picker-to-part systems,
International journal of production research, volume 36, pp.713-732;
- De Koster, Tho Kees, 2006, Design and control of warehouse order picking: a literature review, Erim
report series in research management;
- Frazelle, World-class warehousing, pp. 229-264.
- Marchet G., Melacini M., Perotti S., (2011) “A model for design and performance estimation of pick-and-
sort order picking systems”, Journal of Manufacturing Technology Management, Vol 22, No 2, pp. 261-
282
- Dallari F, Marchet G, Melacini M (2009) Design of order picking system. Int J Adv Manuf Technol 42,(1-
2):1-12

CHALLENGE QUESTIONS
- Why is the picking process key to distribution systems performance?
- What’s the difference (system configuration and performance) between picker-to-parts and parts-to-
picker systems?
- Which are the main design parameters of a picking system? What is the Forward-reserve Problem about?
What is the main choice, in a nutshell?
- What are the three basic management policies? What is the main goal of these policies?

CRITICAL THINKING
How to assess picking costs? How to make an informed guess about picking costs?
ex: food e-commerce, for which 1 order consists of 50 pieces -> ? = cost of picking one order [€/order]

Assumptions:
- Pickers-to-parts system (in food industry is very difficult it is automated)
- Item picking (no batches)
- Packaging cost excluded

Costs components:
- Equipped pickers (cost of labour)
- Equipped space (cost of utilities)

Data (we will assume them)

- Labour cost
o Cost of picker-hour [€/hour] -> we assume 20 €/h
o Productivity [pieces/hour] -> how many pieces you can pick in one hour?
[pf/h; f = flow] -> how many pieces flow from storage and picking area to the warehouse?
it is related to the expected picking time EPT -> we assume 200 pieces/h
- Equipped space
o Unitary cost of equipped space [€/m2 * y] -> we assume 150 €/m2 * y
o AUR [pieces/m2] -> how many pieces we can store in one m3
[ps/m2; s = stock]
NB: remember that there is a trade-off between AUR and productivity
We consider the average volume of one piece -> we assume 250 pieces/m2
o ITR = turnover/inventory =pf/ps [1/y] -> we assume 12 rotations/year
- Order profile ex: lines/order * pieces/line = pieces/order
- Item profile ex: volume/piece [dm3/piece] -> we assume 2 dm3/piece
NB: it is useful to understand the AUR

104
 07. Order Picking and Assembly: systems and design principles

20 €/ℎ
𝐿𝑎𝑏𝑜𝑢𝑟 𝑐𝑜𝑠𝑡 = 200 𝑝𝑖𝑒𝑐𝑒𝑠/ℎ = 0,1 €/𝑝𝑓

150 €/𝑚2 ∗𝑦
𝐸𝑞𝑢𝑖𝑝𝑝𝑒𝑑 𝑠𝑝𝑎𝑐𝑒 𝑐𝑜𝑠𝑡𝑝𝑒𝑟 𝑝𝑖𝑒𝑐𝑒 𝑝𝑒𝑟 𝑦𝑒𝑎𝑟 = 250 𝑝𝑠 /𝑚2
= 0,6 €/𝑝𝑠 ∗ 𝑦
1
𝐸𝑞𝑢𝑖𝑝𝑝𝑒𝑑 𝑠𝑝𝑎𝑐𝑒 𝑐𝑜𝑠𝑡 = 0,6 €/𝑝𝑠 ∗ 𝑦 ∗ 12 𝑦 ∗ 𝑝𝑠 /𝑝𝑓 = 0,05 €/𝑝𝑓

➔ 𝑂𝑟𝑑𝑒𝑟 𝑐𝑜𝑠𝑡 = 0,15 €/𝑝𝑓 ∗ 50 𝑝𝑓 /𝑜𝑟𝑑𝑒𝑟 = 7,5 €/𝑜𝑟𝑑𝑒𝑟

This example is very important since it links very clearly the main features of the picking system, and the models
we used to analyse the different picking solutions
And this example is very useful to understand that e-commerce of food is very sustainable

What are the pros/cons of automation of picking? An in general warehousing?

Talking about automation, we have a lot of examples
- AS/RS
- Automated loading/unloading trucks
- Sorting machines
- Dispenser
- Parts-to-pickers (ex: Kiva Robots)
- …

PROS
- Increase in labour productivity -> we can reduce the cost per item picked
- Increase in space productivity (AUR)
- Increase in safety and security level, in terms of storing the goods avoiding stealing
- Increase in quality level of the service
- Decrease in variability

CONS
- Increase in investment cost (CAPEX)
It should be compared with the sum of labour and space costs (OPEX)
- Decrease in flexibility (in terms of reconfigurability)
There are many dimensions for flexibility:
o Storage capacity
o Throughput capacity
o Ratio between storage and throughput capacity
We cannot change these dimensions independently one from the other
o Item profile (characteristics of the goods: size, fragility, weight…)
- Reliability/availability: it is more reliable of the other systems, but when we have a problem you have the
block of the overall system

105
07. Order Picking and Assembly: systems and design principles

106
 08. Order Picking and Assembly: Management policies

08. Order Picking and Assembly:

Management policies
0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples:
- The main types of management policies in a picker-to-part system:
o Batching policies
o Routing policies
o Storage policies

- apply theories, models, principles and criteria to:

o model and assess the “picking time” in a picker-to-part picking system
o compare, in a picker-to-part system, the main management policies

THE REFERENCE CASE (FOR THE MANAGEMENT POLICIES)

- Picker-to-Parts picking system
- Picking from the “forward” stock
- Multi-aisles low level configuration
In a “low level” system the time required for the
horizontal transfer between each picking position
and the next one is quite more relevant than the
vertical transfer time between the same positions
NB: Picking on the two sides of the aisle

AGENDA
1. Batching policies
2. Routing policies
3. Storage policies
4. Assessment of the picking time

1 BATCHING POLICIES
- Order Picking: The operator has to fulfil one order per mission only (Picking List = 1 customer order)
- Batch Picking: The operator has to fulfil more than one order (a “batch” of orders) per mission (Picking
List = N customer orders = Warehouse Order)

BATCH PICKING PROS AND CONS

Pros:
- Increase in the picking density, decrease of the distance between two consecutive picking locations
In general, the batching policy is used in order to increase the picking density by economies of scope,
meaning that you use the same system/mission for many orders
- Reduction in the average number of lines on the picking list (overlapping) and therefore reduction in
picking stops and picking time per order if more pieces can be extracted together (working on both travel
and retrieval time, because the retrieval picking is more efficient)
- (Reduction in the number of missions and hence reduction in the fixed set-up times)

Cons:
- Sorting is required -> we need an additional activity, so we have more costs and time
- Information-intensive approach: batching is required (i.e. the formation of batches)

107
 08. Order Picking and Assembly: management policies

1.1.1 Pros1: Increase in the picking density

The picking density has increased and the expected travel distance (per order) has decreased
ex: in this case, we have put together 2 different orders: order 1 has 7 orders locations, while order 2 has other 6
locations; according to the different travel policy we use, putting them together, the sum of the length of the two
separated travels is higher than the length of a single travel of a single mission

1.1.2 Pros2: Overlapping of the order lines

More “customer order” lines are grouped in one
single “warehouse order” line

If we overlap different customer orders, may happen

that some order lines are overlapped
In this case, the order line XYZ is repeated in the 4
initial orders, so with the batch picking we have just
a single order line which considers the sum of the
initial quantities of the item XYZ
-> 12 + 13 + 8 +20 = 53

Of course, this second way of managing the order is

used when there are a lot of customers orders and
the batches are big enough
MAIN ISSUES
- batch size (Nopt) -> it depends on picking and sorting costs, there is a trade-off
between them: if we increase the size of the batch, we are increasing the sorting
cost and decreasing the picking cost
- batching algorithms -> the main criteria you should follow to build up the batches
o proximity, putting together orders with item very close from each other
o overlapping of items, which means to put together orders with a lot of equal items
o delivery time, putting together the orders which have to be prepared in a specific time window
- sorting -> which architecture we can use to manage the sorting activity
- (increase warehouse lead time: it is possible to have it, but not always)

1.2.1 Sorting
It is a process through which a batch of orders is sorted into single customer orders
There are two kinds of sorting:
- sorting concurrent to the picking (sort while you pick)
Picked goods are immediately sorted to different boxes/sections of the picking cart
depending on the customer order -> reduction of the identification time, but the size of
the batch and the volume of the orders are limited (N < 10)
- sorting following the picking (pick and then sort)
Picked goods are transferred to a sorting area which could be sorted out
manually or automatically by a sorter fed by the pieces coming from the batch
picking
o manual sorting
o automated sorting, when N is very high

108
 08. Order Picking and Assembly: Management policies

Sorting concurrent to the picking Sorting following the picking: Sorting following the picking:
Manual Sorting Automated Sorting

The objects extracted

are immediately sorted
to different
boxes/sections of the
It is based on a sorting machine fed by
picking cart depending
the pieces coming from the batch picking.
on the customer order
The machine sorts out the pieces
according to the original customer orders

ZONE PICKING
The picking area is divided into different zones which can be fixed (always the same) or dynamic (specific for each
order).
A picking list is split among the different zones and the parts can be picked “in parallel” (following consolidation is
needed) or “in sequence”
Main objectives:
- Technical Requirements, i.e. to use different techniques for different classes of products
- Optimization objectives, i.e. higher efficiency and/or effectiveness
o Reduction in the order throughput time
o Reduction in congestion
o Increase in the batch volume

1.3.1 Zone Picking with pre-defined areas

The picking system has been divided into different areas managed by
different pickers

NB: the selection of a batching policy has not too much to do with the warehouse size, but it is more linked to the
order profile: if for example we have a warehouse with small orders, it is very useful to apply a batching policy,
both for pickers-to-part and parts-to-picker solutions (ex: Artsana)
Of course, we should make the calculations before

2 ROUTING POLICIES
Routing policy: the sequence according to which the items are picked (=picking mission)
The main principle is to find the right sequencing of operations, in order to minimize the paths between 2
consecutive picks

Pros:
- Reduction of travel time in a general sense, considering the people/operators (picker-to-parts approach)
or of the handling effort (parts-to-picker approach)
- Right packaging, which depends on the sequence in which you pick the items
ex: it is important to put the more fragile items over the less ones when preparing the packaging

Cons: it seems that there are not issues, because this idea is always useful
NB: we should remember that for the more sophisticated routing policies, there is operational complexity on the
fields, which is not only a technological complexity at the decision level, but it regards also how to tell the
operators how to perform the activities

109
 08. Order Picking and Assembly: management policies

MAIN ISSUES
Optimality vs pragmatism, which means to solve the trade off between operational complexity and benefits,
because going for an “optimal” solution, we of course increase the benefits, but on the other hand we also
increase a lot the complexity

2.1.1 The problem

Picker-to-part solution
Objective: to minimize the travel of the picking
mission (Input - Locations - Output)
This problem is an instance of the TSP (traveling
salesman problem) with the constraint that the
picker has to move within the aisles

2.1.2 The approaches

1. Order by order: Analyze each order and choose the best picking tour in order to get the minimum travel
(solving the TSP)
o Pros: The travel is minimized
o Cons: A specific picking tour has to be determined per each order -> not standardized path
each order has a specific tour, and they have not to follow specific rules
2. Pre - defined policies: A pattern is chosen once for all and is used for all the orders irrespectively of the
specific travel-minimizing path, meaning that we do not get all the benefits but only part of them,
accepting to have other kind of benefits
o Pros: The routing policy is more stable and easier to memorize
Easier to apply the storage policies, because it is very complex to apply storage policies
without knowing how the pickers move within the system
o Cons: The travel is not minimized

POSSIBLE SOLUTIONS
NB: we are considering a picker-to-parts approach

2.2.1 Traversal policy

The picker enters the aisle, walks through all its length stopping when requested
and gets out of the other side
Sort of “snake-route” along the aisles, so you never do a U-turn; it is easier to
apply this procedure in order to manage possible interferences between pickers

2.2.2 Return policy

The picker enters the aisle from always the same side, picks on one side first,
reaches the farthest requested pick location, then comes back picking on the
other side, getting out of the same side he entered
It is a U-turn approach by definition; it is very clear and standard, so also much
predictable in terms of behaviour of pickers

2.2.3 Mid-point return policy

The picking area is divided into two parts (through the cut of the working aisles).
In each area the picker adopts a return policy. The picking mission is completed
by two traversal paths (one in the first and one in the last aisle to be visited)

2.2.4 Largest- gap return policy

For each aisle to be visited the “largest gap” is determined. The “largest gap” is
defined as the maximum distance between the following distances:
- from the beginning of the aisle to the first picking location
- between each picking location and the next one
- between the last picking location and the end of the aisle

110
 08. Order Picking and Assembly: Management policies

2.2.5 Optimal solution

It is the best one, obtained as a combination of all the policies we have seen

The pragmatic solutions are very predictable, and very applicable in combination with
a storing activity, since knowing the routing policy is useful and necessary in order to
decide the allocation for the items

ASSESSMENT OF THE DIFFERENT ROUTING POLICIES

In this graph we can see the expected travel within one aisle in percent of the aisle length in relation with the
number of picks per aisle

2.3.1 First case

Assumptions:
- Only 1 narrow aisle
- Random storage within the aisle

- Transversal solution we always travel all the

aisle length, and it is always the same, so the
expected travel is always equal to the length -
> percentage = 1
- Return policy
o If we consider just one item to be picked, on average we move
half of the aisle -> travel/aisle length = 0,5 + 0,5 (go and back) = 1
o If we consider more than one item, we should go more in “depth”
along the aisle, so the percentage increases
-> the return policy is always dominated by the transversal policy (considering the assumptions)

2.3.2 Second case

Assumptions:
- only 1 narrow aisle
- 70% of the picks are in the first 20% of the aisle -> different storage policy (AI-based)
For the traversal, nothing changes, but considering the
return policy we have advantages, because we are
increasing the likelihood of having items with high AI
nearest, we need to go to the initial part of the aisle,
not more to its middle point
For a certain number of picks per aisle, the return
policy is better than the traversal

➔ the best routing policy depends on the

possibility to apply the AI-based storage policy

3 STORAGE POLICIES
Main principle is very similar to the one of batching policy, because we want to increase the picking density, but
with a different idea, by proximity, concentrating the majority of picking activity in part of the system smaller than
the whole area

Pros: reduction in the expected picking mission time, considering a reduction in both travel and retrieval time
(the second one through a better management of ergonomics of picking)
Cons: same of routing policy -> increase of operational complexity
we could have to re-allocate items over time within storage and picking areas frequently over time, since
the AI is a dynamic parameter, and this is of course complex and expensive

111
 08. Order Picking and Assembly: management policies

MAIN ISSUES
The storage policies can be divided into three main classes:
- Policies aiming at reducing the picking travel: Access Index Based Storage
-> items with highest AI are located in the best positions
- Policies aiming at reducing the picking time: Correlated Storage
-> items that could easily be in the same order are closer
- Policies driven by operational needs:
o Point of Sale Storage needs
o Compatibility between different products
o Specific needs for the preparation of the Unit Load

Random Allocation means that none of the basic criteria has been used. It might be assumed that there is the
same probability to visit the different areas of the picking system.
ex: in some warehouses (Amazon) this could be the third choice, for different reasons, for example:
- To reduce interferences in a single point of the warehouse
- To reduce the numbers of mistakes due to put close items very similar
- To increase the probability to find the item you are looking for very near to the last one you picked
To do this, basically you split the overall quantity of a single item and position them around the warehouse
NB: this criterion has to be explained, it should have a very clear advantage

ACCESS INDEX - BASED STORAGE

The items with a high ratio between picking frequency/popularity and dedicated space are located closer to the
I/O point
- Pros: reduction in the expected picking time travel (most of the picks are close to the I/O point)
- Cons:
o It could cause congestion and interference within the picking system as all the most requested
items are stored in a few aisles
o Information-intensive approach: AI is a dynamic parameter and allocation should be changed
accordingly

3.2.1 Access index in the picking system

𝑃𝑖𝑐𝑘𝑖𝑛𝑔 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦
𝐴𝐼𝑖 = 𝑝𝑒𝑟 𝑒𝑎𝑐ℎ 𝑖𝑡𝑒𝑚 𝑖
𝐷𝑒𝑑𝑖𝑐𝑎𝑡𝑒𝑑 𝑠𝑝𝑎𝑐𝑒
- Picking frequency: The order lines requiring the specific item in a specific period
∑𝑖(𝑜𝑟𝑑𝑒𝑟 𝑙𝑖𝑛𝑒𝑠)𝑖
∑𝑗(𝑜𝑟𝑑𝑒𝑟 𝑙𝑖𝑛𝑒𝑠)
-> sum of the order lines related to the item i over the overall number of order lines
- Dedicated space:
o It is related to the volume dedicated to the item within the picking systems
o It is the outcome of the Forward Reserve Problem -> 𝑆𝑖

Order lines/year Picking frequency Si AI

NB: the access index is NOT just the picking
Item 1 10 1/1000 1 1/1000
frequency, but it is related to the space
Item 2 100 1/100 0,1 1/10
it is a sort of density of picking frequency
Item 3 500 5/100 1 5/100
ex: with 𝑆𝑖 ≠ 1 → 𝑃𝐹 ≠ 𝐴𝐼 … all items … 10.000

3.2.2 ABC curve “Picking frequency – Space”

In order to apply the AI idea, we need to perform the ABC curve,
ranking the items according to their AI, from the highest to the
lowest
This is the typical curve, which means that the majority of the order
lines are located in a very small part of the overall picking area

NB: the application of this idea is different according to the routing

policy we use, the best location depends on the routing policy

112
 08. Order Picking and Assembly: Management policies

AI-BASED STORAGE - ROUTING POLICIES

The implementation of the AI-based storage depends on the routing policy:
- Traversal policy: the allocation based on the AI aims at reducing the number of aisles to be visited.
Hence, the items with high AI are allocated to the aisles closer to the I/O point
- Return policy: the allocation based on the AI aims at reducing the expected travel within each aisle.
Hence, the items with high AI are allocated to the locations closer to the I/O point of each aisle

3.3.1 AI-based storage with a traversal routing policy

The highest AI items should be stored in the aisles which are nearer to the I/O point
In this example, is not important where you locate the item within the aisle, because when you enter an aisle you
will pass all its length, while is very important the allocation of the aisles in relation with the I/O point

3.3.2 AI-based storage with a return routing policy

The higher the AI the nearer an item should be to the front-end aisle
NB: The partition in three classes is only for illustrative purposes
Using the same layout but different routing policy, the allocation is very different
➔ The application of the AI based storage must be done in relation to the routing policy, otherwise it is
impossible to say which is the best solution

CORRELATED STORAGE
The items that are likely to be picked together (correlated items) should be grouped into “clusters”
The clusters of the correlated items are allocated on the bases of the average AI
- Pros: It reduces the pick stops and as a consequence the picking time (substantial benefit in the part-to-
picker systems, because you move the batch in which you collect the parts less times)
- Cons:
o Information-intensive approach: a complex correlation analysis is required, and it can change
frequently over time
o (The ABC curve based on the AI is not that marked as the ABC curve of the single items)

113
 08. Order Picking and Assembly: management policies

4 ASSESSMENT OF THE PICKING TIME

The four steps to understand the assessment of the picking time Step The main components of the picking time
1. The main components of the picking time
The identification of the main components and the calculations of the fixed time
2. The travel time and the travel distance
The identification of the main components of the travel time/distance
3. The expected travel time and distance
The assessment of the expected travel time/distance in different scenarios
4. Definition of a general framework
A general framework to assess the picking time

1. THE MAIN COMPONENTS OF THE PICKING TIME

The picking time
- Set-up time
o Time to get and read the picking list
o Time to carry the empty picking box and deliver the loaded picking box
- Travel time
NB: We will focus on the travel time since it usually is the most significant and difficult to assess
component
- Retrieval time
o Read and check the picking list
o Position in front of the picking location
o Pick the items from the racks and put them into the box
- (Waiting time: congestion, stock-out, …)

𝑃𝑇 = 𝑆𝑈𝑇 + 𝑇𝑇 + 𝑅𝑇 𝑅𝑇 = 𝑁 ∗ (𝑅𝑇𝐿 + 𝑁𝑃𝐿 ∗ 𝑅𝑇𝑃)

- PT = Picking time - N = Number of lines per tour
- SUT = Set-up time per picking tour (time to - RTL = Fixed retrieval time per line (read and check the
get the picking list and the picking box, …) picking list, position in front of the location, …)
- TT = Travel time - NPL = Number of pieces to retrieve per line
- RT = Retrieval time - RTP = Retrieval time per piece (time to pick the piece)

[from tutorship session]

If we have batching and line overlapping, this has an impact on the retrieval time
𝑅𝑇 = 𝑅𝑇𝐿 ∗ 𝑁𝐿 + 𝑅𝑇𝑃 ∗ 𝑁𝑃𝐿 ∗ 𝑁𝐿 = 𝑁𝐿 ∗ (𝑅𝑇𝐿 + 𝑅𝑇𝑃 ∗ 𝑁𝑃𝐿)
The only way in which the retrieval time is impacting is the first component: if we have the batching overlapping,
we cannot use the second version of this formula

ex: if we have an order picking based on 2 orders, 4 lines/order, 3 pieces/line

If we do not have any overlapping, we will have to pick 8 lines
If we have a line overlapping for both the order, the overall lines we have to pick are 7
The number of pieces we need to retrieve does not change (it is always multiplied by 8), while the number of lines
changes (from 8 to 7 when there is overlapping)!

2. THE TRAVEL TIME IN A PICKING TOUR

𝑇𝑟𝑎𝑣𝑒𝑙 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒
𝑇𝑟𝑎𝑣𝑒𝑙 𝑇𝑖𝑚𝑒 =
𝑆𝑝𝑒𝑒𝑑𝑃𝑖𝑐𝑘𝑒𝑟 𝑡𝑟𝑢𝑐𝑘

𝑇𝑟𝑎𝑣𝑒𝑙 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒
= 𝑊𝑖𝑡ℎ𝑖𝑛 𝐴𝑖𝑠𝑙𝑒𝑠 𝑇𝑟𝑎𝑣𝑒𝑙 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒
+ 𝐴𝑐𝑟𝑜𝑠𝑠 𝐴𝑖𝑠𝑙𝑒𝑠 𝑇𝑟𝑎𝑣𝑒𝑙 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒

114
 08. Order Picking and Assembly: Management policies

4.2.1 The travel components (in the narrow-aisle configuration)

In a narrow-aisle configuration the extra-travel to cross the working aisles (in
order to pick the goods in the front rack) is negligible with respect to the travel
distance measured along the centreline of the aisle

4.2.2 The within aisles travel distance

The within aisles travel distance depends on the number of aisles to enter and
the routing policy within the aisle

4.2.3 The across aisles travel distance

The across aisles travel distance depends on the farthest (from the I/O point)
aisle to enter
Twice the distance of the farthest aisle from the I/O point

4.2.4 Travel distance (Example)

Determine the travel distance for an order made by 10 lines
assuming that:
- the items to be picked are in the red locations
- a traversal routing policy is adopted
- the configuration is narrow-aisle

Within aisles travel distance = 4x15 = 60 m

Across aisles travel distance = (3x4) + (2x4+4) + 5 + 5
= 2x17 = 34 m
➔ Travel distance = 60 + 34 = 94 m

3. THE EXPECTED TRAVEL DISTANCE IN A PICKING TOUR

𝐿(𝑁)
𝐸𝑇𝑇(𝑁) = 𝐿(𝑁) = 𝐿𝐼 (𝑁) + 𝐿𝐸 (𝑁)
𝑆𝑃𝑇
- ETT: Expected travel time for N picks in a tour - LI(N): Expected within aisles travel distance for N
- L: Expected total travel distance for N picks in a picks in a tour
tour - LE(N): Expected across aisles travel distance for N
- SPT: Speed of the picker truck picks in a tour
- N: Expected number of picks in a tour

𝐿(𝑁) = 𝐿𝐼 (𝑁) + 𝐿𝐸 (𝑁) depends on the following:

- N value, which depends on
o Customer order profile
o Batching policy
- Routing policy
- Storage policy
- Layout configuration
-> We will see 3 specific cases stemming from the adoption of different policies

115
 08. Order Picking and Assembly: management policies

4.3.1 Case A
Assumptions The system
- Layout type -----------------> configuration
- Routing policy: traversal - lI = length of
- All aisles have the same probability to the working
be visited (typical in case of random aisle
storage) - wE = width of
- Narrow-aisle the cross
- 1 picking list = 1 picking tour (the aisles
amount of material to be picked is less - wI = width of
than the capacity of the picking box) the module

𝑤𝐸
The expected within aisles travel distance: 𝐿𝑡𝐼 (𝑁) = (𝑙𝐼 + 2 ∗ 2
)∗ 𝑣
- LtI(N): Expected within aisles travel distance (for the traversal policy)
- v: Expected number of visited aisles with N independent picks per mission
- lI: Length of the working aisle
- wE: Width of cross aisles

1 𝑁
The expected number of visited aisles: 𝑣 = 𝑎 ∗ [1 − (1 − 𝑎) ]
- v: Expected number of visited aisles with N independent picks per tour
- a: Number of stocking aisles
1
- 𝑎
Probability that the pick is in the specific aisle
1
- (1 − ) Probability that the pick is not in the specific aisle
𝑎
1 𝑁
- (1 − 𝑎
) Probability that N independent picks are not in the specific aisle
1 𝑁
- 1 − (1 − 𝑎) Probability that at least one pick out of N is in the specific aisle

Example - The expected number of visited aisles

Example Aisles N 15 30
1 8
12 aisles 𝑣 = 12 ∗ [1 − (1 − ) ] = 6,02 𝑎𝑖𝑠𝑙𝑒𝑠 20 10,73 15,7
8 picks 12 30 11,95 19,15
The higher the number of picks per tour (N), the higher the expected number of aisles to be visited

The expected across aisles travel distance 𝐿𝑡𝐸 (𝑁) = 2 ∗ (𝑓 − 1) ∗ 𝑤𝐼 + 2 ∗ 𝐷𝐼/𝑂

- LtE(N) Expected across aisles travel distance (for the traversal policy)
- f The farthest couple of aisles to enter from the I/O point, with N picks per tour
- wI Width of the module
- DI/O Distance between the input/output point and the centreline of the nearest aisle

𝑎/2 2𝑗 𝑁 2𝑗−2 𝑁
The expected farthest couple of aisles to enter 𝑓 = ∑𝑗=1 {𝑗 ∗ [( 𝑎 ) − ( 𝑎
) ]}

- f The farthest couple of aisles to enter from the I/O point, with N picks per tour
2𝑗
- 𝑎
Probability that the pick is in the first j couples of aisles (i.e. in a the first 2j aisles)
2𝑗 𝑁
- ( ) Probability that N independent picks are in the first 2j aisles
𝑎
2𝑗 𝑁 2𝑗−2 𝑁
- [( 𝑎 ) −( 𝑎
) ] Probability that the farthest aisle to enter is one of the two aisles WI (j-1) from the
I/O point. (with N picks per tour)
𝑗 𝑁 𝑗−1 𝑁
NB: if we do not have couple of aisles, we have to eliminate each 2 𝑓 = ∑𝑎𝑗=1 {𝑗 ∗ [(𝑎) − ( 𝑎
) ]}

116
 08. Order Picking and Assembly: Management policies

Example - The expected farthest couple of aisle to enter

N 5 15
Aisles
If the number of picks N is high, we can assume f equal to the (a/2) (with a =
20 8,79 9,75 number of aisles in the warehouse)
30 12,9 14,48

4.3.2 Case B
Assumptions
- Layout type
- Routing policy: return
- All aisles have the same probability to be visited (typical in case of
random storage)
- Narrow-aisle
- 1 picking list = 1 picking tour (the amount of material to be picked is less than the capacity of the picking
box)

Differences with case A

- The expected across aisles travel distance (the same)
- The expected within the aisle travel distance (changing)
o v (number of aisles to be visited) (the same)
o The expected travel distance within one aisle (changing)
This is the only component which is affected by the change in the routing policy

The expected travel distance in a pick tour 𝐿𝐼 (𝑁) = 𝑣 ∗ (𝑤𝐸 + 2𝑙𝐼 ∗ 𝑅(𝑛))

- LI(N): Expected within aisles travel distance (for the return policy)
- R(n): Ratio of the expected value of the farthest pick location within the aisle to the length of the aisle
𝑛 𝑁
𝑅(𝑛) = 𝑤ℎ𝑒𝑟𝑒: 𝑛 =
𝑛+1 𝑣
- v: Expected number of visited aisles with N independent picks per tour
- N: Number of lines per tour n Expected number of picks per aisle visited

4.3.3 Case C
Assumptions
- Layout type
- Routing policy: traversal
- Each aisle has a specific probability to be visited depending on the storage policy
- Narrow-aisle
- 1 picking list = 1 picking tour (the amount of material to be picked is less than the capacity of the picking
box)

Difference with the baseline (case A)

- (*) The expected across aisles travel distance (changing)
- The expected within aisles travel distance (changing)
o (*) v (number of aisles to be visited) (changing)
o The expected travel distance within one aisle (the same)
(*) They change since the “non-random” storage policy makes the probability to visit/enter the aisles different
aisle per aisle

Probability to enter the aisles

The item allocation to the aisles gives the probability to enter each aisle (pi for i =
1,…, a, with a=#aisles)
pi = (annual* picks within the i-th aisle)/(annual overall picks within the system)
*or in another time window

117
 08. Order Picking and Assembly: management policies

The expected number of visited aisles 𝑣 = ∑𝑎𝑖=1(1 − {1 − [𝑝𝑖 ]}𝑁 )

- v Expected number of visited aisles with N independent picks per mission

- a Number of stocking aisles
- pi Probability that the pick is in the i-th aisle
- 1 − 𝑝𝑖 Probability that the pick is not in the i-th aisle
- {1 − [𝑝𝑖 ]}𝑁 Probability that N independent picks are not in the i-th aisle
- 𝑁
(1 − {1 − [𝑝𝑖 ]} ) Probability that at least one pick out of N is the i-th aisle

𝑎/2 2𝑗 𝑁 2𝑗−2 𝑁
The expected farthest aisle to be visited 𝑓 = ∑𝑗=1 (𝑗 ∗ {[𝐹 ( 𝑎 )] − [𝐹 ( 𝑎
)] })

- f The expected farthest couple of aisles to enter from the I/O point, with N picks per tour
2𝑗
- 𝐹 (𝑎) Probability that the pick is in the first j couples of aisles
2𝑗 𝑁
- [𝐹 ( 𝑎 )] Probability that N independent picks are in the first 2j aisles
2𝑗 𝑁 2𝑗−2 𝑁
- [𝐹 ( 𝑎 )] − [𝐹 ( 𝑎 )] Probability that the farthest aisle to enter is one of the two aisles WI (j-1)
from the I/O point. (with N picks per tour)

2𝑗 2𝑗
How to get the F(2j/a) 𝐹 ( 𝑎 ) = ∑𝑖=1 𝑝𝑖

4. DEFINITION OF A GENERAL FRAMEWORK

4.4.1 Customer order profile (P(N)) 4.4.2 The expected travel distance 4.4.3 The expected travel distance
It is the frequency distribution of the in a pick tour in a picking tour
customer orders in terms of number - L: Expected travel distance in a picking
of lines per order tour
- LI: Expected within aisles travel distance
- LE: Expected across aisles travel distance
- P(N): frequency distribution of the
customer orders in terms of number of
lines per order
- L(N): expected travel distance with N
picks in a tour
- N: number of picks in a tour

5 CONCLUSIONS

SUPPLEMENTARY READINGS
- Perego, 1997, Pick quick papers, Logistics Europe, pp. 54-60;
- Caron, Marchet, Perego, 1998, Routing policies and COI based storage policies in picker to part systems,
International journal of production research, volume 36, pp.713-732;
- De Koster, Tho Kees, 2006, Design and control of warehouse order picking: a literature review, Erim
report series in research management;
- Frazelle, World-class warehousing, pp. 229-264.

CHALLENGE QUESTIONS
- Which are the main components of the picking time?
- Which are expected to be the most relevant ones?
- What does the travel time is affected by? How can I reduce it?
- How does the batching policy impact the picking time? Which are its cons?
- How does the Access Index based storage impact the picking time? Which are its cons?
- How do the routing policies impact the picking time? Which are the main alternatives we saw?

118
 09. Transportation: general overview

09. Transportation: general overview

0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to:
- knowingly discuss the following topics and apply them to different real-world examples:
o The role of transportation in the logistic channel
o The main transportation modes and their technical and economic features:
▪ Road
▪ Rail
▪ Water
▪ Air
▪ Intermodal
- apply theories, models, principles and criteria to:
o compare different transportation modes

AGENDA
1. Transportation modes
2. Road transportation
3. Rail transportation
4. Water transportation
5. Air transportation
6. Intermodal transportation

1 TRANSPORTATION MODES

THE LOGISTIC CHANNEL

The logistic channel includes:
- the distribution network, whose nodes are the plants,
the warehouses (central and regional warehouses, transit
points) and the points of sale
- the transportation systems (connections) which connect
the nodes of the network

IMPORTANCE OF TRANSPORTATION IN SUPPLY CHAINS

Transportation is important because it mainly affects
- Costs
o Transportation cost, it is a very expensive activity, it can be up to 5% of the turnover
o Inventory carrying cost, thinking about different types of inventories
▪ SS safety stocks, those taken in order to manage the unpredictability of the demand and
of the lead time -> higher the unpredictability, higher the SS should be, so if the lead time
of the transportation mode is not certain, the SS increase
-> it is highly impacted by the unpredictability
▪ CS cycle stocks, those taken in order to decouple different
activities in the supply chain -> it is highly impacted by the
frequency
▪ ITS in-transit stocks, they have different costs according to
the transportation mode -> it is highly impacted by lead time
o [Packaging cost and other costs]

119
 09. Transportation: general overview

- Service
o order cycle time, it is one of the main activities affecting it
o punctuality (ex: depending on the transportation mode, we can be more or less punctual)
- New trends
o Delocalisation, especially when production and distribution are in very far countries
o E-commerce, transportation is really crucial in order to be very efficient and effective for small
orders made by single person
o Sustainability, since transportation is highly responsible on the CO2 emissions
NB: usually costs and service are in trade-off

1.2.1 Transportation has an impact on service

Transportation is the physical movement
of goods to fulfil the customers demand; it
is the activity that concludes the order
cycle

1.2.2 Transportation has a major impact on costs

…and, as a consequence, on the financial performance of business, as
well as on service level

Transportation expenditure as % of sales

Source: Aberdeen Group (September 2006)

1.2.3 The impact of transportation on costs is growing

Due to both the
increasing cost of fuel
and the adoption of Green Supply Chain practices

Source: European Logistics study 2008-2009 (ELA – AT Kearney)

1.2.4 The importance of freight transportation is growing

Source: Source: PRIMES; year 2000 = 100

COMPLEXITY OF TRANSPORTATION
Transportation is a very complex process
It is possible to identify:
- several modes of transportation
- a number of solutions for each modes of transportation
- many types of transportation providers
- different solutions, based on the product features (e.g. temperature-controlled transportation,
transportation of dangerous products, …)
- different legislations among the countries (e.g. on weight capacity)
Some factors not related to the company (e.g. external factors related to the country such as infrastructures or
traffic level) can affect transportation performance

120
 09. Transportation: general overview

DEFINITION
Transportation includes all the services for the physical movements of a certain quantity of goods between a
point of origin POO to a desired point of destination PD at the required time, (in terms of both delivery cycle time
and punctuality)
- using different modes of transportation
o using load units or vehicles based on the product features and the type of services
▪ ex: for economic reason, we can use big or small trucks
▪ ex: according to product needs, for example if it needs a controlled temperature
o Different providers, since usually it is an outsourced activity, not done in-house
- using the infrastructures -> parts, stations, hubs, …
- being compliant with the legislations (e.g. constraints on the total amount of operating hours of truckers,
safety regulations, …), since transportation is an activity done in a lot of different countries, and each one
has its own rules -> high complexity, and it is one of the main reasons to go for outsourcing

1.4.1 Units of measure

1. Tons [t] : it defines the total volume shipped (also measured in m3)
2. Tons per kilometre [t x km]: it combines weight and distance; it represents the traffic volume or, in other
words, the work done (work equal to force times distance)
Example: 100 tons x km = (100 t) x (1 km) = (1 t) x (100 km)
3. Value [€] : economic value of transportation costs
4. Transportation unit: it refers to the type of load shipped (e.g. pallet, TEU = Twenty-feet Equivalent Unit)

VARIABLES DEFINING A TRANSPORTATION PROBLEM

What are the main variables defining a transportation problem?
- Customer needs (demand) -> we mean a company which requires transportation service
- Logistic server market

1.5.1 Variables related to the customers demand

1. Type of goods (high-volume products, dangerous products, perishable goods, …) [explained later]
2. Service level (e.g. required % on-time delivery, …) [explained later]
3. Terms of sale, for example:
o FOB free on board, which means that the loading cost is already included in the price
o EXW ex-works, transportation is not included in the price of the product
These terms of sale are included in the contract of the transportation service, and we need to look very
carefully at it; all these terms of sale are decided by a norm called Incoterms
4. Shipment characteristics (quantity per trip, frequency, which are in trade-off, parcel service or not, …)
ex: a ship transportation is very good if you have low frequency and high quantity

1. Type of goods
- Density: since it is measured in kg/m3, we can saturate the density working on volume and weight; if we
are able to saturate them both, we can achieve the optimal density
The optimal density depends on the transportation mean (single transportation vehicle), which has a
𝑀𝑎𝑥 𝑊𝑒𝑖𝑔ℎ𝑡
maximum value both of payload and volume -> 𝑜𝑝𝑡𝑖𝑚𝑎𝑙 𝑑𝑒𝑛𝑠𝑖𝑡𝑦 = 𝑀𝑎𝑥 𝑉𝑜𝑙𝑢𝑚𝑒
The type of goods have an impact on the ability to achieve the optimal density; it is not always possible to
achieve it with a single product, so we can try to combine different types of product
o weight to volume ratio < 250/300 kg/m3 -> high-volume products
o weight to volume ratio > 250/300 kg/m 3
-> heavy products
- Perishability/Obsolescence: if we have to transport perishable products (or products that become
obsolete soon, like clothes), it is important to not use transportation like ships which needs months
o frozen products ( -30°C to -20°C )
o fresh products ( -0°C to -4°C )
o temperature-controlled products ( 5°C to 15°C )
o room temperature products ( 15°C to 18°C )
- State: products solid, liquid or gaseous state products

121
 09. Transportation: general overview

- Value: economic value of goods (€/kg, €/m3); it is fundamental due to the impact that transportation cost
has on the value -> we cannot afford an expensive transportation mode if the product has a low value
- Packaging: bulk (with no packaging), pallet, parcel, hung clothes, slip sheets
- Dangerousness/Security: flammable, explosive or toxic products
These factors have an impact on both the mode of transportation and the vehicle type, as well as on the
transportation rates

2. Service level
- Lead time: from 24-48 hours (for couriers) to weeks (for sea transport)
- Punctuality/Reliability: ability to meet the time constraints (time windows)
ex: transportation by ship has to wait a lot of time/days in order to have the required water condition to
enter in the port -> is not always punctual
- Service regularity: ability to meet the required delivery frequency, since there are some transportation
modes that needs to achieve a certain volume to start, so it is not good if you want that your service is
always available
- Traceability:
o tracking (visibility on goods location) -> we want to know the exact current position of the good
o tracing (visibility on shipment routing) -> we want to know exactly the path the good follows
- Safety and security of goods: damages or thefts
- Freight management: monitoring of goods conditions during transportations
- Emergency management

1.5.2 Variables related to the logistics service market

For each mode of transportation, we
- Mode of transportation (road, railroad, sea freight, …)
need to select the right vehicle and
- Vehicle type/transportation unit
transportation unit, and also the right
- Type of logistics service providers
service provider, which can be
- Specialization by industry, production and supply chain, …
specialized by different tasks and also
- Outsourcing decision (transportation outsourcing vs in-house) outsourced

MODES OF TRANSPORTATION
1.6.1 Intermodal transportation 1.6.2 Comparison of transportation mode performances
- Road What are their main performances?
o Full truck loads (FTL) Cost Line haul Delivery time
Accessibility
Environment
o Less than truck loads (LTL) [$/(ton*km)] transit time variability al impact
- Rail Road 3 2 2 1 2
Rail 2 3 3 2 1
- Water
Water 1 4 4 4 3
o Coasting/internal waterways
Air 4 1 1 3 4
o Transoceanic shipments
1= best, 4= worst
- Air

1.6.3 Adoption of transportation modes inside the EU

Freight transportation in EU 25 countries, 2005 (% of ton-km)
BE CZ DK DE EE EL ES FR IE IT CY LV LT LU
Road 70,3% 71,5% 76,6% 63,8% 35,4% 96,9% 92,1% 74,6% 98,4% 86,3% 100 % 26,6% 48,5% 92,6%
Rail 13,0% 24,5% 6,6% 19,6% 64,6% 2,6% 4,6% 14,8% 1,6% 9,3% 62,7% 38,1% 4,2%
Internal waterways 13,8% 0,2% 13,2% 3,2% 3,2%
Pipeline 2,9% 3,8% 16,8% 3,4% 0,4% 3,4% 7,3% 4,4% 10,8% 13,4%
Total 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 %
HU MT NL AT PL PT SI SK FI SE UK BG RO
Road 64,6% 100 % 61,3% 56,4% 59,6% 94,7% 77,5% 60,3% 76,5% 64,0% 83,4% 69,4% 65,4%
Rail 23,3% 3,6% 29,0% 26,7% 5,3% 22,5% 25,3% 23,3% 36,0% 11,1% 25,2% 21,1%
Internal waterways 5,4% 30,7% 2,7% 0,2% 0,3% 0,2% 0,1% 3,9% 10,7%
Pipeline 6,7% 4,3% 11,9% 13,5% 14,1% 5,4% 1,5% 2,8%
Total 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 % 100 %

122
 09. Transportation: general overview

➔ The most used transportation mode inside the countries is road [Source: EU energy and transport]
Road is definitely the most diffused transportation
mode, even considering different countries. This is
because it is the most accessible one and gives good
performances in terms of time and availability, which
have a high impact on the choice, even if it is more
expensive compared to rail and water
Moreover, usually the availability is definitely a
constraint, so using road is the only possible solution

In order to reduce the environmental impact of road

transportation, a solution can be electric solution, but
to do the same for water and air we need a lot of time

2 ROAD TRANSPORTATION
General characteristics:
- Door to door delivery -> this is not possible with the other transportation modes; this is the main reason
for going to road transportation, its flexibility and accessibility
- Limited risk of damage, since goods are handled only at the beginning and at the end of the delivery (*)
This is not possible with other modes,
- Quick loading/unloading, we do not need specific handling systems for this activity
NB: making the comparison between different transportation modes, we should consider all the
activities, for example also the handling activities (ex: for air transportation it is not negligible)
- Low transit time and good reliability (limited impact of weather conditions)
- High accessibility both in terms of roads/highways and in terms of supply (if the company chooses a for-
hire solution)
(*) actually it depends heavily on the carrier’s capability
It has a lot of advantages, for this reason it is the most used

ROAD – FULL TRUCK LOAD

Definition: a full truck load (FTL) shipment is a transportation in which a semi-trailer (or a trailer) is dedicated to a
shipper who fills it with freight for one destination.
The shipment is usually picked up from one point of origin A and delivered to one point of destination B
-> point to point transportation. Between these two points we have a truck which is full, in terms of volume, or
weight or both if we have the optimal density
Once loaded the goods are not handled anymore until they reach their final destination.

Transportation unit load: trailer or semi trailer

Handling equipment: no particular equipment is needed to load-unload the pallet
loads in the trucks, just the normal forklift trucks/pallet jacks and loading platforms;
the handling activity is done only at the point of origin and destination, not during the
transportation activity; the loading and unloading are negligible with respect to the
transit time

Main application fields: a full truck load shipment is usually for medium-long haulage
for medium/high quantities in order to reach the full saturation of the truck (in terms of cube or weight
saturation)
- plant warehouse -> central warehouse
- central warehouse -> regional warehouse -> transit point
- supplier warehouse -> production plant

123
 09. Transportation: general overview

Main managerial issue: saturation of the truck in the direct haul and especially in the back-haul
If we get a full saturation, we cannot be more efficient than this -> to reach this, we should try to reach the
optimal density

Saturation: the transportation unit load (semi-trailer or trailer) has a maximum capacity in terms of weight and in
terms of volume
The ratio between the maximum weight
capacity and the maximum cube capacity
gives the “optimal” freight density, i.e. the
density which saturates both the weight
and the cube/volume capacity
We should always have in mind the value
of optimal density, since if we reach it, we
are minimizing our cost of transportation

Rate structure
The transportation rates depend on both the
distance to be travelled and the weight to be
transported
Fares: money paid for a journey on public transport
[€/km*t]; they are evaluated according to the value
of distance and weight
- If we transport a full truck for the maximum distance, we have the lowest fare
- If we transport a small weight for the smallest distance, we have the highest fare
€
𝑇𝐶 = 𝑄 [𝑡] ∗ 𝑑 [𝑘𝑚] ∗ 𝑓 [ ]
𝑘𝑚 ∗ 𝑡

ROAD – LESS THAN TRUCK LOAD

Definition: a less than truck load (LTL) transportation is a transportation for quantities lower than a truck load,
usually substantially lower, hence the alternative name of “small shipment transportation services”

The transportation means are usually a combination of small trucks/lorries (for the pick-up and delivery) and large
trucks/articulated vehicles (for the line haul) -> we need for sure handling activity
A LTL shipment may be handled only once, or it may be handled multiple times before reaching the final delivery

Transportation unit load: pallet loads or cases aboard a small truck

Handling equipment:
- No particular equipment is needed to load-unload the pallet loads in the
trucks, just the normal forklift trucks
- A network of transit points/platforms is needed to carry out the
consolidation/sorting activities

2.2.1 Main phases of the LTL transportation

Each lorry has to start from the main
warehouse, to go in its area of shipping
and touch point by point all POD (shops,
houses…) it has to serve, and then to
come back to the central warehouse
It is very different from a FTL travel, so it
has different constraints: in this case the
main constraint is the time, since you
have to touch has many points as you
can in order to saturate the time you have
➔ Routing problem

124
 09. Transportation: general overview

Example of a domestic logistics network of an express courier in Italy

- 160 branches
- 9 warehouses

Headquarter (Bologna)
Hub Parcel
Hub for postal logistics
Branches
Warehouses

2.2.2 Typical logistics network of an express courier

Inbound flow: h 18:30-19:30
Flow between the hubs: h 20:00-22:30
Outbound flow: h 4:00-7:00

NB: both processes and systems need to be automated in

order to have short lead times

Main application fields: shipment of small-medium volumes

- from the point of origin to a single point of delivery through a carrier’s network (consolidation with other
shipments)
- from a single point of origin (usually a network warehouse) to many points of delivery: local distribution

Main managerial issue: the minimization of the pick-ups, the minimization of local distribution distances through
routing algorithms (the critical resource is time, i.e. how many pick-ups and deliveries can be done in the time
window)
The main issue is (as said before) to touch as much point as possible in the time of a delivery tour, and the
minimum number of points during a picking tour
NB: the time spent in a delivery area is not negligible with respect to the transit time

2.2.3 Local distribution

- Distribution in urban or small geographical areas,
since the volumes are not so huge:
o Central warehouse -> Big Point of Sales (e.g. supermarkets)
o Regional warehouse/ transit point -> Traditional Point of Sales
- The trucks stop more than once in a delivery trip
- The trucks are smaller than those used for FTL -> express courier
- The fixed times (loading/unloading, controls ...) are comparable with the variable times of a delivery trip

ROAD - SERVICE PROVIDERS

The road transportation can be carried out according to two main options:
- private transportation, i.e. the carrier is also the owner of the goods shipped
- for-hire services
o low fixed costs and high variable costs, resulting in low capital expenditure to enter the market
(especially in the FTL services)
o hence, there are plenty of small carriers (100.000 in Italy, 500.000 in the USA). They usually work
as subcontractors of large carriers/logistic service providers offering FTL or LTL services

ROAD - CHARGEABLE WEIGHT

The concept of Chargeable Weight has been adopted by the transportation industry worldwide because shipping
costs have historically been calculated based on the gross weight (kg, t, q), and not on the volume (m3).
Rates based only on the weight make lightweight and low density packages unprofitable for freight carriers due to
the amount of space required in the vehicle in proportion to their actual weight.
Hence, the transportation rates are calculated based on the Chargeable Weight instead of the real weight.

125
 09. Transportation: general overview

2.4.1 Notations
Parameters related to the vehicle
- CV = Total cost for the vehicle (€)
- PWV = Vehicle capacity in term of weight (q or kg)
- PVV = Vehicle capacity in term of volume (m3)
- DM = PWV / PVV = Minimum density to have a full vehicle in term of PWV (kg/m3)
- CUV = CV / PWV = Reference unit cost for the vehicle (€/q)

Parameters related to the shipment

- WR = Real weight on the vehicle (q, t or kg)
- VR = Real volume on the vehicle (m3)
- DR = WR / VR = Real density (kg/m3)
- UVR = WR / PWV (if DR > DM); VR / PVV (if DR < DM) = Real vehicle utilization

𝐷𝑀 𝑜𝑝𝑡𝑖𝑚𝑎𝑙 𝑑𝑒𝑛𝑠𝑖𝑡𝑦𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛

𝐶ℎ𝑎𝑟𝑔𝑒𝑎𝑏𝑙𝑒 𝑊𝑒𝑖𝑔ℎ𝑡 = 𝑅𝑒𝑎𝑙 𝑊𝑒𝑖𝑔ℎ𝑡 ∗ = 𝑟𝑒𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡 ∗
𝐷𝑅 𝑟𝑒𝑎𝑙 𝑑𝑒𝑛𝑠𝑖𝑡𝑦
The ration can be lower or higher than 1
- Ratio < 1 -> high density, -> saturation in weight
the real density of the goods you have to transport is higher than the optimal density
- Ratio > 1 -> low density -> saturation in volume
We apply the CW only when we have a low density, to avoid the loss of the opportunity to charge more weight

2.4.2 Example
Small truck
- Capacity = 6 t, 30 m3
- CV = 200 €/day
- Real density: Case A = 300 kg/m3; Case B = 100 kg/m3

𝑃𝑊𝑉 6.000
𝐷𝑀 = = = 200 𝑘𝑔/𝑚3 -> optimal density; it is the same for both the cases
𝑃𝑉𝑉 30

Case A 𝐷𝑅 = 300 𝑘𝑔/𝑚3

𝐷 200
𝐷𝑅 ≥ 𝐷𝑀 -> 𝐷𝑀 = 300 < 1 -> Full vehicle in term of weight, we do not apply the Chargeable Weight
𝑅
→ 𝑄𝑊 = 300 𝑘𝑔/𝑚3 ∗ 10 𝑚3 = 3 𝑡 = 3000 𝑘𝑔
200
𝑇𝐶𝑘𝑔 = = 0,33 €/𝑘𝑔 → 𝑇𝐶 = 3000 ∗ 0,33 = 1000 €
6000

Case B 𝐷𝑅 = 100 𝑘𝑔/𝑚3

𝐷 200
𝐷𝑅 < 𝐷𝑀 -> 𝐷𝑀 = 100 > 1 -> We apply the Chargeable Weight
𝑅
𝐷𝑀 200
𝐶𝑊 = 𝑅𝑒𝑎𝑙 𝑊𝑒𝑖𝑔ℎ ∗ = [100 𝑘𝑔/𝑚3 ∗ 10 𝑚3 ] ∗ = 2 𝑡 = 2000 𝑘𝑔
𝐷𝑅 100
𝑇𝐶 = 2000 𝑘𝑔 ∗ 0,33 €/𝑘𝑔 = 667 €

NB: in this case, we do not pay for one transport 1 ton, but we pay for 2 tons
If we do not apply the CW, 𝑇𝐶 = 1000 ∗ 0,33 = 334 €
In this way, the service provider is not losing money: the main idea is that if we have a case in which we saturate
by volume and not by weight, they do not ask you to pay for the real weight you have to transport, but for the CW
The truck is full in volume, but you could transport other product with higher weight, otherwise is like you are
transporting “air” -> there is a fare which refers to the CW in order to take into account not only the weight but
also the volume occupied by that specific product

NB: we have to apply the CW when the service is provided by an external supplier who decides to apply it,; when
we have to internally pay for the transportation mode we are paying all the capacity of the truck, even if it is not
saturated (in volume and/or weight)

126
 09. Transportation: general overview

3 RAIL TRANSPORTATION

RAIL - TRADITIONAL SERVICE

Definition: in the railroad traditional service, the carrier rents space on a train, usually a common carrier, less often
a contract carrier, it not 100% private, they follow some governments economic rules
We can have different configuration of rail transportation, as for the road one:
- a unit train, i.e. the shipper rents a whole train going from one point of origin to one point of destination.
It is similar, from the managerial point of view, to FTL road transportation
Pretty uncommon, since you should always have a huge quantity of products to be shipped; the costs
depends on the specific case, and of course on the different kind of train that you are taking into account
- a full car load
- less than a car load (pallets or cases)
For carload or less than carload shipments (like LTL in road transportation) there are consolidation points, in
which the trains are formed by joining cars going to the same destination, and sorting points, in which a train is
split and the cars go forming other trains

3.1.1 Main characteristics

- Low cost (cost per ton*km is pretty low), especially for long distances and large shipment quantities (high
fixed vs low variable cost structure)
The parameter which affects most the cost is the pay-load, which in this case is linked to
o thousands of tons that can be transported
o very long distances -> high fixed costs
- Low accessibility: the loading/unloading points must be on the railroad, exceptionally large production or
distribution facilities have rail sidings -> it is not always possible
It is pretty uncommon having rails which arrive directly at the warehouses or POS
- Slow speed/High transit times and Low reliability: mainly due to the high and variable handling times at
the consolidation stations (unit trains are an exception), it is not due to the fact that the train is slow, but
to the long transit time
- Damage risk: variable depending on the type of service (full car load vs less than a car load)
ex: considering a less than a car load, you need to sort the goods inside a lot of times, increasing the
probability to break them

Transport unit load: different types of cars depending on the freight characteristics
Handling equipment: we need specific handling equipment
- to load/unload cars Forklift trucks (common), cranes (more expensive, and requires
higher handling time), …
- to combine/sort cars into trains Sorting terminals (stations) with dedicated rails and handling
equipment (locomotives, very expensive)

3.1.2 Different types of cars

- Box car, in which we can put the PL as in the road transportation mode
- Flat car, in which we can put raw materials (ex: lumbers)
- Dumpcar, in which we can put pallet stacks
- Hooper

127
 09. Transportation: general overview

3.1.3 Main application fields

- transportation of bulk raw materials (lumber, cereals, coal, etc.) and generally low-value goods for large
quantities and long distances, otherwise this could not be economically sustainable
- when time performances (both speed and reliability) are not critical
-> the expected order cycle time is pretty long, due to the long transit time
- the cost gap (between rail and road) increases for large shipments over long distances: rail carriers have
higher fixed costs than road carriers (large investment in terminals and handling equipment) and these
fixed costs have a lesser impact when spread over larger ton*km, so it is better to split them along higher
distances (in respect to the road)

RAIL – SERVICE PROVIDERS

The railroad transportation is usually a service offered by common carriers who give a “liner service” for less than
a car load and full car load shipments and also organise unit trains for large shippers (these carriers own both
locomotives and cars)

Other options are:

- contract carriers, which operate under deregulation laws, offering full car load (unusual) and unit train
services (they usually own the locomotive and the shipper provides the loaded cars)
- private transportation is very uncommon – i.e. the carrier (owner of the train) is also the owner of the
goods shipped

MODES OF TRANSPORTATION RAIL – EFFICIENCY

Indicator US Europe
Length of the train 2,000 m 600 m The efficiency and other performances in
Axial weight 31.8 ton 22.5 ton general of rail mainly depends on
Goods weight 5,000 ton 1,300 ton infrastructures
Type double “high-cube” “high-cube”
Speed 75 km/h 75 km/h In the US we have a fair which is lower
Average travel distance 2,500 km 700 km than the half of the fair we have in Europe,
Net train capacity 600 TEU 100 TEU because in US the length is higher and so
Performance the weight could be higher, and this has a
Annual revenue 42 mld€ 17 mld€ very high impact on the costs
“Unit revenue” [fair] 0.016 €/ton*km 0.038 €/ton*km

4 WATER TRANSPORTATION
Main characteristics:
- Low cost compared to the other transportation modes -> cheapest transportation mode
The pay-load of a ship is much higher than the pay-load of the train, and they travel longer distances
- The only option, apart from air, in the intercontinental shipments
- Very low accessibility: the loading/unloading must take place in seaports (that require specific handling
systems), thus requiring land positioning through other ground transportation modes (road or rail)
- Low speed/High transit times, due to limited haul speed -> slowest transportation mode
It is slow itself and travel time is very variable
- Acceptable reliability: the main critical aspects are the dependability on weather conditions (both when
the ship is outside and near the seaport) and the time variability of port handling operations

128
 09. Transportation: general overview

TYPES OF SHIPS
We have different types of ship according to what we need to transport [we do not have to know these numbers]
In general, when we talk about ships, we are talking about a very high capacity (different thousands of payload)
Ship Transport unit load Capacity Handling equipment
General cargo Dry bulk, cases, pallet loads, 1,000 – 50,000 tons Cranes, on board derricks
containers, cars
Bulk and multipurpose Dry bulk, cases, pallet loads 10,000 – 300,000 tons Cranes
Tankers, LNG carriers Liquid, gas bulk 1,000 – 500,000 tons Pipelines
Container ship Containers 7,000 – 50,000 tons Gantry cranes, straddle carriers, reach
1,000 – 9,000 TEUs (*) stackers
RO-RO ship Trailers, semitrailers, cars 2,500 – 20,000 tons None (**)
(Roll On - Roll Off)
(*) TEU = Twenty-feet Equivalent Unit (ISO 8x8x20 feet container), very diffuse unit of measure
(**) The trucks go in the ship by themselves (like on a ferry boat), they do not need any specific handling system

ISO CONTAINERS
We have 2 different types of containers that are most diffused: 40 feet length and 20 feet length; if we need to
know the dimension of different types of container, we can rely on these tables

EXTERNAL DIMENSIONS (MAX) GROSS CAPACITY INTERNAL DIMENSIONS (MIN) # PALLET LOADS/TIER
LENGTH WIDTH HEIGHT [kg] LENGTH WIDTH HEIGHT 1,200X 800 1,200X1,000
TYPE (mm) (feet) (mm) (feet) (mm) (feet) (mm) (mm) (mm)
1 A 12,192 40 2,438 8 2,438 8 30,480 11,998 2,330 2,197 22 20
1 AA 12,192 40 2,438 8 2,591 8.6 11,998 2,330 2,350 22 20
1B 9,125 30 2,438 8 2,438 8 25,400 8,931 2,330 2,197
(*)
1 BB 9,125 30 2,438 8 2,591 8.6 8,931 2,330 2,350
1C 6,058 20 2,438 8 2,438 8 24,000 5,867 2,330 2,197 11 9
1 CC 6,058 20 2,438 8 2,591 8.6 5,867 2,330 2,350 11 9
(*) Used for bulk materials and liquid transportation

TYPES OF WATER TRANSPORTATION

4.3.1 Coasting / Internal waterways
Definition: coasting (short-sea shipping) and internal waterways refers to
shipping goods for short-medium distances between points on the same
coast or along internal navigable waterways
We stay more or less in the same continent

4.3.2 Transoceanic shipments

Definition: transoceanic shipments are international transportations
(from one continent to another one) made by big ships on:
- liner service -> most diffused one
o routes and frequencies are arranged in advance
o ports and routes are chosen by the line
o consolidation of many shipments
o the line might offer a “door-to-door” service, taking
charge also of the road transportation, in the sense
that who is in charge of the water transportation
could offer also the management of the rest of
transportation made by road to/from the seaports
o the line is in charge of the road transportation if
there are “house” conditions
- contract ad hoc services -> not common

Line-haul services: example

129
 09. Transportation: general overview

Main application fields: thanks to its low cost water transport is advantageous for commodities with a low value-
to-weight relationship or for commodities in which transportation cost is a significant portion of the selling price,
for instance
- crude oil, refined petroleum products
- coal, grain, sand, lumber, chemicals
- low-value containerized goods (ex: Alibaba, AliExpress)
… and consider that in many intercontinental transportations the only viable option is air, which is more
expensive, it is not convenient if the products have a low value density
ex: another example is the postal service, since the marginal profit on the goods is very low

WATER – SERVICE PROVIDERS

The water transportation is usually a service offered by:
- carriers that give a “liner service” (both coastal and transoceanic routes) – the usual choice for general
cargo and containerized shipments
- carriers that provide “ad hoc” specific services for one or more shippers
- private transportation – i.e. the carrier is also the owner of the goods shipped – possible for bulk
materials and oil products

5 AIR TRANSPORTATION
Main characteristics:
- Very expensive compared to the other transportation modes, since the payload of air is very low
- Low accessibility: the loading/unloading must take place in an airport, hence road transportation is
usually needed to provide door-to-door service
- High speed/Low transit times, mainly dependent on the speed of the airport-based handling operations
Handling is not easy at all, since you have to prepare everything to the single payload
- Good reliability, even though transit times are affected by weather conditions
- This is very different from ship travels, since we are talking about hours of delay, not days
Compared to the water transportation, we are talking about very different performances but also different costs

AIR – DIFFERENT AIRPLANE TYPES

- freight only (All Cargo; Freighters)
o Wide Body (es. B747 e A900): special pallet loads and special containers
o B747/200F: ~ 108 t payable load; ~ 690 m3 [FTL has a 28t payable load]
With an airplane, you can move quite the same amount of 4 trucks
For each different transportation mode, we are talking about different magnitude of tons
o A300/B4: ~ 41 t payable load; ~ 296 m3 -> optimal goods density: ~ 150 kg/m3
- Passengers + freight (belly only ; combi):
o B747/40: ~ 400 sits + ~ 13 t payable load

5.1.1 Airplane load: narrow body-wide body

Section of the airplane body
We should adapt our payload to the different
shape of the airplane

130
 09. Transportation: general overview

5.1.2 Airplane load: example

Airplane: Airbus A330-223

Maximum weight: 37,578 kg

Available volume: 155.5 m3

Number of unit loads: - Compartment 3: 2 pallet

- Compartment 1: 8 containers - Compartment 4: 6 containers
- Compartment 2: 2 pallet - Compartment 5: bulk

AIR TRANSPORT UNIT LOADS

- Pallet (decks loaded with bulk goods covered by a metal net) ---->
- Specific containers to maximize the loading capacity
We have lot of handling operations before the flight: it requires time, it is
not cheap at all and the risk of damage increases

Pallet Container
- IATA code: P6P, PQP, PMC,PMP - IATA Code: LD2 - DPE - IATA Code: LD9 – RAP
- IATA code: 2-2Q - IATA Class: 8 - ATA Class: 5
- Interchangeability: 74F/D1F upper deck - Interchangeability: 767/ 777 Interchangeability: 74F/ D1F/
- Cube capacity: 20.5 (2) – 11.4 (2Q) m3 - Cube capacity: 3.4 m3 743/ 767/ 777/ M11/ 330/ 340
- Maximum net weight: 4,516 – 6,694 kg - Maximum net weight: - Cube capacity: 7.0 m3
- Maximum height: 160 cm 1,162 kg - Maximum net weight: 5,603 kg

Main application fields: long distance (international and sometimes domestic) transportation of high-value
products and/or when transit time is critical (ex: obsolescence level is high), since the impact on costs is very high
- goods with high value-to-weight relationship
- perishable goods, such as fruits or vegetables
- fashion goods -> high obsolescence
- short life cycle products (electronic components)
- emergency transportation (spare parts for instance)
… air transportation higher costs should be traded off with lower inventory carrying costs

AIR – SERVICE PROVIDERS

The air transportation is usually a service offered by:
- carriers that give a “liner service”– the usual choice for passengers and freight airplanes, possible also for
cargo-only airplanes
- carriers that provide “charter” services for one or more shippers – the usual choice for cargo-only
airplanes
- carriers that provide express door-to-door service, where the air transportation is only part of the overall
transportation service (typically the international express small package carriers)

131
 09. Transportation: general overview

6 INTERMODAL
Definition: Intermodal transportation services refer to the use of two or more carriers of different modes in the
movement of a shipment, without changing the transport unit load, known as Intermodal Transportation Unit
(ITU) -> the other cases in which you use different transportation modes and you change the unit load are called
multi-transportation modes
It requires Intermodal Terminals, i.e. dedicated facilities where the ITU are transferred from one mode to the other
It leverages the best features of each transportation mode, transforming some “labour intensive” activities into
“capital intensive” activities
Two main intermodal transportation modes:
- Road + rail (also known as piggyback)
- Road + water (also known as fishyback)

INLAND TERMINAL
ex: the body of the truck can be moved to the truck of the train (piggyback)
Inland terminals allow shifting from a mode of transportation to another and storing
containers. It is also possible to transfer containers from a train to another (in this
case, the terminal functions as “gateway”).

INTERMODAL – PIGGYBACK
Definition: part of the route is made by truck, part by rail
Main characteristics:
- Door-to-door service (like the road); the most common path is road-rail-road
- Low costs, higher than only-rail but usually lower than only-road, especially for long distances
Higher the distances, lower the travel costs -> we have of course additional costs for the terminal and for
the handling systems, but we assume that the reduction of transportation costs is higher
- High transit times, mainly dependent on the railroad transit times and the handling times in the
intermodal terminals, and usually low reliability
- Low damage risk, ITU are never “opened” during the transportation, we do not touch the product, just
the container

The ITU is formed at the begin, in the POO,

and then it never changes along the overall
path

6.2.1 Intermodal Transport Unit (I.T.U.)

- Container (COFT – Container on flat car) -> container is moved
- Swap Body
- Semitrailer (TOFT – Trailer on flat car)

6.2.2 Loading
According to the handling system, we need different
handling equipments:
- The vertical loading must have a specific handling
system
- The horizontal loading does not necessary require
a specific handling equipment, since the previous
transportation mode can load itself on the next
transportation mode

6.2.3 Handling equipment for containers

- Straddle carriers
- Reach stackers
- Gantry cranes

132
 09. Transportation: general overview

INTERMODAL – ROAD VS INTERMODAL

For the road, the curve is not linear since, apart from the fixed
costs, we have a fare which depends on the distance and
weight of product we have: longer the distance, lower the fair

The slope of the rail is of course lower than the one of the
roads since it is cheaper; we can see a cost increase due to the
fact that we need handling activities (which are fixed costs), and
we need to do them twice

We can compute the cost savings for the two cases

NB: we are mapping the overall costs along the transportation path from a POO to a POD, we are not evaluating
the costs according to the increase of distance
between the two points
Road transportation
Intermodal transportation for case A (X kilometres of
route by rail)
Intermodal transportation for case B (Y kilometres of
route by rail)

The more the distance by rail, the more the cost saving of the intermodal transportation, because we travel more
distance with a transportation mode which is cheapest than the other
NB: the road costs in this case seems linear since the distance you travel on road in the intermodal transportation
is very short, so we can take this approximation
-> we can use the intermodal travel when the distance makes it economically convenient, and we can understand
which the minimum distance is drawing this kind of graph

NB: the application area of intermodal

transportation is wider if the origin point is
located in an inland terminal, because we do not
have road transportation at the beginning

- The cost structure illustrated is only valid if the forward and return flows are balanced
- The slope of the curves heavily depends on external factors (e.g. tolls, labour and fuel costs, constraints
on speed limit and number of operating hours), especially in the road transportation case

FISHYBACK
Definition: part of the route is made by truck, part by ship -> road-water-road
Main characteristics: -> mainly as the piggyback
- Door-to-door service (like the road)
- Low costs, higher than only-water but usually lower than only-road especially for long distances
- High transit times, mainly dependent on the water transit times (and the low frequency of many routes)
and the handling times in the seaports -> higher than the piggyback ones, the reliability is lower
- Low damage risk, ITU are never “opened” during the transportation

6.4.1 Intermodal Transport Unit (I.T.U.)

1. Container (container ship)
o Special handling equipment to load/unload the ship
o Long time to load unload/the ship -> longer handling activity
o Better cube utilization
2. Semitrailer (RO-RO ship)
o No handling equipment to load/unload the ship
o Short time to load unload/the ship (few hours)
o Worse cube utilization, because of the internal structure

133
 09. Transportation: general overview

7 FOCUS ON FTL ROAD TRANSPORTATION

INTRODUCTION

7.1.1 Learning objectives

- The cost structure of road transportation
- The importance of the economies of scale and density
- The optimization of road transportation
o key variables
o main policies

7.1.2 Agenda
1. Cost Structure
2. Optimization

COST STRUCTURE OF ROAD TRANSPORTATION

Transportation costs are affected by:
- Economies of scale
- Economies of density

Economies of scale exist when an increase in the fleet capacity (or in the facilities) leads to a reduction in the unit
transportation cost. Also the increase in the company size allows a reduction in the unit transportation cost; this
also is a benefit related to economies of scale.
We have the opportunity to reduce the transportation costs by increasing the size of the transportation means,
since we are increasing the scale of our resource (the transportation means); this is true when the marginal costs
to pass from a small truck to a bigger one is lower than the increase in the marginal increase in the capacity
NB: of course, in reaching this we should always saturate our mean; saturation is not the reason of economies of
scale, it is a requisite

Economies of density exist when, despite an increase of the flow, the transportation capacity (means + facilities)
remains the same, thus reducing the unit transportation cost.
An increase of number of POD that a single mean has to touch can be considered an economy of density, since we
are maintaining the same flow and reducing distance between different points, so we are increasing the
saturation, which allows a decrease in transportation costs

7.2.1 Economies of scale

Example
The adoption of bigger trucks generates economies of scale (only if
they are saturated); the unit cost decreases the bigger the truck is

7.2.2 Key variable

In order to understand how optimize transportation costs, it should be noted that they heavily depend on the
cost/truck*trip (no matter the structure of the fares if transportation is outsourced)
In the next paragraphs, an analysis of the transportation cost structure has been performed considering a truck
with allowed total weight ranging from 38 to 44 t. In the example, values have been taken from a study
conducted by Albo Nazionale Autotrasportatori (July, 2006)

134
 09. Transportation: general overview

7.2.3 Example
Cost items
- Truck
o Purchase (Amortization)
o Insurance
o Ownership tax
o Maintenance and repair
- Tyres
- Fuel
- Driver
- Tolls

The most important/significant parts of the

cost structure are more or less its 70%: driver,
fuel and investment amortized ≈ 33% + 20% + 18%

Incidence of the cost items

- Fixed costs (i.e. that do not depend on the truck utilization and/or travel distance) → 24,03%
o Truck amortization
o Ownership tax
o Insurance
- Semi-fixed costs (i.e. that partially depend on the travel distance) → 33,86%
o Driver
- Variable costs (i.e. that depend on the travel distance) → 42,11%
o Tyres
o Fuel
o Maintenance and repair
o Tolls

The key message is that almost the 60% of the transportation costs is not variable, and this means that the better
we use the resource (transportation mode), the better it is: considering this, we can associate our transportation
mean to a production system, in which the majority of the costs are fixed

7.2.4 In brief
The truck must be considered as a production system, in which the efficiency increases with the truck utilization
It should be noted that a shipment from an origin point A to a destination point B implies that the truck is not
available since it enters A and exits B (i.e. the loading/unloading time is included)
The main KPIs of the efficiency are the total number of kilometres travelled and the operating hours
→ The total number of kilometers travelled and the operating hours are the parameters to be maximized in order
to have a correct planning

OPTIMIZATION OF ROAD TRANSPORTATION

Based on above mentioned
importance of each cost item, the
following ways of optimizing the
transportation costs can be identified:

(*) Hp: truck (28 t) and annual travel

distance equal to 100,000 km/year

7.3.1 Key variables

The key variables to optimize road
transportation are the travel distance
and the truck utilization

135
 09. Transportation: general overview

Annual travel distance per truck, that depends on:

- Total number of operating hours per driver
- Working load of the truck
- Loading/unloading time (including the waiting times) compared to the total time per trip
- % of backhaul trips
- Average travel speed (related to the limits of the country)

We will see many options to improve the utilization rate of our mean,
using the maximum of
- Annual travel distance per truck
o Waiting time for loading/unloading
There are many cases in which this time is pretty high
(even more than 2 hours), so we are missing the opportunity to use these hours to travel
o Backhaul trips, in which we travel with truck not fully saturate, so we are missing the opportunity
to bring higher UL with one travel
- Truck capacity utilization
o Optimal density, it depends on the specific product, and sometimes it is bad
o Not enough volumes

Waiting time for loading/unloading

It affects negatively the travel distance, since if we spend a lot of time for loading/unloading, we have less time to
travel
The truck waiting time is at least 2 hours

Backhaul trips
Diffusion of empty backhaul trips
There is a large amount of trucks that go back to the POO without any item
inside, so we are losing money, since we are losing the opportunity to use
this distance and time to transport items
→ we should increase both the travel distance and the saturation, not only
one of them

Causes of empty backhaul trips:

- Flows not balanced
ex: flows from Northern to Southern Italy are not balanced with the flows from Southern to Northern Italy
Vehicles (> 3.5 t) Source: Istat (1998)
o Flow from Northern to Southern Italy: 16.380.000 t
o Flow from Southern to Northern Italy: 11.941.000 t
→ The difference (4.439.000 t) implies many backhaul trips from Southern to Northern Italy
- Concentration of points of origin (e.g. industrial districts)
- Difficulty in sharing trucks among companies (for instance, due to incompatibility of goods and/or
strategic reasons) [we are considering the point of view of a service provider]

Truck capacity utilization: keeping the €/km fixed, the higher the truck capacity utilization, the higher the cost
saving (€/km*t)
In some cases, the truck utilization can be low due to the products features (e.g. non palletized unit loads)
Hence, it is necessary to define a target level that takes into account the products features
The truck utilization can be improved by changing the dimensions of the unit loads
In local distribution, when the order size is small, the truck utilization is affected by the maximum number of
stops in a trip

The utilization of the truck floor depends on the number of pallets on the floor; non-standard pallets do not allow
to achieve the best truck capacity utilization

136
 09. Transportation: general overview

The utilization of the weight and volume capacity depends on:

- product density (weight to volume ratio); it depends on specific products
- pallet height (a pallet higher than 2.4 m is not common)
- possibility to overlap pallets

Example: a trailer (28 t; 80 m3)

- maximum number of pallets on the floor: 33 (floor space utilization = 97%)
- optimum weight to volume ratio : 28 t / 80 m3 = 350 kg/m3 (if the average pallet density is lower than 350
kg/m3, the volume capacity is the constraint fully satisfied; on the contrary, it is the weight capacity that is
fully satisfied
- considering 33 pallets, the maximum utilization of the weight capacity is achieved with an average pallet
weight equal to 850 kg (= 28 t / 33)

WAYS TO OPTIMIZE
The ways to optimize the costs per truck can be summarized as follows
[Important role of logistics (in the sense of the flow management)]

7.4.1 Solutions that affect the annual travel distance per truck
- Reduction of loading/unloading times (e.g. collaboration with other supply chain members, scheduling of
the appointments for truck loading/unloading, automation of the loading/unloading activities)
[explained later]
- Triangulation, utilisation of the same truck to do the overall path; it is not always possible, the distances
have to allow that [explained later]
- Continuous movement, the trucks always work; we should be very organized in terms of drivers
management, since they can work/drive for a maximum amount of hours
- Balance between the forward and return flows
- Intelligent Transportation Systems and other new technologies
(e.g. trip planning -> everyone in the warehouse is ready to receive the goods)

Reduction of loading/unloading times

- Organizational changes
o arranging appointments for loading/unloading activities
Operators should be ready to unload/upload; we can go for this
solution if the truck is punctual
o separating the trailer from the tractor (only the trailers, loaded or empty, wait)
The trailers can be quickly downloaded and unloaded, and then managed while the tractor is
already travelling; this is done in order to be more flexible and save a lot of waiting time
- Adoption of new technologies
dispatch advice (DESADV), i.e. the electronic transmission of the list of goods on the
truck, so when the trucker arrives, he has not to wait that the load is controlled by the
operators in the warehouse, since the documents already arrived in
- The automation of the loading/unloading activities is particularly convenient in case of shuttling

Triangulation
Triangulation is possible with limited distances that allow the
return to the plant during the day

With triangulation, the kilometres increase and, therefore, the

costs €/km ton decrease; basically, we can do whit one truck
what previously we were doing with two trucks

Continuous flow
Continuous movement, i.e. truck “always” works
(most of all with long distances)

137
 09. Transportation: general overview

7.4.2 Solutions that affect the truck capacity utilization

- Division of the trailer in compartments to ship different types of goods together
- Increase the truck capacity utilization (e.g. by using shelves on the truck)
in order to use the overall space even if the goods are not stackable
- Reduction of the height/volume of tertiary packaging on the truck (e.g. increasing the
pallet height, use of slipsheet instead of pallet, …) but we need specific handling/fork
system to manage them
- Merging orders (in terms of time or space), if we can: the order cycle time should not be
very short, or we could have no time to separate/sort the orders
- Multi-client warehouses [explained later]
- Using different types of trucks (e.g. adoption of both tractors and trailers) [explained later]

Merging orders
Merging orders (based on the delivery date in term of time)
- Very often orders have not to be delivered on a specific date, but in a wider time
window (i.e. 2-3 days). In this case, a schedule that includes all the orders (i.e. not
only those expiring on that day) should be done.
- When the delivery has to be done on a specific day, those points of delivery close to each other can be
clustered and, by establishing the same delivery date, multi-drop deliveries can be made.

- Multi-pick -> collection of different orders for the same receiver from different warehouses
made in order to saturate big truck when we have not enough volume
- Multi-drop -> travel to different receivers with orders picked from the same warehouse
made in order to have the trucks full, if we have huge volumes in one point, but the orders are not so big

Compared to traditional multi-drop deliveries, the key points are:

- analysis of the orders to group those point of sales with similar delivery frequency and orders that, if
assembled, can saturate the trucks
- collaboration with retailers to have lower truck unloading times and, as a consequence, more points of
delivery in a trip

Multi-client warehouses
Multi-producer/retailer warehouse implies the complete integration
of different producers/retailers with respect to the following
activities:
- warehousing
- distribution

Through these models, benefits and economies of

scale can be achieved thanks to:
- a higher asset rotation
- a reduction of the operative costs
- a common management of the distribution
activities

It should be noted that multi-producers warehouses

have been already adopted, whereas multi-retailer
ones have not been implemented yet.

138
 09. Transportation: general overview

8 CONCLUSIONS

SUPPLEMENTARY READINGS
- Ballou, Transport Fundamentals, Business logistics Management, pp.135-183.
- Coyle, Langley, Gibson, Novack, Bardi, Supply Chain Management: A logistics perspective, pp. 410-459.

- Ballou, Transport Fundamentals, Business logistics Management, pp.135-183.

- Marchet, G., Perego, A., Perotti, S. (2009) "An exploratory study of ICT adoption in the Italian freight
transportation industry", International Journal of Physical Distribution & Logistics Management, Vol. 39
Iss: 9, pp.785 – 812
- Perego, A., Perotti, S., Mangiaracina, R. (2011) "ICT for logistics and freight transportation: a literature
review and research agenda", International Journal of Physical Distribution & Logistics Management, Vol.
41 Iss: 5, pp.457 – 483

CHALLENGE QUESTIONS
- What are the main pro and cons of each transportation mode?
- What is intermodal transportation? What are its main advantages? When piggyback intermodal
transportation becomes interesting with the respect to road only?
- Assume you are a Fast Moving Consumer Goods manufacturer. Which types of transportation modes
would you recommend to consider for distribution in Europe? What would drive your choices?

- What’s the difference between economies of scale and economies of density?

- It is required to discuss the following statement: even if a company outsources the transportation
activities, it can influence the efficiency of such activities.
- Which are the key variables to optimize the road transportation?

139
09. Transportation: general overview

140
 10. Logistics & B2C eCommerce

10. Logistics & B2C eCommerce

0 AGENDA
1. General overview
2. Key features and performances
3. Design of logistics strategies
4. Last mile delivery
5. Multichannel strategies

1 GENERAL OVERVIEW

DEFINITION OF B2C E-COMMERCE

B2C e-commerce refers to:
- sales -> orders should be online, the order generation is online
- of products and services, in the sense that you purchase something that let you to experience a service
(train or concert tickets)
NB: we do not include:
o home banking, because you have access to your own account and you purchase nothing
o online games, it is entertainment, you put some money and then play, you purchase nothing
o downloading of digital contents
- via Internet /online (i.e. shopping cart or web form, no email)
- to end consumers -> not companies, otherwise we address to B2B e-commerce, and it is very different

B2C ECOMMERCE DIFFUSION

Magnitude in absolute value (not volume):
- world -> 3000 bln€ [2019] (+20% [2018])
- Europe = 670 bln€
- USA = 680 bln€
- China = 1300 bln€

In order to better understand the real magnitude, we

should compare the online channel to the physical one,
the traditional one → penetration rate

1.2.1 Importance of the online channel

The incidence of online sales on the overall turnover of a traditional Italian retailer ranges between 0.5% and 15%
(more or less 7%); this value can be even doubled for foreign retailers (e.g. American, English, French, German)
We buy something online, but not so often (considering grocery and family budget)

This percentage depends on:

- Retailer commitment
- Size of the potential market not reached by traditional stores
- Synergies with the offline channel
- …

Today, the online channel represents a huge opportunity for traditional companies (multi-channel strategy)
- It is growing, in every country, and it is still going to grow; from the latest 90s till today, it never stopped
working → it is a very important channel for every company
- It can support omnichannel strategies, a traditional retailer can have a lot of benefits if it is able to
manage the right synergies between traditional and online channel
ex: it gives the opportunity to the consumers to go both in store and online, for example when the
product is not available in the store

141
 10. Logistics & B2C eCommerce

1.2.2 Types of merchants

eCommerce web-site

2 KEY FEATURES AND PERFORMANCES

LOGISTICS AS A CRITICAL SUCCESS FACTOR

What are the most important factors for a B2C e-commerce initiative? Is logistic important?
ex: considering a web site, what are the main factors that allow you to transform visits into orders?

2.1.1 Factors
Service elements:
- Delivery time; in this case, logistics is important for:
o picking and order preparation
o last mile delivery
- Returns policy
- Tracking
- Shipment costs (everybody try to not
have the additional costs for delivery)

Other elements:
- Trust
- Price
- Information (and reviews)
- Product range availability
- User experiences

2.1.2 Logistic impact

Logistics affects all the service elements, and we have of course a trade-off between efficiency and effectiveness
- If you are efficient at the logistics level, you will positively affect the cost of delivery
- If you are effective at the logistics time level, you will positively affect the service level
NB: there are some cases in which they are not in trade-off

ex: considering the last-mile delivery, the final customer is always worried about the missed deliveries
if we miss a delivery, we have negative impact in both effectiveness and efficiency, because the customer is not
happy, since he does not receive the orders, and the transportation cost is higher because if the courier does not
find you at home for the delivery, he has to re-try another time, and this is an additional cost
→ finding something to reduce the missed delivery has a positive impact on both effectiveness and efficiency

142
 10. Logistics & B2C eCommerce

- Trust and reviews→ effectiveness of logistics performance

- Price → efficiency of logistics activities: the more they are efficient, the more they can reduce the price
- Product range availability → having a huge quantity of different products means to split it in many
different parts, and this means to have a very huge overall demand variability
Long tail rule: they do the majority of turnover with those
products they sell in a limited number, but for which the
demand forecasting is very high, because if you miss 1 order of
a product that you sell 4/5 times in a moth, this has a huge
impact on the demand → safety stocks
(The tail is the part below)
[for the traditional retailers we have the Pareto rule, for which
on average the 20% of the products make the 80% of the overall turnover]
- User experience considering the overall experience, not only the one on the website/application

→ an e-commerce company is first of all a logistic company

LOGISTICS PROCESS IN B2C E-COMMERCE

What are the main features of logistics

in B2C e-commerce?

Outbound logistics for B2c eCommerce entails a higher complexity (and costs) compared to the traditional offline
channel due to different reasons:
- small order size: we talk about single piece, or few pieces, not a lot; on average we order 1.x pieces/order,
especially if you have no constraints about shipment costs
NB: this is not true for all the sectors, it is an average; for example, in the grocery sectors we usually order
about 50-60 pieces per order
- the small size of units handled, not pallet loads/boxes, but single pieces. Logistics activities (i.e.
transportation and warehouse activities) are related to single orders
The inventory of an e-commerce company is 90% a picking inventory, where you store and pick up a
single piece, the picking and storage area are quite the same
- very high service level, from many points of view; high expectations of customers in terms of service level
o cycle time, if we overcome the 48h it is a big problem
o punctuality
o completeness
o accuracy
NB: the service level is really pushed at is extreme level
- additional activities: some activities (e.g. last mile delivery, picking and packing) are carried out by the
merchant and not by costumers (as it happens in the traditional process)
ex: in a supermarket, is the consumer which carries out the picking activity, while in B2C e-commerce an
operator does it for us -> we have an additional activity
ex: delivery, in sense that when we go for in-store shopping
- product range: huger offer to be managed

We can understand that the distribution problem is much more complicated than the traditional one

2.2.1 Order profile

Grocery Case
Often a single order has a small size, Mean order value: € 130
since it is composed of a few # lines-per-order: 50
numbers of lines and pieces per line # pieces-per-line: 1.3
(high percentage of orders mono-
line and mono-piece) Consumer Electronics Case
Mean order value: € 240
# lines-per-order: 1.2

143
 10. Logistics & B2C eCommerce

2.2.2 Point of delivery

Generally speaking, Point of delivery
delivery happens in - Home delivery Impact on:
- Optimisation of delivery trips
a location that is not - Delivery at office
- Delivery at specific locations (e.g. pick up - Rationalisation of the stops
used to accomplish
points, post offices, parcel lockers, …) at customers
this task
2.2.3 Customer service
The main factors of the customer service level in B2C e-commerce are:
- Order-cycle time
- Convenience of delivery
- Delivery timeliness
- Delivery accuracy
- Information about the availability of products and / or delivery times
- Availability of prompt information about tracking e tracing
- Quality of post-transaction processes (after-sales services, return management, management of
unsuccessful deliveries)

Convenience of delivery
The convenience of delivery is the ability of the provider to adapt the delivery mode to end customers needs
Some examples are:
- Schedule of the delivery time: in the morning, late in the evening after dinner, on Saturday afternoons, on
Sunday
- Delivery mode: by appointment, attended/ unattended etc.
- Points of delivery: office, home, to the doorman, etc.
- Speed and cost of delivery: use of express couriers to deliver fastest, etc.

Convenience of delivery time

- Depending on the industries, customers can choose the delivery date and/or even the time window:
o a specific day
o a specific delivery time window (agreed with merchant/courier)
o a specific delivery time window of a certain day
- The reduction of the time window increases the costs related to the last mile delivery
- Time windows agreed beforehand leads
to a reduction of the number of
unsuccessful deliveries (due to the
absence of the customers)
Appointment accuracy

Example - Convenience of delivery time schedule – Grocery

Some virtual supermarkets allow costumers to choose the delivery time window. The slots could vary based on
the geographical area

Delivery accuracy
The delivery accuracy is the ability of the provider to be compliant with the conditions agreed with the costumer:
- Consistency between ‘‘ordered’’ and ‘‘delivered’’ products
- Compliance of the unit loads (packaging)
- Compliance of the documents
In the B2C e-commerce, two aspects are very critical:
- Consistency between ‘‘ordered’’ and ‘‘delivered’’ products, due to the very high costs in case of
“nonconformity”
- Compliance of the unit loads (packaging), in order to both resist through the “distribution process’’ and to
be convenient to the costumer
The accuracy of both order transmission and the order entry processes has to be higher than the one in the
traditional sales channel

144
 10. Logistics & B2C eCommerce

Information about product availability

Information provided to potential costumers is related to the product availability and/or the delivery lead times
4 scenarios could be identified:
1. The online catalogue is “static”. The merchant postpones the product availability after the order issuing
and notifies it via e-mail and/or phone. (This option is nowadays very uncommon)
2. The online catalogue shows only available products. The merchant makes the availability check before the
order issuing of the customer
3. The online catalogue (made of both available and not available products) contains real time (or
periodically updated) information about product availability:
o Qualitative information (e.g. status)
o Semi-quantitative information
o Quantitative information (e.g. number of pieces available)
4. Customer issues the order and the merchant checks the product availability before the final order issuing.
Product availability is then notified to the customer. The information provided is:
o Products availability status, or
o Delivery lead time
Examples:
- Eataly - Only ‘‘ready for delivery’’ products (Case B)
- Mango – Product availability updated in real time (Case C)
- ePrice – Qualitative information about product availability (Case C)
- Amazon marketplace- Quantitative information about average delivery time (Case D)

2.2.4 Stock out management

Two different approaches can be pursued to face stock outs:
- “Ex – ante” approach: the objective is to avoid that customers place orders of unavailable products. This
approach entails a careful and continuous monitoring of the product availability.
ex: unavailable products are not displayed in the online store and as a consequence the stock-out
probability is minimised
- “Ex – post” approach: customers can order all the products, no matter of their availability. This approach
requires the management of the stock-outs
ex: list of substitute products that can suit the customer needs

2.2.5 Tracking e tracing

“Tracking and tracing” refer to the capability of the merchant to provide accurate information about both the
order fulfilment and delivery time.
Information can be provided according to two approaches:
- “pull”. Customers request for information about their orders by filling an online form or by calling the call
centre; this service can imply that
o customers interact only with the merchant
o customer requests are forwarded to the courier website (e.g. DHL, UPS, …)
- “push”. The merchant sends information to customers via email or SMS with a certain frequency or when
there are news about the progress of the order fulfilment

Management of unsuccessful deliveries

The merchant is required to quickly solve (and ideally without additional costs) the unsuccessful deliveries.
Reasons of unsuccessful deliveries:
- absence of customers
- mistakes in the delivery address
- courier could not find the delivery address

Preventive actions: Corrective actions:

- careful definition of the delivery features (e.g. - defining clear procedures in case of
point of delivery, delivery time window) unsuccessful delivery (both for customer and
- forcing the customer to provide accurate courier)
delivery details - …

145
 10. Logistics & B2C eCommerce

Management of returns
“Management of returns” refers to the merchant capability to adequately support customers in case of returns.
The merchant has to provide this service to customers:
- quickly
- easily
- without additional costs, if the returns occur for non-conformity reasons
- at a very low cost in all the other cases

Example of customer service issues:

- Maximum period to return products
- Convenience for the customers
- Costs of returns
- Required packaging/labelling for the return
- …

3 DESIGN OF LOGISTICS STRATEGIES

KEY CHOICES
- System configuration
o Inventory ownership, if we want to own the inventories or to go for a more
flexible structure, leaving the inventories to my supplier and picking them
when we need
o Distribution network configuration (# of echelons; # warehouses;
overlapping with the network of the offline channel)
o Outsourcing decisions
- Logistics activities: we can build our economical sustainability on these
o Picking and packing
o Deliveries to the end customers → last mile delivery
o Stock-out management
o After-sales activities (e.g. management of returns, after-sales service, …)

INVENTORY OWNERSHIP: MODELS

Inventory could be located at the merchant or supplier warehouses

We can have
- Merchant without inventory
- Merchant with inventory
In both cases, the merchant has a key role, in the sense that it works for developing the platform of the website
and application, it takes care of all the marketing activities of generating trafic, for providing consumers with a
good customer experience very well structured

146
 10. Logistics & B2C eCommerce

3.2.1 Merchant without inventory

Key features:
- inventories are located at the suppliers’ warehouses
- orders are received by the merchant and then split and forwarded to the different suppliers
- picking activities are performed by suppliers and in most of cases boxes are directly sent to the end
customers
- outbound flows are outsourced to door-to-door service providers (one or more)
ex: this is very common in the consumers electronic industry, in which the suppliers are mainly distributors
→ pure drop shipping

Distribution models
Direct shipment from suppliers to end customers:
Parcel delivery from the point of origin (supplier warehouse) to the point of destination (customer). The order
picking and preparation are carried out by the supplier

Direct shipment from suppliers to end costumers through hubs:

The products coming from different suppliers but belonging to the same orders are consolidated and sorted in
hubs; therefore, the transportation process is split into different parts that are typically outsourced to service
providers
In this variation of the merchant without inventories, we have a transit point between the supplier and the
customer, in which they do the picking and the order preparation, and then there is the real delivery

In both the cases we have

- low inventory costs
- longer delivery time
- no control of the process

3.2.2 Merchant with inventory

Key features:
- inventories are located at the merchant warehouses
- products are delivered using the merchant distribution network (that can have different configurations)
- inbound warehouse flows are typically characterized by a high product quantity per shipping

In this way, we have

- high inventory costs
- shorter delivery
- process under control

→ the trade-off is always between costs (inventory and carrying costs) and service
NB: we do not consider the picking costs since we assume that it is outsourced, it is not done by the merchant but
by the supplier

DISTRIBUTION NETWORK DESIGN

What is the typical network structure of
a Dot Com pure player?
A Dot Com player is a B2C e-commerce
merchant born to sell on the internet,
and it is a pure player in the sense that
its main channel is internet, it does not
have any physical store

European distribution network of a pure player

In each company they have thousands of
suppliers, while the consumers are over millions

147
 10. Logistics & B2C eCommerce

There the distribution problem is very complex, at three different levels:

- Product
o Huge product range
o Obsolescence
- Demand
o Order is very small
o Points of delivery distributed
- Service
o Very short delivery time
o High punctuality

For these reasons, especially for the service level, we have three main problems we need to manage:
- Inventories should not be far from the consumers, otherwise is not possible to deliver the goods in less
than 24-48 hours
- In terms of transportation, the problem is very complex since we transport single parcel units to single
small point of delivery (houses) -> biggest problem, high importance of transportation costs
- Inventory and carrying costs are important, as well as transportation costs
NB: if we want to be close to our consumer, we should have inventory on more levels of my network, but in this
way the carrying costs increase

We have very high value in all the three levels of problem, and we need to manage them all
- High service level → we must have a certain number of warehouses for each country, at
least one for each country (sometimes more than one, according to the country extension)
- High transportation costs → leverage the express couriers, we should understand that it is
impossible to move straight from one single warehouse to the final POD, because of the too high
distances → we need more echelons between the central warehouses and the PODs
→ network of express couriers
- High inventory carrying cost → use of transit points/hubs to split the overall demand

NB: according to the delivery time constraint, we should consider different options
- Local delivery is carried out by express couriers, in order to be efficient in the transportation service level,
since they are very quick in processing (usually with an order cycle time lower than 36h)
- If the delivery time is very short (ex: 2 hours) we cannot go through an express courier, we should have a
warehouse very close to the customers: in this way, we have a network with 2 levels, in which the second
level is composed by warehouses with inventories inside

2-echelons distribution network

- 1st level of warehouses with inventory
- 2nd/3rd levels of Transit Points

DELIVERY NETWORK ALTERNATIVES FOR TRADITIONAL RETAILERS

What are the main alternatives of delivery network for a traditional retailer?
- Network offline business: between the different suppliers and consumers, we can have a central
warehouse which serves the shops, and then the customer goes to the shop
- Shared network (with respect to the offline business): inside the warehouse they dedicate a picking area
to the e-commerce, linked to a transit point of the express couriers; it is a share network since the
warehouse is shared between offline business and e-commerce one
- Dedicated network, which means to have a warehouse dedicated only to B2C e-commerce, linked to the
express couriers to deliver (totally opposite to the previous solution)

Considering these alternatives, the shared network needs less investments, but on the other hand the capacity
and the performances are lower than the dedicated solution

148
 10. Logistics & B2C eCommerce

Shared network Dedicated network

- Less investment - High investment
- Capacity related problems - Dedicated capacity
- “Low” performances - High performances

The transition from shared to dedicated is a matter of critical mass: if the

performance level justifies the high investment in new infrastructures,
we definitely go for the dedicated network, because it means also higher
efficiency, since all is devoted and specialized to B2C e-commerce

The ways a distribution network can be designed are multiple and - Dedicated network
can change over time - Picking Warehouse (PW) in the
Nonetheless, four main systems to design a distribution network can Distribution Centre (DC)
be identified. These solutions, which differ in terms of costs and - PW based network
service, can be either implemented alone and combined together - Points of sale (PoS) based network

Dedicated network
Features - Stock shown on the B2c
- No synergies with traditional channels eCommerce website:
stock in a dedicated
- No conflicts with traditional channels
warehouse
- Investment in dedicated logistics facilities - Stock of online sales:
- Typically one warehouse per country dedicated
- Order allocation: static
Examples
- Industry: consumer electronics
- The company has a warehouse dedicated to online commerce.
- Warehousing is accomplished by employees.
- Home delivery is outsourced to couriers: express couriers are selected for items that are characterised by
limited weight and dimensions; specialised couriers are conversely preferred for managing voluminous or
fragile items or those requiring additional services (e.g. installation, delivery to a specific floor)

PW in the DC
- Stock shown on the
Features
eCommerce website:
- Limited investment in dedicated logistics facilities stock PW
- Synergies in the procurement phase - Stock of online sales:
- Sharing of the inventory in the distribution centre dedicated
- Order allocation: static
- No synergies in the distribution phase
(dynamic if warehouses
are more than 1)
Examples
- Industry: beauty
- The company has an area within the warehouse which is specifically dedicated to stocking and picking of
products of the eCommerce channel
- The warehouse is managed by the same supplier the traditional channel
- Home delivery is carried out by couriers

PW based network
Features - Stock shown on the
eCommerce website:
- Investment in structures specifically devoted to stock PW
online channel - Stock of online sales:
- Synergies in the procurement phase dedicated
- Stock sharing at the warehouse level - Order allocation: static
(dynamic if warehouses
- Possibility to localise the PW close to the market are more than 1)
- Possibility to use ad hoc network (not the courier)

149
 10. Logistics & B2C eCommerce

PoS based network

Features
- Sinergies in the distribution phase that
The point of sale can be
lead to the minimization of the costs to
characterised by different
reach the market areas
«logistics roles» in the
- No investments in dedicated logistics
process of executing
facilities
eCommerce orders
- Possibility to collect and return the items
in the store
- Renewal of the point of sale’s role

Examples - PW and PoS based Network

- Industry: grocery
- The company operates in the Northern and central Italy with a range of more than 10,000 product types
- Online orders can be prepared in local distribution centres, frequently replenished by a central
warehouse
- Delivery is carried out by third parties through customized transportation means (with double
temperature to preserve fresh and frozen products)

3.4.1 Distribution network design: alternatives

PoS based network
The «logistics role» of the point of sale can
be determined based on:
- The delivery mode to the consumer
- The preparation of the order

POS Point of Sales

It is what makes the real difference between
a traditional retailer and e-commerce, there
are a lot of different opportunities to use it

1. Picking and packaging strategy: we have both the picking and the order
preparation in-store
The supplier picks the items of the order and prepares it in the shop, then
the customer comes and takes what he ordered
ex: grocery industry; consumer electronics: you select what you want from
the online channel and then decide the store from which you want to
retrieve it, because they use the inventories they have in-store
1. Full fulfilment strategy: they prepare an order in the shops and then it is
home-delivered (ex: Esselunga)
We can have inventories close to the customer; it is a great advantage in
terms of service
2. Consolidation strategy: consumers take process under control; they can
decide when and where they want to pick what they need
3. Consolidation and delivery strategy: [pretty uncommon]
ex: Ikea, kitchen delivery, the item is built up in a warehouse according to
the customer preferences and then delivered at home; the shop is a point to
visualize the final product, like a transit point

150
 10. Logistics & B2C eCommerce

3.4.2 Picking system configuration

The main choice about picking activity is between manual and automated picking; the right level of automation,
depends a lot on the different business we are working in, according to sectors, and the level of specialization:
higher this level is, lower complex the application of automatic measures is
ex: for Amazon is very difficult to think about a completely automated picking activity, because they have a very
huge range of products, we should need a very flexible automation, which is quite difficult
→ KIVA robots, they automated the travel time, but there is still a manual picking
ex: YNAP is a very specialized player, and for this reason it can use a highly automated system
ex: Ocado, grocery industry, a lot of robots moves over a lot of different boxes with specific items inside; unique
case in the world in which the true picking activity is performed by robots

4 LAST MILE DELIVERY

Last mile delivery is a very complex part of the process, it accounts for about 50% of the overall logistic process
We should consider 3 main elements: Efficiency, Effectiveness and Sustainability
There are many projects in order to improve these characteristics in last mile delivery:
- Parcel locker, way to improve efficiency, since you move different orders to the same point and then is
the customer which goes there to pick his own order, and also the effectiveness, since the customer
never miss the delivery (ex: Amazon locker)
- Crowdsourcing, there is the possibility to be more flexible, you are more able to saturate your trucks
- In trunk delivery, you can receive your order in your trunk, the courier may deliver the pack and open
your trunk of your car; you have no more missed deliveries, increase of density if the delivery point is full
of trunks to be filled
- In-home delivery, the operator may also open your home and you can have a camera to control what the
operator is doing
- Robots, the robot is operating the delivery and it is 100% automated
- Dynamic pricing, prices change in specific times to increase the demand in certain moments of the days
- 1-hour delivery
- Real-time tracking
- Geo-localised delivery
- Green vehicles
- Innovative packaging

5 CHALLENGE QUESTIONS
- What are the main differences in terms of logistics needs between traditional sales channel and online
sales channel?
- What are the main choices to define the logistics strategy of the B2c eCommerce channel?
- Assume you are a retailer in the consumer electronics industry. What are the options in terms of multi-
channel strategies considering both the front- and back- end processes?

151
10. Logistics & B2C eCommerce

152
 11. Distribution Network Design

11. Distribution Network Design

0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different contexts:
- The main elements of the Distribution Network Design Problem: objective function, decision variables,
data & inputs, influencing factors
- The Profile of the Distribution Problem: key drivers affecting the structure of the network
- The Types of Distribution Networks and the corresponding performances
- The key Principles of Distribution Network Design
- A set of real-world examples

AGENDA
1. Distribution system
2. One-echelon distribution networks
3. Two-echelons distribution networks
4. Mixed Distribution Networks
5. Optimization function and decision variables

1 DISTRIBUTION SYSTEM (RECAP)

DEFINITION
Distribution system: all the (inter-company and extra-
company) resources and structures involved in the
transportation and sale of the goods and the associated
services to the end customers

All the resources and structures allowing:

- the physical flow from the manufacturer to the end customer
- the information flow from the manufacturer to the end customer and vice versa
- the reverse logistics of packaging and disused products
- the technical support to the after sales service

All the resources and structures used to transfer the property of goods from the manufacturer to the end
customer. Types of intermediaries:
- Functional intermediaries: branches, one-firm agents, agents working for different firms, brokers, …
- Full intermediaries: wholesalers, distributors, mono and multi-brand dealers, …
- …

1.1.1 Trade channel and logistic channel: inter-relationships

The management and the design of the logistic channel must depend on the management and design of the trade
channel:
- interactions between different design choices: some specific logistic choices are linked to specific trade
choices. i.e.
o configuration of the trade channel (trade channel length, number and typology of intermediaries)
o organization of the trade function (responsibility, division in different Business Units)
- specialization of the logistic channel different logistic channels for different trade channels

153
 11. Distribution Network Design

1.1.2 Logistic channel

The logistic channel includes:
- the distribution network, whose nodes are the plants, the
warehouses (central and regional warehouses, transit
points) and the points of sale
- the transportation systems (connections) which connect
the nodes of the network

1.1.3 Logistic channel: functions

The logistic channel has the following functions:
- consolidation/sorting/transport optimization: collection from upstream of large lots (in terms of both
quantity and time) and preparation of the materials that are required downstream
- product mixing: gathering together all the different product lines which are manufactured in different
plants
- customer service: reduce the order cycle time and increase the order cycle time reliability
- efficient inventory holding: keep efficiently the safety and cycle stocks which are deemed necessary
within the supply chain (to tackle uncertainties and to smooth operations)

[from tutorship session]

Differences
- Distribution system -> we consider both the logistic channel and trade channel
o Logistic channel -> transfer of physical goods
o Trade channel -> transfer of ownership of goods
- Logistic channel it is composed by
o Nodes/infrastructures, which constitute the distribution network
o Connections, which constitute the transportation
- Distribution network is composed by the nodes between the points of origin to the points of
destination

PRINCIPLES OF DISTRIBUTION NETWORK DESIGN

We will address the design of the distribution network
ex: let us consider a consumer electronics operating in Europe market; it has 4 main options to design its logistics
network, and we should find out which one is the best solution
1. Direct delivery
2. 1 central WHS
3. 10 regional WHS
4. 10 regional WHS + 200 transit points -> very expensive in logistic cost, it must be justified
At the end of this topic we will find the right answer

Two main parts of study: For the first topic, we will address 5 main steps:
- Principles of Distribution network design 1. Problem setting & decision framework
-> qualitative evaluation 2. Distribution problem profile
we will be able to describe a distribution problem, to 3. Typologies of distribution network
know the main options we have to solve it 4. Design principles -> main point
- Methodologies & models -> quantitative evaluation 5. Examples -> inductive approach, in order to
understand the point 4

154
 11. Distribution Network Design

1. PROBLEM SETTING AND DECISION FRAMEWORK

A logistic network is composed by nodes + transport connections. After these, there are two kinds of networks:
- the supply/procurement network (or inbound logistics network, meaning in the manufacturing plan)
- the distribution network (or outbound logistics network, meaning from the manufacturing plan to the
final customer)

Product Structure
Supply Raw material Convergent: from many nodes to a
network components fewer number of nodes
Distribution Finished products Divergent: from few numbers of
network POO to many POD

The design of distribution network is more complex than the supply one
NB: the design principles are very similar between distribution network and supply/procurement networks; we
will focus only on the first one, but we will quite be able to address them both

1.3.1 Distribution design problem (in a nutshell)

Having many points of origin, where the flow starts, and other points of
destination, which are the final layer, we should understand which is the best
structure of the network. The points of origin and destination are external
variables, which describe the problem, not the solution, while the nodes and
arc connecting them are the internal structures/layers, they are the internal
variables of the problem

1.3.2 Contextual variables [explained later]

- unity of analysis or the scope of the system (ex: manufacturer, retailer, extended supply chain, industry…)
- sales and marketing strategies + sourcing strategies, since they determine the number and meaning of
the points of origin and destination
- types of product and the unit loads you use to move them within the distribution network
- objective function, how we compare different options in terms of network structure
-> cost, service level, combination of the two

1.3.3 Design decisions

- number and type of nodes
- transport network and modes, so how we connect the nodes
- inventory policy, how many inventories we want to have in the network, and where we want to locate
them (considering the network perspective, not at the node level)

1.3.4 Unit of analysis & Sales and Marketing strategies + Sourcing strategies
It is the main perspective we should have when entering in the problem, and they concur in defining the point of
origin POO (dependent on sourcing strategies) and destination POD (depend on sales and marketing strategies)

Example:
- Network of a manufacturer
o POO = manufacturing plants
o POD = different types -> retailers, wholesaler distributors, other manufacturers, final consumers
ex: if we have a channel directed only to the intermediate costumer, you cannot reach the final

- Network of a retailer
o POO = usually they work directly with manufacturer, and we mainly source our products from the
manufacturer warehouses, so the POD of the manufacturer network
Or, we can source directly from the retailers’ plants
o POD = POS or final consumers; the distribution problem is completely different between these
two PODs

155
 11. Distribution Network Design

1.3.5 Product types/features

- Level of variety <-> product range (items, stock keeping units)
- Density [kg/m3]
- Density of value [€/kg or €/m3] -> it is a more important concept than the value itself
- Obsolescence/perishability -> the risk of a quick value-loss of a product in the network; if we have very
high level of obsolescence/perishability, we should move them quickly

1.3.6 Objective function

We should solve the trade-off between costs and service levels, the problem is a multi-dimensional function
- min (logistics costs) s.t. service level -> we target the service level, and
basing on that, we minimize the logistic costs; in this way, considering
different service levels, we build the costs to serve function
- max (service level) s.t. logistic costs -> we provide a constraint in terms
of logistic costs
- min (logistic costs, including “poor service costs”) -> we include in the logistic costs the cost for providing
a very poor/basic service; they could be real or figurative costs, for ex:
o lost sales (figurative) -> sales lost due to a bad service
o cost of bad quality/accuracy -> if we increase the accuracy level in the system, we are increasing
the service level and decreasing the costs of poor service

1.3.7 The decision framework

1. distribution problem profile -> description of the problem
2. typologies of distribution networks -> description of the
main elements coming from the solutions of the problem
3. design principles -> how to connect the distribution
problem and the distribution network -> most important part

2. DISTRIBUTION PROBLEM PROFILE

What are the variables you would use to describe the problem? It is not important the quantity of the variables
(which could be 10/12), but their quality. The main categories of variables are:
- Points of origin, in terms of type, so in terms of
o Number (10, 100, 1000)
o Location (country, continent)
o Specialization (focused vs general) -> we need to distinguish between very specialized and
general POO, since we have different level of complexity according to this
- Points of destination, in terms of type (as the POO), so in terms of
o Number (10, …, 1.000.000, 10 million)
o Location (geographical distribution)
o Order size -> customer order profile (pieces, cartons, pallet loads, full truck loads)
- Products
o Product range (100, 10.000, 1.000.000 SKUs)
o Value density (1 €/kg, …, 100, …, 1000 €/kg)
o Obsolescence risk <-> market life cycle of the product
- Service level -> we should understand what our customer expects from us
o Order cycle time (weeks, days, hours)
o Punctuality
o Accuracy
o …

We will describe our distribution problem

considering these four areas

156
 11. Distribution Network Design

3. TYPOLOGIES OF DISTRIBUTION NETWORK

Distribution networks can be classified on the basis of three main variables:
- number of echelons (or number of levels) of nodes → length of the network
o direct shipment -> 0-echelon, no nodes between POO and POD
NB: we still have inventories, but they are located at the POO and
not in the distribution network
o 1-echelon -> 1 level of nodes that stands between POO and POD
o 2-echelons
o 3-echelons
o …
NB: each time we add an echelon, we are adding warehousing costs, so
they need to be justified → we will describe the different values of different echelons
- number of warehouses/nodes in each echelon → width of the network
o narrow distribution networks (1, 2, … , a few)
o wide distribution networks (many)
NB: they are very different in terms of cost but also performances
- type of warehouses in each echelon; they have very different functions
o warehouses with inventories -> storage + flow management
o transit points without inventories -> flow management
In the transit point we do not have inventories

These three elements, combined together, creates a lot of different possibilities

ex: a 2-echelons network of warehouses is very different from a 1st-echelon of warehouses and a 2nd-echelon of
transit points, because at the second level we do not have any inventories

1.5.1 Direct shipment

4. DESIGN PRINCIPLES & 5. EXAMPLES

We will address the different types of network according to these 4 main steps
1. Question formulation -> what is the role of this level of the network?
2. Conceptual answers -> what are the possible functions of this level?
3. Empirical evidence -> real example, describing both
o Distribution problem profile of some companies
o Distribution network selected by the companies
and we will analyse if they are consistent considering the conceptual answers
4. Design of principles statement

We will address the 1-echelon and the 2-echelon of a network, comparing them with the previous level
- 1-echelon compared to the baseline, so with the direct shipment L1 vs L0
- 2-echelon compared to the 1-echelon L2 vs L1

157
 11. Distribution Network Design

2 1-ECHELON DISTRIBUTION NETWORK

ROLE OF CENTRAL WAREHOUSES IN A 1-ECHELON NETWORK

With respect to a direct delivery distribution network, the 1-echelon network (more generally, the first tier of
warehouses in a distribution network) provides the following functions:
- reduction of the order cycle time (warehouses are nearer the market than the plants/suppliers)
considering this function, 1-echelon network is useful in the case in which
o the POO are far from the market
o the target OCT is stringent \left[OCT_0>OCT_1\rightarrow1\div2\ months>3\div5\ days\right]
- product mixing, if suppliers only focus on a small part of the product range
-> increase in order completeness; it is important when:
o we have many POO that are very specialized, so it is important to have at least one node to reach
the order completeness
o when the completeness is an important service level KPI
- optimization of transports from plants to delivery points (-> reduction of transportation costs), thanks to
o the reduction in the number of connections/arcs -> we reduce ≈ 10 times
o the ensuing increase in trucks utilization/saturation -> we increase ≈ 10 times
NB: the other driver of transportation costs is distance, but in this case, we have an increase in
distances, which is an increase in the costs, triangulation is longer than direct shipment
ex: considering M=10 POO and N=1000 POD, we have
o M*N connections for L0 = 10.000 connections
o M+N connections for L1 = 1010 connections
It is important when
o we have many connections, so many POO and POD
o we have a high incidence in transportation costs, which depends mainly to product value density
and order size
- [centralization of safety stocks]

158
 11. Distribution Network Design

EMPIRICAL EVIDENCE - EXAMPLES

2.2.1 Example: Bassetti
Product range: Quilts, Bed linen, sheets pillows and pillow cover

European distribution network

- Company: Bassetti S.p.A.
- Industry: textile-apparel
- Product: household linen
- 1-echelon distribution network (high degree of centralization: 1 Central
Warehouse for all Europe)

2.2.2 Example: 3M Europe

Product range
Road signals, tapes, filters, abrasives, stationery, stethoscopes, electromechanical
systems

Distribution network
- Company: 3M Europe
- Business Units: Office, Manufacturing and Industry, Safety security and
protection, Displays and graphics, Health & care…
- Products: stationary, abrasives, stickers, electromechanical systems,
display parts, reflectors, stethoscopes… (50,000 SKUs)
- 1-echelon distribution network (5 Central Warehouses)

[Star: Central warehouses: 5]

The areas served by each warehouse are marked with different colours

2.2.3 Example: Unilever Home & Personal Care

Europe Product range
Tooth paste, soap bars, detergents, shampoos and liquid soaps,
deodorants

European distribution network

- Company: Unilever
- Home and Personal Care
- Industry: health and personal care
- Products: detergents and personal care
- 1-echelon distribution network (12 central warehouses)
*: goods pass through transit points only for deliveries to the islands

[star: Central Warehouses: one for each main European Country]

There are 2 Central Warehouses in Italy, France and UK

159
 11. Distribution Network Design

EXAMPLES SUMMARY
Assumption: considering the European distribution problem, we are considering the magnitude of variables, not
the exact value
Bassetti 3M Unilever
Number 10 man. plants 15 man. plants 30 man. plants
POO

Location Italy + Eastern EU Europe + other continents (USA, Asia) Europe

Focus Very focused (*) Very focused (***) General (*****)
Distribution Problem

Range width Very wide, 10.000-20.000 SKUs (**) Very wide, 100.000 SKUs Narrow, 500 SKUs
Prod

Value density Medium-high, 50-100 €/kg Very high, > 50/100 €/kg Very low, 1-10 €/kg
Obs. risk Medium-high Medium-high (****) Very low

Number 10.000 shops/distributors 5000-10.000 distributors/retailers 3000-5000 retailers

POD

Location Europe Europe Europe

Order size Cartons/Pallet loads Pallet loads Full truck loads (or 1/2 and 1/3)
Service

Order cycle Not so stringent, 1-2 weeks

Not so stringent, 4-6 weeks Very short, 2-3 days
time [It depends on the type of POD]
Solution

Distribution
L1 network, with one warehouse L1 network, with 5 warehouses L1 network, with 12 warehouses
Network

(*) each plant is dedicated to 1/10 to the overall product range

(**) we should consider that for each season we have different kind of products, and each product can have different sizes
and colours
(***) they are very focused since they have a very big product range
(****) it depends on the type of product
(*****) the product range (considering only the home & personal care products) is not so huge, so each plant can produce at
least ½ of the overall range

Conclusions:
We have 3 different cases in which the companies chose the L1 network, but they have different distribution
problem profiles, so they have different number of warehouses in the middle level

Common points
We should understand why they selected this solution; for each of these 3 examples, we have at least one of the
three main potential functions that is very important

Bassetti 3M Unilever
Important
Very important
↓ OCT Not important We have production plants very far
The OCT is very stringent
from the market
↑ Order Very important
Not important
Completeness Huge product range and very focused man. plants
Very important
Very important
We have a lot of POD compared to
↓ Transport Incidence of transportation costs is
Not important the POO, so we should decrease the
Costs very high since the value density is
number of connections to increase
very low
the saturation

DESIGN PRINCIPLES
We can find the first main design principles considering the 3 main potential functions of L1 network, which are
- Reducing the order cycle time -> related to service level optimization
- Increasing order completeness -> related to service level optimization
- Reducing transportation costs -> related to cost optimization

160
 11. Distribution Network Design

0. Directly delivery is a baseline solution when none of the 3 potential functions of L1 network are very important
-> baseline solution is not typical at all, is very rare compared to other solutions, because in order to make sense
we need to have in place a lot of assumptions

1.1 If at least one of these functions is very important, the L1 network should be considered as the baseline option
[NB: we are not saying that we do not consider the other options, but this is the baseline]

1.2 Nodes at L1 should be WHS if reducing OCT is an important function of L1, otherwise consider TPs
We can fulfil a reduction of transportation costs both with transit points and warehouses, and the same is valid
for the increase of order completeness, while the reduction of the order cycle time can be achieved only through
warehouses

1.3 The number of nodes at L1 depends on a trade-off between inventory-related costs and transportation costs
subject to service level constraints [explained later]

2.4.1 Number of central warehouses

The number of central warehouses depends on:
- the trade-off between transportation and inventory costs.
The higher the relevance of the transportation cost with
respect to the inventory cost, the higher the number of
central warehouses
- the order-delivery cycle time customers accept. The shorter
the delivery time, the higher the number of warehouses

Trade-off between transportation and inventory costs

Increasing the #WHS, we are
- increasing the inventory carrying costs because
o We are increasing the amount of inventories, considering
SS (and also CS if we want to reduce the frequency of
replenishments)
o We are increasing the unitary warehousing cost because
of the reduction of the economies of scale
- Increasing the handling costs, because of the reduction in
economies of scale
- Decreasing the transportation costs, because we have more
opportunities to optimize the flow to reduce the distances; this is
true till a certain point, because if we fragment too much the
transportation flow the costs will for sure increase

The combination of these patters creates the cost curve, and its minimum corresponds to the optimal number of
warehouses we need, in terms of overall distribution cost reduction
NB: it is not always at the intersection of the two curves

If we want to optimize the ICC and HC and not the TC, we can have a number of warehouses smaller than the
optimal one (usually when we have high value density, huge product range and/or high obsolescence risk);
otherwise, if we want to optimize the TC, we will have a greater number (usually when we have low value density,
or high number of POO)

If we have high incidence of inventory-related costs, we should go for

solutions that would reduce them, even if the transportation costs, so we
will have few quantity of #WHS/L1; otherwise we will have many WHS in
order to increase the transportation costs
ex: considering our examples, Bassetti and 3M need a narrow level, so
they have just 1 warehouse at the L1, while Unilever needs a wider
echelon

161
 11. Distribution Network Design

Focus on the service level -> OCT

Increasing the #WHS, we are decreasing the OCT, so increasing the
service level, we can go from 10 days with 1 warehouse to 2-3 days with
20 warehouses
If we have a target OCT, we cannot consider all the solutions that do not
allow us to reach it, it is unacceptable from a market point of view

Objective function
The objective function is a multi-objective function, which aims at both minimizing the distribution costs and
maximise the service level; we can simplify this complexity minimizing the distribution costs with a constraint
based on a target service level

ex: we cannot accept a service level with OCT higher than 5 days; it means that we cannot consider all the
solutions with OCT higher than the target one, even if our optimal solution considering the minimization of the
distribution costs is in this range

In this case the minimum is in the acceptable part of the curve, so it is still valid; otherwise, we would go for the
first solution we find on the acceptable curve, which will not be the not-constraint minimum
In this latter case, the two objectives are in trade-off

ex: for Bassetti and UniLever the previous

solution based on the costs is still good also
consider the service level, while for 3M we
should have another kind of curve

In this case, we will introduce the constraint of

OCT for 3M, considering only a portion of the
initial curve of 3M: in this way, the target OCT
excludes the real minimum of costs, so we will
go for the first viable minimum

162
 11. Distribution Network Design

3 TWO-ECHELONS DISTRIBUTION NETWORKS

We will address only the additional benefits the L2 compared

to L1, not to L0

ROLE OF REGIONAL WAREHOUSES (2ND TIER OF THE DISTRIBUTION NETWORK)

The 2nd tier of warehouses in a 2-echelon distribution network provides the following supplementary functions:
- higher service level in terms of both cycle time and punctuality (thanks to the inventory in the regional
warehouses) -> high reduction of OCT, because we have inventories even closer to the customer
It is important when we have a very short OCT (compared to the one we can fulfil with L1), we mean
hours compared to days
- optimization of transport to the end customers (local distribution)
-> reduction of secondary transportation costs, which are the costs to connect the L1 to the POD/market
[primary transportation costs are between POO and L1]
It is important when we have many points of delivery ordering small quantities very frequently, so the
company is receiving many small orders [we will understand the magnitude of “many”]

How we can reduce the secondary

transportation costs?
Considering just L1, we start from L1, go to the
market, perform a delivery tour and then go
back to the L1

Adding a L2, the transport between L1 and L2 is just for

replenishment, with FPL trucks; then we are closer to the market,
and we can increase the number of POD during a delivery tour
since we have “more time” compared to the L1 solution -> in this
way, the unitary cost of transportation decreases

The delivery tour is the most important factor when we have many
POD, since we would like to serve the highest quantity

We are adding a replenishment activity, but it is less expensive in terms of value density [€/kg] since we have full
truck loads, we have a higher saturation, an then between the L2 and the POD (secondary transportation) we can
use smaller trucks in order to have higher saturation of them

163
 11. Distribution Network Design

EMPIRICAL EVIDENCE - EXAMPLES

3.2.1 Example: Ferrero Italia
Product range
Praline, Bakeries, Nutella, Tea drinks, Snacks. Chocolate bars

Distribution network
- Company: Ferrero
- Industry: food
- Products: bakeries, creams,
pralines, snack, chocolate, drinks…
- 2-echelon distribution networks

3.2.2 Example: Iveco Europe Distribution

network of spare parts
- Company: IVECO
- Industry: automotive – truck and
commercial
- Products: original spare parts (300,000
SKUs)
- 2-echelon distribution network

Star: central warehouse

Triangle: regional warehouse

3.2.3 Example: Nestlé Italia (bakeries)

Product range
Panettone + bottle, one of the many different kinds of panettone

Distribution network
- Company: Nestlé Italia Plant
- Industry: food
- Products: bakeries
- 2-echelon distribution network

164
 11. Distribution Network Design

EXAMPLES SUMMARY
Ferrero Iveco Artsana
Range width Not huge but still high, 1000 SKUs Very high, 300.000 SKUs High, 5000 SKUs
Prod

Value density > 10-50 €/kg Very high, 100-1000 €/kg Medium-high, 10-100 €/kg
Distribution Problem

Obs. risk High (on average) Very high (***) Medium-high

Number > 100.000 30.000-50.000 30.000-40.000
POD

Location Italy (*) Europe Italy

Order size Very small, cartons - pallet Very very small, pieces Small, cartons-pallet
Service

Order cycle
Very short, 1-2 days (**) Very short, hours – 1 day 3-5 days
time

L2, 1 WHS (L1) + 4 WHS (L2)

Solution

Distribution
L2, 1 WHS (L1) + 15 WHS (L2) Very narrow compared to Ferrero L2, 1 WHS (L1), 15-20 TP (L2)
Network
(****)

NB: we do not consider the POO because they are not important at this level of the solution, we should consider
them only for the 1-echelon problem

(*) we are considering only Italy, but a similar complexity is applicable for many different countries; the company
strategy is to keep a direct contact with the customers, without passing through distributors, since they want to
keep control of the freshness of the product, the final market and the order size
(**) since we have a very short shelf-life, so they have to pass through the network very quickly
(***) very difficult forecasting products, most difficult forecasting problem, we should forecast for 10 years
(****) we compute a European network for Ferrero duplicating the Italian one for the other countries, in order to
compare it with Iveco

Ferrero Iveco Artsana

Very important Not important
Very important
↓↓ OCT We are talking about hours OCT not so short, we are talking
It is just 1-2 days
considering Europe about days (**)
Very important
A lot of small Very important (*) Very important
↓ Secondary
deliveries and Not so many POD, but they Lot of customer ordering small
transport Costs
they order very order very frequently quantities very frequently
frequently
(**) This is the reason why we can use TP instead of warehouses

If we need both a consistent reduction of OCT and a reduction of secondary transportation costs, we must go for
a second level of warehouses, otherwise, if we only need a reduction in secondary transportation costs, we can go
for just transit points at the second level

NB: the number of nodes at the second level is very linked to the optimization of secondary transportation costs,
but when you have a very complex problem in terms of products you cannot spread your inventories, but you
should try to concentrate them, otherwise the inventory costs would explode -> trade-off
ex: Iveco has only 4 warehouses at the second level because they use very costly and fast transportation means,
and they want a very narrow transportation network

DESIGN PRINCIPLES
2. Level 2 of the network has 2 main functions:
- Reducing the order cycle time (in addition to L1 reduction)
- Reducing secondary/last mile transportation costs

2.1 If both functions are very important, then consider L2 with WHS
2.2 If it is important to reduce secondary transportation costs but the OCT is not a constraint, then consider L2
with TPs

165
 11. Distribution Network Design

4 1-ECHELON + TRANSIT POINT DISTRIBUTION NETWORK

ROLE OF TRANSIT POINTS

It aims to optimize transports to the end customers (local distribution)… … by accepting a longer cycle time (+ 1 or
2 days on average) than the 1echelon distribution network
Transit points adopt at least two different operative models:
- only sorting of Transport Units which have been completely picked upstream
- receipt of a massive delivery (= aggregation of different customer orders), picking and consolidation of
Transport Units for each customer

EXAMPLES
4.2.1 Example: Chicco Artsana Italia
Product range

Distribution network
- Company: Chicco Artsana
- Industries: baby products, health
products, cosmetics and beauty products
- 1-echelon + transit point distribution
network

4.2.2 Example: Nestlé Italia (dry food)

Product range
Coffee substitutes, Stock powder and sauces

Distribution network
- Company: Nestlé Italia
- Industry: grocery
- Products: pasta and packaged food
- Products: pasta and point distribution
network

4.2.3 Example: Unilever Italia - Food

Product range
Distribution network
- Company: Unilever Bestfood
- Industry: grocery
- Products: food (oil, tea, mayo,
margarine, …)
- 1-echelon + transit point distribution
network

166
 11. Distribution Network Design

5 MIXED DISTRIBUTION NETWORKS (UP TO 2 ECHELONS)

Mixed networks are networks with
a promiscuous use, used for
different purposes

ex: L2 network is also used as L1

network or direct delivery

A long network can be always used

in a shorter and narrowed network

It makes sense when we have many

different distribution sub-problems
at the same time

“THE RATIONAL” FOR MIXED NETWORKS

Mixed networks permit to tailor the management of goods flow according to the:
- order/delivery profiles
- service level
- spatial-temporal distribution of the demand
- ….
In other words, they give the flexibility that allows the management of more than
one distribution problem at the same time

5.1.1 Example - Barilla Italy

- few manufacturing plants
- 1-echelon with 6 warehouses
- 2-echelon with 15-20 transit points
- Final market, which is composed by two very different types of POD

Barilla is facing two different problems in terms of distribution, and they need two different solutions in term of
distribution network to be fulfilled
- When they have to serve big retailers, they should have more than one warehouse per country since the
OCT is very short, but not so many because the order size is big enough, you do not need a second level to
reduce transportation costs
Sometimes they even can go for a direct shipment, when the OCT is not so stringent
- When they have to serve HORECA channel, we need a second level of the network because the order size
is very small, so they need to reduce the transportation costs increasing the saturation of the means, but
they do not need warehouses because the OCT is not stringent, so we will have a 2nd level of transit points

Big retailers HO.RE.CA.

Number 500-1000 5000-10.000
profile
Order

POD
Order size Big, FTL, 1/2 FTL, 1/3 FTL Smaller, Pallets
Service level OCT 2-3 days 1 week
Sol. Distribution network L1, 4-6 WHS L2, 4-6 WHS (L1) + 15-20 TPs (L2)

This kind of solution is very diffused, because it is very rare that a single company faces a single distribution
problem, but it is a combination of different requirements; for this reason, they put together the problems and
come out with a single solution usable at the same time for all the different problems

167
 11. Distribution Network Design

5.1.2 Example: Nestlé Purina Pet Care Italia

Product range
Pet food (cans or packet, Pet dry food (boxes)

Distribution network
- Company: Nestlé
- Italia
- Industry: grocery
- Products: food and products for pet care
- 1-echelon + transit point mixed distribution
network

5.1.3 Example: Cola Italia

Product range
Softdrink (0.5-2l PET bottles), Softdrink (0.2-1.5 glass
bottles), Softdrink (cans)

Distribution network
- Company: Coca Cola Bevande Italia
- Industry: Grocery
- Products: beverage
- 1-echelon mixed distribution network

168
 11. Distribution Network Design

6 IDEAL BASE LINE OPTION NETWORK FOR A COMPANY

STEPS
1. Distribution Problem(s) Profile -> we should 2. Ideal Network structure
consider that there can be different distribution o L1? No, Yes with TPs, Yes with WHS
problems Width
o POO o L2? No, Yes with TPs, Yes with WHS
o Prod Width
o POD o …
o Service o Mixed network?

EXAMPLE – CONSUMER ELECTRONICS PRODUCER (SONY, SAMSUNG, …)

[example considered at the beginning of the chapter, pg. 154]
We will consider some assumptions because we do not know the exact values, but the important is to understand
the magnitude of the different variables

6.2.1 Distribution Problem(s) Profile

- POO 10 manufacturing plants (for parts)
EU + Asia
focused on some products families
- Prod high, 1000-2000 SKUs
high value density, 50-100 €/kg
high risk of obsolescence
- POD 10.000-20.000 overall; the number depends on the channel
o retailers (both general and specialized)
o shops
o distributors (B2B channel)
Europe
Small order size, cartons-pallets
- Service OCT = 2 days – 1 week, depending on the type of customer, from big retailer that are highly
demanding, to the others that are less demanding

169
 11. Distribution Network Design

6.2.2 Ideal Network structure

Level 1: it is necessary, or we can go for a direct shipment? Considering the 3 main functions of L1
- ↓ OCT -> very important
- ↑ Order Completeness -> important
- ↓ Transport Costs -> important (inventory transportation costs could be higher than
transportation costs)
All the three functions are important, one is fundamental -> we should go at least for L1, with WHS

The width depends on the trade-off between transportation and inventory related costs
In this case, inventory related costs are more important, so we should go for 1-2
WHS, but we are subject to OCT constraint, since it is very short (ex: as 3M) -> we
should go for more than 2 WHS, since in order to satisfy the target OCT we need at
least one WHS per country (number closer to 10 than to 1)
→ minimization of the distribution costs constrained by the OCT

Level 2: considering the 2 main functions of L2

- ↓↓ OCT -> NO, we do not need an additional level to reduce more the OCT
- ↓ Secondary transport costs -> it depends, since we have two kinds of customers
o Big retailers -> big orders not so frequently -> NO
o Shops/distributors -> small orders more frequently -> L2 is very important, with TPs

➔ Ideal Network: L2, 10 WHS (L1) + 200 TSPs (L2) used as a mixed network
o L1 for big customers or in general for big orders
o L2 for small customers or in general for small orders
[200 TPs since we consider 1 node for 100 km, and we should count both going in and going back]
NB: this is just the baseline, it is a qualitative evaluation, not a quantitative one! But it is fundamental in order to
have a starting point

ex: in order to reduce OCT, we can have two different options, that are very similar in terms of impact on OCT:
a wide L1 or a narrow L1 and a wide L2

We could compare these 2 types of network in the second

phase; we should remember that going for the second option
we are increasing the impact on costs, since the transportation
costs are increasing

7 3-ECHELON DISTRIBUTION NETWORK

170
 11. Distribution Network Design

7.1.1 Example: Galbani Italia

Product range

Distribution network
- Company: Galbani
- Industry: grocery, fresh products
- Products: diary products and salami
- 3-echelon distribution network (2 echelons made of
warehouses + 1 echelon made of transit points)

8 OPTIMIZATION FUNCTION AND DECISION VARIABLES

PROBLEM DEFINITION
Objective: identification of the network structure and the
management policies to achieve the service level objectives
with the minimum overall distribution costs

8.1.1 Distribution costs

Transportation costs:
- Primary transportation (from suppliers to the central warehouses and from the central warehouses to the
regional warehouses/transit points)
- Secondary Transportation/Local distribution (usually from the distribution network nodes to the end
customers)
Inventory costs (they include storage, capital and depreciation costs)
- Cycle stocks
- Safety stocks
- In-transit stocks
Handling costs, due to the handling activities in the warehouses and the transit points of the distribution network
(loading/unloading, picking, etc.)
Other costs (e.g. order management, packaging, reverse logistics, …)

The cost categories should consider how these costs vary with the main features of the distribution network.

INVENTORIES IN THE DISTRIBUTION SYSTEM

- Cycle stock: these inventories deal with the different operative rhythm of two following stages in the
supply chain
- Safety stock: these inventories deal with the uncertainty of both the demand and the replenishment lead
times
- In transit stock: these inventories deal with the uncertainty of both the demand and the replenishment
lead times)

171
 11. Distribution Network Design

8.2.1 Distribution costs

Cost type Main drivers
Primary transportation FLOW (number of shipments and shipment size = vehicle load utilization factor),
distance
Secondary FLOW (number of tours), tour average distance
Transportation
Safety stocks Demand and lead time uncertainties, average lead time
Cycle Stocks Re-order frequency between nodes along the distribution network
In-transit Stocks FLOW and total in-transit time
Handling FLOW (number of pallet loads, number of line items to pick, …)

8.2.2 Customer service

Order cycle time Delivery punctuality Delivery completeness
Time elapsed between the order issue To deliver the order in the All the goods included in the order are
and the delivery of goods time window (ΔT) that has delivered in a single delivery
e.g.: 98% of the order lines deliveredbeen arranged with the e.g.: 95% of the deliveries is complete
in 24 hours customer
e.g.: 90% of the delivery is Delivery flexibility
Delivery frequency within the 2-hour time To satisfy urgent/unexpected needs of
Number of deliveries scheduled in the window arranged with the the customer
reference time window (day, week, customer e.g.: customer can change the ordered
…) e.g. daily quantity within two days after the order
issue

OPTIMISATION FUNCTION: HOW TO MERGE COST AND SERVICE OBJECTIVES

There are three main approaches:
- Identification of the network structure capable to achieve the service level objectives with the minimum
overall cost (cost optimization)
- Identification of the network structure capable to get the maximum service level with a determined
overall cost (service optimization)
- Identification of the network structure capable to get the maximum profit (profit optimization)
NB: the third is both the most comprehensive and the most difficult to implement.

8.3.1 The profit optimization approach

(*) Revenues minus other non-logistic variable costs

8.3.2 The cost of poor service and the total cost

optimization
The cost of poor service represents the marginal impact of
service on gross margin

172
 11. Distribution Network Design

MAIN DECISION VARIABLES

CONSTRAINTS
It is rather unusual that the distribution network design process starts from “green field”. More often the design
consists of a re-configuration of the network due to threats/opportunities in the external environment and/or
within the company, concerning for instance the following:
- Market requests (new service levels, new markets, …)
- Points of sale needs
- Product/packaging features
- External conditions (Trade and custom barriers, Transportation infrastructures, Evolutions of the logistic
services, New Technologies, …)
N.B.: There are constraints which bound the degrees of freedom (in terms of number, type and scope of the
decision variables)

9 RECAP OF DISTRIBUTION NETWORK

We addressed two main topics

Different types/alternatives of Distribution Network Distribution Problem
- Direct shipment - Supply side
- 1-echelon distribution network o Points of origin
- 2-echelon DN with TPs that aim at a reduction of o Products
transportation costs - Market side
- 2-echelon DN with WHS that aim at both o Point of destinations
reduction of transportation costs and reduction o Service level
of order cycle time

Factors describing the distribution network

- # echelons → number of infrastructures through which the goods have to pass through
- Types of nodes, according to the presence of inventories
o Warehouses with inventories
o Transit point
- # nodes per echelon → it tells something about the level of centralization of the network

At the end we addressed the principles that guides the designer in define the distribution network

173
 11. Distribution Network Design

10 AMAZON CASE
We would like to apply all the topics we addressed to the Distribution Network of Amazon, considered as the
most important company operating as B2C e-commerce in Europe

DESCRIBE THE DISTRIBUTION PROBLEM

10.1.1 POO
- Number = hundreds of thousands, k * 100.000, considering both Amazon and third parties, because
Amazon gives them all the logistics services
- Locations = everywhere in Europe
- Focus = medium, the same product can be found from very different suppliers

10.1.2 Products
- Number = 170 mln (overall) / 20.000 items (prime now)
- Value density = medium-high (considering the average order value)
- Density = medium
- Obsolescence = medium-high, considering that many items are electronics

10.1.3 POD
- Number = n * 100 mln
- Location = everywhere, each house can be considered as POD
- Order size = 1.x, very low, especially for prime member, because there are not shipment additional costs
- Frequency = a few per month
- Seasonality = high, there are lot of events which increase seasonality

10.1.4 Service
- OCT = 24 hours (in general) / 1 hour (prime now, for a small selection of products)
- Punctuality = very high, true for all the cases
- Completeness = high

Considering these elements, we can see that we have at least 2 different distribution problems: one of the
common products, one for the prime now service; with the latter, we should consider more variables

174
 11. Distribution Network Design

In general, this distribution problem looks very complex, in terms of supply, because we have a lot of products
and POO, but it is much more complex looking at the market level, because we have the highest number of POD
and expected service level
NB: in this case, the Pareto rule does not make sense anymore, the long tale rule comes first. Amazon makes
return on the long tale, so on the product with low volume but highly requested; in this sense, is very difficult to
make the forecasting

DESIGN THE DISTRIBUTION NETWORK

10.2.1 Do I need a 1-echelon?
We should look at the 3 main features of L1
- TC reduction -> very important
- SL improvement -> very important
- Completeness -> important
-> we of course have to address all these three, so we should have a L1

10.2.2 How many Warehouses do I need at L1?

The most important inventory costs are both ICC and TC, since we have high
complexity for both transportation and logistics management, so we are in the
middle of our curves
We have an OCT constraint, and it is very short, so we should have high number
of warehouses, at least 2-3 WHS per country, we need high capillarity at the
first level

10.2.3 Do I need a 2-echelon?

- TC reduction -> very important
- SL improvement -> important (only for prime now products)
If we have to deliver in 24 hours and we have a high number of warehouses at the L1 which are able to
satisfy this OCT, we should have other warehouses just for items from prime now, since they have an OCT
of 1 hour
In general, we have TP at the 2-echelon, usually made by express couriers, they can have more than 1 level of TP,
in order to have even higher capillarity; it depends on the orders density; then in some cases they Amazon has its
own TPs which go directly to the POD, and for prime now products Amazon owns warehouses

175
 11. Distribution Network Design

HOW WE CAN DEFINE THE PROBLEM

Network structure Management policies
- # echelons - Location of the infrastructures - inventory alternatives
- Type of nodes - Allocation of the flow - transportation
- # nodes

We should take into account two different objectives, and we can address them in 3 different ways:
- minimization of the costs, with fixed service level
- maximization of the service level, with fixed costs
- maximization of the profit

The best approach is the third one, which considers the maximisation
of the 𝑝𝑟𝑜𝑓𝑖𝑡 = 𝑟𝑒𝑣𝑒𝑛𝑢𝑒𝑠 (𝑟𝑒𝑙𝑎𝑡𝑒𝑑 𝑡𝑜 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑙𝑒𝑣𝑒𝑙) − 𝑐𝑜𝑠𝑡𝑠
It is a very theoretical approach since it is very difficult to draw these
curves, because the relation between service level and revenues is not
always so clear

Our problem can be defined considering the first approach, because

we already included the service level addressing the distribution network problem

→ the network structure can be defined having a fixed value of the service level that we want to achieve and
minimizing the costs considering this constraint

10.3.1 What are the main cost items we have to consider?

We have three main parts of the distribution costs, but we should split them into different items/sections
because they are impacted by different factors; according to this, the same decision we take on the structure of
the network can have different impacts on the sub-sections of the main costs, so the overall impact can be
understood only if we split it
ex: if we decide for example to increase the number of warehouses on a specific echelon, the in-transit stock may
be not impacted, while the safety stock may increase -> different overall impact on ICC

- Transportation costs
o Primary distribution -> FTL, mainly affected by saturation
o Local distribution -> affected by distance to delivery areas and density of POD in the area
- Inventory carrying costs
o Cycle stocks -> depends on the order frequency and the demand
o Safety stocks -> unpredictability of demand and LT, and the combination of the two
o In transit stocks -> flow (demand) and LT of transportation
- Handling costs -> we do not split it since it is impacted only by the unitary handling cost

[from tutorship session] ICC computation

We have seen two different ways to compute the ICC
- Based on the AIL
- As the sum of CS + SS + ITS

In general, we have to split the CS and SS from the TS, because

considering the percentage of holding costs, computing the ITS we
do not have to consider the storage part: the goods we are
considering are not stored in the warehouse as the CS and SS, but
they are on the means

For this reason, the formula based on the AIL is related only to the
CS and SS, because in this case the %hc is considering all the
components
𝐼𝐶𝐶 = 𝐴𝐼 ∗ %ℎ. 𝑐.∗ 𝑉 = (𝑆𝑆 + 𝐶𝑆) ∗ 𝑢𝑖𝑐𝑐

176
 11. Distribution Network Design

10.3.2 What are the main performances of customer service?

They are all related to the structure of the distribution network: as a general principle, the
- Order cycle time
more decentralized the network is, the more the inventories are closer to the customer,
- Punctuality
the more echelons we have and the more warehouses we have for each echelon… the
- Frequency
higher the service level is, because all these factors increase
- Completeness
- Flexibility
The approach we decided to keep is to maintain fixed all these factors, so the service level,
- … and to minimize the costs

10.3.3 How is the methodology structured?

After having seen the approaches, we will address the concrete methodologies
We have a lot of different alternatives: starting from just the number of POO and POD we can have a lot of
different combinations in terms of number of echelons and nodes per echelon, and we can also mix up the
different common structures; if we also add the definition of the exact positions, the problem becomes even
more complicated, quite impossible to be resolved at the first time, because the number of different alternatives
is very wide

The main problem is that usually we would like to start immediately from the mathematical models (ex: mixed
integer linear programming), but we have too many alternatives, and the problem will appear too complex
It becomes impossible because we have to consider all the nodes, and in doing so we should collect a huge
amount of data

We should start before from a qualitative evaluation, in order

to have a narrow number of alternatives, and then we can go
for the quantitative evaluation
-> we are going through two different funnels
1. Strategic DND → qualitative models
considering principles, network solution matrixes,
qualitative cost analysis
2. Operative DND → mathematical models
considering centre of gravity model, linear
programming, mixed integer linear programming

We will be able to set and define the problem in order to select

the best alternative → the rational remains always the same,
through both the qualitative and the quantitative evaluation

11 CONCLUSION

SUPPLEMENTARY READINGS
- Ballou, The Network planning process, Business logistics
- Management, pp.542-603;
- Ballou, Facility Location Decisions, Business logistics Management, pp.483-531;
- Ballou 1995, Logistics Network Design: Modelling and Information Considerations, International journal of
logistics management, volume 6, pp.39-54.

CHALLENGE QUESTIONS
- Compare the performance of the main types of distribution networks in terms of delivery time and
distribution costs
- When should a direct delivery option be of interest?
- What are the main decisions to be taken when designing a distribution network?
- How can a “distribution problem” be described? What are the main factors that drive the choice of the
distribution network structure?
- Assume you are a consumer electronics manufacturer. Which types of distribution networks would you
recommend to consider for distribution in Europe?

177
11. Distribution Network Design

178
 12. Distribution Network Design: design methodology and models – Part 1

12. Distribution network design: design

methodology and models – Part 1
0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different contexts:
- A general framework to design a distribution network that splits the problem into two main phases:
o Strategic planning
o Operative planning
- A focus on the strategic planning phase with two methodologies that aim at finding the “right fitting”
between the distribution problem and the structure of the distribution network

1 METHODOLOGY

DESIGN METHODOLOGY FRAMEWORK

DESIGN METHODOLOGY MAIN PHASES

Strategic planning Operative planning
- Objective: Identification of the most promising - Objective: Best alternative selection
configuration of the distribution network (just a limited and fine tuning
number – e.g. 3-5 alternatives) - Methodologies: Quantitative
- Methodologies: Strategic/qualitative methodologies methodologies
- Data required: Limited number of aggregate data - Data required: Huge amount of
detailed data

DATA ANALYSIS
The main objectives of the data analysis are:
1. Represent the distribution problem in order to drive the main strategic choices (choices on the system
configuration)
2. Collect the main data (only the required ones) with the right detail level to feed the models supporting
the strategic and operative planning

179
 12. Distribution Network Design: design methodology and models – Part 1

In order to describe the distribution problem we should consider different factors/elements (main drivers) that
are related to:
- Products features (PRODUCT)
Density (kg/m3), Value (€/kg), Shelf-life (days)
- Space-temporal demand features (DEMAND)
Delivery frequency (deliveries/day), Seasonality
- Space-temporal supply features (SUPPLIES)
Number of plants, Distance between plant and customer (km)
- Customer service level (SERVICE)
Order-delivery cycle time (days), Delivery accuracy (correct orders/overall orders)

1.3.1 Models
- Strategic models
o Network selection matrixes
o Qualitative trade off analysis of the distribution costs
o …
- Quantitative cost models
o Quantitative trade off analysis of distribution costs (optimization approach)
o Cost analysis of different network designs (simulation approach)
o …
- Location/allocation models
o Center of gravity
o Linear programming
o Mixed-integer linear programming
o …
- Specific optimization models
o Sizing and allocation of the safety stocks
o Local distribution optimization
o …

2 NETWORK SELECTION MATRIXES

OBJECTIVE OF THE NETWORK SELECTION MATRIXES

The Network Selection matrixes aims at finding a rough Distribution Network structure that fits the specific
Distribution Problem; they are the concrete application of the principles we have studied till not about the
distribution network
In the middle of this schema we have the principles and the network selection matrixes
NB: these are only two examples of methodologies, but are the main ones we will address
The rationality behind these two components is the same

Distribution problem
Distribution network
- Product type
- # echelons
- Supply features
- Type of warehouses (RWs and/or TPs)
- Demand features
- # warehouses per echelon
- Service required
- …
- …
o Principles
o Network selection matrixes

180
 12. Distribution Network Design: design methodology and models – Part 1

NETWORK SELECTION MATRIXES

1. Data collection -> we come out with the distribution problem description
A determined set of data has to be collected per each distribution problem (a company deals with more
than one distribution problem)
2. Distribution problem representation from a graphical point of view
It has to be a compact representation of the distribution problem (i.e. a representation that includes all
the drivers at the same time)
3. Positioning on the matrixes (of the distribution problem) it is another graphical tool
Each distribution problem will be positioned on a set of matrixes depending on the values of the different
drivers
4. Macro-decision on the network structure
The matrixes give the designer some recommendations about the structure of the distribution network
(number of echelons, rough suggestions about the number of warehouses per each echelon, …)

2.2.1 1. Data collection

Per each distribution problem we have to collect the data of the drivers in the table below
The drivers are always the same, we addressed them in the previous lessons
Production Market
Supply drivers: POO Demand drivers POD
- # plants - Average order dimension [full truck load,
- Average distance between plants pallet, case]
and markets [km] - Density of orders [orders/day]
- Focus level of production plants [1-5] - Seasonality [1-5]

Product drivers: Product Service drivers: service level

- Product range [# items] - Cycle time order-delivery [days]
- Density [kg/m3] - Delivery punctuality [1-5]
- Value [€/kg] - Delivery accuracy [1-5]
- Expiration [days] - Service level [Case fill rate%]
NB: the above list is a selection of a larger amount of drivers that impact the distribution network structure

2.2.2 2. Distribution problem representation

In order to compare the different drivers that define a distribution problem, they should be represented on the
same scale (i.e. a relative scale)
For each macro factor that defines the distribution problem we would like to have a value that tells us something
about the complexity of the distribution problem referring to that specific factor
ex: considering the product, we will find out the complexity values of the range, the complexity and the
obsolescence, and then we will compute the overall value, the one related to the product

We should define the scale and convert the real value in the relative scale value, in order to reduce the
complexity of compute the average value for the single value drivers, which is composed by different factors;
otherwise, we should go for a correlation analysis between the different factors
NB: if we want to be more precise, we should understand the weight on complexity for the different factors, but
of course it is much more complex than the simple average

181
 12. Distribution Network Design: design methodology and models – Part 1

e.g. the product density [kg/m3] E.g. Product:

A product with a 250 kg/m3 gets 3 Density 3
Then we can get a unique value per each driver (Product, Supply, Demand, Service) through a Range 2
simple average Value 4
Expiration 3
PRODUCT 3

We will represent in blue the area referring to the production and in the yellow the area referring to the market;
this representation will immediately give an idea of where the complexity of the distribution problem is

Bassetti (Europe)
→ Predominance of production drivers
Customers: retail
Products: household linen
In this case, the complexity of product is
very high, and mainly for this reason the
complexity is more in the blue area, so in
the production area
NB: we should address to the relative size
of the areas, not to the real area

3M (Europe) Lever Fabergè (Europe)

→ Predominance of production drivers → Market and production drivers are balanced
Customers: wholesalers and distributors
Products: detergents and personal care

Reckitt Benckiser (Italy) Nestlé (Italy) – bakeries

→ Market and production drivers are balanced → Predominance of market drivers
Customers: distribution centers, hypermarket, Customers: distribution centers, hypermarket,
wholesalers wholesalers, retail
Products: detergents (apparel and dishes), personal Products: bakeries
care

182
 12. Distribution Network Design: design methodology and models – Part 1

Ferrero (Italy) Artsana (Italy)

→ Predominance of market drivers Customers: Italian retail
Customers: distribution centers, retail and bars Products: baby products, health products, cosmetic and
Products: bakeries, snacks, pralines, creams, … beauty products

Nestlé (Italy) – dry food Unilever Bestfood (Italy)

Customers: wholesalers, large retailers Customers: catering
Products: pasta and dry food Products: food

Nestlé Purina Pet Care (Italy) Coca-Cola (Italy)

→ Market and production drivers are balanced → Market and production drivers are balanced
Customers: wholesalers, large distributors, normal Customers: wholesalers, large distributors, normal
trade trade
Products: food and products for pet care Products: beverage

Galbani (Italy)
→ Predominance of market drivers
Customers: retail and catering
Products: diary products and salami

183
 12. Distribution Network Design: design methodology and models – Part 1

2.2.3 3. Positioning on the matrixes - Example

- Production: product and supply
- Market: demand and service level

For the market we do the simple

average between demand and
service level values; for the
production axis the simple average
between products and supply

We put on this matrix the different solutions applied to the previous problems, according to different colours
From this matrix, we can see that the more we move from the bottom-right to the top-left, the more the network
is decentralized -> the more the complexity is moving from the production part (yellow area) to the market part
(blue part), the more the network is decentralized, so the number of echelons increases, and inventories are
closer to the customers

2.2.4 4. Macro decisions on the network structure

From the previous step we should understand the type
of network we have, but with this matrix we can
understand their complexity level and have a lot of
information before going in detail with mathematical
computations

Network typology

Selection of the number of warehouses in 1-echelon

distribution networks
Assuming that we got solution of only 1-echelon, considering the
same matrix, we can highlight different areas according to the
number of warehouses, which are showing the increase of
complexity

The matrix is just a way to immediately show the complexity of our

network

184
 12. Distribution Network Design: design methodology and models – Part 1

Structure of 2-echelon distribution networks with Structure of 2-echelon distribution networks

transit points with regional warehouses

3 COST MODELS
We will address two different cost models, and we will consider two examples, one for each model
In this lesson we will go through the qualitative evaluation of the problem, while in the practical lesson we will
compute and compare the real costs
Qualitative cost models
They aim at analysing the relationships between the main costs and a specific parameter of the Distribution
Network (e.g. number of warehouses, number of echelons, etc.) by identifying the cost curves (i.e. curves that
show how the costs vary by varying the parameter).
[See the example in the forthcoming part of the lesson (number of regional warehouses in a 2-echelon
distribution network)]

Quantitative cost models

They aim at calculating the costs of one (or a limited number of) specific Distribution Network configuration in
order to assess the most convenient alternative. They require a huge amount of data and can be used only after
the most promising alternative selection. [See the example in the practical lesson]

3.1.1 Number of warehouses

We will consider that we have a specific problem
Assuming that the number of echelons has already
been chosen, we want to determine the number of
warehouses that minimize the overall distribution cost
for a given service level

2-echelons network, with a central warehouse and

some different regional warehouses
How do the relevant cost items vary by changing the
number of warehouses with inventory in the 2nd
echelon of a 2-echelon distribution network?

THE COST ANALYSIS

Objective: Considered costs: Assumption:
Determine the curves - Transportation costs The replenishment frequency (and
that show how the o Full truck load transportation as a consequence the time lapse
distribution costs vary costs between two consecutive orders)
by varying the number o Local distribution costs does not change
of regional warehouse - Inventory costs
o Cycle stock
o Safety stock
o In transit stock
- Handling costs

185
 12. Distribution Network Design: design methodology and models – Part 1

Transportation costs
- Primary transportation -> having the overall demand remaining the same, the saturation cannot increase,
since the same volume is split between many different arcs from the warehouse till the POD, so if you do
not have enough volume, maybe you are not able to reach saturation
→ primary transportation costs increase
- Local distribution -> the distance between a warehouse and its local area is reduced,
so we can spend more time in the delivery area and so touch more points
→ local distribution costs decrease, till an asymptote
We have an asymptote because at a certain point the warehouse will be inside the
delivery area, but we will still have the distances to touch the POD

Inventory carrying costs

- Cycle stocks → remain the same, since it does not depend on
how we split the demand, it depends on the demand and the
frequency of ordering, so if they remain the same, the cycle stock
are just split by the number of warehouses and then summed up as the begin
In reality, the cost for the cycle stock increases due to the economies of scale, because increasing the
number of WHS, the management costs in the different warehouses are higher, since in a single
warehouse you have half quantity

- Safety stocks: 𝑆𝑆 = 𝑘 ∗ σ𝐷,𝐿𝑇 = 𝑘 ∗ √𝐿𝑇 ∗ σ2𝐷 + 𝐸𝐷 2 ∗ σ2𝐿𝑇

Considering the forecast error, it is higher, since we have more WHS which can be linked to more
different demands
ex: assuming no correlation between two different demands, the variation of the overall demand is lower
than the sum of the two variations σ𝐷𝑇𝑂𝑇 = √σ2𝐷1 + σ2𝐷2 → σ𝐷𝑇𝑂𝑇 < σ𝐷1 + σ𝐷2
BUT adding WHS is like splitting the overall demand, so the variation is exactly the sum, so it is higher
than the demand we could have with one single WHS; for this reason, the overall amount of safety stocks
will increase to cover the overall unpredictability of the demands [it is proportional to the √N]
NB: in case we have economies of scale, it will be something like ICCCS

- In transit stocks: 𝐼𝑇𝑆 = 𝐹 [𝑢𝑛𝑖𝑡𝑠/𝑑𝑎𝑦] ∗ 𝐿𝑇 [𝑑𝑎𝑦𝑠]

→ they remain the same, since they are not affected by the number of warehouses
o the demand is always the same, no matter how many warehouses we put in the 2nd level
o the average transportation LT does not change, considering the overall flow: adding warehouses,
you can dedicate more time to the primary transportation, but then you have also the secondary
transportation

Handling costs 𝐻𝐶 = 𝐹𝑦𝑒𝑎𝑟 ∗ 𝑢ℎ𝑐

The handling costs remain the same, since the demand does not change, and the unitary handling cost it could be
impacted only if there are economies of scale [right answer, depends]
If we have volume big enough, we can decide to put an automated handling system, which has a lower uhc with a
higher investment at the beginning, which can be justified only with a high volume

NB: this graph is representative for only the shapes of these

curves, do not look at their relative position, since for the right
evaluation we should consider the specific case
- PTC increase
- LDC decrease till an asymptote
- ICCCS remain the same (increase with eco. of scale)
- ICCSS increase (increase as ICCCS with eco. of scale)
- ICCITS remain the same
- HC remain the same (increase as ICCCS with eco. of scale)

186
 12. Distribution Network Design: design methodology and models – Part 1

4 COST ANALYSIS - IN DETAIL

TRANSPORTATION COSTS
4.1.1 Full truck load transportation costs

What happens if the number of regional warehouses increases?

- Full truck load transportation costs hold steady until transports are full truck loads
- After a determined N (whose value depends on the flows in the specific situation), transportation costs
will increase because the transports will not be full truck loads anymore. As a consequence:
o the shipment size will decrease
o the # shipments per year will increase
o the fare will increase

→ The overall effect will be a transportation cost increase

4.1.2 Local distribution costs

The reference model

What happens if the number of regional warehouses increases?

- dRWi→Ak” will start decreasing
- “Ak” will start increasing because the vehicles can spend more time to serve the areas and each vehicle
can serve more customers (N). (In other words the delivery area grows and the number of vehicles
decreases)
- “# del” will start decreasing as a consequence of the previous point

This process will keep until the distances between regional warehouses
and delivery areas will be irrelevant compared to the distances travelled
in the delivery area (dA).

187
 12. Distribution Network Design: design methodology and models – Part 1

Length of trips estimation

Random and uniformly distributed N points within an area
of any shape:
- N = number of delivery/picking point (i.e.
customers)
- S = size of the delivery area
- E = N/S = customer density
- D = average travel distance per trip composed of N
points to visit
- d = average distance between two customers

Rectilinear distances: 𝐷 = 𝑁 ∙ 𝑑 = 𝑁 ∙ [1,15 ∙ 𝐸 −1/2 ]

Euclidean distances: 𝐷 = 𝑁 ∙ 𝑑 = 𝑁 ∙ [0,90 ∙ 𝐸 −1/2 ]

Length of trips estimation – Example

Problem
Considering the following information:
- 100 customers uniformly distributed within an area of 8,000 km2
- “commercial velocity” (v) equal to 50 km/hour
- fixed time per stop equal to 12 minutes

and assuming the following hypothesis:

- you have to consider only the time constraint disregarding the truck capacity
- time required to reach the delivery area (TRW→Ak) can be assumed negligible
- available time per day (TA) equal to 8 hours
- for each customer, it is consider one shipment per day (CASE1) or one shipment per week (CASE2)

It is requested to calculate the number of trucks needed to satisfy the customer demand for both cases

Solution
- Average distance between two customers (𝑑) = 0,90 ∗ 𝐸 −1/2 = 8 𝑘𝑚
𝑇 −2∗𝑇𝑅𝑊→𝐴𝑘 8
- Number of stops (i.e. customers) per trip = 𝐴 𝑑 = 12 8 = 22 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠/𝑡𝑟𝑖𝑝
𝐹𝑇− +
𝑣 60 50

100
- Case 1 (1 shipment/day): 22
= 5 𝑣𝑒ℎ𝑖𝑐𝑙𝑒𝑠
- Case 2 (1 shipment/week): 1 𝑣𝑒ℎ𝑖𝑐𝑙𝑒

INVENTORY COSTS

188
 12. Distribution Network Design: design methodology and models – Part 1

4.2.1 Cycle Stocks

If the replenishment frequency does not vary (i.e. time lapse
between two consecutive orders does not change), the cycle stocks
will not change by varying the number of regional warehouses
(If the replenishment frequency changes, the overall cycle stock in
the system will vary – it is proportional to the time between
consecutive orders –)

4.2.2 In-transit stock

The overall amount of the in-transit stocks only depends on the
replenishment lead time and the demand. Neither the lead time
nor the demand depend on the number of warehouses.

4.2.3 Safety stocks

The safety stocks in the system depend on different factors:
- service level
- demand seen by each warehouse and its standard deviation
- geographical correlation of the demand
- replenishment LT and its standard deviation
- inventory policy
- …
Safety stocks deal with the unpredictability of both demand and LT.
The more the demand is split among warehouses, the more the safety stocks in the system.

Reorder point model

𝑆𝑆 = 𝑘 ∗ 𝜎𝐷,𝐿𝑇 𝜎𝐷,𝐿𝑇 = √𝐸𝐿𝑇 ∗ 𝜎𝐷2 + 𝜎𝐿𝑇2
∗ 𝐸𝐷 2
σD,LT: standard deviation of the demand during the LT (or standard deviation of errors in the forecast system)
k: coefficient that takes into account the desired Service Level (the higher the coefficient, the higher the service
level)

Periodic review model

𝑆𝑆 = 𝑘 ∗ σ𝐷,𝐿𝑇+𝑇 𝜎𝐷,𝐿𝑇+𝑇 = √(𝐸𝐿𝑇 + 𝑇) ∗ 𝜎𝐷2 + 𝜎𝐿𝑇 2
∗ 𝐸𝐷 2
σD,LT+T : standard deviation of the demand during T+LT (or, in the forecast system, standard deviation of the error)
k: coefficient that takes into account the desired Service Level (the higher the coefficient, the higher the service
level)

The rule of the √N: safety stocks are proportional to the

square root of the # of warehouses

NB: the relative position of the various curves in the

above graph is arbitrary and is actually application-
specific
.

189
 12. Distribution Network Design: design methodology and models – Part 1

4.2.4 Cost per unit

The inventory costs per unit (icunit load) is subject to economies of
scale. The higher the number of warehouses, the higher the cost
per unit. The effect on the curves of the inventory costs (SS, CS,
ITS) is the following:

HANDLING COSTS

The handling cost per unit (the cost of loading/unloading the

vehicles, handling, picking, etc.) is subject to economies of scale
(for instance automated warehouses only for high flows). The
handling cost curve is shaped as follows:

TOTAL DISTRIBUTION COSTS

The main costs in choosing the number of
warehouses

NB: the relative position of the various curves in the

above graph is arbitrary and is actually application-
specific.

5 CONCLUSION

SUPPLEMENTARY READINGS
- Ballou, The Network planning process, Business logistics
- Management, pp.542-603;
- Ballou, Facility Location Decisions, Business logistics Management, pp.483-531;
- Ballou 1995, Logistics Network Design: Modelling and Information Considerations, International journal of
logistics management, volume 6, pp.39-54.

CHALLENGE QUESTIONS
- Which are the main macro-phases of the Distribution Network Design process? How would you cope with
them?
- What do the “Network selection matrixes” aim at? Which are their strengths? Which their limitations?
- Assume you are a manufacturer of car parts. How would you structure your European Distribution
Network?
- Which are the main differences between the qualitative and the quantitative cost models in terms of
objectives/field of applications and information required?

190
 13. Distribution Network Design: design methodology and models – Part 2

13. Distribution network design:

design methodology and models – Part 2
0 INTRODUCTION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different contexts:
- Definition of the location/allocation problem
- The main models to deal with the location/allocation problem
- A framework to assess the application fields (pros/cons, completeness, effort, etc.) of the different
models
- An in-depth analysis of the main models and a few practical examples

AGENDA
1. Location and allocation problem
2. Single warehouse location
3. Allocation
4. Multi-warehouse location

1 LOCATION AND ALLOCATION PROBLEM

Type of problems we can address considering the quantitative models:
Assuming that the number of echelons and the number of warehouses
for each echelon have already been decided, we have to
- roughly determine the location of each warehouse
(warehouse location) → mathematical model
- then to find its exact position (site selection)
→ we will consider other factors

ALLOCATION
Assuming that the number of echelons and the number of warehouses for each echelon have already been
decided, we have to determine:
- the flow to be allocated to each warehouse (i.e. the quantities on the “arcs” of the network)
flows both going in and going out the warehouse
- the customer allocation to the warehouses: it can be served by one single warehouse or by more than
one warehouse, and this can happen for different reasons

NB: the dimensions of the arrows are

proportional to the flows
ex: when each warehouse does not have all the
products: each customer could be served by many
warehouses, according to what he orders
NB: it is a little bit strange that one customer will be
served by more than one WHS for the same product

191
 13. Distribution Network Design: design methodology and models – Part 2

LOCATION
The warehouse location process is based on two main steps:
- the determination of the geographical position of the warehouse, rough design of the WHS position
- the review of the results stemming from the quantitative techniques by taking into account more specific
factors; some factors are very related to the very specific position (ex: if we should position the
warehouse near to the rail or not), and other are wider factors (ex: in which country put the warehouse)

Quantitative techniques Choice factors

1. Single-warehouse location - Proximity to suppliers/plants
o Center of gravity - Proximity to customers
2. Multi-warehouse location - Presence of transport infrastructures
o Heuristic methods - Cost of the area and of the public utilities
o Linear programming - Costs of in/outbound transports
(simple/ mixed-integer) - Cost and reliability of labour
o Simulation - Tax relieves/restrictions
- Proximity to other companies
- Meteorological conditions / quality of life

1.2.1 Single vs multi-warehouse location

Single warehouse location
The only issue is where to locate one warehouse. The most relevant costs are transportation costs
We have to decide only the coordinates of the warehouse, there is no allocation problem, only the location one
→ location = allocation

Multi-warehouse location
It is a problem of both location and allocation (location-allocation problem), because depending on where we
locate the warehouses, the flows allocation can change, because it could be cheaper to allocate it another way,
considering changes in transportation and inventory costs

A multi-warehouse location model has to deal with the following issues:

- Multi-warehouse location, since it will definitely impact the allocation
- Product and production capacity allocation to both plants and warehouses
- Customer allocation to the warehouses

MODEL CLASSIFICATION
Types of problem:
- Location → only one warehouse to be located
- Allocation → more than one WHS, but they are already located
- Location + Allocation → more than one WHS

The combination of the two problems will

increase the complexity of our problem and
also of the mathematical model we should
use

Considering the different kinds of problem

we can have, we can use different models
to solve them

192
 13. Distribution Network Design: design methodology and models – Part 2

- Location → Centre of gravity

The centre of gravity model will consider the distances between WHS and POD, because it aims to only
minimize the transportation costs
- Allocation → Linear programming
Linear programming can be used at different levels of complexity, which refer to different costs; it makes
sense to include different costs when they vary according to the type of warehouse, because if they are
the same for each of our warehouses, this will not have impact the allocation of the flow
- Location + Allocation → Mixed-integer programming
The model will take into account all the main costs (transportation, handling and inventory costs), and we
can also add also the service level as a constraint, in order to be compliant with a target OCT
The objective function will minimize the sum of the three types of cost
Having all the data according to these types of costs according to the different warehouse possible
positions, we can assign to each warehouse a value: 0 if that infrastructure is close, 1 if it will be open
- Location + Allocation + #WHS → Mixed-integer programming
With this model, we can have another level of decision, which is the number of warehouses, in the sense
that considering the infrastructure opened/closed, the model will tell how many warehouses we need

2 SINGLE WAREHOUSE LOCATION

CENTER OF GRAVITY
Application field: Center of gravity model can be used for the location of a single warehouse
Main managerial issue: (X,Y) = coordinates of the center of gravity, which represent the warehouse position

Model parameters
- (Xi,Yi) = coordinates of both the points of origin
and destination
- Fi = inbound (for the point of destination) and
outbound (for the point of origin) flows; it is
forced, every flow has to pass through the
warehouse, we have no allocation problem
NB: in this model, no matter if the flow goes in or
out, so we will consider the POO and POD as they
are the same; in the graph they are differently
coloured just to understand the context
- Ri = transportation rate per unit [€/(km*t)]
(it depends on the weight and the distance)
The fares are in a table, defined according by distances and quantities
The longer the distances and the higher the quantities, the higher the fare
NB: fare is different by the cost, to define the cost we should include the distance and quantity values

Objective function
Minimization of the transportation costs by finding the coordinates of the center of gravity
𝑚𝑖𝑛 (𝑡𝑜𝑡𝑎𝑙 𝑐𝑜𝑠𝑡) = 𝑚𝑖𝑛𝑥,𝑦 ∑[𝐹𝑖 ∗ 𝑅𝑖 ∗ 𝑑𝑖 (𝑥, 𝑦)]
𝑖
= 𝑚𝑖𝑛 [𝑓𝑙𝑜𝑤 ∗ 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 ∗ 𝑓𝑎𝑟𝑒] [𝑡𝑜𝑛𝑠] ∗ [€/𝑘𝑚 ∗ 𝑡] ∗ [𝑘𝑚] = [€]

Transportation costs linearly depend on the distance and the transported quantity of each flow

193
 13. Distribution Network Design: design methodology and models – Part 2

2.1.1 Steps
In order to find the center of gravity at least two steps have to be made:
1. First step - Centroid
A first approximation of the real position of the center of gravity has to be found (i.e. the “centroid”)
We will compute a weighted average of the coordinates based on the flows and the fares
NB: it is an approximation because the real fares are not known at the beginning, as they depend on the
distances to be travelled; for this reason, in this first step, we will consider the fares always the same,
for every flow [Ri = 1 €/km*t] → the average will be based only on the flow

∗
∑𝑖 𝐹𝑖 ∗ 𝑅𝑖 ∗ 𝑋𝑖 ∗
∑𝑖 𝐹𝑖 ∗ 𝑅𝑖 ∗ 𝑌𝑖
𝑋𝐶𝑂𝐺 = 𝑌𝐶𝑂𝐺 =
∑𝑖 𝐹𝑖 ∗ 𝑅𝑖 ∑𝑖 𝐹𝑖 ∗ 𝑅𝑖

∗
∑𝑖 𝐹𝑖 ∗ 𝑋𝑖 ∗
∑𝑖 𝐹𝑖 ∗ 𝑌𝑖
𝑋𝐶𝑂𝐺 = 𝑌𝐶𝑂𝐺 =
∑𝑖 𝐹𝑖 ∑𝑖 𝐹𝑖

The lower and the more concentrated the flows, the farthest the result is from the real one
Thanks to the coordinates of the COG, we can compute all the distances between it and the POO/POD,
and with these we can compute the real coordinates of the COG (second step)

2. Second step – “Exact” center of gravity

- Starting from the position of the “centroid” (X*,Y*) we have to proceed with a further calculation of the
position of the “center of gravity” according to the following formulas:

𝐹𝑖 ∗ 𝑅𝑖 ∗ 𝑋𝑖 𝐹𝑖 ∗ 𝑅𝑖 ∗ 𝑌𝑖
∑𝑖 ∑𝑖
∗∗ di ∗∗ di
𝑋𝐶𝑂𝐺 = 𝑌𝐶𝑂𝐺 =
𝐹𝑖 ∗ 𝑅𝑖 𝐹𝑖 ∗ 𝑅𝑖
∑𝑖 ∑𝑖
di di

di is the distance of the point (Xi, Yi) from the center of gravity (X*, Y*)
It is still a weighted average, but we have the real fares (from the table) and the distances

- Then, we have an iterative procedure: the more the number of iterations the more precise the solution
(quite often the second step is enough)
This because at this moment distances (and so the fares) are based on the (X*,Y*) coordinates of COG,
which were an approximation
→ starting from (X**,Y**), we will compute new distances (and fares) and then (X***,Y***)
We will stop when the values of the coordinates are very close to each other between two steps,
and the cost does not change

2.1.2 Distances computation

The distances “di” can be got as Euclidean distances or as rectilinear distances;
we should consider the best one according to the context in which we are working
ex: for countries and continent the Euclidean distances are better
These are approximation: in order to cover this gap, we can take into account a correction factor (coefficient c =
1.2 - 1.3) which help us to approximate the real distance in efficient way; it depends on the specific
context/situation

Euclidean distances 𝑑𝐴−𝐵 = c ∗ √(𝑥𝐴 − 𝑥𝐵 )2 + (𝑦𝐴 − 𝑦𝐵 )2

(outside urban areas) → we are underestimating a little bit the real distances -> 𝑐 > 1

Rectilinear distances 𝑑𝐴−𝐵 = 𝑐 ∗ [|𝑥𝐴 − 𝑥𝐵 | + |𝑦𝐴 − 𝑦𝐵 |]

(in the city), we have roads like a grid → we are overestimating the real distances -> 0 < 𝑐 < 1

194
 13. Distribution Network Design: design methodology and models – Part 2

2.1.3 Key messages

- The result does not depend on the origin (0,0) of the graph
- In the first step, it is strongly recommended to use a constant Ri (e.g. Ri=1 ꓯi)
- It is a very useful method to find a rough location even though it does not include some costs (e.g.
handling costs, storage costs, etc.)
- It can be used also to find iso-cost curves in order to assess the trade off with the other costs

Through the same model, we can compute the

COG coordinates and all the costs (TC, ICC, HC),
then we can draw the iso-cost curves
According to the different areas, we can say
that within a specific area the cost does not
change a lot, otherwise it will for sure increase

3 ALLOCATION

REFERENCE SCENARIO
We will take into consideration the allocation of the
flow between the POO/POD and the warehouses, and
in order to do that we will use the linear programming
- i = index related to the plants
- j = index related to the warehouses
- k = index related to the delivery areas

Application field: allocation of the plant flows to the warehouses and/or allocation of the customers to the
warehouses

Model variables → what we want to determine

- xij = quantity to be delivered from the plant i to the warehouse j
- yjk = quantity to be delivered from the warehouse j to the customer k

Model parameters
- Rij = Transportation fare per unit from the plant i to the warehouse j [€/unit]
- Rjk = Transportation fare per unit from the warehouse j to the customer k [€/unit]
- Qi = production capacity of the plant i
- Qj = operative capacity of the warehouse j
- Qk = demanded quantity by the customer k
NB: since the warehouses have already been located, the distances are known, but the fares are still considered
parameters because quantities have to be determined → we are computing an iterative process
In the formulation of this problem the distances are already included in the fares, to simplify the problem

Objective function
Minimize the transportation costs by determining the optimal flows from the plants to the warehouses and from
the warehouses to the customers
𝑚𝑖𝑛 (𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑠𝑡𝑠) = ∑ ∑ 𝑅𝑖𝑗 ∗ 𝑥𝑖𝑗 + ∑ ∑ 𝑅𝑗𝑘 ∗ 𝑦𝑗𝑘
𝑖 𝑗 𝑗 𝑘
Constraints
- Constraints related to the production capacity of the plants (Qi)
- Constraints related to the operative capacity of the warehouses (Qj)
- Constraints related to the forecast demand of each customer (Qk)

∑𝑗 𝑥𝑖𝑗 ≤ 𝑄𝑖 (∀𝑖) ∑𝑖 𝑥𝑖𝑗 ≤ 𝑄𝑗 (∀𝑗) ∑𝑘 𝑦𝑗𝑘 ≤ 𝑄𝑗 (∀𝑗) ∑𝑗 𝑥𝑗𝑘 ≤ 𝑄𝑘 (∀𝑘)

195
 13. Distribution Network Design: design methodology and models – Part 2

3.1.1 Example (1)

Company that has to supply 4 regional warehouses from 3
plants. The transportation costs per unit (€/unit) for each
plant/warehouse connection are shown in the following table

If we define xi,j as the quantity to be delivered from plant 𝑚𝑖𝑛 [0,8𝑥1,1 + 0,9𝑥1,2 + 1,1𝑥1,3 + 1,6𝑥1,4 +
i to warehouse j (with I = 1,2,3 and j = 1,2,3,4), the + 1,2𝑥2,1 + 0,7𝑥2,2 + 0,5𝑥2,3 + 0,8𝑥2,4 +
objective function can be written as follows: + 1,4𝑥3,1 + 1,0𝑥3,2 + 0,6𝑥3,3 + 0,7𝑥3,4 ]

Constraints related to the availability of the products Constraints related to the forecast demand for each
in the plants [with xi,j ≥ 0; i = 1,2,3; j = 1,2,3,4] warehouse:
𝑥1,1 + 𝑥2,1 + 𝑥3,1 ≤ 90
𝑥1,1 + 𝑥1,2 + 𝑥1,3 + 𝑥1,4 ≤ 50
𝑥1,2 + 𝑥2,2 + 𝑥3,2 ≤ 70
𝑥2,1 + 𝑥2,2 + 𝑥2,3 + 𝑥2,4 ≤ 80
𝑥1,3 + 𝑥2,3 + 𝑥3,3 ≤ 40
𝑥3,1 + 𝑥3,2 + 𝑥3,3 + 𝑥3,4 ≤ 120
𝑥1,4 + 𝑥2,4 + 𝑥3,4 ≤ 50

3.1.2 Example (2)

The distribution network is made of 2 plants, 3
warehouses and 5 customers. The
transportation fares per unit (€/unit), the
product availabilities in the central warehouses,
the ordered quantities by the customers and
the capabilities of the regional warehouses are
shown in the following table
[H = very high, in order to avoid the direct
shipment from plants to customer, and also
transportation between a warehouse and
another warehouse]

If we define xi,j as the quantity to be delivered from plant 𝑚𝑖𝑛[3𝑥1,𝐴 + 8𝑥1,𝐵 + 6𝑥1,𝐶 +
i to the warehouse j (with i=1,2 and j=A,B,C), and yj,k the +8𝑥2,𝐴 + 3𝑥2,𝐵 + 7𝑥2,𝐶 +
quantity to be delivered from the warehouse j to the +9𝑦𝐴,1 + 9𝑦𝐴,2 + 10𝑦𝐴,3 + 11𝑦𝐴,4 + 13𝑦𝐴,5 +
customer k (with j=A,B,C and k=1,2,3,4,5), the objective +13𝑦𝐵,1 + ⋯ + 9𝑦𝐶,4 + 10𝑦𝐶,5 ]
function can be structured as follows:

Constraints related to production Constraints related to the operative Constraints related to the forecast
capacity of the plants: capacity of the warehouses: demand of each customer:
𝑋1,𝐴 + 𝑥2,𝐴 ≤ 40 𝑦𝐴,1 + 𝑦𝐵,1 + 𝑦𝐶,1 = 10
𝑋1,𝐴 + 𝑥1,𝐵 + 𝑥1,𝐶 ≤ 20 𝑋1,𝐵 + 𝑥2,𝐵 ≤ 40 𝑌𝐴,2 + 𝑦𝐵,2 + 𝑦𝐶,2 = 6
𝑋2,𝐴 + 𝑥2,𝐵 + 𝑥2,𝐶 ≤ 20 … …
𝑌𝐴,1 + 𝑦𝐴,2 + 𝑦𝐴,3 + 𝑦𝐴,4 + 𝑦𝐴,5 ≤ 40 𝑌𝐴,5 + 𝑦𝐵,5 + 𝑦𝐶,5 = 6
with xi,j ≥ 0; with yj,k ≥ 0 i = 1,2; j = A,B,C; k = 1,2,3,4,5

196
 13. Distribution Network Design: design methodology and models – Part 2

4 MULTI-WAREHOUSE LOCATION
I want to understand
- How many warehouses
- The location of each warehouse
- The flows for each warehouse

MODELS FOR MULTI-FACILITY ALLOCATION

- Heuristic algorithms/Clustering methods: warehouses are located as centers of gravity of each clusters in
which the distribution problem has been split
- Optimization criteria (e.g. mixed-integer linear programming): they aim at determining the optimal
problem solution, by assessing both the optimal location and the distribution plan (at the same time)
- Simulation (what if analysis): it evaluates the impact of the different network alternatives on the
objective functions (costs and service level). It is a Decision Support System (DSS) that allows the designer
to make a few preliminary decisions and to dynamically evaluate the efficiency of the different solutions
(if the problem variables vary)

CLUSTERING METHODS
Clustering methods allow the designer to determine the location of many warehouses and the allocation of a
delivery area to each warehouse
We will not go in detail of this model, because it is of course complex at the formulation level, but it is even more
complex at the collection of data level, because we have a lot of variables, and for each variable we can have
different values

Steps
- Allocate one warehouse to each customer and assess the cost of this solution
- Reduce by 1 the number of warehouses and group together the two nearest customers
- Determine the center of gravity in which the new warehouse will be located
- Get the cost of the new network configuration
- Repeat steps 2, 3 and 4 by reducing the number of warehouses by one
The process will be over if:
o service level is too scarce
o the cost of the configuration starts increasing
o only one warehouse remains

Reference scenario
- j = index related to the plants
- i = index related to the potential locations of the warehouses
- k = index related to the delivery areas
- l = index related to the product categories

Model variables
- xi,j,l = the quantity of the product category “l” that is produced in the plant “i” and delivered to the
warehouse in the area “j”
- yj,k.l = the quantity of the product category “l” that is delivered in the delivery area “k” from the
warehouse in the area ”j”
- zj = integer variable that can be only 1 or 0. It is 1 if a
warehouse has been set in the area j, 0 if it is not so.
This parameter shows us that we have to consider many
alternatives for warehouses, and we do not know how many
of these variables we will take into consideration
→ The model is able to show us both how many warehouses
we should open, and also which warehouses

197
 13. Distribution Network Design: design methodology and models – Part 2

Model parameters
- ai = annual production capacity of plant “i”
- ai,l = required effort of production capacity to manufacture 1 unit of product category “l” in the plant “i”
- pi,l = production cost per unit of product category “l” in the plant “i”
- fj = annual fixed cost for a warehouse in the area “j”
- bj,l = required effort of operative capacity to prepare 1 unit of the product category “l” in the warehouse
in the area “j”
- bmj, bMj = minimum and maximum capability of the warehouse that is localised in the area “j”. This
capability can be defined as the material flow through the warehouse
We need at least a minimum volume to run/open a warehouse
- Vj = variable cost per unit of material flow passing through the warehouse in the area j
Handling cost
- ci,j,l = transportation cost per unit of the product category “l” from the plant “I” to the warehouse in the
are “j”
- hj,k,l = transportation cost per unit of the product category “l” from the warehouse in the area “j” to the
delivery area “k”
- dk,l = forecast demand for the product category “l” in the delivery area “k”
- tj,k,l = average delivery time of the product category “l” to the delivery area “k” from the warehouse in the
area “j” → useful to consider the service level
- Tl = maximum delivery time for the product category “l”

Objective function (to be minimized)

∑𝑖,𝑗,𝑙 𝑝𝑖,𝑙 𝑥𝑖,𝑗,𝑙 (= production costs) +
+ ∑𝑗 𝑓𝑗 𝑧𝑗 (= fixed costs of warehouses) +
+ ∑𝑗 𝑣𝑗 ∑𝑘,𝑙 𝑏𝑗,𝑙 𝑦𝑗,𝑘,𝑙 (= variable costs of warehouses) +
+ ∑𝑖,𝑗,𝑙 𝑐𝑖,𝑗,𝑙 𝑥𝑖,𝑗,𝑙 (= distribution costs of full pallet truck load from plants to warehouses) +
+ ∑𝑗,𝑘,𝑙 ℎ𝑗,𝑘,𝑙 𝑦𝑗,𝑘,𝑙 (= costs of local distribution from warehouses to end customers)

Constraints
- ∑𝑗,𝑙(𝑎𝑖,𝑙 𝑥𝑖,𝑗,𝑙 ) ≤ 𝑎𝑖 ∀i
(constraint related to the production capacity of the plant “i”)
- ∑𝑖(𝑥𝑖,𝑗,𝑙 ) = ∑𝑘(𝑦𝑗,𝑘,𝑙 ) ∀ j,l
(the inbound and the outbound flows in the warehouse in the area j must be equals)
- ∑𝑗(𝑦𝑗,𝑘,𝑙 ) = 𝑑𝑘,𝑙 ∀ k,l
(demand of the product category “l” in the delivery area “k” must be fulfilled)
- (𝑏𝑚𝑗 𝑧𝑗 ) ≤ ∑𝑘,𝑝(𝑏𝑗,𝑙 𝑦𝑗,𝑘,𝑙 ) ≤ (𝑏𝑀𝑗 𝑧𝑗 ) ∀j
(constraint related to the minimum and the maximum value of the material flow that passes through the
warehouse in the area “j”)
∑𝑗,𝑘(𝑡𝑗,𝑘,𝑙 𝑦𝑗,𝑘,𝑙 )
- ∑𝑘(𝑑𝑘,𝑙 )
≤ 𝑇𝑙 ∀l
(constraint related to the delivery times for the product category “l”)
- 𝑥𝑖,𝑗,𝑙 ≥ 0 𝑦𝑗,𝑘,𝑙 ≥ 0 ∀ i,j,k,l 𝑧𝑗 = 0, 1

198
 13. Distribution Network Design: design methodology and models – Part 2

5 PIRELLI PRACTICAL LESSON

Once collected all the data, we will try to understand if the structure they already have is consistent with the
distribution problem

PRODUCTION SIDE MARKET SIDE

[two different areas + OEM(*)]
Points of origin Points of destination
- Number: 19 plants - Number: 250.000 (worldwide), 150.000 (EU)
- Location: all over the world [2000 wholesalers, 148.000 small customers]
- Focus: very high, they produce a specific - Order size: 500 pc wholesalers, 8-12 pc small
product in a specific plant customer
- Order frequency: weekly wholesalers, daily
small customer
- Seasonality: high
Product Service level
- Range: not so important, 3000 SKUs - Order cycle time: 48h wholesalers, 24h small
- Density: very low, 120 kg/m3 customers (or some days for the 5% of the
it is ≈ 1\3 of the optimal density (350 kg/m3) flow)
→ we can get saturation on volume, not on - Punctuality: high
weight - Accuracy: high
- Value: 10 €/kg
- Obsolescence: 24-48 months, not so
important

The complexity on the production side is increasing since we are facing some new trends
- Increase in the product range: more cars and more alternatives for each model
- Increase in the value, because the size of the tyre is increasing (≥ 18 inches)
→ overall, considering the production side, it has medium-high complexity level

(*) We should add another POD, the car vendors, OEM channel (original equipment manufacturers)
- Order size: larger than the wholesalers
- Order frequency: once a week/every two weeks, not so high
- OCT: not so relevant

We can rank the order complexity of the POD in this way:

1. OEM
2. Wholesalers
3. Final customers

→ High complexity on the market side, not so high on the production side

1. Do we need a 1-echelon of the network?

Car vendors (OEM) Wholesalers Small customer
Decrease of No, since they are No, since they order Yes, we are delivering
transportation costs ordering huge enough quantities small orders
quantities → easy
saturation of the FTL
Decrease of No, they do not need Yes, because they want to receive
order cycle time the good immediately the order very quickly
Completeness No, no as transportation Yes Yes
costs, they order huge
quantities
→ in conclusion, we need a 1-echelon for both wholesalers and small customers, not for the car vendors

199
 13. Distribution Network Design: design methodology and models – Part 2

2. How many central warehouses on the 1-echelon?

In general, it depends on the trade-off between transportation costs and inventory
carrying cost, and it can be constrained by the service level
In this case, we have transportation costs higher than the balanced situation,
considering the wholesalers and the small customers

Then we need to take into account the fact that they have to reach a very high service level
- 24h for small customer
- 48h for wholesalers
→ no matter about the trade-off, for sure they need at least 1 warehouse for each main country they serve

3. Do we need a 2-echelon? If yes, TPs or RWs?

Wholesalers Small customers
TPs, RWs Transportation costs No, they are receiving Yes, the order size is
already full trucks very small
RWs Order cycle time No No, a certain number
of WH is Europe is
enough
→ we need a 2-echelon of transit points only for small customers

In conclusion, according to the distribution problem we need:

- Direct delivery for OEM
- 1-echelon DN for the wholesalers
- 2-echelon DN with TPs for small customers

5.1.1 Their solution

Focusing on Europe, we have
- 2-echelons DN for small retailers (*)
o 12 central warehouses at the 1st echelon
o 4 transit points at the 2nd echelon
- 1-echelon DN for wholesalers → 12 central warehouses (**)
- Direct delivery for OEM

(*) this is true for the 95% of the small retailers’

demand, while for a 5% they do not need the level
of TPs since the service level is not so stringent
(**) this is true for the 80% of the wholesalers’
demand, for the remaining 20% we have direct
delivery; we do not consider that in the
distribution problem since we are considering the
average value

This flow considers those wholesalers who order a

huge quantity of goods, for which they can deliver
FTL

→ the network is consistent

with the distribution problem

200
 13. Distribution Network Design: design methodology and models – Part 2

MIXED INTEGER PROGRAMMING

→ Project of redesign of the network
This is something that companies should take into account every 2/3 years (according to the demand variations
along time), in order to check if the distribution problem has changed and if their solution is still consistent with it

Taking into account the European distribution network, they want to check
- The number of central warehouses
- Their location
- The allocation of the flow
→ allocation+location problem, we have to minimize the sum of the relevant costs, with a given service level

We will use the variables they already have from

the as-is option, in order to understand if starting
from here it is possible to have a better solution

Variables
- p = 19 different plants
- h = 24 possible location for central
warehouses
- j = 30 possible demand areas of final
customers
NB: the customers have been clustered in order to
reduce the problem complexity; the position of that
area has been computed as the COG of the POD

Relevant costs
- Primary transportation costs → between plants and WHS
- Local distribution costs → from WHS to delivery area
- Inventory carrying costs → in the WHS
- Handling costs → in the WHS

Objective function
[considering the costs one by one] 𝑇𝐶 = 𝐹 ∗ 𝑑 ∗ 𝑅
NB: we assume that each area is characterised by a specific demand, and it can be fulfilled by more than one
plant at the upstream level, since they are very focused on different specific types of product
We also assume that each area is served by one central warehouse only

- 𝑃𝑇𝐶 = 𝐹 ∗ 𝑅 = ∑𝑝 ∑ℎ(∑𝑗 %𝑝,𝑗 ∗ 𝐷𝑗 ∗ 𝑘ℎ,𝑗 ) ∗ 𝑐𝑝𝑝,ℎ

o F = flow, we should consider the percentage of the general demand coming from different plants
→ %p,j = %Dj produced in plant p and distributed passing through warehouse h
o k = 1 if the connection between the central warehouses and the demand area exists
k = 0 if it does not exist
o cpp,h = fare, which includes also the distances d [€/unit*trip]

- 𝐿𝐷𝐶 = 𝐹 ∗ 𝑑 ∗ 𝑅 = ∑ℎ ∑𝑗 𝐷𝑗 ∗ 𝑘ℎ,𝑗 ∗ 𝑑ℎ,𝑗 ∗ 𝑐𝑠ℎ,𝑗

- 𝐼𝐶𝐶 + 𝐻𝐶 = 𝑐𝑤 ∗ 𝐹 = ∑ℎ ∑𝑗 𝑐𝑤ℎ ∗ 𝐷𝑗 ∗ 𝐾ℎ,𝑗
o cw = cost of warehouse, it includes both the unitary inventory carrying and handling costs

Constraints
- Each area must be served by one warehouse ∑ℎ 𝑘ℎ,𝑗 = 1 ∀𝑗
- If a CW h is in area j, that CW serves the area j 𝑘ℎ,𝑗 = 1 ∀ℎ, 𝑗 if CW belongs to the area
- The number of CWs activated for a certain area should not be high, to not overcrowd a certain set of
areas; this number should be lower than a value n 𝑎ℎ = 𝑚𝑎𝑥𝑗 (𝑘ℎ,𝑗 ) → ∑ℎ 𝑎ℎ ≤ 𝑛

201
 13. Distribution Network Design: design methodology and models – Part 2

Service level constraint

- iso-time curves for each CW [computed through software] → 24h and 48h
- for each CW, we can have different areas:
o Entirely within 24h
o Partially within 24h
o Entirely within 48h
o More than 48h
o [Partially within 48h]
To consider the service level, we will use the boolean variables, considering as 1 the areas which are
consistent with the required service level, and with 0 otherwise

CONCLUSIONS
With this example we faced three different levels of distribution network design
- Rough design of the distribution network, at a strategic level
- Quantitative evaluation of the distribution network, at a concrete level
- Iteration of the model, at a continuous improvement level

This method is useful in addressing a continuous improvement of the network, necessary due to the continuous
changes from both the market and the production sides
D&A → software that computes the centre of gravity of each area, from which they can check if the central
warehouse is in the correct position, or if they have to move the central warehouse
It is something they can address in a dynamic way, making it extreme they could use it every day

6 CONCLUSIONS

SUPPLEMENTARY READINGS
- Ballou, The Network planning process, Business logistics Management, pp.542-603;
- Ballou, Facility Location Decisions, Business logistics Management, pp.483-531;
- Ballou 1995, Logistics Network Design: Modelling and Information Considerations, International journal of
logistics management, volume 6, pp.39-54.

CHALLENGE QUESTIONS
- What are the main macro-phases of the Distribution Network Design process? How would you cope with
them?
- What do the “Network selection matrixes” aim at? Which are their strengths? Which their limitations?
- How would you describe the location problem? And the allocation one?
- What are the main models to cope with them? Which are their pros and cons?
- What are the main techniques to deal with the multi-warehouse location?

202
 14. Logistics Outsourcing

14. Logistics Outsourcing

0 INTRODUCTION
Till now we have understood how to design a distribution network, but we never took into consideration if the
logistic service is directly owned by the company or outsourced to a logistic service provider
→ we should include this variable

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples:
- The main logistics outsourcing strategies
- Structure and actors of the logistics services supply chain
- Pros and cons of logistics outsourcing
- The value creation tree of logistics outsourcing

AGENDA
1. Definitions & Strategies
2. Types of providers and market size
3. Profile of logistics outsourcing activities
4. Value creation, risk & cost and providers strategies
5. Setting outsourcing relationships

Outsourcing of the logistics process; we will address this topic following 3 steps
- Step 1: understanding the definition of outsourcing and what are the main outsourcing strategies (how
can we answer to the problem?) → what and how
- Step 2: contract logistics (services) supply chain outsourced → from who we buy this service
- Step 3: risks & costs and benefits/value → why of logistics outsourcing

1 DEFINITIONS & STRATEGIES

GENERAL OVERVIEW
Outsourcing = the action or practice to obtain goods or services by contract from an outside source

According to the Italian law: ‘‘Outsourcing The outsourcing can be related to different business areas:
consists in the formalization of a contract - Production
through which the company “outsources” to a - Information systems
third party (outsourcer) the conduction of some - Wages and taxes
processes (and activities) necessary to achieve - Logistic
the company’s goals.” - …

Characteristics
- It is a make or buy decision about the logistic process WHAT AND HOW
o Logistic process: there are a lot of activities that can be
outsourced, at different levels/subproblems of the logistic
process of a company (ex: we used the three stacks model in
order to understand them)
→ what part of these subproblems we could outsource?
o Make or buy: they are the two extremes, but there is a huge
spectrum of logistics outsourcing strategies between the full
in-house and the full outsourcing
→ how we can outsource the activities, how we move along
this spectrum?

203
 14. Logistics Outsourcing

- It is a strategic decision, because WHY

o It has a strong impact on the corporate performances (ex: ROA)
Changing from full in-house to full outsourced or going for a solution in the middle change for
sure the cost/revenues structure, or the denominator of the ROA
o Long term impact, since it is not easy to move from one decision to the other
→ long term irreversibility due to very high switching costs
NB: it is always considered as a strategic decision, even if we go for a full in-house strategy; it has the
same impact of decisions taken on the distribution network

- It is based on a trade-off, between WHY

o Value: what is the additional value that a logistic provider can give to
your company, value that you cannot achieve on yourself
o Risks and costs: going for a “buy” solution we have additional costs,
which are for sure transaction costs, and others
The best solution finds a balance between these two components

LOGISTICS OUTSOURCING STRATEGIES

Assets & People → E Activities/functions → E → (E)/P (*) →P →D
What Direct Strategy
Ownership Transport WHS Distribution/logistics Inventory
control of &
of assets management management management management
How manpower Control
Full in-house I I I I I I I
Very uncommon, you think to be very good in both value and costs, while doing other things
Mix E/I E/I E/I E/I I I I
We can decide what we want to outsource, but typically you keep internally the planning and decisions stacks
Full outsourcing E E E E E I/E I
- At the inventory management level, we can find different options, because it is very linked to the production part
- Talking about strategy and control, it is very unlikely to find companies which leave the design level to other companies
Three - E = execution
- I = internally The main difference between a full
stacks - P = planning
- E = externally outsourced and a mix is at the planning level
model - D = design
(*) Logistic management is the integration of transportation and warehousing activities, so it can be considered between the
execution and the planning activities, it is the core of the logistic process

- Strategic outsourcing = full outsource → contract logistic

Outsourcing of the entire logistics process, at least in a region or for a BU, to a few integrated logistics
service providers (i.e. 3PL = third party logistics providers) through long-term contracts
The company and the service providers become strong partners, and the performances depend a lot on
the performance of the other part
It is characterised by:
o High switching costs
o Full exploitation of logistics outsourcing potential benefits
o Control and innovation of the logistics process must be shared
→ you want to exploit the capability of your providers in order to innovate your logistic process

- Commodity outsourcing = all options between full in house and full outsourcing
Outsourcing of single logistics subprocesses, activities, functions or assets to many interchangeable
specialized/functional logistic service providers through short-term contracts or single orders
NB: these providers are different from the ones mentioned before, since they are focused on some
specific activities, not all the process; the integrator is the company itself
It is characterized by:
o Low switching costs
o Exploitation of functional benefits (not full)
o Control and innovation of the logistic process are mainly internal

The common strategy is a combination of these two, according to different regions or part of the supply chain

204
 14. Logistics Outsourcing

PROFILE OF LOGISTICS OUTSOURCING ACTIVITIES

Two main types of logistics outsourcing can be identified:
- Commodity outsourcing: “Tactical” decisions about outsourcing the execution of “traditional” logistics
activities; outsourcing of:
o Full truck loads shipments
o Material handling activities
o Local distribution in a certain geographic area
→ Short-medium term reversibility and ‘‘Local’’ optimisation (both space and time)
… but keeping the logistics process innovation ‘‘levers’’ inside the company

- Strategic outsourcing: “Strategic” decisions about outsourcing the whole logistics process or a large part
of it, i.e. including at least transportation and warehousing activities
Outsourcing of the entire outbound (or inbound) logistics to a single provider (considering in case the
distribution channel specialisation or the geographic area covered)
→ Co-management of the innovation ‘‘levers’’ (between shipper and provider)
… but short-medium term irreversibility (then outsourcing has a strategic importance)

→ Outsourcing of the entire warehouse management

1.3.1 Outsourcing percentage per industry in Italy

The percentage of the outsourced logistics activities (both in

general terms and in terms of Strategic Outsourcing) varies
depending on the industry considered
Source: Observatory of Contract Logistics, companies with
revenues exceeding € 50 million

1.3.2 Innovative services

The logistics outsourcing could include not only “traditional”
logistics activities

EXAMPLES
Strategic outsourcing
- All the distribution process along all the distribution system (from factories’ WHS to customers) in a
region, to one or a few country-focused/specialized providers/3PLs
- Distribution process as before, and also the inventory planning process in the distribution network; all the
keys of the distribution system are given to the 3PLs (ex: McDonald’s)
- Global transportation connecting factories and distribution facilities
- All the logistic process (supply and distribution) of a retailer in a region: connections, POO/POD and
warehouses are all managed by an external party

Commodity outsourcing
- Long-haul primary transportation to many road carries or a route-by-route basis
- Local distribution in an area (TP and redistribution)
- The operations of a WHS to a warehousing service provider
- The management of the workforce/people working in the warehouse to cooperative of workers or staffing
firms

205
 14. Logistics Outsourcing

1.4.1 Real examples

Assets & People → E Activities/functions → E → (E)/P →P →D
What Direct
Ownership Transport WHS Distribution/logistics Inventory Strategy
control of
of assets management management management management & Control
How manpower
AB InBev E E E E E I I
Barilla E E E/I E E I I
Artsana I/E I/E E I I I I
Comifar I I/E I I I I I
Pirelli I/E, depends on the country
Amazon Mainly I, they are moving to full in-house

- AB InBev: distribution in Italy outsourced to two 3PLs (one specialized for HORECA channel, and one for
the retail); we do not put this to a full outsourcing, because the strategy and control is still internal, but it
has the whole distribution process outsourced → strategic outsourcing
- Barilla:
o all the distribution in Italy is outsourced to one 3PL (number one logistic), which is a spin-off of
logistic system of Barilla
NB: it is very common that the origin of a logistic provider is as a spin-off of bigger companies,
which thought that they would be more efficient even providing the logistic service to other
companies
o Assets & people, and activities and functions are externally managed
o Transportation management is partially kept internally
▪ for some internal areas, in order to control tariffs/fares
▪ for international transportation
▪ for raw material transportation: they decided to go greener using train
→ for the majority of elements, Barilla is very similar to AB InBev, despite the incidence of the
transportation management → commodity outsourcing, near to strategic one
- Artsana:
o Assets & people: they are made mainly internally, if we consider the Italian warehouses; for some
subsidiaries worldwide, they of course use external management
o The main commodity that is bought from the market is the transport management (primary
transportation + local distribution) [and some people and building for some subsidiaries]
→ the core here is closer to a full in-house than a full-outsource → commodity outsourcing
- Comifar: local distribution is its core; it is all managed internally despite the fact that they do not directly
employ the drivers, they are typically independent → very close to full in-house
- Pirelli: it has a mixed solution, applying different approaches in different regions
→ strategic outsourcing approach in many countries (Europe, US), commodity outsourcing approach in
others [more selective that the Barilla’s one]
The decision depends on the maturity of the market, on the Pirelli presence and often they do not go for
outsourcing because they do not want to lose control on the logistic process
- Amazon: it is moving towards the full in-house direction, considering the last 20 years
They pick a provider when they think that is better in performances; otherwise, if service providers could
not follow the peaks in Amazon market, they do it in-house, they build internal competences

206
 14. Logistics Outsourcing

OTHER DEFINITIONS
1.5.1 General drivers 1.5.2 Core competencies
Generally speaking, in order to have a “Core competence” in any given area
(e.g. logistics), three things are necessary
- Expertise
- Ability to invest
- Strategic fit
The absence of any one or more of
these may suggest that outsourcing is
appropriate

1.5.3 Strategic fit

Companies outsource “non-core-competency” activities (i.e. those activities where they do not excel), in order to
focus on their "core competencies" (what they do well) TRUE

…but a misleading understanding of this concept is…

Companies outsource only non-key activities in their competitive strategy (i.e. that do not contribute to achieve a
competitive advantage) FALSE

…and in logistics terms…

Companies outsourcing logistics activities do not consider logistics as a source of competitive advantage FALSE

- A competitive strategy is based on both ‘‘core competencies’’ (i.e. distinctive skills) and other important
skills. The latter could be outsourced
- Also relational skills (i.e. based on the value stemming from the relationship) could be very important
- Contract Logistics Industry is not a ‘‘perfect’’ market. Since providers differ in terms of performances, the
provider selection could be a source of competitive advantage (disadvantage)

Logistics Outsourcing is a possible competitive strategy whose benefits/advantages must be “assessed”

Logistics Outsourcing should be adopted when:
- Companies do not have internal strong logistics competencies
- Third parties could be able to perform logistics activities in a "better" way (e.g. achieving economy of
scale/scope)
- “Costs of relationships” (e.g. integration costs for strategic and operational planning) and “risks” (e.g.
dependence on a third party) do not overcome the benefits

2 TYPES OF PROVIDERS AND MARKET SIZE

SUPPLY CHAIN OF LOGISTIC SERVICES

At the final layer we have the final customer, which is buyer of logistic services
Behind this, we have a very long supply chain
- At the beginning, we have the providers of assets and people, so of the production factors, the very basis
of the logistic process, the elementary factors you need to provide logistics services
- Then we can have the functional services providers, which provide basic activities/processes
(transportation, warehousing…)
- At the third level, we have integrated logistics providers, also called 3PL or 4PL (if they work also in the
planning part of the logistic process)

207
 14. Logistics Outsourcing

- Strategic outsourcing: final customer buys from the integrated logistic providers, who buy a lot from the
functional services providers and the providers of assets and people: integrated logistic providers do not
do all internally → we have a lot of transactions, considering the whole supply chain
- Commodity outsourcing: the final customer buys directly from providers of assets and people and from
functional services providers

The conclusion is that there are a lot of transaction within the supply chain
- Upper arrows are typical of the commodity outsourcing
- Lower arrows are typical of the strategical outsourcing

Moreover, we have also different connections within a single provider

ex: some functional services provided by a specific company are interrelated and need to be integrated with other
functional services provided by another specialized company

EXAMPLES OF DIFFERENT PROVIDERS

Providers of assets/workforce (production factors)
- Real estate developers
- Equipment
- ICT services
- Temporary jobs/staffing companies
NB: they can have different types of Business Model → they can sell or rent both products and services

Functional services providers

- Transport companies
o road carriers
o railway/air/ship companies
o intermodal operators → integration of different types of transportation modes
o couriers/express couriers → specialization in small parcels transportation
NB: express couriers are different from road carriers, since the first collect many parcels and then deliver
wherever you want (passing through many transit points), while road carriers are for short tracks
- Freight forwarders (focus on import/export services, intercontinental transportations)
- Local distribution, specialized in distribution in a specific area
- Warehousing/handling operators
There are tens of thousands of different possibilities of functional services providers, very big magnitude
For this reason, we are talking about commodity outsourcing, we are buying some ingredients of the overall
logistic process; we have the possibility to easily change the providers, if the switching costs are not so high

Providers of integrated logistics services (3PL/4PL)

They are considering also part of the planning activities of the logistic process; the difference is that the 4PL
considers all the planning activities, so it has a lot of experts, for example in redesign the network
If you want to outsource you have to choose if you want a specialized provider or a more general and global one
- Global players, at least at a continental level; they are not a lot → 15-20 overall
Their force is that they have a wide range in product categories and industries, they are asset light, they
tend to offer a broader category of services (VAS Value Added Services like final assembly, repackaging,
after sales, customer clearance, tracking and tracing, etc…)
ex: Ceva, DHL, XPO, Kuehne Nagel, DSV, DB Schenker, Geodis,…

- Country/region/channel/product… specific players → 5-10 players for a specific task (ex: food in Italy)
They are more local, so they can have a more capillary network, they are more focused, more expertise
on a certain area, they have a higher assets structure, they usually own the assets, in order to have higher
control of performances (asset heavy)

208
 14. Logistics Outsourcing

2.2.1 Third Party Logistics Providers

Furthermore, 3PL providers differ from each other
depending on their strategy in terms of services
provided

Four axes:
- Asset ownership
- Service
- Geography
- Product categories/industry

Gradual shift from executors to

planners of the physical and
informative flow

From basic to advanced services

2.2.2 The Contract Logistics Industry 2.2.3 Contract Logistics Net Revenue
The ‘‘Contract Logistics Industry’’ is the set of players Contract Logistics Net Revenue is: “the revenue of the
offering logistics services Contract Logistics service providers, without supply
chain internal trade”

209
 14. Logistics Outsourcing

CLASSIFICATION
- Road carriers – large companies
Road freight transport operators, mainly FTL mode
‘‘Small independent carriers’’: small fleet owners, even just 1 or 2 trucks (usually self-employed drivers)
- Road carriers – small companies
Operators providing parcels and/or envelops delivery services, mainly LTL mode
- Couriers/Express Couriers
Operators providing mainly storage and material handling activities
- Intermodal terminal Operators
- Warehouse Operators
This category includes also porters cooperatives (specifically Italian reality)
- Railway carriers and Intermodal rail-road Transport Operators
Railway carriers and Intermodal railroad Transport Operators
- Freight Forwarders
International transport companies that arrange and combine all modes of transportation
- Third Party Logistics Providers
Providers of integrated logistics services (warehousing and road transportation activities, mainly at a
national level)
Innovative services (provided mainly by big players)

EXAMPLE OF LOGISTIC SERVICES SUPPLY CHAIN IN ITALY

The typical final customer of the logistic services supply
chain spends more or less 110 bn €; if we would like to
split this value (not in terms of quantities):
- 56% is managed in-house
- 9% is bought through strategic outsourcing
→ mainly related to 3PL/4PL
- 35% is bought through a commodity
outsourcing approach
→ mainly related to functional services
providers; a small part is related to 3PL/4PL if
they decide to provide also functional services
through an internal division

The sum of the latter two (44%) could be considered as the value of the market of logistic services
This example is useful to understand the incidence of outsourcing of logistics (44%), that in other counties could
be higher (ex: UK = 70%), and we also understand that the percentage for strategic outsourcing is very low, so
there is huge space for the increase of 3PL/4PL role, since it is still a minority

210
 14. Logistics Outsourcing

3 VALUE CREATION, RISK & COST AND PROVIDERS’ STRATEGIES

LOGISTICS COST STRUCTURE

3PL providers create value for their customers pursuing several business
strategies that affect the logistics cost structure of the customers
This ability can be used to provide a higher service level at the same costs
or the same service level at lower costs (or a combination of the two
previous cases)
The way this ability is used depends on the logistics strategy of the
customers (i.e. what the customer wants from the logistics process)

NB: we are talking about potential sources of value, but this does not mean that going for outsourcing will for
sure provide these benefits → we should make some analysis in order to compare them with our current situation

We will talk about two different categories to create benefits:

- Direct/concrete value creation, and then value sharing
- Shift of costs and complexity from a player to another one

FACTORS FOR CREATE VALUE

A 3PL provider can create value through 4 different factors:
- Economies of scale
- Learning economies
- Ability to introduce innovation
- Flow management capabilities

3.2.1 Value Tree

We can create value by increasing the expected ROA,
through 3 main possible paths
- working on the ROA numerator, moving the cost-
to-serve curve to a more favourable position in that
curve
- decreasing the capital invested (ROA denominator)
- reducing the expected standard deviation of the
ROA, so its variability/uncertainty

We will address the main ways to going through these 3

paths, but not in detail, since they can be exploited in
different ways

Increase the ROA numerator -> moving the cost-to-serve curve:

- Economies of scale (mainly relates to an increase in volumes): going for outsourcing, the provider is able
to aggregate volumes from many customers → increase in economies of scale
o increase in bargaining power (ex: buy production factors at a lower cost) → shift of value
o increase in productivity (mainly related to a better use of resources) → value creation
▪ increase in critical mass/density, increasing density of activities in a portion of the
network (ex: increase in truck saturation or flow balance)
▪ economies of scope, with sharing resources (ex: multi-customer WHS, Pirelli-Goodyear)
▪ flexibility (ex: dynamic re-allocation of workforce between sites)
- Economies of learning (mainly related to an increase in knowledge)
o Experience, related to an increase of cumulative volumes which gives an increase in productivity
o Replication and cross-fertilization, idea of contamination of “good practices” between sectors
- Innovation management
o Process new picking processes
o Assets ICT systems, automation
o People talent development, workforce training

211
 14. Logistics Outsourcing

Reducing the ROA denominator (capital invested):

- Assets transfer to 3PLs/logistic providers (ex: investment in technology, buildings, equipment)
- Cash-flow optimization:
o delivery accuracy and collection of payments
o inventory control, like on perishable products, or lead times

Reduction of the ROA variability:

- Resilience, short-term flexibility: how the system is able to react/adapt without problems of continuality
ex: logistics network visibility increase, working on “control towers”, where data are constantly updated
- Reconfigurability, long-term flexibility: how the system is able to adapt itself to a redesign of the network
- New markets/channels: you can reduce the time to market to enter in a new market (time to market);
you can increase the reversibility of entering in a new market that you still do not know

3.2.2 Economies of scale

3.2.3 Learning economies 3.2.4 Ability to introduce innovation

3.2.5 Flow management capabilities

NB: For achieving each factor, a 3PL provider can

identify and pursue different business strategies

212
 14. Logistics Outsourcing

RISKS AND COSTS

Costs:
- Transactional costs: relationships costs, joint management and control, ICT/process integration
- Switching costs: costs to switch from one logistics service provider to another one; the highest costs are
when you switch from outsourcing to go back to in-house logistics process

Risks:
- Dependency on service providers performances
- Continuity risks: you have a dependence on providers continuity of operations; if they have an
interruption somewhere in the network you pay for that interruption
- Loss of competences: risk of progressively losing touch and internal competences
- Loss of control: risk of losing visibility on the main operative KPIs

The decision about the logistic outsourcing needs to be

taken with a balance between the potential value, that a
third party can create compared to the actual one achieved
with full in-house approach, and the risks and costs

There are contextual variables that affects also this decision:

1. Maturity of the logistics services providers → more mature providers maybe can give you more value
2. Internal competences that companies have → if you have more logistic competences, maybe you need
less 3rd party providers
3. Volumes (in the full in-house approach) → if a company already has big volumes, outsourcing may not
give lot of benefits

PROVIDERS’ STRATEGIC APPROACHES

Based on the most important factors,

the main strategic approaches of a 3PL
provider are:

Volume-oriented
‘‘We manage our customers exploiting synergies (mainly in transportation) in order to both optimize the routing and
maximize the quantity transported. Routing is carried out in dynamic way (delivery times are fixed, routes change) […] We try
to exploit the backhauls in order to pick up the goods from another supplier in the same routing […] Based on the time series
of the demand that we have for the majority of our customers, we can make accurate aggregate forecasts. This let us plan
and manage our resources on a weekly base”.

Process-oriented
‘‘We try to be efficient managing the synergies between customers in the same supply chain (e.g. transit points in order to
both arrange multi-drop deliveries and conduct cross docking activities” […] In addition to the warehouse performance
measurement, we conduct a benchmark between all the warehouses in order to identify the best in class and transfer the
best-practices to the other facilities […] The transfer of the best practices is a structured process. Periodic meetings with all
the operational mangers are arranged aiming to: 1) identify the best practice, 2) assess the compatibility and check the
applicability to the other facilities and processes”.

Innovation-oriented
‘‘We try to develop problem solving abilities through a structured process in which errors are analyzed (e.g. at the end of the
month, based on the classification of the errors we have developed, we compile a report in order to both identity the error
that impact the performances the most and implement corrective actions to solve it) […] Moreover we promote different
initiatives to involve all the people in the company (i.e. from top managers to operatives) in order to push them propose
ideas to improve customer service level (e.g. every 15 days we make “operative meetings” in which the “as-is” activities are
analyzed and possible improvements are proposed) […] Partnerships with technology providers are essential in order to
innovate processes. Thanks to one of these partnerships, we can trace the goods all along the supply chain’’

213
 14. Logistics Outsourcing

4 SETTING OUTSOURCING RELATIONSHIPS

MAIN IMPLICATIONS
For a shipper, the logistics outsourcing entails:
- Reconfiguration of the company staff
- Reorganisation of the company areas
- Transferring of goods (in case of warehouse change)
- Development of interfaces with the information systems of the providers
- Set of an outsourcing agreement (with the definition of the rates and the bonus/malus systems)

4.1.1 ‘‘Baby sitter’’ syndrome

The main features of a logistics service provider that has to be selected for a Strategic Outsourcing are:
- Beautiful (infrastructure, technologies)
- Forever young (innovation)
- Smart (quality of management; planning)
- Well-bred (stable shareholders)
- Rich (corporation stock, profitability)
- Honest (audited financial statements; transparency)
- Loyal (must not have preferences for competitors)

OUTSOURCING PROCESS FRAMEWORK

4.2.1 Step 1: AS-IS situation assessment

The AS-IS situation assessment is
based on several factors
characterising the logistics
process

214
 14. Logistics Outsourcing

Optimum conditions (for outsourcing)

- Strong variability of demand (seasonality, recurrences, ….)
- Market extensiveness
- Limited flows per customer combined with:
o Need for high frequency and/or on-time deliveries
o Low value per unit delivered
o Dispersion of final customers
- Financial difficulties
- Industry “Dynamism” (-> competition)
- A high number of Logistics Service Providers (LSPs) that are able to satisfy shippers needs at competitive
prices

Potential Risks
- Loss of business opportunities
- Loss of specialised skills
- Lower performance control
- Less timeliness for reactions in case of anomalies or unusual contingencies (higher potential risks related
to “cohabitation”)
- Difficult reversibility (provider switching or back sourcing)
- Low service level

4.2.2 Step 2: Competitive advantages related to logistics outsourcing

Companies with a Strategic
Outsourcing approach seem
to give more importance (*)
– after cost factors and
flexibility increase – to
customer service level
improvement and
innovation capability
improvement
(*) by asking respondents to assign
a percentage from 1 to 100%
depending on the importance
assigned to each competitive
advantage

4.2.3 Step 3: LSP assessment and selection

215
 14. Logistics Outsourcing

5 CONCLUSIONS
Nowadays, logistics outsourcing is a key competitive strategy, whose success depends, not only on the providers
selection, but also on the relationship setting and management
A company must assess how the logistics outsourcing can create value and then adapt the possible solutions to its
own reality
The choice can change over time depending on the corporate strategies
FIAT Spare part logistics
Summarizing, Logistics Outsourcing is an opportunity:
- To be discussed
- To be managed
- To be controlled

SUPPLEMENTARY READINGS
- Aghazadeh, S. M. (2003), “How to Choose an Effective Third Party Logistics Provider”, Management
Research News, Vol. 26, No. 7, pp. 50-58.
- Belcourt, M. (2006), “Outsourcing – The benefits and the risks”, Human Resource Management Review,
Vol. 16, pp. 269-279.
- Hertz, S. and Alfredsson, M. (2003), “Strategic development of third party logistics providers”, Industrial
Marketing Management, Vol. 32, pp. 139-149.
- Langley, C. J. and Capgemini (2014), “2014 Third-Party Logistics Study: The State of Logistics Outsourcing.
Results and Findings of the 19th Annual Study”, available on-line at: http://www.3plstudy.com/.
- Wong, C. Y. and Karia, N. (2010), “Explaining the competitive advantage of logistics service providers: A
resource-based view approach”, The International Journal of Production Economics, Vol. 128, pp. 51-76.
- Zacharia, Z. G., Sanders, N. R. and Nix, N. W. (2011), The Emerging Role of the Third Party Logistics
Provider (3PL) as an Orchestrator”, Journal of Business Logistics, Vol. 32, No. 1, pp. 40-54

CHALLENGE QUESTIONS
- What is the core business of a Logistics Service Provider (LSP)? Which types of LSP can be identified? How
can these be classified? According to which dimensions?
- What is the possible value of Logistics Outsourcing?
- What are the main differences between Commodity Outsourcing and Strategic Outsourcing?
- What are the elements a customer should consider in the logistics outsourcing assessment?
- Assuming you are a small company, focussed on the domestic market with a single warehouse (factory
warehouse), how would you define your logistics outsourcing strategy?

216
 15. Advanced Inventory Planning

15. Advanced Inventory Planning

[the following chapter is composed only by notes; after that, there will be supporting slides]

1 WHAT IS ADVANCED INVENTORY PLANNING?

It is an extension of traditional/standard inventory planning on different dimensions:
- Number of items compared to traditional logistic system, we consider much more items
→ from a single-item context to a multi-items context
- Number of warehouses
→ from a single-warehouse context to the entire distribution network
- Service level
→ from the capacity to avoid stockout to the order cycle time
The main difference here is the much more complex context in which the advanced inventory planning operates,
that is closer to the reality

1.1.1 What is the meaning of inventory planning?

The inventory planning basically consists They answer to the request of how much, when and where
of two main activities: we have to order
- Inventory sizing For each product we are managing, we have to decide all
- Flow management these three points

2 RECAP ABOUT INVENTORY PLANNING

WHAT IS THE ROLE OF INVENTORIES WITHIN THE SUPPLY CHAIN?

We keep inventories with a double role
- Lever of efficiency
o Transport related inventories: the presence of inventories can reduce the transportation costs,
because we can fulfil a non-completely full truck
ex: if we order something from the supplier, maybe we are not ordering enough goods to reach
the saturation of the truck, so we spend a lot of money in transportation (the fare will be high); if
we decide to order a bigger quantity to saturate the truck, we will have less transportation costs,
but of course we will have more inventories, because we will do not immediately use all the
goods
→ trade-off with transportation costs
o Speculative inventories: the cost of purchasing the goods can be really different, so we can buy
them when their price is very low; we try to exploit the price, but of course then we have the
inventory carrying costs
→ trade off with cost of the goods
o Lots related inventories, considered in order to reduce the production costs, for example
producing in lots to reduce the cost of setup, but if the lots are very big, we will have inventories
→ trade off with production costs
o Seasonal inventory, we use the manpower all the year at its full capacity, even if the demand is
concentrated in few months of the year, instead of having an extra-capacity in these periods,
since this second option is not convenient, so we keep inventories to cope with the picks of the
demand
→ trade off with production costs

- Lever of effectiveness
o Uncertainty related inventory we will see that are basically the safety stocks
o Quality related inventory
ex: wine and cheese → if you want to have a product of a better quality you have to keep a
certain product for month, you can sell the product at a higher price

217
 15. Advanced Inventory Planning

WHAT ARE THE MAIN TYPES OF INVENTORY WITHIN THE SUPPLY CHAIN?
- Inventories in the warehouses
o cycle-stocks kept in order to decouple the different levels of the supply chain
o safety-stocks kept in order to cope with the uncertainty and unpredictability of the
demand and the replenishment LT
- Inventories in the transportation means
o in-transit stocks they are the inventories present on the transportation mean for the
transportation delivery time

2.2.1 What are the main drivers to size these stocks?

- In-transit stocks 𝐼𝑇𝑆 = 𝐹 ∗ 𝐿𝑇 they depend on:
o Flow F → demand (for example daily, pc/day)
o Lead time LT → transportation lead time (for example in days)
NB: they must be expressed with the same unit of measure

𝐷∗𝑇
- Cycle stocks 𝐶𝑆 = 2
[considering a periodic review model]
𝑄
𝐶𝑆 = [considering a fixed order point model]
2
→ they depend on demand D and order frequency T

ex: assuming to have suppliers -> central warehouse -> customer,

and a demand of 2000 PL/month

2000
o Case A, T = 1 month 𝐶𝑆 = 2
= 1000
1000
o Case B, T = 2 weeks 𝐶𝑆 = 2
= 500

- Safety stocks 𝑆𝑆 = 𝑘 ∗ √(𝑇 + 𝐿𝑇) ∗ 𝜎𝐷2 + 𝐸𝐷 2 ∗ 𝜎𝐿𝑇

2

o Demand D
o Variability of the demand σD
o Replenishment lead time, the one of the supplier LT
o Variability of the lead time σLT
o Service level
▪ Probability to avoid stock out P(DLT < OP) [OP = order point]
▪ Item fill rate IFR

WHAT ARE THE MAIN KPIS OF INVENTORY PLANNING?

2.3.1 Efficiency
- Average inventory level → weighted average of all the quantities
∑𝑖 𝑞𝑖 ∗𝑇𝑖
𝐴𝐼𝐿 = ∑𝑖 𝑇𝑖
the smaller the quantities, the more efficient we are
It is the first way we have to measure for the efficiency
- Inventory turnover ratio
𝐹
𝐼𝑇𝑅𝑖,𝑇 = 𝑖,𝑇 [F = outgoing flow; i = item; T = period]
𝐴𝐼𝐿𝑖,𝑇
NB: we are considering the outgoing flow, which is not necessarily equal to the demand
ex: 𝐹𝐼,𝑇 = 2000 𝑃𝐿𝑠/𝑚𝑜𝑛𝑡ℎ 𝑇 = 1 𝑚𝑜𝑛𝑡ℎ 𝐴𝐼𝐿𝑖,𝑇 = 500 𝑃𝐿𝑠
2000
→ 𝐼𝑇𝑅𝑖,𝑇 = 500 = 4 𝑡𝑖𝑚𝑒𝑠/𝑚𝑜𝑛𝑡ℎ
The bigger the ITR, the more efficient we are, because it means that we are satisfying a certain level of
demand with a small level of inventories; the lower it is, the less efficient we are, because we are
satisfying the same level of demand with highest level of inventories
NB: it is really important because it finds a relation between the demand and the inventory level, while
the AIL does not consider the demand level, but only the inventory level

218
 15. Advanced Inventory Planning

- Inventory carrying costs 𝐼𝐶𝐶𝑖,𝑇 = 𝐴𝐼𝐿𝑖,𝑇 ∗ 𝑉𝑖 ∗ %ℎ𝑐𝑖

o Vi = value of item i [€/unit], the money we spend in order to keep the inventory in the warehouse;
it could be composed by different things:
▪ Purchasing + transportation costs if we acquire the item from a third party
▪ Production costs if we internally produce it
o %hci = percentage of holding costs of item i [%/year, if T = year]
It considers:
▪ Storage
▪ Cost of capital
▪ Obsolescence
▪ Depreciation
→ 𝑖𝑐𝑐𝑢𝑛𝑖𝑡 = 𝑉𝑖 ∗ %ℎ𝑐𝑖 unitary inventory carrying cost

2.3.2 Effectiveness
How good we are in fulfil the customer demand
#𝑢𝑛𝑖𝑡𝑠 𝑓𝑢𝑙𝑓𝑖𝑙𝑙𝑒𝑑 𝑓𝑟𝑜𝑚 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑖𝑒𝑠
- Item fill rate 𝐼𝐹𝑅 = #𝑢𝑛𝑖𝑡𝑠 𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑒𝑑
NB: considering the number of orders instead of the units is a measure for the service level,
the order fill rate

WHAT ARE THE MAIN INVENTORY PLANNING MODELS?

…and how can we cluster them?
- Pull inventory model
- Push inventory model

2.4.1 Pull inventory model -> independent work

The flow of this system is activated by the downstream node, by the request of the previous nodes
In this way, we can see the demand at each level, each node is working independently from the other ones,
forecast its own demand, minimizing its own relevant costs

2.4.2 Push inventory model -> dependent work

The final demand is forecast, and based on it all the other demands are computed, and then there is one only
independent demand, and all the rest of the flow is based on the first node; all the nodes know the overall
demand, and the main activities are made concurrently for all the nodes
- MRP material requirement planning
- DRP distribution requirement planning

219
 15. Advanced Inventory Planning

WHAT ARE THE MAIN PULL INVENTORY MODELS?

[We will focus only on the pull system, in particular on 2 main models]
- Re-order point model
- Periodic review model
We will answer to the main requests: how much and when we have to order; we should consider also the where
and what, but we assume that we are going to order 1 item from one specific warehouse

2.5.1 Re-order point model

- How much → fixed quantity Q
- When → when quantity in the warehouse reaches the
Order Point (OP)
When we order, we have to wait the replenishment LT of the supplier to
receive our product, so in this period of time we should have enough
items to cope with the demand

We add a layer of safety stocks in order to cope with

- the variability of the demand → if we have higher demand than
the forecast one, we have enough product to satisfy the demand
during the LT
- the variability of the replenishment LT of the supplier → if it is
late, we have still enough goods to satisfy the demand

𝑄
𝐴𝐼𝐿 = + 𝑆𝑆
2

2.5.2 Periodic review model

- How much → demand during T on average
We can compute the quantity we have to order as
𝑄 = 𝐴𝑇 − 𝐼𝐿(𝑡) = 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡𝑎𝑟𝑔𝑒𝑡 – 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑦 𝑙𝑒𝑣𝑒𝑙 𝑖𝑛 𝑡ℎ𝑎𝑡 𝑚𝑜𝑚𝑒𝑛𝑡
- When → every period T (period between two points of
review), it is fixed

We can set the availability target as 𝐴𝑇 = 𝑆𝑆 + 𝐷𝐿𝑇 + 𝐷𝑇

𝑄 = 𝑆𝑆 + 𝐷𝑇+𝐿𝑇 − 𝐼𝐿(𝑡) ≈ 𝐷𝑇 [≈ on average]

We add a layer of SS for the same erasons of the reorder point model,
but in this case we should consider σ of the demand during both the
period T and the LT

2.5.3 What are their main variables/parameters?

Reorder Point model
- Q = quantity to be ordered (fixed)
- OP = order point (variable)
→ 𝑂𝑃 = 𝑆𝑆 + 𝐷𝐿𝑇

Periodic review model

- T = time between 2 consecutive orders (fixed)
- AT = availability target (variable, set when we issue the order, to decide
the quantity to be ordered)
→ 𝐴𝑇 = 𝑆𝑆 + 𝐷𝐿𝑇 + 𝐷𝑇

220
 15. Advanced Inventory Planning

HOW CAN THESE VALUES BE OPTIMIZED?

How much? When? We want to optimize
Re-order point 𝑄 𝑂𝑃 = 𝑆𝑆 + 𝐷𝐿𝑇 these two models, and
𝑄(𝑡) = 𝐴𝑇 − 𝐼𝐿(𝑡) it means to minimize
Periodic review 𝑇
𝐴𝑇 = 𝑆𝑆 + 𝐷𝑇+𝐿𝑇 some costs

2.6.1 Re-order point

First, we have to determine the optimal value of Q, the one which minimizes the sum of the relevant costs, which
are the ones that have an impact on our decision, so on how much we have to order, for example:
- Cost of goods 𝐴𝐶𝐺 = 𝐷𝑦 [𝑝𝑐/𝑦𝑒𝑎𝑟] ∗ 𝑃 [€/𝑝𝑐]
[NB: assumption, the price is constant, it does not vary according to the quantity, there are no discounts]
𝐷𝑦
- Procurement costs 𝐴𝑃𝐶 = #𝑜𝑟𝑑𝑒𝑟𝑠 [𝑜𝑟𝑑𝑒𝑟𝑠/𝑦𝑒𝑎𝑟] ∗ 𝑢𝑝𝑐 [€/𝑜𝑟𝑑𝑒𝑟] = ∗ 𝑢𝑝𝑐 𝑄
𝑄
- Inventory carrying costs 𝐴𝐼𝐶𝐶 = 𝐶𝑆 [𝑝𝑐] ∗ 𝑉 [€/𝑝𝑐] ∗ %ℎ𝑐 [%/𝑦𝑒𝑎𝑟] = 2
∗ 𝑉 ∗ %ℎ𝑐
[NB: assumption, no safety stocks]
- (Transportation costs)
- …

→ 𝐴𝑇𝐶 = 𝐴𝐶𝐺 + 𝐴𝑃𝐶 + 𝐴𝐼𝐶𝐶

𝜕
𝐸𝑂𝑄 = 𝑚𝑖𝑛 𝐴𝑇𝐶 → (𝐸𝑂𝑄) =0
𝜕𝑄
∂
→ (𝐴𝐶𝐺 + 𝐴𝑃𝐶 + 𝐴𝐼𝐶𝐶) = 0
∂𝑄
𝐷𝑦 1
→ 0 − 𝑄2 ∗ 𝑢𝑝𝑐 + 2 ∗ 𝑉 ∗ %ℎ𝑐 = 0
2∗𝑢𝑝𝑐∗𝐷𝑦
→ 𝑄 = 𝐸𝑂𝑄 = √ %ℎ𝑐∗𝑉

- The bigger the unitary procurement costs, the better is to increase the quantity we order
- The bigger the percentage of holding costs, the better is to reduce the quantity we order

This computation could be of course more complicated, for example

- If the cost of the goods is not constant (ex: we have quantity discounts)
- If we consider also transportation costs, that depends on the quantity

2.6.2 Periodic review model

We should first compute a theoretical value of EOQ,
and then we have to find the optimal T
ex: EOQ = 2000 pieces Dw = 1000 pieces/week
𝐸𝑂𝑄 2000
→ 𝑇 = 𝐷 = 1000 = 2 𝑤𝑒𝑒𝑘𝑠
𝑤

NB: we compute a theoretical value of EOQ and not an optimal one

because every time we reach the moment for issuing the order, the
quantity to be ordered varies according to the demand during T

3 SAFETY STOCKS
At this point, we optimized both Q for re-order point model and T for the periodic review model, which are fixed;
now we have to address the variables and find out what we can optimize
Order point OP depends on Availability target AT depends on
- DLT, which cannot be optimized, since the demand is - DT+LT, which cannot be optimized: the DLT for the
an exogenous factor, and the LT depends on our same reasons as before, the DT because we already
supplier determined the optimal T (so it is fixed)
- SS, which is our decision, so can be optimized - SS, which is our decision, so can be optimized
→ Optimal SS is the minimum quantity that satisfies the service level required

221
 15. Advanced Inventory Planning

HOW CAN SAFETY STOCKS BE DETERMINED?

3.1.1 Re-order point case - example
𝑆𝑆 = 𝑘 ∗ σ𝐷,𝐿𝑇
- 𝜎𝐷,𝐿𝑇 = √𝐿𝑇 ∗ 𝜎𝐷2 + 𝐸𝐷 2 ∗ 𝜎𝐿𝑇 2
It combines demand, lead time and their variability/unpredictability
- k = service level, which can be defined through two different ways:
o Probability to avoid stock out during LT
#𝑢𝑛𝑖𝑡𝑠 𝑓𝑢𝑙𝑓𝑖𝑙𝑙𝑒𝑑 𝑓𝑟𝑜𝑚 𝑖𝑛𝑣𝑒𝑛𝑡𝑜𝑟𝑖𝑒𝑠
o Item Fill Rate 𝐼𝐹𝑅 = #𝑢𝑛𝑖𝑡𝑠 𝑟𝑒𝑞𝑢𝑒𝑠𝑡𝑒𝑑
NB: these two ways are very different; we cannot compute the same k with them

Example
DLT 0 1 2 3 4 5 6
P(DLT) 5% 10% 20% 25% 20% 15% 5%
LT = 1 month, Q = 4 units

Questions:
- determine the re-order point to have a probability to avoid stock out during the LT at least 90%
- determine the re-order point to have an item fill rate at least 90%

Case A
If the demand is higher than expectations, and higher than the order point itself, we will have a stockout
𝐸𝐷𝐿𝑇 = ∑ 𝐷𝐿𝑇𝑖 ∗ 𝑃𝑖 = 5% ∗ 0 + 10% ∗ 1 + ⋯ + 5% ∗ 6 = 3,1 𝑢𝑛𝑖𝑡𝑠
Since the expected demand (on average) is around 3 units, we start with a ROP = 3

- 𝑅𝑂𝑃 = 3 → 𝑃(𝐷𝐿𝑇 ≤ 3) = 5% + 10% + 20% + 25% = 60%

With a reorder point equal to 3 we have a probability to avoid stockout of 60%
- 𝑅𝑂𝑃 = 4 → 𝑃(𝐷𝐿𝑇 ≤ 4) = 60% + 20% = 80%
- 𝑅𝑂𝑃 = 5 → 𝑃(𝐷𝐿𝑇 ≤ 5) = 80% + 15% = 𝟗𝟓%
- 𝑅𝑂𝑃 = 6 → 𝑃(𝐷𝐿𝑇 ≤ 4) = 95% + 5% = 100%
→ ROP = 5, for having a probability higher than 90% to avoid stockout we need to have a ROP equals to 5

Case B
USOY = number of units in stockout during the year

𝐷𝑌 −𝑈𝑆𝑂𝑦 𝑈𝑆𝑂𝑦 𝑈𝑆𝑂𝐿𝑇 𝐷𝑦

𝐼𝐹𝑅 = =1− =1− and 𝑈𝑆𝑂𝑦 = 𝑈𝑆𝑂𝐿𝑇 ∗ #𝑜𝑟𝑑𝑒𝑟𝑠/𝑦𝑒𝑎𝑟 = 𝑈𝑆𝑂𝐿𝑇 ∗
𝐷𝑦 𝐷𝑦 𝑄 𝑄
𝑈𝑆𝑂𝐿𝑇
→ IFR = 1 − 𝑄

(4 − 3) ∗ 20% + (5 − 3) ∗ 15% + (6 − 3) ∗ 5%
𝐼𝐹𝑅(𝑅𝑂𝑃=3) = 1 − = 83,7%
4
If the ROP = 3, we will go in stock out if they order 4, 5 or 6 units, respectively of (4-3), (5-3) and (6-3) units; we
have to weight this by the different probabilities of having that specific demand

(5−4)∗15%+(6−4)∗5%
- 𝐼𝐹𝑅(𝑅𝑂𝑃=4) = 1 − 4
= 93,75%
(6−5)∗5%
- 𝐼𝐹𝑅(𝑅𝑂𝑃=5) = 1 − 4
= 98,75%
- 𝐼𝐹𝑅(𝑅𝑂𝑃=6) = 100%

Comparison
𝑃(𝐷𝐿𝑇 ≤ 𝑅𝑂𝑃) 𝐼𝐹𝑅 We are talking about two different performances of the service level
3 60% 83,7% Which is the best one?
4 80% 93,75%
5 95% 98,75%
6 100% 100%

222
 15. Advanced Inventory Planning

3.1.2 Re-order point case - General case

Probability to avoid the stock out

𝐷𝐿𝑇 − 𝐸𝐷𝐿𝑇
𝑃(𝐷𝐿𝑇 < 𝑅𝑂𝑃) = 𝑃(𝐷𝐿𝑇 < 𝐸𝐷𝐿𝑇 + 𝑆𝑆) = 𝑃(𝐷𝐿𝑇 < 𝐸𝐷𝐿𝑇 + 𝑘𝜎𝐷,𝐿𝑇 ) = 𝑃 ( < 𝑘) = 𝑃(𝑍 < 𝑘)
𝜎𝐷,𝐿𝑇

We are not going to have stockout if the expected demand during the lead
time will be lower than the order point, which is equal to the expected
demand during the LT and the safety stocks
𝐷𝐿𝑇 −𝐸𝐷𝐿𝑇
𝜎
= 𝑍 → normal distribution
𝐷,𝐿𝑇

Having the probability we have to achieve we can find the correspondent

value of k from the table of Normal distribution

Item fill rate

𝑈𝑆𝑂𝐿𝑇
𝐼𝐹𝑅 = 1 −
𝑄
NB: we are in a continuous domain, so we will pass through an integral
∑ 𝑈𝑆𝑂𝐿𝑇 (𝑖𝑓 𝑤𝑒 ℎ𝑎𝑣𝑒 𝑎 𝐷 ∗ 𝑃(ℎ𝑎𝑣𝑒 𝑡ℎ𝑎𝑡 𝐷
𝐿𝑇𝑖 ) 𝐿𝑇𝑖 )
∞
→ 𝑈𝑆𝑂𝐿𝑇 = ∫𝑅𝑂𝑃[𝑌 − (𝐸𝐷𝐿𝑇 + 𝑆𝑆)] ∗ 𝑓(𝑦) 𝑑𝑦 [𝑌 = 𝐷𝐿𝑇 ]

Change of variable:
𝑌−𝐸𝐷𝐿𝑇
𝑍= → 𝑌 = 𝑍 ∗ σ𝐷,𝐿𝑇 + 𝐸𝐷𝐿𝑇
σ𝐷,𝐿𝑇
∞
→ 𝑈𝑆𝑂𝐿𝑇 = ∫𝐸𝐷 [𝑍 ∗ σ𝐷,𝐿𝑇 + 𝐸𝐷𝐿𝑇 − (𝐸𝐷𝐿𝑇 + 𝑘 ∗ σ𝐷,𝐿𝑇 )] ∗ 𝑔(𝑧) 𝑑𝑧
𝐿𝑇 +𝑘∗σ𝐷,𝐿𝑇 −𝐸𝐷𝐿𝑇
σ𝐷,𝐿𝑇
∞
→ 𝑈𝑆𝑂𝐿𝑇 = ∫ (𝑍 − 𝑘) ∗ σ𝐷,𝐿𝑇 ∗ 𝑔(𝑧)𝑑𝑧
𝑘
∞
→ 𝑈𝑆𝑂𝐿𝑇 = σ𝐷,𝐿𝑇 ∗ ∫ (𝑍 − 𝑘) ∗ 𝑔(𝑧)𝑑𝑧 NB: do not study the demonstration, to
𝑘
→ 𝑈𝑆𝑂𝐿𝑇 = σ𝐷,𝐿𝑇 ∗ 𝐼(𝑘) solve the exercises is only needed the I(k)
(1−𝐼𝐹𝑅)∗𝑄
formula 𝐼(𝑘) = σ𝐷,𝐿𝑇

𝑈𝑆𝑂𝐿𝑇 σ𝐷,𝐿𝑇 ∗ 𝐼(𝑘)

𝐼𝐹𝑅 = 1 − =1−
𝑄 𝑄
(1 − 𝐼𝐹𝑅) ∗ 𝑄
→ 𝐼(𝑘) = 𝑙𝑜𝑠𝑠 𝑖𝑛𝑡𝑒𝑔𝑟𝑎𝑙
σ𝐷,𝐿𝑇

- The higher the service level IFR you want to offer,

the lower the I(k), the bigger is the k (black)
- The bigger the Q,
the bigger the I(k), the lower the k (blue)
- The bigger the sigma,
the lower the I(k), the lower the k (red)
[the higher the variability, the higher the quantity you need
to cope with that]

223
 15. Advanced Inventory Planning

We know all the variables: as soon as I(k) is computed, we check the table of
loss integral and come out with k
- If we defined the service level as the IFR, we have an idea of the
expected units of stock out
- If we defined the service level as the probability
𝑃𝑆𝑂 = 𝑃(𝐷𝐿𝑇 < 𝑅𝑂𝑃), we have no idea of the units of stock out,
but just if we will go or not in stock out

The I(k) is something we can put in relationship with k, having the IFR, the quantity and the std deviation of the
demand during the lead time

The best way to define the service level depends on how the stock out cost is computed
- If I have defined the service level as the item fill rate IFR, we can have an idea about the USOY; this
information is important if our annual cost of stock out depends on the average value of USO
𝐴𝐶𝑆𝑂 = 𝑈𝑆𝑂𝑌 ∗ 𝑈𝐶𝑆𝑂 [annual cost of stock out = units of stock out in a year * unitary cost of stock out]
We should consider this value because it is also through it that we decide the service level value we want
to assess
- If the service level is defined as the probability to avoid stock out PSO, I have no idea about the USOY
We only know the probability to have or not the stock out; this computation is better than the previous
one for example if we have a production line, because if all the line will stop due to a problem, we will go
for sure in stock out -> the cost will not be related to how many units go in stock out, but only if we go or
not in stock out

Considering the IFR, we have a trade-off between ICC (considering the impact
on only the SS, not on the CS), ITS and stock out costs: the higher the IFR, the
higher the ICC and the lower the CSO
To solve this trade-off, we will minimize the sum of both ICC and CSO, in order
to find the best IFR

3.1.3 Periodic review model

𝑆𝑆 = 𝑘 ∗ σ𝐷,𝑇+𝐿𝑇
In this case, the standard deviation considers both T and LT because the effect of our order will be evident only
after a period equal to this sum: we are exposed to stock out to all this period because we do not control the level
of stocks as in the previous model

σ𝐷,𝑇+𝐿𝑇 = √(𝑇 + 𝐿𝑇) ∗ σ2𝐷 + 𝐸𝐷 2 ∗ σ2𝐿𝑇

NB: we do not add T and its variability because we assume that it is fixed, so it has no variability

224
 15. Advanced Inventory Planning

4 DISTRIBUTION NETWORK DESIGN: INVENTORY ALLOCATION

LEARNING OBJECTIVES
At the end of the module the students will be able to knowingly discuss the following topics and apply them to
different real-world examples
- The main alternatives to allocate the safety stock in a 2-echelon distribution network
o the independent system
o the coupled system
- The relationship between the safety stock and the number of regional warehouses

SAFETY STOCK IN THE DISTRIBUTION NETWORK

Allocation: assuming that the number of echelons, the
number and the capacity of the warehouses, the
location of the plants etc. have already been defined,
we can now take into account the different stock
allocation strategies that allow the minimization of the
overall costs and the achievement of the established
service level

4.2.1 Safety stock in a 2-echelon distribution network

- Coupled System: Safety stock is allocated to
the RWs (stock in the RWs “depends” on the
production system)
- Independent System: Safety stock is allocated
to both the RWs and the CW (stock in the RWs
doesn’t depend on the production system)

Example: 2-echelon distribution network with 1

CW and n RW

VARIABILITY
4.3.1 Coupled system
If safety stock is allocated to the Regional Warehouses only, it has to face the demand of all the period, so it faces:
- The demand variability σDi (downstream variability):
o during the replenishment lead time of the Central Warehouse (LTc)
o during the replenishment lead time of the Regional Warehouses (LTrw)
- The lead time variability (upstream variability):
o the replenishment lead time of the Central Warehouse (LTc)
o the replenishment lead time of the Regional Warehouses (LTrw)

225
 15. Advanced Inventory Planning

4.3.2 Independent system

If safety stock is allocated to both the Regional Warehouses and the Central Warehouse, is like we split the
problem in two different demands, so it faces:
- In the CW the variability of:
o the downstream demand (σDC) during the Lead Time LTc
o the upstream replenishment Lead Time LTc (i.e. σLTc)
- In the RWs the variability of:
o the downstream demand (σDi) during the Lead Time LTrw
o the upstream replenishment Lead Time LTrw (i.e. σLTrwi)
NB: if we have a direct shipment between the CW and a delivery area, the safety stocks needed to cope with the
demand variability of this area are stored

SAFETY STOCKS CALCULATION

4.4.1 Assumptions
- Replenishment lead time of the CW: 𝐿𝑇𝑐 ≈ 𝑁(𝐸𝐿𝑇𝑐 , σ𝐿𝑇𝑐 )
- Replenishment lead time of the RW “i”: 𝐿𝑇𝑟𝑤𝑖 ≈ 𝑁(𝐸𝐿𝑇𝑟𝑤 , σ𝐿𝑇𝑟𝑤𝑖 )
- Demand seen by the warehouse “i”: 𝐷𝑖 ≈ 𝑁(𝐸𝐷𝑖 , σ𝐷𝑖 )
𝑐𝑜𝑣(𝐷𝑖 ,𝐷𝑗 )
- Correlation between the demands of the warehouses i and j: ρ𝑖,𝑗 = σ𝐷𝑖 ∗σ𝐷𝑖

COUPLED SYSTEM
𝑆𝑆𝐶𝑆 = ∑𝑁
𝑖=1 𝑆𝑆𝑅𝑊𝑖 The overall safety stock in the system is the
sum of the safety stocks in each of the N RW
NB: we are assuming we use the reorder point model
𝑆𝑆𝑅𝑊𝑖 = 𝑘 ∗ σ𝐷𝑖,(𝐿𝑇𝑐+𝐿𝑇𝑟𝑤𝑖) [k can be computed as both IFR and
probability to avoid stock out]
𝑆𝑆𝑅𝑊𝑖 = 𝑘 ∗ √(𝐸𝐿𝑇𝑐 + 𝐸𝐿𝑇𝑟𝑤𝑖 ) ∗ σ2𝐷𝑖 + 𝐸𝐷𝑖2 ∗ σ𝐿𝑇𝑐+𝐿𝑇𝑟𝑤𝑖
We have to compute this formula for each Regional Warehouse,
and then to sum the values to obtain the overall SS
NB: it should be noticed that all the terms in the formulas above
must be referred to the same period of time (days, weeks, …)
We take as reference the unit of measure of the demand

4.5.1 Independent system

𝑆𝑆𝐼𝑆 = ∑𝑁
𝑖=1 𝑆𝑆𝑅𝑊𝑖 + 𝑆𝑆𝐶𝑊 The overall safety stock in the system is the sum of the safety stocks in each RW
and the safety stock in the CW

𝑆𝑆𝑅𝑊𝑖 = 𝑘 ∗ 𝜎𝐷𝑖,𝐿𝑇𝑟𝑤 = 𝑘 ∗ √𝐸𝐿𝑇𝑟𝑤𝑖 ∗ 𝜎𝐷2𝑖 + 𝐸𝐷𝑖2 ∗ 𝜎𝐿2𝑟𝑤

𝑖 𝑖

𝑆𝑆𝐶𝑊 = 𝑘 ∗ σ𝐷𝑐,𝐿𝑇𝑐 = 𝑘 ∗ √𝐸𝐿𝑇𝑐 ∗ σ2𝐷𝑐 + 𝐸𝐷𝑐 2 ∗ σ2𝐿𝑇𝑐

𝐸𝐷𝑐 = ∑𝑁
𝑖=1 𝐸𝐷𝑖

The demand seen by the central warehouse is the sum of the demands seen by the RWs
- If the demands are not correlated 2
𝜎𝐷𝑐 = ∑𝑁 2
𝑖=1 σ𝐷𝑖
- If the demands are correlated σ2𝐷𝑐 = ∑𝑁 2 𝑁−1 𝑁
𝑖=1 σ𝐷𝑖 + 2 ∗ ∑𝑖=1 ∑𝑗=𝑖+1 ρ𝑖,𝑗 ∗ σ𝐷𝑖 ∗ σ𝐷𝑗

226
 15. Advanced Inventory Planning

SAFETY STOCKS AND NUMBER OF WAREHOUSES

Particular case: independent and identically distributed demands seen by the regional warehouses
- 𝐷1 = 𝐷2 = ⋯ = 𝐷𝑁 = 𝐷
𝜎𝐷1 = 𝜎𝐷2 = ⋯ = 𝜎𝐷𝑁 = 𝜎𝐷 The distribution of the demands is the same for all the RW
- ρi,j = 0 (for each couple RWi, RWj) All the demands are independent (no correlation)
→ 𝐷𝑐 = 𝑁 ∗ 𝐷 σ𝐷𝑐 = 𝑁 ∗ σ2𝐷
- LTc : constant
- 𝐿𝑇𝑟𝑤1 = ⋯ = 𝐿𝑇𝑟𝑤𝑁 = 𝑐𝑜𝑠𝑡(𝐿𝑇𝑟𝑤 )
- k = constant (the same value for both the CW and the RWs)

4.6.1 Coupled system

𝑁

𝑆𝑆𝐶𝑆 = ∑ 𝑆𝑆𝑅𝑊𝑖 = 𝑁 ∗ 𝑆𝑆𝑅𝑊 = 𝑁 ∗ 𝑘 ∗ 𝜎𝐷,(𝐿𝑇𝑐+𝐿𝑇𝑟𝑤) =

𝑖=1

= 𝑁 ∗ 𝑘 ∗ √(𝐿𝑇𝑐 + 𝐿𝑇𝑟𝑤 ) ∗ 𝜎𝐷2 + 𝟎 = 𝑁 ∗ 𝑘 ∗ σ𝐷 ∗ √𝐿𝑇𝑐 + 𝐿𝑇𝑟𝑤

1
σ2𝐷𝐶 = 𝑁 ∗ σ2𝐷 → σ𝐷 ∝ → 𝑆𝑆𝐶𝑆 ∝ √𝑁
√𝑁

The safety stock is proportional to the square root of the number of RW

[demonstration is not required, is just to understand the mathematical reason behind the shape of the curve]

4.6.2 Independent system

𝑁

𝑆𝑆𝐼𝑆 = ∑ 𝑆𝑆𝑅𝑊𝑖 + 𝑆𝑆𝐶𝑊 = 𝑁 ∗ 𝑘 ∗ 𝜎𝐷,𝐿𝑇𝑟𝑤 + 𝑘 ∗ 𝜎𝐷,𝐿𝑇𝑐𝑤 The safety stocks value of CW is

𝑖=1 constant, while the one of CW is
= 𝑁 ∗ 𝑘 ∗ √(𝐿𝑇𝑟𝑤 ) ∗ 𝜎𝐷2 + 𝟎 + 𝑘 ∗ 𝑁 ∗ 𝑘 ∗ √(𝐿𝑇𝑐 ) ∗ 𝜎𝐷𝑐
2
+𝟎 proportional to the square root of the
number of RW
= 𝑘 ∗ (𝑁 ∗ σ𝐷 ∗ √𝐿𝑇𝑟𝑤 + σ𝐷𝑐 ∗ √𝐿𝑇𝑐 )
= 𝑘 ∗ (𝑁 ∗ σ𝐷 ∗ √𝐿𝑇𝑟𝑤 + σ𝐷 ∗ √𝑁 ∗ √𝐿𝑇𝑐 ) → 𝑆𝑆𝐼𝑆 =
= 𝑆𝑆𝐶𝑊 ∝ 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (∀𝑁)
𝐿𝑇𝑐
= 𝑘 ∗ 𝑛 ∗ 𝜎𝐷 ∗ (√𝐿𝑇𝑟𝑤 + √ ) = 𝑆𝑆𝑅𝑊 ∝ √𝑁
𝑁

5 CONCLUSIONS

SUPPLEMENTARY READINGS
- Stenger A.J., 1994, Inventory Decision framework in: Robeson, Copacino (Eds.), The Logistics Handbook,
pp. 352-391;
- Ronald H. Ballou, Inventory policy decisions, Business logistics Management, pp.403-455;

CRITICAL THINKING
5.2.1 What are the main levers to reduce safety stocks?
Safety stocks are a function of
- Service level → we should target it at a reasonable level
- Demand variability/unpredictability
- Suppliers’ lead time
- Suppliers’ reliability
All these elements can be combined in a common formula → σ𝐷𝐿𝑇 = √σ2𝐷 ∗ 𝐿𝑇 + σ2𝐿𝑇 ∗ 𝐷 2

227
 15. Advanced Inventory Planning

We can combine these elements considering

- Internal supply chain → narrow perspective
- External supply chain → broader perspective

σD → SDE LT σLT
Internal Increase in forecast accuracy Suppliers selection Suppliers’ selection
(ex: location) (ex: vendor rating)
External Demand planning collaboration Transport modes Supply planning collaboration
Digitation and integration

5.2.2 What are the levers to reduce OCT?

Assuming a specific structure of the network
We have the order issued from the customer. The
OCTD is considered from the distribution warehouse
to the customer. Then we are going to have an OCT
from the plant to the depot: OCTP→D.

The OCT seen from the customer will be:

𝑂𝐶𝑇 = 𝑂𝐶𝑇𝐷 ∗ 𝑂𝐹𝑅 + (𝑂𝐶𝑇𝐷 + 𝑂𝐶𝑇(𝑃→𝐷) ) ∗ (1 − 𝑂𝐹𝑅)

1. Times
Each OCT is composed by different activities:
- Order receiving and processing
o Digital integration (use ICT systems for dealing with orders)
- Order picking and preparation
o Change the picking system (level of automation)
o Change some policies
- Transportation
o use specific service providers or change them
o I can decide (depending on the specific problem that I have in front of) to change the
transportation mode.
2. OFR
In order to increase the OFR I can:
- Increase the # of items managed in the warehouses
o 0 transit points
o Different coverage of product range
- Increase the IFR increase of IFR → increase of SS

3. Structure of the network

Increase of number of warehouses → reduction of OCT since we are closer to our customer

5.2.3 Centralization/Decentralization of safety stocks in the network (allocation problem)

- Centralization main benefit is risk pooling: reduction of the demand variability/unpredictability when we
pool the safety stock in a low number of central warehouses [+ Reduction of unitary inventory costs
(economies of scale)]
- Decentralization main benefit is the proximity to the market, which means a reduction of OCT and an
increase of on-time delivery

Centralization:
𝐷𝑇 = 𝐷1 + 𝐷2 + ⋯ + 𝐷𝑁 → σ𝐷𝑇 < σ𝐷1 + σ𝐷2 + ⋯ + σ𝐷𝑁
Usually we reduce the variation of the demand if we aggregate the local demands
NB: this is true if there is NOT a strong positive correlation between Di, otherwise we should consider the
correlation indexes

228
 Logistics Management
Prof. Alessandro Perego

Inventory Planning

Learning objectives 2

§ The inventory planning process: objectives, inputs and activities

§ The KPI of the inventory planning process: how to measure
effectiveness and efficiency
§ The roles of inventories and the types of inventories in the logistic
systems: in-transit, cycle and safety stocks
§ The main factors driving the size of in transit, cycle and safety
stocks
§ The planning approaches: distribution requirement planning vs pull
inventory models
§ How to analyze inventories: the sales-inventories ABC analysis

© Alessandro Perego, Marco Melacini

1
 Agenda 3

3. Inventories in 4. Inventory
2. Roles of
the distribution planning
inventories
network approaches

1. Inventory
5. Inventory
planning
Analysis
process

© Alessandro Perego, Marco Melacini

Supply chain planning process 4

Acquired Forecasted Time

orders orders series

Demand planning

Inventory planning

Operative
Distribution planning PLAN

Production planning
Information
about
downstream Procurement planning
constraints

© Alessandro Perego, Marco Melacini

2
Inventory planning 5
process structure

Tools &
Actors Methodologies

Input Output
Activities

Performance
Measurement

© Alessandro Perego, Marco Melacini

Output and input 6

§ Output:
• to define how much, when and where to hold the inventories in the
production-logistic system
• put it in other words, to identify the “inventory management technique +
parameters”, for each item in each node of the distribution network

§ Input:
• demand plan (demand forecast and estimated accuracy) detailed per item
and location (SKUL=stock keeping unit per location),
• service target
• replenishment lead times
• logistic costs structure

© Alessandro Perego, Marco Melacini

3
Activites (I) 7

To define and implement the replenishment policy, which affects:

§ the Product Flow through the replenishment orders (which items,
in what quantity and when)
§ the Inventory Levels: items, quantity and location

THE CHOICE OF THE THE CHOICE OF THE

MANAGEMENT POLICY MANAGEMENT PARAMETERS
(fixed order point, fixed order (order point, replenishment
frequency, …) frequency, target availability … )

BY ITEM AND WAREHOUSE (SKUL)

© Alessandro Perego, Marco Melacini

Activities (II) 8

The replenishment policy has to be determined:

§ Along the supply chain (in depth), e.g. the same item at different supply
chain levels
§ Along the different warehouses/plants (at the same level – in width), e.g.
the same item in different regional warehouses

© Alessandro Perego, Marco Melacini

4
Inventory planning “effectiveness” 9

Fill Rate
An order (or line, or case) is completely fulfilled if all the items requested
are available from stock
The Fill rate:
§ is related to the number and range of the items at stock and to the
inventory level per item
§ directly affects customer service level only if the customer required
order cycle time is zero (e.g.: supermarkets)
§ otherwise, it influences the average order cycle time: the higher
the fill rate, the lower the average cycle time

© Alessandro Perego, Marco Melacini

Inventory planning “effectiveness” 10

Emphasis on the idea of “customer satisfaction”

Order ORDERS FULFILLED FROM STOCK

X 100
fill rate ORDERS RECEIVED + (POTENTIAL)

Useful also for single-item measures

Line LINES FULFILLED FROM STOCK

X 100
fill rate LINES RECEIVED + (POTENTIAL)

Case CASES FULFILLED FROM STOCK

X 100
fill rate CASES RECEIVED + (POTENTIAL)

© Alessandro Perego, Marco Melacini

5
Inventory planning “efficiency” 11

WHAT
Absolute Relative
value value
Physical

Inventory Turnover
Average
units
Ratio (ITR)
inventory
Days of Supply
level (AIL) (DOS)
HOW

Inventory Turnover
Monetary

Average Ratio (ITR)

units

value of the Days of Supply

inventory (DOS)

© Alessandro Perego, Marco Melacini

Average Inventory Level 12

The average inventory level is the weighted average of the inventory

levels, the weights being the length of the period for each inventory level

I1 ⋅ T1 + I2 ⋅ T2 + I3 ⋅ T3 + ... + In ⋅ Tn
AIL =
T1 + T2 + T3 + ... + Tn
§ j = 1 to n : elementary period with constant inventory level
§ Ij : inventory level in the j-th period
§ Tj : duration of the j-th period (days, hours)

Inventory Ij
Level
AIL

Time

Tj
© Alessandro Perego, Marco Melacini

6
 Inventory Turnover Ratio 13

OUTGOING FLOWi,T
ITRQi ,T =
AILi,T
For the i-th item in the T period

§ It tells how many times the inventory of an item “turns” in the time
period (usually it is calculated on an yearly basis)
§ The unit of measure is [turns/period] or [1/period]
§ Outgoing Flow and Average Inventory can be measured as
preferred: number of pieces, cases, pallet loads, m2, m3, liters, kg
…

© Alessandro Perego, Marco Melacini

Inventory Turnover Ratio 14

VALUE OF THE OUTGOING FLOW (€)i ,T

ITRVi ,T =
AVERAGE VALUE OF THE INVENTORY (€)i ,T

For the i-th item in the T period

§ It can be used to measure the turnover of product families or the
whole warehouse
§ The ITRV values are affected by the monetary values (prices,
costs, …) used to assess the outgoing flow and for the inventory
§ It is easy to compute in aggregated terms since these data can be
obtained from the general accounting systems

© Alessandro Perego, Marco Melacini

7
 Example 15

Item 1 - Sugar Item 2 - Flour

Outgoing flow : 40 pallet loads/month Outgoing flow : 40 pallet loads/month

Replenishment frequency : once a month Replenishment frequency : once a week
Replenishment Lot : 40 pallets loads Replenishment Lot : 10 pallet loads
Safety stocks : 5 pallet loads Safety stocks : 5 pallet loads

AIL = 25 pallet loads

AIL = 10 pallet loads

ITRQ1 = (40 / 25) x 12 = ITRQ2 = (40 / 10) x 12 =

= 19,2 turns/year = 48 turns/year

© Alessandro Perego, Marco Melacini

Days (or weeks, months,..) of Supply 16

AILi,T
DOSi, T =
AVERAGE DAILY CONSUMPTIONi,T
AILi,T
MOSi,T =
AVERAGE MONTHLY CONSUMPTIONi,T
For the i-th item in a definite T period

§ It is the reciprocal of the ITR (DOS=1/ITR)

§ It measures the expected period of autonomy (coverage) of
the warehouse in case it were not replenished anymore
§ It is equal to the average “duration of stay” of the goods within
the warehouse

© Alessandro Perego, Marco Melacini

8
 Inventory Carrying Costs 17

Average value of
the inventories

ICCi, T = AILi,T ⋅ Vi ⋅ phc i [€ /T]

Inventory Carrying
Inventory Carrying Costs Costs as a
i-th item
for the i-th item and for percentage of the
Unitary Value
period T item value [%/T]

Inventory Carrying Costs comprise:

§ PHYSICAL STORAGE COSTS
§ ASSET COSTS
§ DEPRECIATION
§ OBSOLESCENCE

© Alessandro Perego, Marco Melacini

Physical storage costs 18

Physical storage (or warehousing) costs include the following:

§ SPACE : proportional to the average inventory (random storage)

or to the maximum inventory (dedicated storage) [e.g. pallet loads]
and to the occupied space per pallet load [e.g. pallet loads/m2].
There are 2 main components:
• Land costs + storage equipment
• Services: heating, light, security costs

§ INSURANCE (fire and thefts): proportional to the value insured

which in turn is connected to the average or maximum value of the
inventories

§ MANAGEMENT (controls, inventory accounting, etc.)

© Alessandro Perego, Marco Melacini

9
Asset costs 19

§ It depends on:
§ the capital tied up in inventories, which should be valued
considering only the costs that have been anticipated due to the
inventory policy
§ the “cost of capital” for the company which is affected by internal
and external factors
• The interest rate of the investment the company is missing because of the
capital tied up in inventories (assuming the company has no cash and no
open credit line)
• The interest rate of the bank if the company is borrowing the money tied up
in inventories
• The lending interest rate if the company could lend the money tied up in
inventories, if available
• …

© Alessandro Perego, Marco Melacini

Depreciation 20

It is the loss of value caused by the “flow time” of the inventories across
the distribution network.
It depends on:
§ loss of value due to factors outside the logistics system (e.g. price
reduction trend in the electronic components market, whose effect
is proportional to the delay from production to sale)
§ loss of value due to internal factors (e.g. damages of the
packaging, which are more likely when products spend a lot of
time within the system)

Depreciation costs are function of the quantity of goods that are subject
to depreciation and the average loss of value

© Alessandro Perego, Marco Melacini

10
Obsolescence 21

Obsolescence can be caused by:

§ SPOILAGE, e.g. perishable foods or medicines
§ TECHNOLOGICAL DEVELOPMENT, e.g. a new processor for
notebooks
§ NEW MODELS, e.g. the restyling of a car
§ FASHION, e.g. apparel products
Obsolescence costs are function of the quantity of goods that become
obsolete, in turn depending on the average amount of inventories within
the system, and the average loss of value
Example - clothes that are obsolete could be :
§ Sold at lower prices (lower margins)
§ Sold through different channels or in other markets (e.g. further
logistics cost and lower margins)
§ Destroyed (no incomes + destruction costs)

© Alessandro Perego, Marco Melacini

Agenda 22

3. Inventories in 4. Inventory
2. Roles of
the distribution planning
inventories
network approaches

1. Inventory 5. Inventory
planning process Analysis

© Alessandro Perego, Marco Melacini

11
Inventory roles 23

Inventories can be:

§ a Lever of Efficiency in order to decrease the costs of other

processes/parts of the production/logistic system (production,
purchasing, transportation, …)

§ a Lever of Effectiveness, in order to better satisfy the customer

needs (e.g. to achieve the target service level even though the
environment is uncertain)

© Alessandro Perego, Marco Melacini

Inventory roles - Efficiency 24

Examples:
§ Seasonal inventories: they are used to decouple a phase with
storng seasonality (e.g. the market demand) from a more
balanced one (e.g. production) (tradeoff: production costs –
inventory carrying costs)
§ Opportunistic inventories: they allow to buy at low prices for large
quantity lots (tradeoff: purchasing costs – inventory carrying costs)
§ Transportation-related inventories: they seek to reduce transport
fares thanks to Full Truck Load shipments (tradeoff: transportation
costs – inventory carrying costs)
§ Inter-operational inventories – they decouple subsequent
production phases (trade off: production costs – inventory carrying
costs)
§ …

© Alessandro Perego, Marco Melacini

12
Inventory roles - Effectiveness 25

Examples:
§ Uncertainty-related inventories: they seek to face the demand
uncertainty and the supply uncertainty (with regard to both
conformity and capacity)
§ Quality-related inventories: they improve the quality of the
products (e.g. food ripening or ageing: cheese, ham, wine, etc. or
quarantine for the pharmaceutical products)
§ ….

§ In these cases the trade-off is between inventory carrying costs

and the value of increased effectiveness (service level, quality, …)
© Alessandro Perego, Marco Melacini

Agenda 26

3. Inventories in 4. Inventory
2. Roles of
the distribution planning
inventories
network approaches

1. Inventory 5. Inventory
planning process Analysis

© Alessandro Perego, Marco Melacini

13
Inventories in the logistic system 27

In transit Inventory Inventory in the warehouses

In process
Inventory (receiving,
sorting, loading, …)

© Alessandro Perego, Marco Melacini

In Transit Inventory/Stock 28

These are the inventories on the transport arcs (on the vehicles) or in-
process within the nodes. Their average value depends on the average
product flow and the lead time of the transport or process activity

Average product flow between the point of origin and the point of
destination

ITS = F * LT
Average “In transit” Average lead time
stock
The hydraulic simile
F [m3/s]
F . LT = volume of
liquid in the pipe
LT [s]

© Alessandro Perego, Marco Melacini

14
In Transit Stock: examples 29

goods : mobile phones (750 € /kg)

route : MILAN - ROME (600 km)
duration of the journey : 1 day
shipment quantity : 32 PL ITS = 1 * (5 / 2)= 2,5 mln €
shipment value : 5 millions euro
frequency : 1 journey every two days

goods : paper reels (0.75 €/kg)

route : Helsinki - Milan
duration of the journey : 10 days
shipment quantity : 1000 reels (800 t) ITS = 10 * (0,6 / 15)
shipment value : 600.000 euro = 0,4 mln €
frequency : 1 journey every 15 days

goods : wheat (0.09 €/kg)

route : Istanbul - Savona
duration of the journey : 20 days
shipment quantity : 10.000 t ITS = 20 * (0,9 / 30)
shipment value : 900.000 euro = 0,6 mln €
frequency : 1 journey a month (30 days)

© Alessandro Perego, Marco Melacini

Inventories in the warehouses 30

There are two “logical” types of inventories within the warehouse that
reflect the two main roles:

Cycle Stocks: these inventories deal with (result from) the different
operative rhythm of two consecutive stages in the supply chain
§ they are a lever of efficiency (the sizing is a trade-off between
logistic costs)

Safety Stocks: these inventories deal with the uncertainty of both the
demand and the replenishment lead time
§ they are a lever of effectiveness (the sizing is a trade-off between
service level and inventory carrying costs)

© Alessandro Perego, Marco Melacini

15
 Cycle inventory/stocks 31

Cycle stocks cope with the discontinuity of the replenishment and

retrieval processes (i.e. the different operative rhythm of two consecutive
stages in the supply chain)

§ production “anticipated” with respect to the demand

• Seasonal production and constant consumption
• Constant production and seasonal consumption

§ production “concurrent” with the demand, but replenishment rates

different from the retrieval rates
• Discreet purchasing/production/distribution lots

© Alessandro Perego, Marco Melacini

Cycle stocks 32
example 1

Production anticipated with respect to the demand

Cycle stock plays as decoupling buffer allowing the production system to
work with a constant rate even though sales are seasonal

40

30

Production
20

Sales
10

0
jan feb mar apr may jun jul aug sep oct nov dec

80

60

Inventory 40
AIL
20

0
jan feb mar apr may jun jul aug sep oct nov dec

© Alessandro Perego, Marco Melacini

16
 Cycle stocks 33
example 2

Warehouse replenishment rates different from the consumption rates

RW1
Central WH Demand dispatched from RW1
500 cases / day

RW2
Demand dispatched from RW2
500 cases / day

1. Weekly replenishment 2. Daily replenishment

The regional warehouses are The regional warehouses are
replenished with a lot equal to the replenished with a lot equal to the daily
weekly demand demand

Average Cycle St. =500 x 5 =1.250cases Average Cycle St. 500

= x1 = 250cases
2 2
© Alessandro Perego, Marco Melacini

Cycle stocks 34
Main drivers

Assumption: constant demand, replenishment with fixed lot equal to Q

Inventory
Maximum Cycle
Stock (MSC)
Average Cycle Q
Stock (ACS)
Time

Assumption: Safety stocks = 0 T

MSC=Q =DxT
§ T =replenishment interval ACS=Q/2 = DxT/2
§ f = replenishment frequency, T=1/f
§ D =average daily demand Main drivers:
§ Q = replenishment lot, Q = D.T n D

n T, f

© Alessandro Perego, Marco Melacini

17
 Safety stocks 35

Safety stocks cope with the demand uncertainty (demand higher than the
forecasted level) and the replenishment lead time uncertainty (higher
than expected)

Inventory
Inventory

Demand as Demand
expected higher than
expected

SS

LT as expected LT longer than expected

© Alessandro Perego, Marco Melacini

Safety stock 36
Main drivers

Assumption: constant demand, replenishment with fixed lots

Inventory

Main drivers:
§ LT = replenishment lead time
§ σLT = replenishment lead time
uncertainty
D.LT § DM =average daily demand
§ σD = daily demand uncertainty
§ Target service level

LT
D and LT are Random
Variables

© Alessandro Perego, Marco Melacini

18
Agenda 37

3. Inventories in 4. Inventory
2. Roles of
the distribution planning
inventories
network approaches

1. Inventory 5. Inventory
planning process Analysis

© Alessandro Perego, Marco Melacini

PULL Inventory Control 38

RM WIP FP FP FP FP

Production system Distribution system

§ The product flow is triggered by the inventory levels at the

downstream nodes and aims to refill the inventory up to a pre-set
target level (from here the term PULL)
§ The model does not consider the relationships between the nodes
at the different levels (each node is independent from the others,
has its own replenishment model, carries out its own demand
forecast activity)
§ The parameters of the inventory management models are set in
function of the demand forecast
© Alessandro Perego, Marco Melacini

19
PUSH Inventory Control 39
(Distribution Requirement Planning)

RM WIP FP FP FP FP

Production system Distribution system

§ The model considers the relationships between the demands at

the different levels (i.e. the demand in the upper levels is the sum
of the requirements from the downstream levels)
§ The product flow is planned on the basis of the demand forecast
from the market and of the actual inventories in all the nodes of
the logistic network (thence the use of the term PUSH)
§ The timing of the replenishments is planned in function of the
lead times

© Alessandro Perego, Marco Melacini

Demand relationships along the supply 40

chain

Production system Distribution system

RM WIP FP FP FP FP

Demand Demand Indipendent

dependent on: dependent on: demand
- Product structure (Bill Of Materials) - Distribution network
- Production lead time - Delivery lead time

© Alessandro Perego, Marco Melacini

20
 Comparison between the approaches (I) 41

Physical flow
In the push
approach “global
PUSH approach information”
PUSHES the flow of
products
Information flow

Physical flow
In the pull
approach the
products are PULL approach
PULLED by “local
information”
Information flow

© Alessandro Perego, Marco Melacini

Comparison between the approaches (II) 42

PUSH approach PULL approach

§ It plans the availability of the § It sees only the next lead time
goods on the basis of the forecasted needs (or it reacts to
forecasted needs on the overall the past demand)
supply chain lead time horizon § It links a dyad of nodes
§ It can link many actors § It requires few local data
§ Much information is required § High operative simplicity
§ High space-time depth

There is an high operative complexity Limited space-time depth

© Alessandro Perego, Marco Melacini

21
 Space-time depth 43

SPACE Overall view of the logistic system, thus

considering the interrelationships between
Supply Chain levels
scope

System
PUSH

Fragmented view of the logistic

system and inability to manage the
interrelationships between levels
Planning
horizon

Here PULL

Now Future TIME

© Alessandro Perego, Marco Melacini

Agenda 44

3. Inventories in 4. Inventory
2. Roles of
the distribution planning
inventories
network approaches

1. Inventory 5. Inventory
planning process Analysis

© Alessandro Perego, Marco Melacini

22
Inventory Analysis 45

What is the “right” level of inventories and how to find out the
“useless” part?

INEFFICIENCY
TOTAL
INEFFICIENCY

Right Safety
Stocks

Stock
INEFFICIENCY

Total
Right Stock
Right Cycle
Stocks

© Alessandro Perego, Marco Melacini

Pareto’s rule 46

“In whatever group of factors to be controlled there is

a small fraction, in terms of number, to which a big
influence on the overall effects can be ascribed.
On the contrary, most of the factors, in terms of
number, has a low impact on the effects”
(Parigi 1848 - Celigny 1923)

© Alessandro Perego, Marco Melacini

23
 Example – Inventory 47
Pareto’s (ABC) curve

100%

90%

80%

70%
Cumulated 60% Curve without the
inventory 50% cumulative effect
40%
B C
30%

20%
A
10%

0
0 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Number of items

20% of the items (A class) determines

80 % of the value of the stocks

© Alessandro Perego, Marco Melacini

Inventory – Sales matrix 48

% Sales
% Inventory 100
100

A B C
A B C 100
100 Sales
From the analysis of the 9 A B C
classes it is possible to get some
A 1 2 3
Inventory

useful indication to correct the

inventory management policies
B 4 5 6

C 7 8 9

© Alessandro Perego, Marco Melacini

24
 Inventory – Sales matrix 49

Possible causes of misalignment between inventories and sales:

§ Low replenishment frequency

§ Minimum lot imposed by the supplier
§ Low reliability in suppliers’ lead time
§ Discounts for the purchases of big quantities
§ Inventory to support New Product launch
§ Speculation
§ Demand forecast errors
§ Old inventory management parameters
§ …

© Alessandro Perego, Marco Melacini

Use of the matrix 50

Congruence Inefficiency area

area
Sales
A B C
Attention area 1
A 2 3
Inventory

B 4 5 6

C 7 8 9

Ineffectiveness
area
Carelessness area
© Alessandro Perego, Marco Melacini

25
 Attention area 51

1. CLASS A-A (high sales, high inventory)

The items in this cell are usually a few, so they
can be analysed one by one. They are both a
risk and an opportunity:
§ Stockouts will cause heavy drops in
sales
Sales
§ No other class offers the same
opportunities in terms of inventory A B C

reduction A 1 2 3

Inventory
B 4 5 6

C 7 8 9

© Alessandro Perego, Marco Melacini

Inefficiency area 52

3. CLASS C-A (low sales, high inventory)

In this cell there are the items that have lower
impact on sales, but which have high inventory
levels, often hard to reduce
It is not enough to wait until they are consumed
but it might be necessary to dispose off them:
Sales
§ Stop the replenishment and the
A B C
production
A 1 2 3
§ Push the sales through promotions
Inventory

§ Alienation
B 4 5 6

C 7 8 9

© Alessandro Perego, Marco Melacini

26
 Ineffectiveness area 53

7. CLASS A-C (high sales, low inventory)

It is a privileged class of items and can be a
good example of inventory management to
be extended to the other classes

It can hide some dangers: if the inventory

level requirements are underestimated there
Sales
is a risk of missing sales
A B C
§ check if there are frequent stock-outs
A 1 2 3

Inventory
B 4 5 6

C 7 8 9

© Alessandro Perego, Marco Melacini

Dynamic use of the matrix 54

The matrix is a dynamic tool that has

to be used over time:
§ To take different “pictures” to Sales
verify how the different cells
A B C
change over time
§ To study the “migrations” in A 1 2 3
Inventory

order to find out and

implement improvement B 4 5 6
actions
C 7 8 9

© Alessandro Perego, Marco Melacini

27
Supplementary Readings 55

§ Stenger A.J., 1994, Inventory Decision framework in:

Robeson, Copacino (Eds.), The Logistics Handbook, pp.
352-391;

§ Ronald H. Ballou, Inventory policy decisions, Business

logistics Management, pp.403-455.

© Alessandro Perego, Marco Melacini

Challenge questions 56

§ Where would you position the inventory planning in the overall

planning process?
§ Which are the main objectives of the inventory planning?
Which are the main decision to be taken?
§ How would you measure the efficiency of the inventory
planning process? And the effectiveness?
§ What are the criteria according to which the different types of
inventories can be classified? What are the functions of the
inventories in the supply chain?
§ What are the in transit stocks? How would you assess them?
§ What is the function/objective of the cycle stocks? What are
the drivers affecting them?
§ What is the function/objective of the safety stocks? What are
the drivers affecting them?

© Alessandro Perego, Marco Melacini

28
 Logistics Management
Prof. Alessandro Perego

Inventory Planning – Models

Learning objectives

An overview of the main types of “pull” inventory planning models

An in-depth focus on the reorder point model and the periodic review
model
A clear understanding of the general methodology to size/to choose the
parameters of the models
§ reorder frequency related parameters: “order quantity” for the
reorder point model and “order interval” for the periodic review
model
§ the safety stock for both the models

© Alessandro Perego, Marco Melacini

1
 Agenda

1. Inventory 2. Reorder point 3. Periodic 4. Safety stock

models model review model determination

5. Advanced
EOQ models

© Alessandro Perego, Marco Melacini

Inventory planning

The inventory planning aims at:

§ sizing both the cycle and the safety stock (what, how much and
where)

§ controlling the flow in the logistic system (what, how much and
when)
Inventory
planning

Flow
Inventory sizing
management

© Alessandro Perego, Marco Melacini

2
 Pull inventory models classification

Order repetitiveness Product range extention

§ Repetitive orders § Single item
§ Single orders § Multi-items

Demand profile Vertical extention

§ Regular and medium- § Single level models
high demand § Multi-level models
§ Low demand
§ Lumpy demand
§ ….. Models that will be taken
into account

© Alessandro Perego, Marco Melacini

Pull inventory models classification

Discrete review Continuous

Single or Multi-item review
replenishment Single-item
Order interval
replenishment
Fixed Variable

Reorder
Fixed
Order quantity

point
model
Variable

Periodic
Switch
review
models
model

© Alessandro Perego, Marco Melacini

3
 Agenda

1. Inventory 2. Reorder point 3. Periodic 4. Safety stock

models model review model determination

5. Advanced
EOQ models

© Alessandro Perego, Marco Melacini

Reorder point model

WHAT?
The model takes into account one item at a time (single-item, single
location model)

HOW MUCH?
A constant order quantity (Q)

WHEN?
When the inventory level goes under the Reorder Point quantity (ROP)

© Alessandro Perego, Marco Melacini

4
 Reorder point model

Quantity
on hand
AVAILABILITY
Order
quantity
Q INVENTORY
Reorder LEVEL
point
ROP
Time

Order Order
place receive

Lead Time
LT
© Alessandro Perego, Marco Melacini

Uncertainty management

Since the demand and the lead time cannot be known for sure (because
of their unpredictability) we have to add to the regular stock (that is
maintained to meet the average demand in the average lead time) the
safety stock

Quantity 1 2 3
on hand
AVAILABILITY

Q Q Q
INVENTORY
Reorder
point LEVEL
ROP
Safety
stock Time
LT LT LT
SS
© Alessandro Perego, Marco Melacini

5
 Reorder point

Quantity LT
on hand

ROP
EDLT
SS
Time
ROP = EDLT + SS
Reorder point ED LT = expected demand during the
[units] replenishment lead time
SS = average value of the safety stock

© Alessandro Perego, Marco Melacini

Average Inventory Level

Quantity
on hand

AIL
SS
Tempo

AIL [units] AIL = SS + ACS

ACS = average value of the Q
cycle stock 2

© Alessandro Perego, Marco Melacini

6
Parameters to be determined

In order to adopt the reorder point model, two parameters must be

determined:

Q (order quantity) – it is connected to the cycle stock

ROP (reorder point) – it is connected to the safety stock

© Alessandro Perego, Marco Melacini

Methodology for the EOQ determination

The order quantity Q (if there are no constraints) can be determined as

the EOQ = Economic Order Quantity in three steps:

Identification of the relevant costs

§ Inventory carrying costs
§ Costs of the goods
§ Procurement costs
§ Transportation costs
§ Out of stock costs
§ ….
Modeling of the relevant costs. A cost function depending on Q has to be
found for each relevant cost
Minimization of the total cost function (for instance by mathematically
differentiating the sum of the relevant costs on Q and equaling the results
to 0)
© Alessandro Perego, Marco Melacini

7
 Basic EOQ Calculation
(particular case)

The economic order quantity is calculated by minimizing the sum of the

cost of the goods, procurement costs and inventory carrying costs

Assumptions:
Relevant costs:
§ Costs of the goods
§ Procurement costs
§ Inventory carrying costs
Safety stock does not depend on Q
Q is delivered outright

© Alessandro Perego, Marco Melacini

Annual Inventory Carrying Costs

Unitary item
Value

Q
AICC = ⋅ V ⋅ phc [€ / year]
2
Annual Inventory
Carrying Costs Inventory Carrying
Costs as a percentage
of item value [%/year]

SS stock assumed equal to 0

© Alessandro Perego, Marco Melacini

8
 Annual Procurement Costs

Annual Unitary
Demand Procurement cost
[€]

Dy
APC = ⋅ upc [€ / year]
Q
Annual Procurement
Costs Order
Quantity

© Alessandro Perego, Marco Melacini

Annual Cost of the Goods

Annual
Demand

ACG = D y ⋅ p [€ / year]

Unitary Price [€]

Annual Cost of the Goods

© Alessandro Perego, Marco Melacini

9
Basic EOQ formula
(Particular case)

Annual
costs TC
[€]
AICC

ACG
APC
Q
EOQ

min ( TC = ACG + APC + AICC )

2 ⋅ upc ⋅ Dy
EOQ =
phc·V

© Alessandro Perego, Marco Melacini

Remarks on the basic EOQ formula

1.The EOQ:
§ grows, if the procurement costs increase (the function is not linear)
§ decreases, if the inventory carrying costs decreases (not in
inverse proportion)

Proportional to the
APC
2 ⋅ upc ⋅ Dy
EOQ = Proportional to
phc·V AICC

2.TOTAL COSTS are almost constant around EOQ

© Alessandro Perego, Marco Melacini

10
 EOQ Extensions

The Basic EOQ formula is valid only under the assumptions previously
explained

It is possible to extend it by modifying some of the assumptions

Two extensions presented in appendix:
§ Economic order quantity with quantity discounts
§ Economic order quantity with non-istantaneus supply

Other extensions available in literature:

§ Bounded order quantity :
• Limited storage capacity
• Limited number of orders in a year
• Maximum investment in stock

© Alessandro Perego, Marco Melacini

Agenda

1. Inventory 2. Reorder point 3. Periodic 4. Safety stock

models model review model determination

5. Advanced
EOQ models

© Alessandro Perego, Marco Melacini

11
 Periodic review model

WHAT?
This model can be applied both to the single item and the multi-item
cases

WHEN?
Orders are placed every T (order interval)

HOW MUCH?
The order quantity aims at reaching a determined availability target

© Alessandro Perego, Marco Melacini

Periodic review model

Quantity
on hand
Availability
target
AT
AVAILABILITY

INVENTORY LEVEL

Order Order Order Time

place place place
i-th i-th+1 i-th+2
LT LT
T T

© Alessandro Perego, Marco Melacini

12
 Uncertainty management

If the lead time and the demand are uncertain, safety stock is required in
order to avoid the stock out during T+LT

1 2 3
AT AVAILABILITY
Quantity
on hand
Q1 Q3
Q2
INVENTORY
LEVEL

SS
Time
LT LT LT

T T

© Alessandro Perego, Marco Melacini

Availability target

Quantity
on hand LT LT
AT
AVAILABILITY

ED (LT+T)
INVENTORY
LEVEL
SS
Time
T T
AVAILABILITY TARGET [units] AT = ED(LT+T) + SS

ED (LT+T) = expected demand during T+LT

© Alessandro Perego, Marco Melacini

13
 Average order quantity and average
inventory level

Order quantity [units] Q (t) = AT - IL (t)

IL (t) = inventory level at the time t

Average order quantity [units] AQ = EDT

Average inventory level

[units]
AIL = SS + ACS
ACS = average value of the cycle stock = EDT / 2

© Alessandro Perego, Marco Melacini

Order interval

The order interval can be determined:

§ for each item separately. In this case, the optimal T can be

determined by taking into account the relevant costs of the single
item

§ for all the items that are jointly purchased (usually from the same
supplier). In this case, the optimal T is determined by minimizing
the sum of the costs of all the items that are jointly purchased (e.g.
each product family)

© Alessandro Perego, Marco Melacini

14
Order interval

Under the same assumptions as for the calculation of the “basic” EOQ
the optimal order interval is:
§ Single item

EOQ 1 2 ⋅ upc ⋅ Dy 2 ⋅ upc

T= = ⋅ =
Dy Dy phc ⋅V phc ⋅V ⋅ Dy
§ Multi-items

2 ⋅ (O + ∑ upci )
O is the common cost
T= i
for procuring an order
phc ⋅ ∑ Vi ⋅ D y ,i
i

© Alessandro Perego, Marco Melacini

Model comparison

Reorder point Periodic review

- The joint reorder of more than
- Inventory level under control
one item can be easily done
(continuous control)
- The inventory level control is
- Minimization of the main costs
easy to be done (only when an
(AICC and APC)
order has to be placed)

- The joint reorder of more than

one item is very hard to be done - On average, the inventory level
(as a consequence many orders is higher than in the reorder
have to be placed – even to the point model
same supplier)
- The inventory level control has
to be continuous and as a
consequence expensive

© Alessandro Perego, Marco Melacini

15
 Agenda

1. Inventory 2. Reorder point 3. Periodic 4. Safety stock

models model review model determination

5. Advanced
EOQ models

© Alessandro Perego, Marco Melacini

Safety stocks in the reorder point model

Safety stock copes with the uncertainty of both the demand during the LT
and the LT, so there are two main parameters to assess, the demand
and the lead time

Quantity
on hand
Reorder
point Expected
ROP value of
EDLT demand in
the LT

Probability
Safety stock distribution of
SS the demand in
the LT
Expected Lead Time Time
Probability distribution
of the LT
© Alessandro Perego, Marco Melacini

16
Demand
Forecast system vs Demand system

Depending on the relationship between the forecast system and the

inventory control system we can have two cases:
§ demand system, the two systems are separate. The average
demand (ED) and the standard deviation of the demand (sD) must
feed the model

§ forecast system, the two systems are joined. The forecasted

demand (in T or T+LT depending on the model) and the standard
deviation of the error (SDE) are the two parameters that must feed
the model

For simplicity’s sake from now on we will use the notation of the first case
(ED, sD), meaning (ED, sD) for the demand system and (forecasted
demand, SDE) for forecast system

© Alessandro Perego, Marco Melacini

Safety stock in the reorder point model

We can define the safety stock level for a specific item through the
following formula:

SS = k ⋅ σ D , LT

σD,LT: standard deviation of the k: coefficient that takes into

demand during the LT (or account the desired Service
standard deviation of errors Level (the higher the coefficient
in the forecast system) the higher the service level)

This is a “composite” standard deviation that includes the

uncertainty (=unpredictability) of both the demand during
the LT and the LT (with reference to the average demand
value)

© Alessandro Perego, Marco Melacini

17
 Safety stock parameters

k σD,LT

Assuming that the demand and the

Lead time are independent:

It depends on the way D : (ED,σD)

the Service Level was
defined (i.e. the way LT : (ELT,σLT)
the stock out is We have:
measured) and on the
target value of the
service level σ D, LT = ELT ⋅ σ D2 + σ LT
2
⋅ ED 2

© Alessandro Perego, Marco Melacini

Safety stocks in the periodic review model

Safety stock copes with the uncertainty of the demand both in the LT and T

Quantity
on hand
Availability
target
AT EDLT+T
Expected value
of the demand
in T+LT

Probability
Safety distribution of the
stock demand in the LT
SS
LT ELT Time
T Probability distribution of the
LT

© Alessandro Perego, Marco Melacini

18
 Safety stock in the periodic review model

We can define the safety stock level for a specific item through the
following formula:

SS = k ⋅ σ D , LT +T

σD,LT+T : standard deviation of the demand

during T+LT (or, in the forecast system,
standard deviation of the error)

σ D, LT +T = (ELT + T )⋅σ D2 + σ LT2 ⋅ ED 2

© Alessandro Perego, Marco Melacini

Service level

We can define the service level as:

§ The probability to avoid the stock out during the lead time:

PSO = Prob [( DLT ≤ ROP )]

§ The item fill rate (IFR): demanded quantity – available quantity ratio

IFR = 1 – USO / Dy

USO = expected annual stock out [units/year]

Dy = expected value of the annual demand

© Alessandro Perego, Marco Melacini

19
Service level

Assumption: the demand during the LT can be represented through a

normal distribution with average EDLT standard deviation sDLT

Average demand = EDLT

Probability
Standard deviation = σDLT
to avoid
stock out Stock out
probability

EDLT DLT
ROP = EDLT + SS = EDLT + k·σDLT

© Alessandro Perego, Marco Melacini

Service level - Probability to avoid the stock

out

By setting the safety stock level, the probability to avoid the stock out has
been set as well

Probability to
avoid stock out:

84,1%
97,7%
99,8%
EDLT
DLT
EDLT+ σDLT

EDLT +2žσDLT
EDLT +3žσDLT

© Alessandro Perego, Marco Melacini

20
Service level - Probability to avoid the stock
out

PSO = Prob ( DLT < OP ) = Prob (DLT - EDLT < SS) =

= Prob ( (DLT - EDLT)/σDLT)< k ) = Prob ( z < k )

z = standard normal distribution with average equal to 0

and standard deviation equal to 1
PSO k
Prob(z)

85.0% 1.04
90.0% 1.28
95.0% 1.64
98.0% 2.06
99.0% 2.33
k Z 99.5% 2.58

© Alessandro Perego, Marco Melacini

Service level - Probability to avoid the stock

out

The steps to get the safety stock level are:

1.set the probability to avoid the stock out
2.find k in the standard normal distribution table (i.e. with average equal
to 0 and standard deviation equal to 1)
3.put k in the formula to calculate the safety stock

SS = k ⋅ σ D , LT SS = k ⋅ σ D , LT +T

© Alessandro Perego, Marco Melacini

21
Service level - Item Fill Rate

The item Fill rate IFR can be calculated as follows (see next slides for the
demonstration – not compulsory)

I (k )⋅ σ DLT
IFR = 1 −
Q

The steps to get the safety stock level are:

1. set the item fill rate IFR
2. compute I(k)
3. read k from the table of the normal loss integral
4. put k in the formula to calculate the safety stock

SS = k ⋅ σ D , LT SS = k ⋅ σ D , LT +T
© Alessandro Perego, Marco Melacini

Service level - Item Fill Rate

Values for the unit normal loss integral

k I(k) k I(k)
0 0,3989 1,6 0,0232
0,2 0,3069 1,8 0,0143
0,4 0,2304 2 0,0085
0,6 0,1687 2,2 0,0049
0,8 0,1202 2,4 0,0027
I(k) 1 0,0833 2,6 0,0015
0,4 1,2 0,0561 2,8 0,0008
1,4 0,0367 3 0,0004

1 2
k

© Alessandro Perego, Marco Melacini

22
 Service level - Item Fill Rate Demonstration

IFR = 1 – USO / Dy
USO = expected annual stock out [units/year]
Dy = expected value of the annual demand

USO = USOLT⋅Dy/Q

USOLT = UNITS IN STOCK OUT DURING THE DELIVERY LEAD TIME

HAVING SS UNITS OF SAFETY STOCK
∞

USOLT = ∫ ( y − ED
EDLT + SS
LT − SS )⋅ f ( y )⋅ dy

f(y) = FUNCTION OF THE PROBABILITY DENSITY OF THE DEMAND

DURING THE LEAD TIME

© Alessandro Perego, Marco Melacini

Service level - Item Fill Rate Demonstration

Average = EDLT
Prob(y)=Prob(DLT) Standard Deviation= σDLT

f(y)⋅dy

y=DLT

EDLT

EDLT+SS dy
y - EDLT - SS

© Alessandro Perego, Marco Melacini

23
 Service level - Item Fill Rate Demonstration

Using the Standardized Normal function: g(z)

z = (DLT - EDLT)/σDLT)= (y - EDLT)/σDLT)

d(z)*g(z)
With probability density g(z)
z
∞ k dz

∫ ( y − ED
EDLT + SS
LT − SS )⋅ f ( y )⋅ dy z-k

EDLT+SS = EDLT+k⋅σDLT y - EDLT - k⋅σDLT f(y)⋅dy

k (z - k)⋅σDLT g(z)⋅dz

© Alessandro Perego, Marco Melacini

Service level - Item Fill Rate

Demonstration

⎛ ∞ ⎞
USOLT = ⎜⎜ ∫ (z − k )⋅ g (z )⋅ dz ⎟⎟ ⋅ σ D, LT = I (k )⋅ σ D , LT
⎝ k ⎠
∞
Where: I(k ) = ∫ (z − k ) ⋅ g (z ) ⋅ dz
k

is the unit normal loss integral

Dy
I (k )⋅ σ DLT ⋅
Q I (k )⋅ σ DLT
IFR = 1 − = 1−
Dy Q

© Alessandro Perego, Marco Melacini

24
 Agenda

1. Inventory 2. Reorder point 3. Periodic 4. Safety stock

models model review model determination

5. Advanced
EOQ models

© Alessandro Perego, Marco Melacini

EOQ Extensions

The “classic” formula is valid only under the assumptions previously

explained
It is possible to extend it by modifying some of the assumptions
Two presented extensions:
§ Economic order quantity with quantity discounts
§ Economic order quantity with non-istantaneus supply

Other extensions available in literature:

§ Bounded order quantity :
• Limited storage capacity
• Limited number of orders in a year
• Maximum investment in stock

© Alessandro Perego, Marco Melacini

25
 EOQ with quantity discounts (1)

In presence of quantity discounts the total cost function become a

discontinuous function

Annual
cost Unit price changes in function
of the order quantity

TC
p1 : 0 < Q < Q1
p2 : Q1 ≤ Q < Q2
ACG p3 : Q ≥ Q2
AICC

APC
Size of the lot Q
Q1 Q2

© Alessandro Perego, Marco Melacini

EOQ with quantity discounts (2)

§ The EOQ has to be calculated in each case (P = P1; P = P2, …)

§ Only the values of the EOQ that fall in the validity interval have to
be considered. If the EOQ is not in the validity interval the
minimum quantity for which the price is valid has to be taken into
account
§ Choose the EOQ for which the overall cost is minimum

© Alessandro Perego, Marco Melacini

26
 EOQ with quantity discounts (3)

Annual
cost TC (1)

TC (2)

EOQ1
TC (3)
EOQ2
EOQ
EOQ3 Size of the lot Q
Q1 Q2

© Alessandro Perego, Marco Melacini

EOQ with noninstantaneus supply (1)

Demand rate
[units/day]
Quantity
on hand
-d
Maximum
Production rate inventory [units]
[units/day] (less than Q)
r

Q’ = T1 * (r – d) = Q * (r – d) / r

Time
T1 = Q / r

Time to replenish
the lot Q [days]

© Alessandro Perego, Marco Melacini

27
EOQ with noninstantaneus supply (2)

§ Maximum inventory is: Q’ = Q * (r – d) / r

§ Therefore the average inventory is: Q’/2 = Q/2 * (r – d) / r

So:

AICC’ [€/year] = V * Q/2 * phc * (1-d/r)

The other costs (APC, ACG) remain unchanged with the

respect of the “basic” EOQ formula

© Alessandro Perego, Marco Melacini

EOQ with noninstantaneus supply (3)

Finally, adding up the other costs (APC, ACG) and differentiating, you
obtain:

2 ⋅ upc ⋅ D y
EOQ’ =
phc ·V ·(1 - d/r)

As d < r, then EOQ’ > EOQ

© Alessandro Perego, Marco Melacini

28
Supplementary Readings

§ Stenger A.J., 1994, Inventory Decision framework in: Robeson,

Copacino (Eds.), The Logistics Handbook, pp. 352-391;

§ Ronald H. Ballou, Inventory policy decisions, Business logistics

Management, pp.403-455.

© Alessandro Perego, Marco Melacini

Challenge questions

§ How would you classify the inventory models?

§ Which are the main features and the fields of application of the Reorder
Point model? Draw the profile of the inventory level with and without the
demand/lead time unpredictability. And what about the Periodic Review
model?
§ How would you determine the safety stocks if you have adopted the
Reorder Point model? What is the difference with respect to the Periodic
Review model?
§ How would you determine the k parameter (which is necessary for the
safety stocks determination) if the service level has been defined as the
probability to avoid the stock out during the replenishment lead time?
How do things change if the service level is defined as the Item Fill Rate?
§ Which are the parameters to be optimized in order to use the models
proposed?
§ How would you determine the Economic Order Quantity in the Reorder
Point model? What if there is a quantity discount? And if there are non
instantaneous supplies?
© Alessandro Perego, Marco Melacini

29