See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/327548481

Mass Customized Manufacturing: Theoretical Concepts and Practical

Approaches

Book · February 2017

DOI: 10.1201/9781315398983

CITATIONS READS

26 409

1 author:

Vladimir Modrak
Technical University of Kosice - Technicka univerzita v Kosiciach
159 PUBLICATIONS 1,125 CITATIONS

SEE PROFILE

All content following this page was uploaded by Vladimir Modrak on 12 June 2023.

The user has requested enhancement of the downloaded file.

 1

Mass Customized Manufacturing:

Theoretical Concepts and Practical Approaches

V. Modrak (ed).
 2

TABLE OF CONTENTS

Section 1
Trends and Success Factors in Mass Customization

1. An Introduction to Mass Customized Manufacturing

VladimirModrak
Technical University of Kosice, Slovakia
Corresponding author contact: [email protected]

2. Designing Assembly Lines for Mass Customization Production Systems

Dominik T. Matt, Erwin Rauch
Free University of Bozen-Bolzano, Italy
Corresponding author contact: [email protected]

3.Role of Information Systems in Mass Customization

Arun Nambiar
California State University Fresno, USA
Author contact: [email protected]

Section 2
Complexity Drivers in Mass Customization

4. Complexity Issues in Mass Customized Manufacturing

Hendrik Van Landeghem, El-Houssaine Aghezzaf
Ghent University, Belgium
Corresponding author contact:[email protected]

5. Modelling of Assembly Supply Chain Structures

Vladimir Modrak, Slavomir Bednar
Technical University of Kosice, Slovakia
Corresponding author contact: [email protected]

6. Variety-induced Complexity Metrics

Vladimir Modrak, Slavomir Bednar
Technical University of Kosice, Slovakia
Corresponding author contact: [email protected]

7. Product Variety Management Assuming Product Configuration Conflicts

Vladimir Modrak, Slavomir Bednar
Technical University of Kosice, Slovakia
Corresponding author contact: [email protected]

Section 3
Management and Sustainability of Mass Customization
 3

8. Product Configuration for Order Acquisition and Fulfilment

Dario Antonelli,Giulia Bruno,
Polytechnic University of Turin, Italy
Corresponding author contact:[email protected]

9. Shoes Configurators: a Comparative Analysis of Capabilities and Benefits

Enrico Sandrin1, Cipriano Forza1, Zoran Anisic2, Nikola Suzic1,2, Chiara Grosso1, Thomas Aichner3,
Alessio Trentin1
1 Department of Management and Engineering, University of Padova, Italy
2 Department of Industrial Engineering and Management, University of Novi Sad, Serbia
3 Faculty of Economics and Management, Free University of Bozen-Bolzano, Italy

Corresponding author contact:[email protected]

10. Empirical Investigation on Implications of Configurator Applications for Mass

Customization
1
Linda Zhang, 2Petri T. Helo,
1
IESEG School of Management, Lille-Paris, France
2
Faculty of Technology, University of Vaasa, Finland
Corresponding author contact:[email protected]

11. Sustainability Assessments for Mass Customization Supply Chains

Carlo Brondi, Rosanna Fornasiero, Davide Collatina
Institute of Industrial Technologies andAutomation, Italy
Corresponding author contact:[email protected]

12. Sustainability Issues in Mass Customized Manufacturing

Arun Nambiar,
California State University Fresno, USA
Author contact:[email protected]
 4

List of Figures

Figure 1.1 Four levels of mass customization

Figure 1.2 Generic concept of identification optimal degree of product
customization
Figure 1.3 Customers' Customization Readiness construct; a) explicit form; b)
implicit form
Figure 2.1 Framework for Mass Customization Assembly
Figure 2.2 Design rules in Product Design for Mass Customization Assembly
Figure 2.3 SMART Reconfigurability approach (Matt et al., 2014)
Figure 2.4 Typical elements of MES systems for Mass Customization Assembly
Figure 2.5 Industry 4.0 applications in the field of Mass Customization Assembly
Figure 2.6 Improvements in Mass Customization Assembly at a manufacturer of
electric motors
Figure 4.1 Casual model of complexity in MCM
Figure 4.2 Objective complexity elements
Figure 4.3 Subjective complexity elements
Figure 4.4 Classification results on 76 workstations
Figure 4.5 Complexity score of 76 workstations
Figure 4.6 ROC curve of both complexity models
Figure 4.7 Impact of complexity on workstation efficiency
Figure 4.8 Operator reaction time to choice complexity
Figure 4.9 Effect of complexity on operator reaction time (after Hanson et al.,
2012) – dark regions are +/- 1 std.dev. around the mean
Figure 4.10 Example of a mixed-model assembly line (Zeltzer, 2016)
Figure 4.11 Workstation’s major parameters
Figure 4.12 a) Workstation’s major statistics: maximum spread per station
Figure 4.12 b) Balance loss versus overload statistics per operator, workstation for and
models
Figure 4.13 a) If the builds are not sequenced, overload occurs
 5

Figure 4.13 b) If the builds are sequenced, no overload occurs

Figure 4.14 Results of the current line balance
Figure 4.15 Results of the optimized line balance
Figure 4.16 Summary of Manufacturing Complexity and Line Balancing Analyses
(Source: Luiza Zeltzer, 2016)
Figure 4.17 Augmented reality on the shop floor (Courtesy of VUZIX)
Figure 5.1 All possible ASC networks based on numbers of original suppliers
Figure 5.2 (a) Model of ASC network, (b) transformation into a labeled graph, and
(c) transformation into ASC structure No. 10 in Class #5
Figure 5.3 Original ASC structure with i=5 and related alternative ASC networks
Figure 5.4 (a) ASC structure with 10 FRs and 2 DPs; (b) ASC structure with 10
FRs and 3 DPs, both transformed into design matrix.
Figure 5.5 Labeled graph incorporating compulsory optional components
Figure 6.1 Case MCA illustration forCL2 SCL2
Figure 6.2 (a) Four possible product configurations for product class CL3 SCL2;
(b) nine possible product variations for the selected product
configurations
Figure 6.3 Distribution of product configurations in calculation for the product
class CL4 SCL3
Figure 6.4 a) Product configurations for Class CL2 SCL42depicted graphically; b)
number of product configurations obtained by enumeration
Figure 6.5 (a) Case model of MCA as bipartite graph; (b) Axiomatic design matrix
of the case model
Figure 6.6 Complexity aggregation principle
Figure 6.7 Model of MCA with CC values for individual nodes and summary CC
value including legend
Figure 6.8 Summary feed SDC values of the MCA model
Figure 7.1 Case of product structure with rules for constraint satisfaction problem
Figure 7.2 Assembly graph of personal computer without component restrictions
Figure 7.3 Proposed approach to transform incidence matrix into product
configurations model
Figure 7.4 Model of 21 possible product configurations respecting the
configuration rules
Figure 7.5 Procedure for the selection of optimal design platform
 6

Figure 7.6 Procedure for the selection of optimal design platform from the three
available
Figure 8.1 Comparison of the two approaches to Product Configuration
Figure 8.2 Product tree of a Computer
Figure 8.3 Example of komponent tree and fiction tree
Figure 8.4 Decomposition of the problem in three main issues
Figure 8.5 Class diagram of a component and application to the mixer example
Figure 8.6 Set of UML 2.1 standard diagrams
Figure 8.7 An example of the class diagram notation.
Figure 8.8 UML class diagram of the PLC reference ontology.
Figure 8.9 Implementation of the PLC reference ontology in Protégé.
Figure 8.10 A RF filter unit produced by the RTTcompany
Figure 8.11 UML class diagram containing the specification of the basic concepts
for RTT.
Figure 8.12 Implementation of theRTT ontology in Protégé
Figure 9.1 Benchmarking of web-based sales configurator (WBSC)capabilities
Figure 9.2 Benchmarking of WBSCs on product-related and customization
experience-related benefits
Figure 9.3 The user-friendly product-space description capability in the Shoes of
Prey WBSC
Figure 9.4 The focused navigation capability in the Reebok WBSC
Figure 9.5 The flexible navigation capability in the Adidas WBSC
Figure 9.6 The benefit-cost communication capability in the Shoes of Prey WBSC
Figure 9.7 The easy comparison capability on the Nike WBSC
Figure 10.1 Major tasks of product configurators and corresponding percentages
Figure 10.2 Main users of product configurators and corresponding percentages
Figure 10.3 Function unites reorganized and corresponding percentages
Figure 10.4 Business process changes and corresponding percentages
Figure 10.5 Changes to legacy systems and corresponding percentages
Figure 10.6 Changes to the number of employees and corresponding percentages
 7

Figure 11.1 Knowledge horizons in modelling environmental impact of the product

chain
Figure 11.2 Integrated model for supply chain evaluation
Figure 11.3 Modular representation of a product chain for a specific customized
product
Figure 11.4 Inventory data for the definition of in- and outgoing flows
Figure 11.5 Model application to the study case
Figure 11.6 Operational drivers for the product customization
Figure 11.7 Impact of improvement in suppliers performance
Figure 11.8 Scenarios for implementation of customization within the product chain
Figure 12.1 Faces of Mass Customization (Modified from Gilmore & Pine(1997))
Figure 12.2 Product and Process Characteristic (Adapted from Hayes
&Wheelwright 1979)
 8

List of Tables

Table 4.1 Direct drivers of complexity

Table 4.2 Overview of 9 methods to quantify complexity (Mattson et al.)
Table 4.3 Workstation characteristics used in the complexity quantification
Table 4.4 Coding rules to transform variables into Likert Scale values
Table 4.5 Complexity-induced cognitive load and mitigating actions
Table 5.1 SCL values of ASC structures for the related topological classes
Table 5.2 Ivd values of ASC structures for the related topological classes
Table 5.3 SDC values of ASC structures for the related topological classes
Table 6.1 Related values of product configurations and variations for Classes
CL2-5
Table 6.2 Summary table with related values of product configurations (ΣConf)
and variations (ΣVar) for product Class CL1
Table 6.3 Summary table with related values of product configurations (ΣConf)
and variations (ΣVar) for product Class CL2-∞
Table 6.4 Fragment of table summarizing product configurations for the three
selection rules (situations), while j=0
Table 6.5 Fragment of complexity indicators NCP and SDC
Table 7.1 Incidence Matrix for Component Restrictions R#1-7
Table 7.2 Computational results of numbers of product configurations
Table 7.3 Computational results of waste entropy for different numbers of
product configurations
Table 7.4 Module ompatibility platform D0 for gears and front drivetrain
Table 7.5 Computational results of numbers of product configurations
Table 8.1 BOM corresponding to the product tree of Figure 8.2
Table 8.2 Individualconnection matrix forthe mixer withthe universal adapter
Table 9.1 Sample of web-based sales configurators (WBSCs)
Table 10.1 Performance improvements acheived by the respondents
Table 11.1 Notations for the equations 1-4
Table 11.2 Manufacturing scenarios
 9

Table 11.3 Product features

Table 11.4 Description of customization drivers and the related variance range
Table 11.5 Simulation scenarios
 10

List of Appendixes

Appendix 5.1 All possible ASC structures in class #6 and class #7

Appendix 5.2 All possible ASC structures in class #8 (Part 1)
Appendix 5.3 All possible ASC structures in class #8 (Part 2)
Appendix 5.4 Summary table of SCL, Ivd and SDC complexity values for all ASC
structures in classes i=2-8
Appendix 6.1 Summary table of avaialble product configurations when k=11-21
Appendix 7.1 Reduced module compatibility platform D1
Appendix 7.2 Reduced module compatibility platform D2
Appendix 7.3 Reduced module compatibility platform D3
Appendix 10.1 Questions in the survey
Appendix 11.1 Global warming potential (GWP100)
Appendix 11.2 Acidification potential (AP)
Appendix 11.3 Eutrophication potential (EP)
Appendix 11.4 Ozone depletion potential (ODP)
Appendix 11.5 Photochemical Ozone Creation (POCP)
 11

About the editor

Vladimír Modrák is Full Professor of Manufacturing Technology at Faculty of Manufacturing

Technologies and currently Deputy Rector of Technical University of Kosice. His research

interests include manufacturing logistics, cellular manufacturing, lean manufacturing, mass

customization and other related disciplines. He lectured as Visiting Professor at University of

Perugia (Italy), Technical University of Applied Sciences Wildau (Germany), University of

Czestochowa (Poland), and held seminars at the Keyworth Institute of the University of Leeds

(UK) and University of Salerno (Italy). Professor Modrak is an editorial board member in

several international journals and also serves as ad hoc reviewer for reputable journals such as

International Journal of Production Research, Journal of Cleaner Production, Applied

Mathematics and Computation, Entropy, Journal of intelligent and Robotic systems, Journal of

Engineering Business Management and others. He was the leading editor and co-author of the

books “Operations Management Research and Cellular Manufacturing Systems: Innovative

Methods and Approaches” (2012) and “Handbook of Research on Design and Management of

Lean Production Systems” (2014). He is Fellow of the European Academy of Industrial

Management (AIM).
 12

Contributors

El-HoussaineAghezzaf Rosanna Fornasiero

Faculty of Engineering and Architecture ITIA-CNR
Ghent University Institute for Industrial Technologies and
Technologiepark 903, CampusArdoyen, Automation
9052 Belgium Italy
[email protected] [email protected]

Thomas Aichner Cipriano Forza

Faculty of Economics and Management Department of Management and Engineering
Free University of Bozen-Bolzano University of Padova
Italy Italy
[email protected] [email protected]

Zoran Anisic Chiara Grosso

Department of Industrial Engineering and Department of Management and Engineering
Management University of Padova
University of Novi Sad Italy
Serbia [email protected]
[email protected]
Petri T. Helo
Dario Antonelli Department of Production
Department of Production Systems and Business Faculty of Technology
Economics University of Vaasa
Polytechnic University of Turin Finland
Italy [email protected]
[email protected]
Hendrik Van Landeghem
Slavomir Bednar Ghent University
Faculty of Manufacturing Technologies with a seat Belgium
in Presov [email protected]
Technical University of Kosice
Bayerova 1, 080 01 Presov, Slovakia Dominik T. Matt
[email protected] Free University of Bozen-Bolzano
Italy
Carlo Brondi [email protected]
ITIA-CNR
Institute for Industrial Technologies and Vladimir Modrak
Automation Faculty of Manufacturing Technologies with a seat
Italy in Presov
[email protected] Technical University of Kosice
Bayerova 1, 080 01 Presov, Slovakia
Giulia Bruno [email protected]
Department of Management and Production
Engineering Arun Nambiar
Polytechnic University of Turin California State University-Fresno
Italy USA
giulia.bruno@polito .it [email protected]

Davide Collatina Erwin Rauch

Institute of Industrial Technologies and Free University of Bozen -Bolzano
Automation Bolzano, Italy
Italy [email protected]
[email protected]
Enrico Sandrin
 13

Department of Management and Engineering Alessio Trentin

University of Padova Department of Management and Engineering
Italy University of Padova
[email protected] Italy
[email protected]
Nikola Suzic
Department of Management and Engineering Linda Zhang
University of Padova Department of Management
Italy IESEG School of Management
Department of Industrial Engineering and Lille-Paris
Management France
University of Novi Sad [email protected]
Serbia
[email protected]
 14

About the contributors

El-Houssaine Aghezzaf is Professor of Industrial Systems Engineering and Operations Research at

Ghent University. He is currently heading the department of Industrial Systems Engineering and

Product Design. His main research interests are in integrated optimization and simulation solutionsto

the design, planning, scheduling, and control problems arising in manufacturing systems and in

logistical and utility networks. He co-authored more than 120 papers, published in major Operations

Research and Industrial Engineering journals and conference proceedings. He is associate editor of

‘International Journal of Production Research’ and member of editorial boards of three other journals.

He is member of the executive committee of the Belgian Society of Operations Research, member of

the International Federation of Automatic Control - Manufacturing and Logistics Systems.

Thomas Aichner is a lecturer in consumer behaviour at the Faculty of Economics and Management,

Free University of Bozen-Bolzano (Italy). He holds a joint PhD in marketing from the University of

Padova (Italy) and ESCP Europe Business School (Berlin, Germany), with the special mention of

“Doctor Europaeus”. His research is primarily focused on international marketing, country of origin,

mass customization, e-commerce and social media. Thomas is the co-author of the 2011 book “Mass

Customization: An Exploration of European Characteristics”.

ZoranAnisic is Professor of Industrial Engineering and Management at Faculty of Technical Sciences,

University of Novi Sad, Serbia. His area of interest is Product Development and Management,

especially Mass Customization and Product Lifecycle Management. He is founder and manager of My

Product - Center for Product Development and Management under who takes point on a number of

actions including the Mass Customization and Open InnovationNetwork (MC-OI Network);

International Conference on Mass Customization and Personalization in Central Europe (MCP-CE

2008 - 2016) and Master studies in PLM. From activities related to cooperation with industry, it should

be emphasized the development of different types configurators for local industry. He published more

than 160 papers in conference proceedings and scientific journals.

 15

Dario Antonelli is Associate Professor of Technology and Production Systems atthePolitecnico di

Torino, Italy. His main research interests concern the development of strategies and methods for human-

robot collaborative working cells and the finite elements simulation and optimization of metalworking

processes. He has been team leader of several EC-funded research projects: ECHORD-FREE, AMICO,

@CARE, ADIUVARE. He is the director of the degree course of Engineering and Management at the

Politecnico di Torino, and member of the Scientific Committee of PRO-VE and VINORG.

Slavomir Bednar holds a Ph.D. in Manufacturing Management and is currently occupying position of

Assistant Professor at the Technical University of Kosice, Slovakia. His research focuses mainly on

mass customization (MC), product/process variety and complexity management, with a particular

emphasis on the ASC complexity in terms of MC productions using graph theory and several kinds of

clustering. He is currently engaged in several industrial projects within his research field for electronic

and automotive industry.

Carlo Brondi is researcher at the Institute for Industrial Technologies and Automation (CNR-ITIA)

with decades of experience in environmental impact assessment. He has been involved in several

European, National and Regional Project on the environmental impact assessment. Main area of

expertise are: the Sustainable Design of innovative production Processes and Products, Life Cycle

Assessment of advanced materials, Implementation of LCA and LCSA approach in industrial process

simulation, Implementation of LCA for company evaluation performance, LCA modularization in

product design phase, Cradle-to-gate and gate-to-grave data integration within certification procedures

(EPD scheme). Research activities concerned different industry areas like automotive manufacturing,

machine and equipment, white goods industry, electronic sector, leather and wood industry, packaging

sector and fashion industry.

Giulia Bruno holds a doctorate in Information and System Engineering and she is currently a post-doc

researcher at the department of Management and Production Engineering of the Politecnico di Torino,

Italy. Her research activity is focused on knowledge management, data mining and production system

analysis. Furthermore, she is working in the fields of human-robot collaboration and product traceability
 16

along the whole supply chain. She was involved in several European projects in the context of product

lifecycle management, healthcare systems and SME networks. Previous research interests also include

the analysis of gene expression data to improve tumor classification and the development of data mining

algorithms for clinical analysis.

Davide Collatina is a research fellow at ITIA-CNR. After an initial formation in the LCA of temporary

structures in an urban area, his expertise focused on environmental performance assessment of

innovative technologies from factory field up to product chain level. Recent application is devoted to

LCA and LCSA implementation in the design and simulation of innovative products facing emerging

industrial paradigms. Application areas are closed-loop recycling and remanufacturing, industrial

symbiosis within industry districts, eco-effectiveness of new technologies and product customization.

Rosanna Fornasiero works as researcher at the Institute of Industrial Technologies and Automation of

the Italian National Council of Research (CNR-ITIA). She has a degree in Economics with a

specialization in international business and her current research interests include Supply Chain

Management, Mass Customization and Sustainability. In the last 15 years, she was involved in several

European projects related to the mentioned topics under the programmes ICT, NMP and FOF. She is

author of more than 60 scientific publications and she is member of the IFIPgroup5.7. She is coordinator

of the Road mapping group of the National Cluster of Intelligent factories in Italy.

Cipriano Forza is Professor of Management and Operations Management at Padova University, where

he serves as coordinator of the PhD course in Management Engineering and Real Estate Economics.

He teaches research methods at the European Institute for Advanced Studies in Management (EIASM)

in Brussels. He served for six years on the board of the European Operations Management Association

(EurOMA). He published in various journals including the Journal of Operations Management, the

International Journal of Operations and Productions Management, the International Journal of

Production Research, the International Journal of Production Economics, Production Planning and

Control and Computers in Industry. His current research focuses on product variety management,
 17

including such topics as mass customization, concurrent product-process-supply chain design and

product configuration.

Chiara Grosso is an experienced researcher in economic sociology with major in technological

changes as social and economic processes. She earned the Ph.D. europaeus at the Universidad

Internacional de Catalunya, (Barcelona, Spain) where she carried out teaching and research activities

as junior faculty member at the Department of Communication Science. In 2014 she joined the

Department of Management Engineering, at University of Padua, as jr. research fellow. Her research

activities include: computer mediated communication (CMC), social network analysis (SNA), user

experience (UX), customer value management (CVM), business digitalization, web based product

configuration strategies.

Petri T. Helo is a Professor of Industrial Management, Logistics Systems and the head of Networked

Value Systems research group, at Department of Production, University of Vaasa, Finland. His research

addresses the management of logistics systems in supply demand networks and use of IT in

operations. Dr. Helo is also partner and board member at Wapice Ltd, a software solution provider of

sales configurator systems and mass customization solutions.

Hendrik Van Landeghem is Full Professor in Operational Excellence at Gent University in Belgium.

He is an expert on design and optimization of manufacturing and logistics systems and their application

within industry. He is also an expert in Lean Management, specifically on implementation methods. He

has written more than hundred articles on these and related subjects and contributed to several books,

among them one on Best practices in Operations management, and recently 3 volumes on Lean

implementation methods and Flow optimization for SME’s. He is Associate Editor for the Engineering

Management Journal. He is Fellow and Board Member of the European Academy of Industrial

Management (AIM) and Fellow of the World Academy of Productivity Science.

Dominik T. Matt is a Full Professor for Manufacturing Technology and Systems at the Free University

of Bolzano, Italy. He studied Mechanical Engineering at the Technical University of Munich and

achieved a Ph.D. in Industrial Engineering at the University of Karlsruhe. In 1999, he entered the
 18

Research and Engineering Center (FIZ) of the BMW Group in Munich. In 2004, he was appointed to

the post of a Professor for Manufacturing Systems and Technology at the Polytechnic University of

Turin, Italy. In 2008, he accepted a call of the Free University of Bolzano to a tenured professorship at

the Faculty of Science and Technology. Since 2010, Professor Matt is also Head of Fraunhofer

Innovation Engineering Center (IEC) in Bolzano.

Vladimír Modrák is Full Professor of Manufacturing Technology at Faculty of Manufacturing

Technologies. His research interests include manufacturing logistics, cellular manufacturing, lean

manufacturing, mass customization and other related disciplines. Professor Modrak is an editorial board

member in several international journals and also serves as ad hoc reviewer for reputable journals such

as International Journal of Production Research, Journal of Cleaner Production, Applied Mathematics

and Computation, Entropy, Journal of intelligent and Robotic systems, Journal of Engineering Business

Management and others. He is Fellow of the European Academy of Industrial Management (AIM).

ArunNambiar is an Associate Professor at California State University - Fresno. His main research

interests include software systems as it applies to mass customization, lean principles and production

scheduling. He also works in the areas of RFID and data analytics. He is currently working on the

applications of big data in manufacturing.

Erwin Rauch is an Assistant Professor for manufacturing technology and systems at the Free

University of Bolzano, Italy. He holds a Master in Mechanical Engineering and in Business

Administration fromthe Free University of Bolzano and the Technical University Munich. He obtained

his Ph.D. degree in Mechanical Engineering from the University of Stuttgart, with “summa cum laude”.

He also has been awarded with the “Overall best Paper Award” at ICAD 2013 and with a “Best Track

Paper Award” at IEOM 2015. His research interests include the application of Axiomatic Design, the

design of changeable and flexible production systems, the design of distributed manufacturing systems,

the design of Industry 4.0 applications for production and production planning in “make-to-order”

(MTO) and “engineer-to-order” (ETO) manufacturing.

 19

Enrico Sandrin holds a PhD in Operations Management and an MS in Management Engineering from

the University of Padova (Italy). His research interests concern mass customization, product

configuration, organization design and human resource management. In his researches, he has used both

quantitative and qualitative methods, especially survey and case study, and a variety of statistical data

analysis techniques. Before enrolling in the PhD school, he had worked as a knowledge engineer, as a

buyer and as a controller in a firm operating in the machinery industry, where he also had had the

opportunity to develop strategic improvement projects supported by leading consultancy companies and

to act as an in-company trainer. Currently, he is a research associate at the University of Padova, where

he also works as a teaching assistant on his topics of expertise.

Nikola Suzicis a lecturer of Production System Design and Decision Making Theory at the Department

of Industrial Engineering and Management, Faculty of Technical Sciences, University of Novi Sad

(Serbia). His main research interest is Mass Customization, including new product development and

production system design for high product variety contexts. The current focus of his research is the

application of the Mass Customization in Small and Medium Enterprises (SMEs). Nikola has

contributed to three international books, published over twenty papers in international journals and

conferences as well as realized a number of projects with companies. Since year 2006, he has been a

member of the Organizing Committee of the International Conference on Mass Customization and

Personalization in Central Europe (MCP-CE).

AlessioTrentin is an Assistant Professor at the Università di Padova (Italy), where he got a PhD in

Operations Management in 2006. In 2007-2008, he was a visiting assistant research professor at the

Zaragoza Logistics Center (Zaragoza, Spain), a joint research center of MIT (US) and Aragona

government (Spain). His research interests include mass customization, form postponement, product

configuration, build-to-order supply chains, sustainable operations management and country-of-origin

effect. His work has been published in Computers in Industry, the International Journal of Operations

& Production Management, the International Journal of Production Economics, the International

Journal of Production Research, the International Journal of Mass Customisation, the International

Journal of Industrial Engineering and Management, and the Journal of Global Marketing.
 20

Linda Zhang is a Professor of Operations Management in Department of Management at IESEG

School of Management (LEM-CNRS), Lille-Paris, France. She obtained her BEng and Ph.D. degrees

in Industrial Engineering from China in 1998 and Singapore in 2007, respectively. Her research interests

include mass customization, design and management of warehousing systems, healthcare service

design, and supply chain management. On these areas, she has published many articles in international

refereed journals, such as Decision Support Systems, IIE Transactions, IEEE Transactions on

Engineering Management, European Journal of Operational Research, International Journal of

Production Economics, etc. Dr. Zhang has the extensive teaching experiences in a number of countries,

including the Netherlands, France, Singapore, and China. She has taught courses at the undergraduate,

graduate, and post-graduate levels.

Preface
 21

Research interests in mass customization clearly show that this business phenomenon is on the rise.

Bearing in mind the extensive amount of literature in this area as well as scientific publications, one

would think that there is not much to say in this direction anymore.And this question occurred to me

when I consideredwhether to start the preparation for this publication. Later, after sortingthrough all

theexisting literature, I came to the conclusion that the literature dedicated to the manufacturing

perspective of mass customization is not excessive at all. This fact has encouraged me to embark on

this journey and to build a wider team of contributors. Some of the invited authors from this research

domain perceived the issue as sufficiently elaborated, and for this reason they have not expressed their

interest in participatingin this project. Those, who saw the gap in the literature, have supported my

intention and sent me proposals for their intendedcontributions. Subsequently, proposed topics for this

book were sorted and arranged into three sections and the project moved to its realization phase. The

editing process was quite challenging to harmonize different research approaches and attitudes into one

common piece of work. The book you are holding in your hands is result of these efforts.

The content of the book can be summarilycharacterized as follows.

The first section presents recent trends and success factors in mass customization. In the introductory

chapter, the concept of mass customization and its success factors are briefly outlined. Subsequently,

recent trends in the design of mass customization production systems in connection with the Industry

4.0 concept are presented. Theconcluding chapter of the first section offers a comprehensive view on

the need for a functioning and effective mass customization information system that encompasses all

activities involved in the customization process.

Section two focuses on complexity drivers, as it is known that products and processes in terms of mass

customization are becoming more complex. It brings higher uncertainty that may negatively affect the

firm's performance results. The first chapter focuses on the role of complexity and its impact on the

human operator within assembly oriented mass customization manufacturing. The second chapter

provides a structural framework to model alternative assembly supply chain structures, which are later

analyzed using selected structural complexity indicators. Subsequently, the authors focused on a

combinatorial-based method to quantify product configurations and variations representing product

variety arising in mass customized manufacturing. A method to assess alternative mass customized
 22

system designs using newly developed measures is then described. The last chapter of this section

presents a method to determine the more suitable degree of customization for a specific product design

platform. Its practicality is demonstrated through a case application.

Section threemainly provides an overview of the product configuration management and its different

approaches. The first chapter presents a meta-model of an interactive product configurator to assist the

customer during the definition of a customized product. Web-based sales configurators are further

discussed in the state-of-the-art analysis for footwear configurators and benchmarked with

configurators forthe fashion industry. Another chapter investigates the implications of product

configurator applications based on a survey and brings answers to questions – for example, how product

configurator applications affect companies’ business activities? The last two chapters are dedicated to

sustainability issues in mass customization manufacturing.

The book provides researchers, scholars, and practitioners with conceptual tools and information about

how to use new approaches to maintain and expand market share in a competitive environment. It is

also intended that the book could encourage conventional manufacturers to adopt some of the types of

mass customization.

Vladimir Modrak

Technical University of Kosice, Slovakia

Acknowledgment
 23

It is my honor and pleasure to acknowledge those who have helped to make this book possible.

I would first like to thank Mrs. Cindy Renee Carelli, Senior Acquisitions Editor, Kyra Lindholm,

Editorial Assistant, and Jill J. Jurgensen, Senior Project Coordinator for their professional support in

managing this book project.

I am grateful to all chapter authors who contributed to the edited book.

My special thanks goes to Dr. Slavomir Bednar for his helpful technical and continuous assistance on

this project. I would also like to thank my doctoral student Ms. Zuzana Soltysova who helped me to

produce figures in my chapters.

Last but not least, I would like to thank my wife Reny for standing beside me throughout my career and

preparing this book. I appreciate her understanding of my absence in many of her precious life moments

during editing the book and writing its four chapters.

 24

Section one

Trends and Success Factors in Mass

Customization

Chapter one
 25

An Introduction to Mass Customized Manufacturing

Vladimir Modrak
Technical University of Kosice, Slovakia

ABSTRACT

In global business environment manufacturing companies are facing several important challenges.

Among them, non-market issues such as environmental and social risk mitigation can be identified.

This chapter focuses on the market-oriented challenge associated with the necessity to continuously

update product offers in order to serve today's markets and remain competitive. Implementation of

appropriate mass customization strategy provides companies with the most effective way to satisfy

individual customers' expectations. This chapter provides a summarization of features and trends of

mass customization and offer new views on the subject.

1.1 About the subject

Mass customization (MC) is likely the future trend of business strategy development. In this context

the questions arise regarding what characterizes current trends of MC, how it differs from previous

manufacturing strategies and what will be the future of manufacturing when it takes a global approach.

Providing accurate answers is not easy and requires considering at least two aspects of the term. If we

comprehend mass customization as a marketing and manufacturing technique that combines

personalized customization and mass production, then we can see manufacturing, and marketing

perspectives in determining optimal overall strategies for firms.

The first view regards a world of manufacturing that is changing as it follows the world of technology.

Technological changes are driven by many factors such as safety and environmental standards, social

demands, diffusion of innovation, etc. Technology is changing very rapidly and the newest

technological developments are reshaping the manufacturing sector in its original form. For example,
 26

additive manufacturing, cloud computing, radio-frequency identification, 5th generation wireless

systems and Internet of Things (IoT) are only few of the new technologies that are driving a paradigm

shift in manufacturing. The umbrella term for this new wave of so called smart manufacturing is

Industry 4.0. This concept originates from Germany, but there is a similar view around the world on the

future of manufacturing. For instance, the Industrial Internet Consortium in the U.S. was founded in

March 2014 by manufacturing, Internet, IT, and telecommunications companies with 212 members as

of 16 September 2015. Successive implementation of smart manufacturing capabilities will allow for

a faster and better response to customer requirements than ever before. Wide adoption of IoT into smart

manufacturing systems will allow improved flexibility and productivity of the production process and

will enable a higher level of mass customization than is possible today. In this way the manufacturing

sector is undergoing a serious transformation process that promises other disruptive innovations

including adopting new business models and to production of mass-customized products with improved

quality and reduced direct costs.

Further development of mass‐customization from the point of view of the consumer will depend on the

willingness of customers to spend time in specifying their preferences and to accept increased price and

delivery time of a customized product. Experiences show that modern consumers desire more and more

customized products. At least one of the reasons that consumers prefer custom-made products relates

to the so called counter-conformity motivation [1]. This kind of motivation is based on the fact that they

want to be recognized from others as having a particular status in their communities. According to Piller

[2], the key element of mass customization from the consumer perspective is that customers are

integrated into value creation as product co-designers by defining, configuring, matching, or modifying

an individual solution.

A good starting point in the identification of differences between the current situation of mass

customization and future scenarios is to outline distinct approaches to mass customization and an

evolutionary development of MC.

1.2 Definitions and approaches to mass customization and its evolutionary development
 27

Since the early 1990s, mass customized manufacturing (MCM) has developed into one of the leading

ideas in the production of goods. Mass customization can be understood as the marketing, development

and production of affordable goods or providing services with such a sufficient diversity and adaptation

that each customer can customize them to suit his or her needs. In other words, the aim is to give the

customer what he wants and when he wants it [3]. Another definition of mass customization was

introduced by Tseng and Jiao [4]. They define MC as the technology and systems producing goods and

services based on individual customer requirements.

MCM is a business strategy of expanding the sphere of influence by particular companies bringing

attractive opportunities to added value to the precisely addressed customer requirements. This is also

thanks to the fact that consumers are often willing to pay more than for mass produced products. On the

other hand MCM is often criticized for the fact that such products do not provide comparable profit for

companies as with traditional customized production. In fact, mass customization is still a new

manufacturing concept that is rarely applied by companies and requires further research and

development.

The most recent classification of MC approaches was offered by Kull [5] who divided mass

customization into two types: configuration and parameterized type of mass customized manufacturing.

If customers can configure selected parts of products individually to their own solution – this is the

configuration type. The second one allows customers to change visual aspects of the product, e.g. shape

and/or size of product components. According to Gilmore and Pine [6] mass customization may appear

as: a) Adaptive Customization, b) Cosmetic Customization, c) Transparent Customization, or d)

Collaborative Customization. When applying Adaptive Customization, products and services are

standardized but with a few customized options. For example, a company offers a package of software

designed to run all activities of small businesses. So, the buyer as a final customer can add e.g. more

accounting functions into the package.

Cosmetic Customization means that company produces products, which are standardized, but the

market offers the products in different ways to different customers. The main difference between

Adaptive and Cosmetic Customization is in the offer of standardized products, where Adaptive
 28

Customization offers only a choice from standardized products and Cosmetic Customization offers the

customer groups of standardized products. It means that the customer could choose from an apparent

variety of offered products. For example, companies offer different sizes of products, different colors

and different appearances, etc.

The transparent form of customization is driven by modularity. Customers as co-designers simply

combine different parts or parameters of a predefined set of components according to their needs.

Products are in this way customized for customers. For example, car producers offer own car

configurators on their webpages, where the customer can choose from groups of component options

and build the car of their specifications.

Collaborative Customization is when the producer conducts a dialogue with the individual customer.

This form of customization allows a customer to select from groups of standardized

component/parameter options and specify undefined features or properties of the product. An example

of Collaborative Customization can be taken from the shoe industry. Customers can customize their

own products online and give them specific lines. These simplified descriptions of the four types of

customization are graphically depicted in Figure 1.1.

 29

Product
change
dialogue with individual customer

groups of
Collaborative component customized
Customization options by product
producer
Design
Manufacturing
group of
components
specified by
customer
high groups of component selection from groups of
options by producer component options

customized
Transparent product
Customization

Design
Manufacturing

groups of
standardized products

components
 option
Cosmetic
Customization

Design for
Manufacturing
customer
low
standardized
components products
 option
Adaptive
Customization
for
Design Manufacturing customer

Level of Mass Customization

FIGURE 1.1 Four levels of mass customization

Lampel and Mintzberg [7] have proposed five strategies between Pure Standardization and Pure

Customization, by which business models leading to a customized-oriented production can be

categorized. In Pure Standardization, there is a dominant design targeted to a wide group of customers.

This strategy conforms to Adaptive Customization. The second strategy, classified as Segmented

Standardization, targets small groups of customers that have better choices than in the previous one.

This business model corresponds well with a category of Cosmetic Customization. The next category
 30

is the so called Customized Standardization, where product configurations are designed by customers.

This business model is more or less identical to the Transparent Customization. Tailored Customization

as the further development of MC offers a customer a generic product model that is tailored according

to the customer's individual needs. This business strategy is more or less similar to the model of

Collaborative Customization. And finally, as the highest level of mass customization is considered Pure

Customization, where the optimal level of personalization is achieved and the customer can build the

product according to his individual specifications. These classification frameworks offer important

insights into the development of that manufacturing policy and practice, but it does not mean that these

classification frameworks will fit any original equipment manufacturer.

1.3 Success factors for mass customization

An elementary condition for application of mass customization is consideration and analysis of

alternative strategies aimed to increase competitiveness through innovative product design and

customer satisfaction management. The well-known fact is that overall customer satisfaction is higher

when the product better matches his or her ideal preference. Long standing strategies, by which this

objective can be achieved, are product customization and product proliferation.

A firm pursuing product proliferation offers many different product types with different features,

functions, etc. When applying this strategy, production and logistic costs can be negatively affected by

the number of different products. If increasing numbers of products are provided by a firm on a

replenishment basis, then its suppliers have to follow to expand their product lines. Such a situation

makes it more difficult to forecast demand and call out for a transition from a ’make-and-sell’ model to

a so called ’sense-and-respond’ organization. The first model is focused on production efficiency and

the second one is customer satisfaction oriented. Moreover, the sense-and-respond organization

business model allows all members of the supply chain to adapt to changing market conditions and to

work together seamlessly [8].

Product customization can be defined as producing physical goods that are tailored to a particular

customer's requirements [9]. In the case of a mass customization environment, production and logistics

cost can by negatively affected by a degree of product customization. According to Zipkin [10]
 31

increasing complexity of mass customization processes can potentially limit the degree to which

customization is beneficial to customers. A reasonable degree of customization depends on several

factors, such as the kind of industry a company is part of, the level of manufacturing flexibility, the

clients' wishes and so on. It is rather difficult for firms to find optimal rates of customization for an

existing or new product due to a wide range of opinions represented by numbers of offered product

configurations. For this purpose, the generic concept identifying an optimal degree of product

customization can be used (see Figure 1.2).

Costs,
revenues,
benefits

Revenues

Costs

Tainter´s curve
of complexity

Variety induced
Optimal degree complexity
Degree of
Reasonable range customization

FIGURE 1.2 Generic concept of identification optimal degree of product customization

As outlined above, there are at least two ways to deliver a higher level of product variety, and mass

customization may not always be the best. Therefore, an early and reliable decision whether mass

customization is the right prescription for a firm or not is a critical step toward achieving sustainable

development objectives. To answer this, the following four fundamental factors can be analyzed:

1. Customers' readiness. According to Da Silveira et al. [11] ’mass customization encompasses the

ability of original equipment manufacturers (OEMs) and their suppliers to provide individually

designed products and services to customers in the mass-market economy’. Notable attention is paid to
 32

the role of OEMs, which are selling their products to personal consumers. Guilabert and Naveen [12]

argue that knowing how significant customization is for potential consumers as well as how it varies by

type of product will help producers to implement one of the customization strategies.

Basically, customers' readiness can occur in explicit or implicit form. Based on this categorization, the

following construct of Customers' Customization Readiness (CCR) can be outlined. So called explicit

customers' readiness for buying customized products can be seen in daily life. For example, most people

prefer customizing their furniture by choosing style, size, and finish to suit their individual needs, rather

than purchasing standard products. In such situations mass customization is directed by customer’s

specific needs and whims (see Figure 1.3(a)).

Request for customized products

Unique
2 requirements
Standard products

Dominant offer of
a) standard products
and
customized products 1

Marketer Customer

Delivery of customized products

Order of customized products

Standard
2 requirements
Customized products
Dominant offer of
customized products

b) and
1
standard products

Marketer Customer

Delivery of customized products

FIGURE 1.3 Customers' Customization Readiness construct; a) explicit form; b) implicit form

So called, implicit customers' readiness for buying customized products has to be revealed through

interaction between the marketer and customer by offering a full range of customization options. When

applying such an approach, then one can say that mass customization is pushed by marketer’s options

(see Figure 1.3(b)). Both the types of CCR are highly pivotal in paving a way toward implementing a
 33

mass customization strategy. The proposed CCR construct differs from the Customer Customization

Sensitivity (CCS) construct developed by Hart [13]. The CCS construct is based on two factors:

uniqueness of customer needs and customer sacrifices. According to him the level of CCS is directly

proportional to the uniqueness of customer needs and/or customer sacrifices.

2. Type of products. According to Da Silveira et al. [11] mass customization will never be possible for

all types of products. In this sense, Duray [14] argues that ’the production of standardized modules is

the key to high volume mass customization’. In addition, Tsang and Hu [15] point out that convenient

products for mass customization are those with short life cycles, and Blecker [16] emphasizes that

’companies have to offer tailored products while ensuring short delivery times simultaneously’. Taken

together, the following factors predispose products to mass customization:

- products can be grouped together into a product family;

- products are designed as modules so that they can be easily assembled into different ones;

- product can be delivered in short times.

3. Market characteristics. Some markets are offering full customized product varieties whereas in other

markets they are mostly available as discrete product varieties. Obviously, under specific conditions,

both types of markets can bring opportunities or indicate threats when a company is intending to

implement one of the mass customization strategies. Cavusoglu et al. [17] identified the following

market categories as critical in choosing an optimal customization strategy: competition, segmentation

and cannibalization.

The intensity of competition significantly influences customization possibilities and predetermines a

choice of the strategic patterns of the players in the market. Pine [18] recommends three ways to shift

to mass customization: incrementally over time, more quickly through business transformation, or by

creating a new business. The incremental path toward a mass customization strategy, apart from other

factors, assumes that competing firms that operate in the same reference market are not delivering

customized products.

However, this can be a slow process if competitors are already effectively issuing mass customization.

For companies facing such competition, rapid transition to a mass customization business strategy helps

them remain competitive in the marketplace. Increased need for new products in a highly dynamic
 34

market environment can be reflected by transforming businesses to a higher level of customer

satisfaction. Pine [18] recommends that transforming business in such way can be achieved by creating

a group of related businesses focused on individual customers needs.

Segmentation is often the key to developing a competitive edge. Research conducted by Jiang, [19]

shows that mass customization is not totally the same as segmenting‐to‐one. In this context it has been

highlighted that firms that aim customization at specific consumer segments may not be optimal [17].

The cannibalization effect is understood as the extent to which a product variety reduces a firm’s profits

from the standardized varieties it produces. Yayla‐Küllü et al [20] argue that the cannibalization effect

dominates in a highly competitive market. A useful insight to this rate is provided e.g. by Selladurai

[21].

4. Firms’ readiness for mass customization can be understood as having the right conditions and

resources in place to support the transformation process. Knowing that mass customization means a

huge variety of products by combining a large number of product modules, companies that want to

follow this path firstly have to analyze their technological, organizational and managerial capabilities

to determine whether they are potentially transformative to this strategy. El Kadiri et al. [22] predict

that in mass customized manufacturing environment intelligent devices technologies thanks to

intelligent sensors and actuators will dominate. It is expected that extension of 5G technologies to the

ICT sector will facilitate automation, which brings for manufacturing a new impetus to foster mass

customization. Rapid technological progress continues in additive manufacturing (AM) technologies.

This paradigmatic change in manufacturing poses significant challenges to enterprises to utilize the

technology for mass customization. For example, laser-sintering machines present an attractive mode

of production that has great potential to widely customize products such as bone implants, prostheses,

medical devices, etc. Equally, direct metal laser-sintering is widely and effectively used e.g. to fabricate

metal prototypes. When looking at economic evaluation, cost comparisons between AM technologies

show that traditional processes are more economically effective than AM technologies in high output

quantities (see e.g. [23], [24]).

Thomas and Gilbert [25] in this context state that these viewpoints come from analyses of well-

structured costs of AM and they add that significant benefits and cost savings in AM may be hidden in
 35

the ill-structured costs. An important advantage of AM technologies is design freedom. Reeves et al

[26] point out that due to these design freedoms assembly operations that were previously required to

build a complex component can be reduced. Moreover, AM technologies remove the risk of the long

lead time for the delivery of tooling [27]. When assuming the need for the combination of the above

mentioned trends and technologies, then in particular, mass customization calls for the development of

entirely new business organizations.

1.4 Some current and future trends in mass customization

A background to the development of the mass customization paradigm is connected with the inception

of its term by Davis [28]. MC was later developed by Pine [3], and from this period till now, the topic

of customization has continued to be at the forefront in the consumer world for many years. The global

entrepreneurial environment, which is constantly changing is presently influenced especially by the

increase in the importance of foreign direct investment. Several studies (e.g. [29], [30], [31]) show that

foreign direct investment helps local firms to respond quickly to production changes, and in yielding

improvements on moving to more customized production.

The most important enablers of the mass customization are web-based product configuration systems.

Taking this fact in account, Blecker et al. [32] developed a customer needs model providing decision-

makers with insights concerning product-variety management problems. The basic idea of this model

lies in eliminating product variants that only correspond to the subjective needs. The authors emphasize

that this principle might be incorporated into online configuration systems.

It is expected that diffusion of digital manufacturing technologies will trigger a transformation from

mass customization to mass personalization. Kumar [33] argues that IT capabilities will drive mass

customization programs toward the mass personalization strategy. Mass personalization differs from

mass customization in many aspects. While mass customization assumes stable product architecture

and product modules, for mass personalization possible changes of the basic design architecture and

product features are typical [34]. However, wider acceptance of this strategy in individual industries

will strongly depend on the availability of attainable digital manufacturing tools and other devices

belonging to the ’smart manufacturing’ concept.

 36

1.5 Concluding remarks

Initially, mass customization has been seen as a contradictory approach that cannot lead to

entrepreneurial success. In spite of its conflicting ideas, the existence of mass customization is a reality

thanks to the advances realized especially in the fields of flexible manufacturing and information

technology. As was predicted in earlier as well as more recent literature ([35],[36],[37],[38]), mass

customization becomes an imperative rather than a choice to success and sustainability across business

sectors.

REFERENCES:

1. Tepper, K., Bearden, W. O., and Hunter, G. L., “Consumers’ Need for Uniqueness: Scale

Development and Validation,”Journal of Consumer Research, Vol. 28, No. 1, pp. 50–66 (2001)

2. Piller, F. T. (2004). Mass customization: reflections on the state of the concept. International journal

of flexible manufacturing systems, 16(4), 313-334.

3. Pine B.J., 1993. Mass Customization: The New Frontier in Business Competition, Harvard Business

Review Press; 1 edition (October 1, 1992), pp. 368

4. Tseng M M, Jiao J. Design for Mass Customization. Annals of the CIRP 1998;45:1

5. Kull H., 2015, Mass Customization: Opportunities, Methods, and Challenges for Manufacturers,

Apress, pp. 148.

6. Gilmore, J., H., Pine, B., J., (1997), "The four faces of customization", Harvard Business Review,

Vol., pp. 91-101.

7. Lampel, J., Mintzberg, H., (1996), "Customizing customization", Sloan Management Review, Vol.,

pp. 21-29.

8. Haeckel, S. H., 1999, "Adaptive Enterprise — Creating and Leading Sense-and-Respond

Organizations," Harvard Business School Press.

9. Blecker, T., Friedrich, G., Kaluza, B., Abdelkafi, N., & Kreutler, G. Information and management

systems for product customization. Vol. 7. Springer Science & Business Media, 2005.
 37

10. Zipkin, P. The limits of mass customization. MIT Sloan Management Review, 42(3):81–87, 2001.

11. Da Silveira, G., Borenstein, D. and Fogliatto, F. S. (2001). Mass Customization: Literature Review

and Research Directions. International Journal of Production Economics 72(1): 1–13.

12. Guilabert, M. B., and Naveen D. (2006). Mass customisation and consumer behaviour: the

development of a scale to measure customer customisation sensitivity. International Journal of Mass

Customisation 1(2-3): 166-175.

13. Hart, C. W. (1995). Mass customization: conceptual underpinnings, opportunities and limits.

International Journal of Service Industry Management, 6(2), 36-45.

14 Duray, R. (2002). Mass customization origins: mass or custom manufacturing?. International Journal

of Operations & Production Management, 22(3), 314-328.

15. Tseng, M.M., and Hu. S. J. Mass customization. CIRP Encyclopedia of Production Engineering.

Springer Berlin Heidelberg, 2014. 836-843.

16. Blecker, T. Mass customization: concepts-tools-realization. GITO mbH Verlag, 2005.

17. Cavusoglu, H., and Raghunathan, S. (2007). Selecting a customization strategy under competition:

mass customization, targeted mass customization, and product proliferation. Engineering Management,

IEEE Transactions on, 54(1), 12-28.

18. Pine, B. J. Mass customization: the new frontier in business competition. Harvard Business Press,

1999.

19. Jiang, P. (2000) "Segment‐based mass customization: an exploration of a new conceptual marketing

framework", Internet Research, Vol. 10 Iss: 3, pp.215 – 226

20. Yayla‐Küllü, H. Müge, Ali K. Parlaktürk, and Jayashankar M. Swaminathan. "Multi‐Product

Quality Competition: Impact of Resource Constraints." Production and Operations Management 22.3

(2013): 603-614.

21. Selladurai, R. S. "Mass customization in operations management: oxymoron or reality?." Omega

32.4 (2004): 295-300.

22 El Kadiri, S., Grabot, B., Thoben, K. D., Hribernik, K., Emmanouilidis, C., von Cieminski, G., &

Kiritsis, D. (2015). Current trends on ICT technologies for enterprise information systems. Computers

in Industry.
 38

23. Ruffo, M., Christopher Tuck, and R. Hague. "Cost estimation for rapid manufacturing-laser

sintering production for low to medium volumes." Proceedings of the Institution of Mechanical

Engineers, Part B: Journal of Engineering Manufacture 220.9 (2006): 1417-1427.

24. Deradjat, D, and Tim M.. "Implementation of additive manufacturing technologies for mass

customisation." (2015). International Association for Management Of Technology IAMOT 2015

Conference Proceedings, pp. 2079-2094.

25. Thomas, D. S., & Gilbert, S. W. (2014). Costs and Cost Effectiveness of Additive Manufacturing.

NIST Special Publication, 1176.

26. Reeves, P., Tuck, C., & Hague, R. (2011). Additive manufacturing for mass customization. In Mass

Customization (pp. 275-289). Springer London.

27. Hopkinson, N., Hague, R., & Dickens, P., eds. Rapid manufacturing: an industrial revolution for

the digital age. John Wiley & Sons, 2006.

28. Davis, S. M. (1989). From “future perfect”: Mass customizing. Planning review, 17(2), 16-21.

29. Barrell, R., & Holland, D. (2000). Foreign direct investment and enterprise restructuring in Central

Europe. Economics of Transition, 8(2), 477-504.

30. Tu, Q., Vonderembse, M. A., Ragu-Nathan, T. S. & Ragu-Nathan, B. (2004) Measuring Modularity-

Based Manufacturing Practices and Their Impact on Mass Customization Capability: A Customer-

Driven Perspective. Decision Sciences 35, no, 2, 147-168.

31. Makino, S., Lau, C.-M., & Yeh, R.-S. 2002. Asset-Exploitation versus Asset-Seeking: Implications

for Location Choice of Foreign Direct Investment from Newly Industrialized Economies. Journal of

International Business Studies, 33(3): 403-421

32. Blecker, T., Abdelkafi, N., Kaluza, B., & Kreutler, G. (2004). A framework for understanding the

interdependencies between mass customization and complexity (pp. 1-15).

33. Kumar, A. (2007). From mass customization to mass personalization: a strategic transformation.

International Journal of Flexible Manufacturing Systems, 19(4), 533-547.

34. Tseng, M. M., Jiao, R. J., & Wang, C. (2010). Design for mass personalization. CIRP Annals-

Manufacturing Technology, 59(1), 175-178.

 39

35. Piller, Frank T., and Mitchell M. Tseng, eds. Handbook of research in mass customization and

personalization. Vol. 1. Singapore: World scientific, 2010.

36. Anderson, D.M., Agile product development for mass customization, Chicago: Irwin, 1997.

37. Kratochvíl, Milan, and Charles Carson. Growing modular: mass customization of complex products,

services and software. Springer Science & Business Media, 2005.

38. Kumar, U., Ahmadi, A., Verma, A. K., & Varde, P. (Eds.). (2016). Current Trends in Reliability,

Availability, Maintainability and Safety: An Industry Perspective. Springer.

Chapter two

Designing assembly lines for mass customization production

systems
Dominik T. Matt, Erwin Rauch,
Free University of Bozen-Bolzano, Italy

ABSTRACT
 40

This book chapter focuses on the design of assembly lines and systems for the production of mass

customized products. It reviews the state of the art in Mass Customization and Mass Customization

Manufacturing (MCM) to give an overview of existing research, actual findings and future trends of

manufacturing and assembly systems for this application. A framework for a systematic design of

assembly for MC products is described. The framework discusses product design requirements, the

design ofMCassembly systems, requirements for MC shop floor management in assembly as well as

opportunities for MC assembly through the fourth industrial revolution (Industry 4.0). Afterwards a real

case study at a medium sized manufacturer of electromotors shows an example of successful assembly

systems design for make to order production with high product variety and mass customized products.

The experience from the case study showsa concept based on the combination of manual assembly and

automated stations as well as the integration of RFID technology in the assembly process. The chapter

closes with a brief summary of the guidelines for assembly systems design for Mass Customization and

the experience form the case study. Further, the conclusions give an outlook to future research activities.

2.1 Introduction

Mass Customization as a combination of mass production and customized production is the response

from the manufacturing industry to the requirements of an increasingly modern and dynamic world.

The society is constantly developing new opportunities and challenges for individuality in consumption.

The modern individualist eats its own cereals selection ordered online, is wearing tailored-made shirts

and drives an individually configured vehicle. The increasing desire for individuality of people

promotes the trend towards Mass Customization. In addition to this consumer-driven trend, however,

there are very often also very practical and pragmatic reasons for the request for customized products.

The producer, which is able to produce a product to low prices and with customer specific characteristics

as quickly as possible, has the highest competitive advantage on the market. Simultaneously, new

technologies such as advanced web technology and Additive Manufacturing technologies have opened

new possibilities for capturing customer requirements and for producing customer-specific products.

This chapter is mainly concerned with the production of customized products with a special focus on

assembly processes and systems. In traditional massproduction, assembly lines are often highly
 41

automated and designed as single model lines or lines with a small number of variants. Due to the high

degree of automation, costs can be reduced to a minimum, but at the same time flexibility is seriously

restricted. As we will see in the chapter, Mass Customization is combining these highly controversial

objectives and allows maximum flexibility producing products with a reasonable cost structure. This

chapter intends to give systems designers a support for the design of Mass Customization Assembly

Systems by means of a systematic framework. In addition to product-related requirements for MC, the

design of MC Assembly Systems design and their operation, also the integration of Industry 4.0 related

topics is treated.

2.2 Theoretical Background

Mass customization has been identified as a competitive advantage strategy by an increasing number

of companies (Da Silveira et al., 2001). The concept of mass customization was first expounded

formally in the book “Future Perfect” by Stanley M. Davis in 1989 (Davis, 1989). Mass customization

means the production of products, which have been customized for the customer, at production costs

similar to those of mass-produced products (Kaplan and Haenlein, 2006). Mass customization allows

customers to select attributes from a set of pre-defined features in order to design their individualized

product, by which they can fulfill their specific needs and take pride in having created a unique result

(Hart, 1995; Schreier, 2006; Stoetzel, 2012). Thus, customization integrates customers with the design

process (Qiao et al., 2006). The primary focus of the product designer should be on providing value to

the end user. To achieve personalization increasing the value for the end user, different authors suggest

a user-centered design approach. (Kramer et al. 2000). The end user, in the sense of Open Innovation

and the Democratization of Design, is more often directly or indirectly involved in the product

development process. Therefore, innovation, in future, will take place not only within the company but

can be seen as an interactive process between the company and the market to generate customized

products (Reichwald &Piller, 2005).

Numerous authors have published articles about mass customization; many of them discussed Mass

Customization from a strategic and economic point of view. Only few research works investigate

technical aspects of manufacturing and manufacturing systems design of mass customized products
 42

(Koren, 2005; Terkaj et al. 2009;Mourtzis et al., 2013; Bednar &Modrak, 2014; Matt et al., 2015).Mass

customization brings radical changes to methods used to operate traditional manufacturing enterprises.

It is changing the way customers make purchases and this has a strong impact on how products are

manufactured (Smirnov, 1999). Mass Customization Manufacturing (MCM) has been gaining

recognition as an industrial revolution in the 21st century. Customers usually can select options from a

predetermined list and request them to be assembled (Qiao et al., 2006). While the manufacturing

industry in the past distributed globally standardized products to keep the production cost and

complexity low, nowadays a customization of products based on customer specific needs is becoming

more and more important (Matt et al., 2015). Simultaneously with this development in the direction of

an increasing number of individual product variants and product configurators, the requests on

productionand the complexity of manufacturing systems increased. Manufacturing systems in a mass

customization environment should be able to produce small quantities in a highly flexible way and to

be rapidly reconfigurable (Qiao et al., 2006; Thirumalai &Sinha, 2011).

In the eighties, the concept of flexible manufacturing systems (FMS) was introduced in response to the

need for mass customization and for greater responsiveness to changes in products, production

technologies, and markets (Wiendahl et al., 2007). Traditional flexible manufacturing system (FMS)

including the popular Japanese lean production manufacturing system, have not reached the flexibility

and adaptability demanded by an MCM system (Koste &Malhotra; 1998, Matt et al., 2014). The concept

of reconfigurable manufacturing systems (RMS) has emerged later in the nineties in an attempt to

achieve changeable functionality and scalable capacity (Koren, 2005). In recent years developed the

concept of Changeable and/or Agile Manufacturing Systems, as the ability of a factory or a production

system to switch from one product family to another, changing the production capacity accordingly

(ElMaraghy & Wiendahl, 2009; Wiendahl et al., 2009). Such a manufacturing system fits the needed

requirements for Mass Customization Manufacturing better than a traditional one.

In addition to the basic design of production systems for Mass Customization several scientists were

engaged in the design of assembly systems in specific. Assembly is one of the most cost effective

approaches to high product variety allowing the production of mass customized products nearly at mass-

production costs (Hu et al., 2011). The aim is a co-design process with an open product architecture
 43

(Koren et al., 2013) in combination with on-demand manufacturing systems, enabling user

participation in design, product simulation, manufacturing, supply and assembly processes that rapidly

meet consumer needs and preferences (Tseng & Hu, 2014). The economy of scale is achieved at the

component level, while economy of scope of high variety is achieved in the final assembly using flexible

and reconfigurable assembly systems (Hu et al., 2011). Modern approaches for designing assembly

lines for mass customization often show two general alternatives for multiple product models: a) a

multi-model assembly line where products are produced on the same line, but in batches for each

product model and b) mixed-model assembly lines, where the product model variants are sufficiently

similar, that they can be assembled simultaneously on the same line (Boysen et al., 2007).Many of these

approaches can be found also in practice. BMW claims that, “Every vehicle that rolls off the belt is

unique” (Zuh et al., 2008). Such a situation presents enormous difficulties in the design and operation

of assembly systems. It has been shown by empirical and simulation studies(Fisher & Ittner, 1999;

McDuffie et al., 1996) that increased product variety has a significant negative impact on the

performance of assembly processes. The higher the number of product variants, configurations or

overall variety, the more complex difficulties in the production design and operational management of

assembly systems or assembly supply chains there are (Modrak et al., 2015).

The latest trend in Mass Customization is digitalization in manufacturing, also known under the term

“Industry 4.0” or “Cyber-Physical-Systems” (CPS). The large potential of Industry 4.0 will be a key

enabler for further developments in Mass Customization Manufacturing (Kull, 2015). Dombrowski et

al., 2013 describe in their work a concept for a Cyber Physical Assembly System (CyPAS) with modular

Cyber Physical Systems (CPS) supporting the handling of a high level customization.Liu et al, 2014

proposes a concept for IoT (Internet of Things) enabled intelligent assembly system, in order to improve

interconnection, perception, efficiency and intelligence of assembly systems. Intelligent, cognitive and

self-optimizing manufacturing systems are able to learn and thereby perform self-determined changes

in production systems (Schmitt et al., 2012). To reach such a next level of changeability it is necessary

to equip manufacturing systems with cognitive capabilities in order to take autonomous decisions in

even more complex production processes with a high product variety (Zäh et al., 2009).
 44

2.3 Framework for the design of highly changeable production systems for Mass Customization

Assembly (MCA)

This book chapter presents a framework for the design of highly changeable production systems for

Mass Customization Assembly. Figure 2.1 illustrates four different stages of the framework, described

later in detail. First of all, product design related aspects for the design of MC products are shown giving

the focus on design aspects for increasing “assemblability” of MC products. Then the design of

changeable assembly systems is described with an approach called “SMART reconfigurability”. A

production system consists not only of the assembly system itself, but also of organizational aspects

about operation and shop floor management. Thus, specific shop floor arguments for Mass

Customization Assembly are explained in the third section. Due to the actual challenge to become

“Industry 4.0”-ready, the last section discusses advanced technologies like RFID, Internet of Things

and other, with the aim to achieve a smart Mass Customization Assembly.

FIGURE 2.1 Framework for Mass Customization Assembly

 45

2.3.1 Product design related aspects for Mass Customization Assembly (MCA)

It is generally estimated that up to 70%–80% of the product lifecycle costs are determined during the

design phase (Whitney, 1988). Therefore, the designer of products should take into account what the

requirements are to simplify and accelerate assembly of mass customized products. Assemblability is

defined as a combination of the following three aspects in assembly (Lefever & Wood, 1996):

1) the number of operations required to assemble the product,

2) the handling of each needed component, and

3) the complexity of the required operations.

A fundamental technique to reduce complexity in manufacturing and assembly processes of products

with a high variety is Postponement. The concept is to design products in such a way that the point of

differentiation into multiple product variants is delayed as much as possible within the production

sequence. Products, where the concept of postponement is properly applied increase the capability of

an assembly system to be flexible and easily reconfigurable (Lee & Billington, 1994). Another concept

in Product Design for MC is modularity. Modularity is integral to a successful implementation of mass

customization since it forms the vital component of flexibility (Nambiar, 2009). According to Pine

(1993), developing modular products is the best method to achieve mass customization. Modular design

can address the need for a high number of product variants and further allow a higher degree of

automation in the assembly line (Salonitis, 2014). A successful best practice example of modular

product design is Scania AB in Sweden. The case includes eight types of cabs with thousands of variants

within each type. All cabs are produced on one line and are built up from a standardized assortment of

modules and components (Eastman, 2012). A further concept for Product Design is “Design for

Assembly” (DFA). The goal of DFA is to help designers explicitly consider the assembly process and

ultimately design products that are assembled with the minimum required number of parts in the most

efficient and economical way possible in order to reduce error and cost (Anisic & Krsmanovic, 2008;

Arnette et al., 2014). Typical practical rules for DFA are (Lefever & Wood, 1996; Boothroyd et al.,

2010; Jakubowski & Peterka, 2014):

- reduce the total number of parts,

- develop a modular design,

 46

- use of standard components,

- design parts to be multi-functional,

- minimize assembly directions,

- minimize complexity of assembly operations,

- minimize handling of parts and products.

An interesting approach for the design of mass customized goods in six steps is shown by Hernandez

et al. (2007):

1. Definition of the space of customization.

2. Formulation of anobjective, which can often be the minimization of cost.

3. Identification of modes for managing product variety such asadjustable controls, modular

combinations and dimensionalcustomization that are used to customize the product.

4. Determination of the number of hierarchy levels and allocation of themodes for managing

product variety to these levels.

5. Formulation of a multi-stage optimizationproblem.

6. Problem Solving with focus on cost-effectiveness, suitability for small or large variety in

product specification and adaptability of the product line.

 47

FIGURE 2.2Design rules in Product Design for Mass Customization Assembly

2.3.2 Design of flexible and reconfigurable Mass Customization Assembly Systems

Based on the concept of flexibility and reconfigurability of manufacturing systems this section outlines

design guidelinesfor Mass Customization Assembly Systems. The concept of “SMART

reconfigurability” will be explained (Matt et al., 2014) where assembly systems should be designed in

way to be quickly adapted to varying product variants. The proposed approach includes the principles

and concepts already applied in industrial manufacturing to design flexible and changeable production

systems and extends them fulfilling the requirements of mass customization oriented manufactures.

These companies have to adapt their manufacturing and assembly systems frequently and temporarily.

Sometimes the requirements for the next reconfiguration are not even known in advance or are

depending of personalized features of the product. Thus, a manufacturing or assembly system, which is

able to fulfil these requirements, may be called also a “smart” system. To obtain a highly flexible and
 48

reconfigurable assembly system for a mass customization environment five major requirements could

beidentified (see also Figure 2.4):

• Scheduled

• Modular

• Adaptable

• Rapid and

• Temporary.

FIGURE 2.3SMART Reconfigurability approach (Matt et al., 2014)

In this concept the term SMART stands for

a) Scheduled and planned changeovers,

b) Modular and decoupled assembly stations,

c) Adaptable stations to a various number of configurations,

d) Rapid and quick ability to change from one configuration to the other and

e) Temporary, due to the very limited time period of validity of assembly configurations.
 49

Scheduled stands for planning and scheduling every major reconfiguration of the assembly system.

Often inefficiencies arise by inadequate planning of the necessary reconfigurations in assembly.Unlike

in a traditional production, the assembly of a variety of different and personalized product variants

requires anaccurate planning, for example with the support of digital planning tools such as

Manufacturing Execution Systems (MES). The information-transfer between order processing and

production about upcoming varying ordershas to be accelerated, so that a suitable planning and

scheduling of necessary reconfigurations can be enabled and carried out in time. Therefore, the

implementation of a precise and a perfectly working production planning is a prerequisite for

reconfiguration in assembly. Also transparency in order sequence and regular coordination meetings

between production and production planning are necessary elements. For this purpose, in the case study,

the role of the manufacturing system designer was introduced. When the production schedule is defined,

the production engineer becomes responsible to define, organize and execute all the necessary changes

of the production system to guarantee a material-flow-oriented layout and the availability of the required

tools and devices for an efficient and ergonomic assembly and material supply.

Modular means that the individual assembly work stations can be decoupled. This leads to an increased

flexibility in the use and reuse of process modules representing the key functionality of modularity.

Increasing flexibility on one side, the reconfiguration needs solutions for quick change-overto

achieveshortorder execution times and setup times. In relation to the growing demand for modular

systems, the producers of assembly systems or cells reacted in the last years developing and offering

such systems. On the German market the first pioneers were producers like teamtechnik with a system

named TEAMOS, the ZBV-Automation with his system named CORACell as well as Paro offering in

2007 the first modular assembly systems (Drunk, 2007). In the sector of mass customization, these

characteristics to adapt an assembly system as quickly as possible to a new product or variant are

indispensable. Nevertheless, the aforementioned systems of industrial assembly systems are suitable

also for Mass Customization Assembly.

Adaptable is equivalent to changeable and describes the ability of the assembly system to be

reconfigured or adapted to produce a high variety of products. Thus, such an adaptable pre-assembly

requires an extremely high degree of changeability and flexibility. A high degree of adaptability for
 50

Mass Customization companies includes not only the adaptability of individual machines or assembly

work stations, but also the ability for a quick and easily reconfiguration and setup of the entire assembly

line before beginning a new variant. This includes the design of universal or variably and adjustable

process elements or fixing systems (external flexibility), as well as a high amount of internal flexibility.

Internal flexibility means, that the assembly system is able to be modified and reconfigured for a next

product variant without any physical change of the layout or of the process elements. A simple example

for internal flexibility is the identification of a new product variant through barcode scanning or reading

of a RFID-tag of a new productconfiguration (operating parameters e.g. torque for wrenches) and the

subsequent automatic setup of the assembly line or cell. Today, flexible programmable industrial robots

or assembly devices as well as collaborative robots with an automatic setup-function through controlled

axes are able to reduce setup times substantially.

Rapid describes the need of the assembly systems to switch as soon as possible in another configuration

reducing waste and increasing efficiency in operation. As we know, setups in the assembly

systemsrestricts the available and value-adding time fundamentally and therefore can have a high

impact on the cost of mass customized goods. Thus, a rapid adaptation of the assembly system has a

direct impact to the success and the profitability of mass customization oriented companies. The system

designer has to be very consequent in implementing maximum mobility (wheels, roles, crane hooks,

integrated apertures for forks, etc.), compatibility (standard connectors) and rapid disassembly and

assembly (e.g. use of quick-release fasteners instead of bolts and nuts). It is important to not only

optimize the hardware, but also to train the employees in methods of Lean and Quick Changeover (e.g.

SMED method), to minimize the downtime.

Temporary characterizes the challenge that a Mass Customization Assembly System is only valid for a

very limited period and has to be reconfigured very often. While reconfiguration of assembly in many

other industries such as the electronics and consumer goods industries often is limited only to a switch

to another similar product variant, in Mass Customization the assembly system can change

substantially, depending on the defined range of customization (see also chapter 3.1). Thus, assembly

processes, layouts, material provision, tools are only temporary in nature and are subject to constant

change. Therefore, a system designer should renounce for providing material through rigid systems and
 51

transfer points using more likely mobile or flexible shelving and trolleys. This represents a challenge

for the system designer as well as for all involved support departments such as Quality Management.

Processes and tools for monitoring process stability and quality of products must be constantly adapted

in parallel with the assembly system. To achieve this, many companies try to install screens at the

workplace integrating digital quality management systems that can be easily changed at the moment of

the reconfiguration of the assembly line.

2.3.3 Operation and ShopFloor Management of Mass Customization Assembly Systems

In the last decade, ShopFloor Management was mainly optimized by methods from Lean Production

achieving significant savings and productivity gains. By the introduction of production monitoring

software all data of production can be processed in real time to provide the needed information for the

operative production management. ShopFloor Management comprises not only the control of orders in

real time, but also optimization of inventory in the production, increase of the OEE (Overall Equipment

Effectiveness), reliable maintenance and assistance as well as just-in-time material supply through

organizational measures and IT systems.

Mass customization manufacturing does not only rely on the advances in manufacturing technology.

The organization of the production process is an important prerequisite to customize products efficiently

(Blecker & Friedrich, 2007).Masscustomization production companies must fight with ShopFloor

uncertainty and complexity caused by a wide variety of product components (Zhong et al., 2013).

Planning and scheduling becomes therefore difficult on shop floor level in the case of make-to-order

industries. Due to many different product variants, mixed model production and the need for Just-in-

Sequence (JIS) the material supply and order management at assembly lines is very complex. Many

components of MC products are one of a kind and need to be supplied at the right time of the assembly

process. While production planning and order processing are interested mainly to deliver products on

time, the traditional aim of Shop-Floor management is to maximize machine utilization and productivity

(Huang et al., 2009). The above mentioned challenges in the operation and management of the Shop

Floor in Mass Customization encouraged many firms to introduce ERP (Enterprise Resource Systems)

and later also MES (Manufacturing Execution Systems). Such manufacturing IT systems deployed on
 52

the shop floor enhance manufacturing precision and process flexibility, thus supporting the

development of MC capability. In addition, they enable fast and efficient manufacturing operations by

an integrated access to production-related data (Peng et al., 2011). Modern MES mainly focuses on

ShopFloor operations like scheduling, execution and control, informing Shop-Floor supervisors in

terms of manufacturing progress, equipment status, material delivery and consumption (Blanc et al.,

2008). The ISA 95 standard shows a common framework and terminology of MES processes and allows

to classify and compare MES software systems (Cottyn, 2011). Receiving the product information of

the MC product, the MES has the information about suitable and available stations looking-up, which

operations can be performed. It decides the operations scheduled on the machines, choosing the best fit

by means of utilization rate, energy efficiency and delivery time. After planning the production, the

MES sends the processing instructions to the selected stations, and hence starts the production (Keddis

et al., 2013).

Figure 2.4 illustrates typical features of MES systems for Mass Customization Production and

Assembly. The picture shows the different system levels in the IT landscape of manufacturing firms.

Business Intelligence (BI) is a strategic tool for transforming general data into meaningful information

for top management decisions. The ERP (Enterprise Resource Planning) is needed for general data

management and for the execution of sales forecast and a rough capacity planning. Between ERP and

the physical machine level, we can find the MES-level for planning, execution and process/order

control. In the figure are shown typical MC Assembly related features of typical MES systems to

support operations and Shop-Floor management. This features range from the simple collection of

production and machine data to performance analysis, resource management, advanced planning and

scheduling functions, order release and control as well as tracking and tracing.
 53

FIGURE 2.4 Typical elements of MES systems for Mass Customization Assembly

2.3.4 Smart Mass Customization Assembly Systems - opportunities through the fourth industrial

revolution

The first industrial revolution was characterized by the mechanization of production while the second

industrial revolution introduced mass production, followed by the third digital industrial revolution by

the use of electronics and IT as well as automation. Industry 4.0 refers to the fourth industrial revolution

with a technological evolution from embedded systems to Cyber-Physical Systems (CPS) and the

introduction of the Internet of Things into production. The main objectives of Industry 4.0 are the

reduction of complexity in industrial production processes by the use of “intelligent" structures and CPS

enhancing decentralized intelligence and networking with the interaction of a real world and the virtual

environment. New digital ICT and web technologies seem to act as enablers to shift from “centralized”

to “decentralized” production where the product communicates with the manufacturing system and the

linked software applications and databases. In the context of Industry 4.0 new ICT technologies act as

enabler of smart, autonomous and self-learning factories. With the technological opportunities of

Industry 4.0 also ShopFloor Management should be become mobile, digitally visualized and even more

smart and intelligent at the same time.Industry 4.0 will only work, if machines can communicate
 54

between each other and material flows are tracked by RFID or similar technologies throughout large

sections of the industry. Self-controlling systems communicate via the Internet among themselves and

with the human operators in assembly and manufacturing systems (Brettel et al., 2014).

Figure 2.5 shows an overview of possible applications of Industry 4.0 technologies in the field of Mass

Customization Assembly. New technologies will have a major impact on how jobs or product data is

generated. Through the creation of online platforms for product design and product configuration items

are individually created and forwarded directly to the producing companies. Smart Connectivity here is

not only the keyword for the exchange of data between customer and company, but also for transparency

in the communication between machines, plants, products, warehouses and tools.

FIGURE 2.5 Industry 4.0 applications in the field of Mass Customization Assembly
 55

Smart Products also make up almost a prerequisite for the use of many industry 4.0 applications in

production and assembly. Through RFID-enabled real-time manufacturing execution systems,

assembly objects can be tracked and traced systematically collecting real-time production data. This

allows to identify disturbances and inefficiency at the assembly line much early than with traditional

instruments (Zhong et al., 2013).RFID with agent-based work flow management system was focused

to improve the shop floor’s configurability and reusability during manufacturing process (Zhang et al.

2014). If products cannot be equipped by RFID it is possible in many cases to create smart containers

or boxes equipped with RFID technology.Another application of advanced technologies in production

and assembly are Smart Sensorics. Through a wireless data acquisition in the assembly process or in

the machine, data can be checked, used and tracked every time and anywhere in the world. Other

applications for smart sensorics can be found in the detection of incorrect products for rejection as well

as identification of preventive maintenance actions in the process or the monitoring of machine

performance and productivity. As just mentioned smart sensorics enables self-diagnostics and thus also

Self-Configuration of Intelligent Assembly Systems. When the system identifies a product change the

assembly system should be able to perform autonomously a quick changeover without the need of

manual setup activities. In most of the cases, a full-automated assembly process will not be suitable or

possible. Therefore,in Industry 4.0 the human operator plays a major role towards productivity

improvement. Through human centred Digital Assistance Systems productivity of employees can be

increased, the failure rate can be reduced and fatigue as well as non-ergonomic processes can be

improved. Such digitally support technologies are pick-by-light or pick-by-vision in material provision

or also the visualization of digital work instructions and information on a screen.Using employee badges

or RFID-tags these information could be visualized in the language of the employee. Location Based

Services can make assembly processes more productive in the future. Tracking the position of

production supervisors or maintenance staff they can be brought via push message and intelligent alarm

notification to the closest decision points or machines, thus reducing unplanned downtime of assembly

systems.In particular, the transparency of real-time data shows great potential in production and

assembly. The information about assembly performance at the end of the shift is only a retrograde point

of view and allows the supervisor not to introduce correction measures. Through a mobile Real-time
 56

Production Monitoring every level of the company - from top management to employees on the line –

at all times, has transparency about the status of the assembly system. This allows problems to be tackled

immediately and thus productivity and OEE can be increased.

2.4 Industrial Case Study

The concepts and design approaches for Mass Customization Assembly have been applied in a project

with a medium-sized manufacturer of electric motors. The company has about 400 employees and is

one of the world's leading manufacturers of electrical drive systems for mobile vehicles. The company's

strength lies not in producing cheap standard motors for standard applications, but in the development

and production of customized drives with individual customer requirements. The company currently

produces approximately 900 different variants of AC and DC motors. Of these, only 20 types are A-

articles with a large demand of up to 10 units per day, while the remaining motors have small to medium

order quantities. Approximately 60% of the motors sold are masscustomized products or small series

with a quantity of 1-10 units per year. Based on this analysis, the company wanted to developed a new

concept for the assembly, providing a separation of the production into a "high runners" area and in a

Mass Customization area. Figure 2.6 summarizes the most important improvements in the Mass

Customization production area. In what follows these improvements are explained more in detail.

In a first step, the project team analyzed the products through an ABC analysis and a detailed analysis

of every single assembly step. The aim was to determine the similarity of products and variants in their

assembly processes, and to cluster them into product families. Based on the results of the investigation

and the determined product families the articles were taken together to develop a new optimized

assembly concept for every product family. In this first phase of the project, various meetings were held

also with engineering to define for each product family the range of individualization. The following

main criteria were defined: number of wire turns, size limits of the motor, length limits of the iron

plates and the individualization limits in terms of customer-dependent optional parts and components.

In a next step, the integration of RFID tags on the product has been discussed with Engineering. By

applying RFID labels, the product can be equipped with intelligence: production data can be saved

during the production process, the product can be traced throughout the entire production process and
 57

at the same time, internal logistic processes can be synchronized with the real-time order status. In

addition to the logistics optimization in the production process thereby also the logistic process in

shipping and in the preparation of transport documents can be realized more efficient. The work effort

for counting the motors manually, booking them out of the warehouse system and then assigning them

to the transport documents can be eliminated or reduced to a minimum. With the integration of RFID

gates in the shippingarea, the forklift truck driver can drive through the gate without waiting, attaches

the automatically printed pallet label and then transport documentswill be automatically generated. At

the same time also the error rate is reduced, as the system performs an automatic comparison between

the goods on the truck and ordered goods.

Every assembly station is equipped with a digital touchscreen monitor.When the productpasses, the

related information such as drawings, work instructions or notes (ultimate happened errors) is visualized

on the screen identifying the product with the RFID tag.

In the assembly line, the workpiece carrier was standardized for every product family. If standardization

was not possible, the carrier was designed such that employees can switch easily and rapidly between

the individual sizes by folding workpiece holders. In the past, the needed pressing processes were

executed in different assembly stations with up to 3 consecutive presses. In the new assembly system,

those processes were aggregated into a fully automated universal pressing device. The new universal

pressing device identifies autonomously the product by reading the RFID-tag starting a setupprocess

when the product is changing. Using electric drives in the device, distances and positions for pressing

are adapted fully automated (self-configuration). Thus, the time for changeover could be reduced to a

minimum without any manual tool change, as was needed in the “old” assembly lines.

In the new assembly process, several smart sensors and cameras were integrated for quality checks and

control of process stability. For example, the automated process control system performs an automated

presence control of locking rings and sealings. Through the help of smart sensorics, the system measures

process parameters as well as environmental parameters (temperature, humidity,…) to investigate long-

term statistics about the influence of environmental changes on process stability and quality. In addition,

data requested by the customer, like the force-distance relation as well as final quality testing results

are stored on the RFID tag.

 58

With the use of a warehouse management software and an Electronic Kanban system also material

provision for assembly could be improved.Purchased parts and components can now be supplied just-

in-time and, if needed, just-in-sequence, checking the actual progress of the product through defined

scanning points in the production and assembly process.

Data is now collected at every single station by a MES system. The MES system supports not only the

collection of machine and process data (utilized for initiatingcontinuous optimization measures on the

shop floor level) but also for planning functions working with a real-time manufacturing execution

system. Due to a high transparency in production data and machine data, the productivity can be

visualized real-time on monitors in production as well as in a smartphone app for supervisors and top

management.

The new assembly concept has already been implemented successfully in parts and shows first

promising results. A full implementation of all mentioned and elaborated measures will take place in

2016. Thus, the full impact and effects of the new assembly concept will be measurable in 2017.
 59

FIGURE 2.6 Improvements in Mass Customization Assembly at a manufacturer of electric motors

2.5 Conclusion and outlook

Due to the increasing individualization of products and increasingly complex customer requirements,

the trend of mass customization will rise even further in the future. Similarly, as shown in the industrial

example, companies have to adapt their production accordingly to manage the realization of customer-

specific products with batch size 1 or smallest lot sizes, without a disadvantageous increase in costs.

Especially in assembly, small batch sizes and a wide range of variants lead to a high degree of

complexity in the design of the assembly system and in its operation. Mass customization causes not

only difficulties in machinery design, but also a certain cognitive complexity for operators in assembly

environments. As indicated in the chapter, Mass Customization Assembly systems have be designed in

a changeable and flexible way. A framework for the design of such assembly systems was introduced.
 60

It was pointed out, that a lot of the problems can be solved already in Engineering by the approach of

Design for Assembly. The approach of SMART Reconfigurability gives system designers a set of

guidelines for the planning and realization of modern assembly systems for MC. In addition, the

framework also shows the importance of the operational shop floor management in Mass Customization

Assembly systems. At this level, manufacturing companies are introducing increasingly specific IT

tools such as Manufacturing Execution Systems. Finally, new opportunities of advanced technologies

were discussed. For the purposes of Industry 4.0 and the Internet of Things, assembly systems should

be designed smart and intelligent. In modern assembly systems, important information can be sent

between machines and humans in real-time and visualized continuously and anywhere with the scope

to improve efficiency.

Future research should examine the possibilities of 4.0 technologies, to support the implementation of

Mass Customization in assembly as well as in the entire production system. Further investigation in the

field of modern and intelligent production and assembly systems thus represents a significant “booster”

for Mass Customization in Industry.

REFERENCES

Anisic, Z., & Krsmanovic, C. (2008). Assembly initiated production as a prerequisite for mass

customization and effective manufacturing. Strojniški vestnik, 54(9), 607-618.

Arnette, A. N., Brewer, B. L., & Choal, T. (2014). Design for sustainability (DFS): the intersection of

supply chain and environment. Journal of Cleaner Production, 83, 374-390.

Bednar, S.& Modrak, V. (2014). Mass Customization and its Impact on Assembly Process' complexity.

International Journal for Quality Research, 8(3), 417–430.

Blanc, P., Demongodin, I., & Castagna, P. (2008). A holonic approach for manufacturing execution

system design: An industrial application. Engineering Applications of Artificial Intelligence, 21(3),

315–330.

Blecker, T., & Friedrich, G. (2007). Guest editorial: mass customization manufacturing systems. IEEE

Transactions on Engineering Management, 1(54), 4-11.

 61

Boothroyd, G., Dewhurst, P., & Knight, W. A. (2010). Product design for manufacture and assembly.

CRC Press.

Boysen, N., Fliedner, M., & Scholl, A. (2007). A classification of assembly line balancing problems.

European Journal of Operational Research, 183(2), 674-693.

Brettel, M., Friederichsen, N., Keller, M., & Rosenberg, M. (2014). How virtualization, decentralization

and network building change the manufacturing landscape: An Industry 4.0 Perspective. International

Journal of Science, Engineering and Technology 8 (1), 37, 44.

Cottyn, J., Van Landeghem, H., Stockman, K., & Derammelaere, S. (2011). A method to align a

manufacturing execution system with Lean objectives. International Journal of Production Research,

49(14), 4397-4413.

Da Silveira, G., Borenstein, D., & Fogliatto, F. S. (2001). Mass customization: Literature review and

research directions. International Journal of Production Economics, 72(1), 1-13.

Davis, S. M. (1989). From “future perfect”: Mass customizing. Planning review, 17(2), 16-21.

Dombrowski, U., Wagner, T., & Riechel, C. (2013). Concpt for a Cyber Physical Assembly System. In

Assembly and Manufacturing (ISAM), 2013 IEEE International Symposium on (pp. 293-296). IEEE.

Drunk, G. (2007). How much modularity needs the assambly? Ten years of modular platforms for

assembly systems (in german: Wie viel Modularität braucht die Montage? Zehn Jahre modulare

Plattformen für Montageanlagen), Report xpertgate.

Eastman, C. M. (2012). Design for X: concurrent engineering imperatives. Springer Science & Business

Media.

ElMaraghy, H. A., & Wiendahl, H. P. (2009). Changeability–an introduction. In Changeable and

Reconfigurable Manufacturing Systems (pp. 3-24). Springer London.

Fisher, M. L., & Ittner, C. D. (1999). The impact of product variety on automobile assembly operations:

Empirical evidence and simulation analysis. Management science, 45(6), 771-786.

Hart, C. W. (1995). Mass customization: conceptual underpinnings, opportunities and limits.

International Journal of Service Industry Management, 6(2), 36-45.

Hernandez, G., Allen, J. K., Mistree, F. (2006). A Theory and Method for Combining Multiple

Approaches for Product Customization. International Journal of Mass Customization 1(2–3), 315–339.
 62

Hu, S. J., Ko, J., Weyand, L., ElMaraghy, H. A., Lien, T. K., Koren, Y., Bley, H., Chryssolouris, G.,

Nasr, N. & Shpitalni, M. (2011). Assembly system design and operations for product variety. CIRP

Annals-Manufacturing Technology, 60(2), 715-733.

Huang, G. Q., Fang, M. J., Lu, H., Dai, Q. Y., Liu, W., & Newman, S. (2009). RFID-enabled real-time

mass-customized production planning and scheduling. In 19th International Conference on Flexible

Automation and Intelligent Manufacturing (FAIM), pp. 1091-1103.

Jakubowski, J., & Peterka, J. (2014). Design for manufacturability in virtual environment using

knowledge engineering. Management and Production Engineering Review, 5(1), 3-10.

Kaplan, A. M., & Haenlein, M. (2006). Toward a parsimonious definition of traditional and electronic

mass customization. Journal of Product Innovation Management, 23(2), 168-182.

Keddis, N., Kainz, G., Buckl, C., & Knoll, A. (2013). Towards adaptable manufacturing systems. In

Industrial Technology (ICIT), 2013 IEEE International Conference on pp. 1410-1415. IEEE.

Koren, Y., Reconfigurable Manufacturing and Beyond (Keynote Paper),in Proceedings of CIRP 3rd

International Conference on Reconfigurable Manufacturing, Ann Arbor, Michigan, USA, 2005.

Koste, L. L., & Malhotra, M. K. (1999). A theoretical framework for analyzing the dimensions of

manufacturing flexibility. Journal of Operations Management, 18(1), 75-93.

Kramer, J., Noronha, S., & Vergo, J. (2000). A user-centered design approach to personalization.

Communications of the ACM, 43(8), 44-48.

Kull, H. (2015). Intelligent Manufacturing Technologies. In Mass Customization (pp. 9-20). Apress.

Lee, H. L., & Billington, C. (1994). Designing products and processes for postponement. In

Management of Design (pp. 105-122). Springer Netherlands.

Lefever, D., & Wood, K. (1996, August). Design for assembly techniques in reverse engineering and

redesign. In ASME Design Theory and Methodology Conference (pp. 78712-1063).

Liu, M., Ma, J., Lin, L., Ge, M., Wang, Q., & Liu, C. (2014). Intelligent assembly system for mechanical

products and key technology based on internet of things. Journal of Intelligent Manufacturing, 1-29.

Matt, D. T., Rauch, E., & Franzellin, V. (2014). SMART Reconfigurability Approach in Manufacture

of Steel and Façade Constructions. In Enabling Manufacturing Competitiveness and Economic

Sustainability (pp. 29-34). Springer International Publishing.

 63

Matt, D. T., Rauch, E., & Dallasega, P. (2015). Trends towards Distributed Manufacturing Systems and

modern forms for their design. Procedia CIRP, 33, 185-190.

MacDuffie, J. P., Sethuraman, K., & Fisher, M. L. (1996). Product variety and manufacturing

performance: evidence from the international automotive assembly plant study. Management Science,

42(3), 350-369.

Modrak, V., Marton, D., & Bednar, S. (2015). The Influence of Mass Customization Strategy on

Configuration Complexity of Assembly Systems. Procedia CIRP, 33, 539-544.

Mourtzis, D., Doukas, M., & Psarommatis, F. (2013). Design and operation of manufacturing networks

for mass customisation. CIRP Annals-Manufacturing Technology, 62(1), 467-470.

Nambiar, A. N. (2009, July). Mass Customization: Where do we go from here?. In Proceedings of the

World Congress on Engineering (Vol. 1, pp. 1-3).

Qiao, G., Lu, R. F., & McLean, C. (2006). Flexible manufacturing systems for mass customisation

manufacturing. International Journal of Mass Customisation, 1(2-3), 374-393.

Xiaosong Peng, D., Liu, G., & Heim, G. R. (2011). Impacts of information technology on mass

customization capability of manufacturing plants. International Journal of Operations & Production

Management, 31(10), 1022-1047.

Pine, B. J. (1999). Mass customization: the new frontier in business competition. Harvard Business

Press.

Reichwald, R., & Piller, F. (2005). Open Innovation: Kunden als Partner im Innovationsprozess

(Customers as partners in the innovation process). Electronically published: http://www. impulse.

de/downloads/open_innovation. pdf, accessed on 18.10.2013.

Salonitis, K. (2014). Modular design for increasing assembly automation. CIRP Annals-Manufacturing

Technology, 63(1), 189-192.

Schmitt, R., Brecher, C., Corves, B., Gries, T., Jeschke, S., Klocke, F., ... & Pyschny, N. (2012). Self-

optimising Production Systems. In Integrative Production Technology for High-Wage Countries (pp.

697-986). Springer Berlin Heidelberg.

Schreier, M. (2006). The value increment of mass‐customized products: an empirical assessment.

Journal of Consumer Behaviour, 5(4), 317-327.

 64

Smirnov, Y. (1999). Manufacturing planning under uncertainty and incomplete information. In

American Association for Artificial Intelligence Spring Symposium.

Stoetzel, M., Exploiting Mass Customization Towards Open Innovation, in Proceedings 5th

International Conference on Mass Customization and Personalization in Central Europe, MCP-CE

2012, Novi Sad, Serbia, 2012.

Terkaj, W., Tolio, T., & Valente, A. (2009). A review on manufacturing flexibility. In Design of

Flexible Production Systems (pp. 41-61). Springer Berlin Heidelberg.

Thirumalai, S., & Sinha, K. K. (2011). Customization of the online purchase process in electronic

retailing and customer satisfaction: An online field study. Journal of Operations Management, 29(5),

477-487.

Tseng, M. M., & Hu, S. J. (2014). Mass customization. In CIRP Encyclopedia of Production

Engineering (pp. 836-843). Springer Berlin Heidelberg.

Whitney, D.E. (1998). Manufacture by design. Harvard Business Review, 66, 83–91.

Wiendahl et al., 2009

Wiendahl, H.-P., Reichardt, J.& Nyhuis, P. Manual Factory Planning - concept, design and

implementation of agile production (Handbuch Fabrikplanung – Konzept, Gestaltung und Umsetzung

wandlungsfähiger Produktionsstätten), Carl Hanser, Munich/Vienna, 2009.

Wiendahl, H. P., ElMaraghy, H. A., Nyhuis, P., Zäh, M. F., Wiendahl, H. H., Duffie, N., & Brieke, M.

(2007). Changeable manufacturing-classification, design and operation. CIRP Annals-Manufacturing

Technology, 56(2), 783-809.

Zaeh, M. F., Beetz, M., Shea, K., Reinhart, G., Bender, K., Lau, C., Ostgathe, M., Vogl, W., Wiesbeck,

M., Englhard, M., Ertelt, C., Rühr, T., Friedrich, M. & Herle, S. (2009). The cognitive factory. In

Changeable and reconfigurable manufacturing systems (pp. 355-371). Springer London.

Zhong, R. Y., Dai, Q. Y., Qu, T., Hu, G. J., & Huang, G. Q. (2013). RFID-enabled real-time

manufacturing execution system for mass-customization production. Robotics and Computer-

Integrated Manufacturing, 29(2), 283-292.

Zhang, Y. F., Zhang, G., Wang, J. Q., Sun, S. D., Si, S. B. & Yang, T. (2014). Real-time Information

Capturing and Integration Framework of the Internet of Manufacturing Things.

 65

International Journal of Computer Integrated Manufacturing, 28 (8), 811-822.

Zhu, X., Hu, S. J., and Marin, S. P., 2008, "Modeling of Manufacturing Complexity in Mixed-Model

Assembly Lines", Journal of Manufacturing Science and Engineering, 130(5), pp. 051013

Chapter three

Role of Information Systems in Mass Customization

Arun N. Nambiar
California State University – Fresno, CA, USA
 66

ABSTRACT

The dawn of the globalization era has been both a boon and a bane for companies and enterprises

engaged in providing goods and services. Globalization opened up new markets thus allowing

companies to expand beyond their traditional market base. It also provided access to new sources of

raw materials, latest technologies and a skilled workforce. However, it also engendered stiff competition

with too many players vying for the same market base. Seeing the shift, customers began to demand

products and services that have been customized to their needs thus making customers truly the kings

and queens. As a result, companies with the help of practitioners and researchers continuously strive to

find new ways to provide the product/service mix that will attract customers while keeping the costs

down. Mass customization is a paradigm that focuses on providing customers with a customized

experience at a reasonably low cost. This paradigm requires significant data collection across the entire

operation of the enterprise and extensive analysis of the collected data to identify information related

to customer preferences and how they translate into product or service features. With the advances in

information technology in the recent years (also a result of globalization), it is possible to leverage

information systems to facilitate this data collection and information processing. This chapter will

identify key elements and features of a holistic mass customization information system that

encompasses all activities involved in the customization process.

3.1 Introduction

Companies world over are continuously striving to generate profits through increased market share and

reduced operating costs. With the help of advances in technology and access to global markets and

workforce, more and more companies are entering the milieu resulting in increased competition. Gone

are the days where customers were happy with mundane and uniform off-the-shelf products. Customers

are becoming ever more demanding and the stiff competition as a result of globalization creates a perfect

market condition for the customer to be truly the king and queen. This serves as an impetus for both

practitioners and researchers to come up with ways to reduce costs and provide better value to the

customer. It has been shown (Agouridas et al 2001) that the value of a product as perceived by the
 67

customer has evolved over the years from being focused on price and quality to include customized

service, ease of placing an order, quick order turn-around among other aspects of the product. Thus,

there is increased emphasis on the value of the product and a renewed focus on the concept of value

chain (Porter, 1985). Focusing on the value chain forces companies to examine each activity that is

carried out within its enterprise in the light of its value to the product or service being offered by the

company.

With customers considering the ability to customize products or services as a significant part of its

value, enterprises quickly adapted to the new scenario by introducing a plethora of choices for

customers to choose from in the hope that some or all of its offerings might attract customers. However,

the profusion of choices only served to confound the customer further since neither did the customer

understand the variety of features offered nor were they packaged appropriately for it to be of value to

the customer. At the other end of the spectrum were companies that were providing individualized

products albeit at an exorbitantly high price thus putting it out of reach for the large segment of the

population.

Joseph Pine, often considered as the father of mass customization, defined mass customization as (Pine,

Victor & Boynton, 1993, Pine, 1999) “ providing tremendous variety and individual customization, at

prices comparable to standard goods and services ...with enough variety and customization that nearly

everyone gets exactly what they want.“ As the definition suggests, the idea is to provide customers their

version of the product. This is a tectonic shift from the typical approach of one size fits all that

companies have been practising so far. The impetus for this change is competition and the need to gain

market share in order to stay profitable. It has been shown (Jiao, Ma & Tseng, 2001) that this approach

provides better value to customers. However, it has also been shown (Svensson & Barford, 2002) that

high operating cost is one huge deterrent.

Computers have been successfully leveraged by many manufacturing companies since the 1950s

(Cooper 1957). Some of the initial uses of computers were in the areas of scheduling and production

planning (Kocchar 1978, Kocchar 1981, Kimber 1988). Computers have been used mainly for brute
 68

force analysis or to solve mathematical models to determine optimal solutions to manufacturing and

scheduling problems. Simulation (Yan & Kou, 2009) has also helped organizations evaluate alternative

solutions for process improvements. As organizations grew and enterprises became more

geographically dispersed, there arose a need for system-level information systems. Solutions such as

Materials Resource Planning (MRP) and Enterprise Resource Planning (ERP) software became more

popular allowing companies to leverage information technology to coordinate its dispersed activities.

In recent years, cloud computing has become widely popular where idle computing resources from

around the world are harnessed to solve computationally intensive problems. No longer need companies

be tied down by the legacy computing equipment or not having the resources to make huge capital

investments in expensive computing equipment. Numerous solutions are offered now as a service over

the cloud thus allowing companies to leverage the network to achieve its processing and computing

needs. Extending the concept of cloud computing over to manufacturing, companies are beginning to

seek out manufacturing resources on a need-basis (Husejnagic & Sluga, 2015) instead of investing in

high-end capital equipment.

Information systems facilitate data capture, storage, analysis and retrieval. All of these features are

critical in a mass customization environment to manage customer preferences and their relation to

product features in a cost-effective manner. In fact, information systems is one of the two pillars of

mass customization (Pine, 1999). Some of the typical features of an information system in mass

customization include knowledge management and design configuration. This chapter will explore the

role of information systems in efficient and successful implementation of mass customization.

Additionally, this chapter will identify some of the characteristics of a typical effective and efficient

information system.

3.2Background

Mass customization is often compared against mass production since the prevailing practice is to

produce goods and services on a large scale to benefit from the economies of scale. The mass production
 69

system was pioneered by Henry Ford (Ford and Crowther 1922) in the 1920s where large quantities of

limited products were manufactured and assembled. Customers had limited, if any, choice in the

products and services. A mass production system is typically vertical in nature where each partner aims

to mass produce goods for the customer downstream. There is very limited information sharing between

the partners in the value chain beyond the immediate supplier-customer pair. This results in information

failure or lack of information which can lead to significant problems. The well-known bullwhip effect

(Lee, Padmanabhan & Whang, 1997) where demand information is distorted as it travels up the supply

chain value chain is a cause for excess production and the resulting excess inventory and increased

waste in many industries.

Mass customization, introduced by Davis (Davis, 1997) and developed by Joseph Pine (Pine, 1999),

requires a significantly different approach. The partners in a mass customization value chain need to be

more cohesively integrated and need to work closely with each other in order to be able to deliver

customized products at mass production prices. The key features of the mass customization paradigm

include product differentiation, cost reduction and responsiveness (Nambiar 2009a, Ngiatedema, 2012).

In order to achieve these features, it is imperative that partners in the supply chain work very closely

with each other. Thus, in stark contrast to mass production and other manufacturing paradigms

(Nambiar 2010), mass customization system is considered to be horizontal with increased collaboration

and information sharing. Thus, the information system needs are vastly different (Dean, Tu & Xu, 2008;

Ngiatedema, 2012) for both these systems.

Understanding the mind of a customer is critical for the success of mass customization. A typical

customer goes through five stages while making a purchasing decision. These include (Dibb, Simkin,

Pride & Ferrell, 2001) identifying the need or problem, searching for more information about various

products, evaluating all alternatives, making a purchase and finally evaluating the product. This is

particularly important for mass customization. If a company is able to interact with the customer at the

first stage of identifying the need or problem, it can offer customized solutions to the need thereby
 70

increasing the probability that the customer would evaluate the proposed alternative favorably and

ultimate make the purchase.

It is also important to understand the factors that influence the customer in making the purchasing

decision. These factors can be broadly classified into four categories (Turban and King, 2003). These

include personal characteristics such as the age of the person, and the level of education. Gender also

plays a significant role in purchasing decisions. Environmental characteristics include the society in

general and what is considered acceptable in the society to which the customer belongs. Some customers

might also be conscious of how they affect the environment and focus on sustainability aspects such as

environmental pollution, recycling, etc. Access to effective customer support is becoming more and

more critical. This together with the logistical aspects of how quickly the product would be delivered

constitute the third category. The fourth category includes marketing and other external factors that

influence the customer. Understanding these four categories of factors influencing the customer have a

significant impact on the success of mass customization initiatives.

It can be seen that information about customer’s decision-making process and factors that influence the

decision-making process is really crucial. All pertinent information needs to be collected and analyzed

to identify patterns which would help the company develop customization options. Information

efficiency has been identified as a critical success factor (Mahajan, Srinivasan & Wind, 2002) for the

retailing sector. With the evolution and ever-increasing popularity of online retailing, information

becomes even more vitally important. It has been shown (Varadarajan and Yadav, 2002) that the online

environment combines the benefits of traditional retailing through its information richness while

lowering the mismatch of information between various players in the supply chain. There is a related

concept called collaborative commerce or c-commerce (Lim, 2003) where partners in the value chain

leverage Internet to improve information exchange.

A system that integrates information from a wide variety of sources such as customer, supply chain

partners including retailers is indispensable (Reichwald, Piller & Moeslein,2000) for a successful mass

customization campaign. This system should be able integrate the different modes of operation as
 71

identified by MacCarthy, Brabazon & Bramham (2003) such as order processing, design, production,

and post-processing. Peng, Liu & Heim (2011) identify four main areas where information technology

can help with mass customization implementation. These include product configuration, product

development, manufacturing and supplier coordination. The needs of such an integrated system has

been nicely summarized by Frutos & Borenstein (2004) as “..provide direct links among main agents

involved in the customization process, namely customer, company and supplier”. The needs of such as

an information system vary based on the type of industry and type of process used. Similar integrated

systems have been proposed for doors and windows (Dean, Tu & Xue, 2008), online retailing

(Vrechopoulos, 2004), housing industry (Shin et al 2008), shoe industry (Dietrich, Kirn & Sugumaran,

2007) and software development (Karpowitz et al 2008).

3.3Mass customization information systems

An information system architecture can be defined (Cook, 1996) as “a conceptual framework that

includes the identification of different components of the information systems environment and their

interrelationships”. However, the needs of the information systems are different in a mass production

environment and a mass customization environment. It has been suggested (Ngiatedema, 2012) that the

information systems architecture for a mass customization environment adopt a more horizontal

structure based on the processes within the organization thus facilitating greater integration. There have

been numerous piecemeal approaches (Chen, Tsai & Tu, 2008; Da Silveira, Borenstein. & Fogliatto,

2001; Dean, Tu & Xue, 2009; Frutos & Borenstein, 2004; Fulkerson, 1997; Piller, 2002; Yao et al 2007;

Zahed & Reddy, 2004) to meet specific information system needs in a mass customization environment.

This is primarily because of the diverse nature of the industry making the development of a single one-

size-fits-all architecture an arduous task. Another interesting approach (Verdouw, 2010) has been to

apply the concept of mass customization itself to developing such systems by plugging together

individual modules as may be the need.. The goals of a such holistic information system architecture

for mass customization (Ngiatedema, 2012) include:

 72

● integrating data across the entire system,

● presenting information to the customer in a clear and lucid manner,

● analyzing copious amounts of data to garner knowledge about customer preferences,

● keeping up with the technological advances

● reducing information distortion

Barranchea (2010) developed a suitability index to determine if a given information system is suitable

to the company’s needs based on a series of criteria. This index allows companies to gage their level of

preparedness vis-a-vis information systems for mass customization initiatives. Despite the subtle

variations across companies and industries regarding information system needs in a mass customization

environment, there are certain key elements of the architecture that are indispensable irrespective of the

industry or size of the company. These have been described below.

3.3.1Customer Relationship Management

The success of mass customization hinges on the ability of companies to capture and maintain customer

preferences. Strauss and Frost (2001) identify the 3 steps involved in a typical customer relationship

management (CRM) process. This includes identification, differentiation and customization. It can been

seen that all the three steps are equally important for the mass customization initiative as well. It is

important to identify the customer needs, differentiate customers into different groups based on their

preferences and customize the offerings for each group. Even though CRM has been around for a really

long time, its application and use in today’s world of Internet and mobile technology is not as

widespread (Feinberg, Kadam, Hokama & Kim 2002). This allows potential developers to integrate

CRM with mass customization systems since both systems are so inter-dependent.

3.3.2Product Configuration System

The very essence of mass customization is the ability to provide highly customized products or services.

However, as the number of product or service features increase, the number of possible combinations

increase exponentially thus making it virtually impossible for a company to provide all of its customers
 73

with the off-the-shelf products or services tailored to their needs. Instead, companies provide customers

the ability to design their own product or service. This provides a truly customized experience for

customers who benefit from receiving a product or service that meets their specific needs. However,

this becomes an onerous task given the plethora of choices that companies are wont to give their

customers. Thus, a configuration system (Forsza & Salvador, 2006) that allows customers and product

designers to design a product or service of their choice based on the available features is an integral part

of the information system needs of a mass customization company. The system needs to not only

provide customers with the ability to design their own products or services but also provide pre-

configured systems generated based on inferred customer choices thus making the customer’s task

easier. The configuration system also needs to provide customers with the relevant information so that

they can make informed decisions regarding their product or service choices. This information needs to

be provided in a clear and lucid format bereft of technical jargon so that it caters to the disparate

backgrounds of its customers.

The scope of such a product configurator (Forza & Salvador, 2006) varies based on the type of the

business practice adopted by the company. At one end of the spectrum is the practice where a company

provides variety with little customization where the customer does not interact with the configurator at

all. For example, consider the case of Tropicana (Tropicana, 2016) orange juice. There is a wide range

of the types of orange juices available from the same company such as juice with pulp, without pulp,

fortified with vitamins or fiber etc. In this case, the customer does not design the product directly.

Instead the company designs the product based on customer input through surveys and questionnaires.

At the other end of the spectrum is a truly customized approach, where the customer is involved right

from the design stage. For example, consider the case of custom-made Rolls-Royce cars (Rolls-Royce,

2016) where the product configurator allows the customers to build their own cars. The success of a

product configurator also hinges on efficiently translating customer needs into features in the product

or service. The more efficient this translation, the smoother is the customer experience.
 74

3.3.3Product Design

Product or service design needs to be based on modularity (Magrab, 1997) thus allowing customization

based on customer preferences. Consider the example of laptop computers. Most companies like Dell,

Apple or Lenovo allow customers to custom build their laptop on their website. Each of the myriad

features of a laptop such as hard drive, memory, screen size, to name a few are individual modules that

can be plugged together to make the customized computer. The need for a good product data

management system has been underscored by many researchers (Agouridas et al 2001, Pan et al 2014).

As the number of features and options increase, the total number of parts and assemblies that need to

be tracked increase exponentially (Jiao & Tseng, 1999) thus making inventory management an onerous

task if carried out individually by the partners. This necessitates a fully networked system that integrates

inventory from all partners in the value chain into a holistic system that allows for more efficient

inventory management.

3.3.4System Integration

Information systems have been around for a long time and the underlying technology has been evolving

at a drastic pace over the years. As new technologies and tools are developed, new systems are designed

as well. Moreover, due to the open-source nature of some of the technologies, it becomes easy for

entrepreneurs to set up companies that are engaged in developing software systems using these

technologies. This proliferation of software offerings and the underlying technologies provides

companies a wide variety of choice to accomplish their information system needs. As the company

grows its information system needs change as well. Moreover, in today’s age, a company often does

not exist by itself. It has numerous partners along its supply chain. This compounds the complexity of

the information system. Thus, there are three main requirements for a successful company wide

integration of information system:

● Legacy Systems Integration - Due to the very nature of technology and the fast pace at which

changes occur in this field, companies are often forced to upgrade to newer systems in order to
 75

keep up with competition and customer needs. As a result, companies are saddled with multiple

versions of software systems that store data in diverse formats. As data is key in mass

customization, it is vital that newer systems be able to interact with these legacy systems and

leverage historical data residing in these systems to provide better customer choice. Efficient

data exchange mechanisms are integral to this seamless interaction with legacy systems. These

data exchange mechanisms make data portable across diverse systems thus enabling companies

to make effective use of all the resident data.

● Cross-technology Integration - It is essential that all partners in the supply chain are able to

communicate and seamlessly exchange information with each other in order for the company

to be truly agile and responsive to changing customer needs. Due to the varying needs of the

supply chain partners, each partner may choose to have its own software system to support its

information system needs. With the plethora of systems available and the diverse nature of

underlying technology, software systems in these supply chain partners have to be able to

communicate across platforms and technologies. This calls for an effective mechanism for

porting data into different formats readable by these disparate systems.

● Seamless Integration - Often times, the end-users of these software systems are not

technologists conversant with the inner workings of these systems. Also, given the multitude

of platforms and technologies, it would be impractical to expect the end-user to be able to

manipulate the innards of software system to allow for communication with legacy and diverse

software systems. Thus, along with data portability and legacy integration, the company’s

information system needs to be able to achieve this integration seamlessly so that the end-user

can focus on data analysis and design development.

3.3.5 Data Capture Mechanisms

Since information is one of the two pillars of mass customization and data is the foundational basis for

information, collecting and storing data is a significant activity in ensuring the success of mass

customization efforts. Data can be collected through various direct and indirect mechanisms. In direct
 76

mechanisms, the product, service or customer is actively involved in the data collection process. Some

of the direct mechanisms include:

● Point of Sale (POS) - As products are being checked out at the store (or purchased in an online

environment), information about the sale is recorded into the system. This information can help

monitor stock, and trigger production and/or deliveries based on the inventory levels.

● Identification Tags - Radio Frequency Identification (RFID) tags are also becoming

increasingly popular to help with inventory management and data collection. These tags have

the unique advantage over barcodes in that these do not require a direct line of sight with the

scanner thus drastically speeding up the process of stock-taking.

● Customer Surveys - Customizing products or services relies heavily on customer input.

Customer input can be obtained in numerous ways such as survey questionnaires, interviews,

sampling or trial offers, etc. The feedback obtained from these instruments can be utilized to

ascertain customer preferences and thus steering product or service design in the right direction.

Indirect mechanisms as the term suggests, do not require active involvement of the product, service or

customer. Customer behavior is closely monitored to ascertain their preferences. These mechanisms

have become more relevant these days with the proliferation of online shopping. Some of the indirect

mechanisms for collecting data include:

● Online visitors - The number of people browsing the company’s product page or news articles

can also be mined to gage the types of customers and their level of customer interest in a new

product/service or a new variation of an existing product/service. It has been shown that

(Turban and King, 2003) understanding the customers and providing a tailored experience is

critical for the success of companies in an online environment. For example, if a company has

released news about its upcoming new product or service, the number of unique visitors or

readers might help gage the level of interest amongst the customers for the new product or

service.

● Customer Profiles - Many websites require customers to open an account and create a profile

before utilizing the services. Every product or service that the customer orders is automatically

collected and stored in the customer’s profile. Through an in depth analysis of this data,
 77

companies can identify customer preferences and tailor their products or services based on these

preferences. Understanding customer behavior (Newman, Yu & Oulton, 2002) is really critical

for the success of the customization efforts. For example if a customer consistently books an

aisle seat at an airline website, the airline company could automatically tailor the choices

offered to be more aisle seats thus customizing the experience based on the preferences.

Before diving into installing mechanisms to collect data however, it is important to analyze (O’Brien,

2002) what data is required and what are the best mechanisms to collect and store this data. It is also

important not to confuse customization with personalization (Strauss & Frost, 2001). Customization is

the ability to provide a mix of products based on the user preferences while personalization is adding a

personal touch to the interaction by including the customer’s name, etc. Moreover, some instruments

might be more appropriate than others depending on the type of interaction within the supply chain such

as supplier-retailer relationship, retailer-consumer relationship, or in some rare cases, even supplier-

consumer relationship (Vrechopoulos, 2004). Using the above mentioned instruments to tackle the

previously mentioned three steps in the CRM process can work wonders in enhancing the success of

the mass customization approach.

3.3.6 Supplier Relationship Management

A big part of today’s value chain in any industry is the array of suppliers and vendors that provide a

wide variety of services such as raw materials, processing, packaging or even distribution. Companies

are off-loading its routine activities to suppliers in order to focus on their niche areas for which they are

known. For example, Apple Co. is known for its immaculately designed and exquisitely engineered

products. Thus, the company focuses on design and engineering while engaging sub-contractors for the

actual assembly. Some companies took this approach overboard and pitted suppliers and contractors

against each other to lower costs thus resulting in an ugly bidding battle leading to corners being cut to

save on cost which in turn engendered quality issues. Toyota Co. which pioneered the Toyota

Production System focusing on elimination of wastes ran its operations with a handful of suppliers with

whom it had a long-standing relationship. This long-term relationship begets trust which in turn allowed
 78

suppliers and Toyota to work together and share information related to product, design and production

automatically leading to lowered costs and improved quality. It has been shown (Liao, et al 2011) that

trust is essential for greater information sharing and increased collaboration which is indispensable for

a mass customization environment.

3.3.7 Data Analytics

Information is key in the success of mass customization and this information is obtained by sifting

through data in the form of sales orders, purchase history, customer preferences and needs. For example,

consider the airline industry. By analyzing the seating choices made by a frequent flyer, the information

system used to book tickets should be able to determine the most likely seating preference for the

customer thus providing the customer with customized seating. The information system should be able

to collect data and leverage the collected data to determine desirable product or service features.

However, this is not a simple task given that in a truly connected system, there is copious amounts of

data being collected every day. Analyzing this data manually is an onerous and practically impossible

tasks. Thus, the information system needs to have the capability to analyze data to identify patterns in

customer preferences.

3.3.8 User Experience

With widespread availability of Internet, and more and more business being conducted on mobile

devices, it is imperative that businesses adapt to this changing scenario and provide stakeholders and

customers the ability to seamlessly interact with the information system irrespective of the device used.

This again represents a dramatic shift from the traditional desktop-based systems. There are two main

approaches to providing this seamless interaction:

● Responsive Design - In this approach, the main portal through which customers interact with

the system reconfigures itself based on the device being used and the resulting screen resolution.

This is achieved through behind-the-scenes web programming that understands the device and
 79

reorganizes the elements on the web page based on the screen resolution of the device. This

approach has the advantage of providing customers a uniform experience across devices. It also

makes the task of system maintenance and upkeep easier since there is only one system to

manage. However, this approach does not fully leverage the capabilities of the mobile devices

and companies often develop a separate stand-alone application that its customers can use to

interact with through their mobile devices.

● Mobile App - In this approach, a stand-alone application is downloaded to the customer’s device

which then provides the necessary interaction. This allows companies to leverage the

capabilities of a truly mobile device. This however compounds the task of maintaining these

systems since there needs to be a separate application for each platform such as Apple’s iOS,

Google’s Android etc. However, despite this hurdle, many companies choose this approach due

to multiple reasons:

○ Customer perception - Customers might perceive that the company to be on the

forefront of technology if it has a mobile app. This helps with building the company’s

reputation and allows the company to leverage this into improved market share.

○ Customer experience - The mobile applications developed for each platform can make

use of the unique features available in these platforms to provide customers with a

better experience.

Apart from these key elements, the mass customization information system also needs to incorporate

the following features:

● Adaptability & Flexibility - Markets are very dynamic in nature and customer preferences tend

to be very fluid constantly changing over time. Thus, it is imperative that the information system

needs to be agile (Nambiar 2009b) and have the capability to be reconfigured to respond

(flexibility) to changes in the market dynamics. It is also important to be proactive and envision

some of the changes in the market or to even create watershed moments that completely

revolutionize the market (think Apple Co.’s iPhone). A truly adaptable system will be able to

seamlessly handle these changes with minimal impact on the existing operations. It has been
 80

shown (Porter et al 1999) that the lack of reconfigurability is often a significant factor for the

reluctance on the part of companies to invest in such systems.

● Scalability - Successful companies grow over time and diversify its product mix to capitalize

on its reputation and enter into new markets. Any chosen information system should be able to

grow with the company and accommodate its information processing needs. However, there

would potentially come a time when the information system is bursting at its seams and unable

to handle the ballooning needs of the enterprise. This is where interoperability becomes crucial.

● Interoperability - As companies come together with the common objective of providing a truly

customized experience for the customer, as is wont these days, they may already have a

functioning information system that handles some if not all aspects of their operations. It would

be redundant to require all partners to reinvest in another completely new system. Thus, any

implementation of information systems needs to be interoperable with other legacy and existing

systems. This facilitates free and seamless information sharing which is an essential facet of

mass customization.

3.4 Future research directinos

The analysis of literature has revealed a renewed interest among practitioners and researchers in this

area of leveraging information systems for a successful implementation of the mass customization

paradigm. Although extensive work has been done in certain areas of the customization value chain

such as developing smart and efficient product configurators, and managing customer relationships to

name a few, there is still work to be done in developing a holistic information management system that

fully integrates all facets of the customization process. Standards (Dean, Tu & Xue, 2008) need to be

developed for coding information so that it can be easily shared across partners. Integrating legacy

systems and coalescing piecemeal technologies into a comprehensive system also needs to be

investigated further. Knowledge management (Nambiar, 2013) is another area that is gaining in

importance in order to manage the resident tacit knowledge about customer preferences and product
 81

features. Integrating this into the mass customization framework (Helms et al 2008) can help leverage

the knowledge to provide a better customized experience.

3.5 Conclusions

The mass customization paradigm holds significant promise for companies stifled with competition and

lack of new products or services. Despite being around for more than two decades, this concept is still

lagging behind in its practical implementations. This is primarily due to the complexities involved in

successfully putting into practice the tenets of mass customization. Information technology can be

leveraged extensively to help overcome some of this implementation hurdle. This has already been

underscored by the numerous piecemeal approaches developed to address one or more aspects of the

process. However, a fully integrated information system with the key elements and features identified

in this chapter can go a long way in improving the success of mass customization implementation thus

allowing this paradigm to really take wing and become more commonplace like its lean counterpart.

REFERENCES

Agouridas, V., Allen, M., McKay, A., Pennington, A., Holland, S. (2001). Using Information Structures

to Gain Competitive Advantage. Proceedings of the 2001 PICMET - Portland International Conference

on Management of Engineering and Technology, Portland, OR, Vol. 2, pp 681 - 692.

Barrenechea, O. L. (2010). An information technology suitability index for mass customization.

University of Texas-Pan American Thesis. ProQuest Dissertations Publishing, MI.

Chen, RS., Tsai, YS., Tu, A. (2008). An rfid-based manufacturing control framework for loosely

coupled distributed manufacturing system supporting mass customization. IEICE Transactions on

Information and Systems, E91D(12), 2834-2845.

Cook, M. (1996). Building enterprise information architectures: reengineering information systems.

Prentice-Hall, Inc.

Cooper, A.H. (1957) , Integration of computers with factory processes, Journal of the British Institution
 82

ofRadio Engineers, vol.17, no.8, pp.431-441, doi: 10.1049/jbire.1957.0044.

Da Silveira, G., Borenstein, D. & Fogliatto, F. (2001). Mass customization: literature review and

research directions. International Journal of Production Economics. Volume 72. No. 1. pp 1- 13.

Davis, S. (1989) From future perfect: mass customizing. Planning Review. Volume 17. No. 2. pp 16 -

21.

Dean, P.R., Tu, Y.L.,Xue, D. (2008) A framework for generating product production information for

mass customization. International Journal of Advanced Manufacturing Technology. Volume 38. Issue

11-12. pp 1244 - 1259.

Dean, P.R., Tu, Y.L.,Xue, D. (2009). An information system for one-of-a-kind production.

International Journal of Production Research. Volume 47. No. 4. pp 1071 - 1087.

Dibb, S., Simkin, L., Pride, W.M. & Ferrell, O.C. (2001). Marketing - concepts and strategies. Fourth

European Edition. Houghton Mifflin, Boston, MA.

Dietrich A. J., Kirn, S. & Sugumaran, V. (2007). A service-oriented architecture for mass

customization - a shoe industry case study. IEEE Transactions on Engineering Management. Volume

54. No. 1.

Feinberg, R.A., Kadam, R., Hokama, L. & Kim, I. (2002). The state of electronic customer

relationship management in retailing. International Journal of Retail and Distribution Management.

Volume 30. Issue 10. pp 470 - 481.

Ford, H. & Crowther, S. (1922). My Life and Work. Doubleday, Page & Company.

Forsza, C. & Salvador, F. (2006). Product Information Management for Mass Customization.

Connecting Customer, Front-office and Back-office for Fast and Efficient Customization. Palgrave

Macmillan, New York, NY.

Frutos, J.D. & Borenstein, D. (2004). A framework to support customer-company interaction in mass

customization environments. Computers in Industry. Volume 54. pp 115 - 135.

Fulkerson, B. (1997). A response to dynamic change in the market place. Decision Support Systems.

Volume 21. Issue 3. pp 199 - 214.

 83

Helms, M.M., Ahmadi, M., Jih, W.J.K., Ettkin, L.P. (2008). Technologies in support of mass

customization strategy: exploring the linkages between eCommerce and knowledge management.

Computers in Industry. Volume 59. pp 351 - 363.

Husejnagic, D. & Sluga, A. (2015). A Conceptual Framework for a Ubiquitous Autonomous Work

System in the Engineer-to-Order Environment. International Journal of Advanced Manufacturing

Technology. Vol. 78, pp 1971 - 1988.

Jiao, J., Ma, Q. & Tseng, M.M. (2001). Towards high value-added products and services: mass

customization and beyond. Technovation. pp 149 - 152.

Jiao, J. & Tseng, M.M. (1999). A methodology of developing product family architecture for mass

customization. Journal of Intelligent Manufacturing. Volume 10. pp 3 - 30.

Karpowitz, D. J., Cox, J.J., Humpherys, J.C., Warnick, S.C. (2008), A Dynamic Workflow framework

for Mass Customization using Web Service and Autonomous Agent Techniques, Journal of Intelligent

Manufacturing, Vol. 19, No. 5, pp 537 - 552.

Kimber, A. (1988), Production Planning and Scheduling on Process Plant, Process Engineering, 69

(2), pp 81 - 82

Kocchar, A. K. (1978), Use of Computers and Analytical Techniques for Production Planning and

Control in the British Manufacturing Industry, Computers and Industrial Engineering, 2 (4), pp 163 -

179

Kocchar, A.K. (1981), Low-cost computers in Production Planning and Control Systems, Production

Engineer, 60 (5), pp 44 - 47

Lee, H. L., Padmanabhan, V. & Whang, S. (1997). Information distortion in a supply chain, the

bullwhip effect. Management Science. Volume 43. Issue 4. pp 546 - 558.

Liao, K., Ma, Z., Lee, J. J., & Ke, K. (2011). Achieving mass customization through trust-driven

information sharing: A supplier's perspective. Management Research Review, Volume 34. Issue 5. pp

541-552.

Lim, J.M. (2003). The impact of information technology on mass customization - an evolving trend

for companies that want to be successful in the e-business arena. Thesis Publication. Rochester

Institute of Technology.
 84

MacCarthy, B., Brabazon, P.G. & Bramham, J. (2003). Fundamental modes of operation for mass

customization. International Journal of Production Economics. Volume 85. pp 289 - 304.

Magrab, B.E. (1997). Integrated product and process design and development. CRC Press, New

York, NY.

Mahajan, V., Srinivasan, R. & Wind, J. (2002). The Dot.com retail failures of 2000: Were there any

winners? Journal of the Academy of Marketing Science. Volume 30. Issue 4. pp 474 - 486.

Nambiar, A.N. (2009a), Mass Customization - Where do we go from here?, Proceedings of the 2009

World Congress on Engineering, London, U.K, July 1 - 3, Vol I.

Nambiar, A.N. (2009b), Agile Manufacturing: A Taxonomic Framework for Research, Proceedings

of the 2009 International Conferences on Computers and Industrial Engineering, July 6 - 9, pp 684 -

689. Nambiar, A.N. (2010), Modern Manufacturing Paradigms - A Comparison, Proceedings of the

2010 International Multiconference of Engineers and Computer Scientists, Hong Kong, March 17 -

19, vol III.

Nambiar, A.N. (2013). A Current Perspective of Knowledge Management in a Global Economy,

International Journal of Soft Computing and Software Engineering, Vol. 3, No. 3, pp 236 - 239.

Newman, A.J., Yu, D.K.C. & Oulton, D.P. (2002). New insights into retail space and format planning

from customer-tracking data. Journal of Retailing and Consumer Services. Volume 9. Issue 5. pp 253

- 258

Ngniatedema, T. (2012). A mass customization information systems architecture framework. The

Journal of Computer Information Systems, Volume 52. Issue 3. pp 60-70.

O’Brien, J.A. (2002). Management information systems - managing information technology in the e-

business enterprise. 5th edition. McGraw Hill/Irwin, New York/Homewood, IL.

Pan, X., Zhu, X., Ji, Y., Yang, Yi., & Wu, Y. (2014). An information integration modelling

architecture for product family life cycle in mass customisation. International Journal of Computer

Integrated Manufacturing, 27(9), 869-886.

Peng, D.X., Liu, G., Heim, G.R. (2011). Impacts of information technology on mass customization

capability of manufacturing plants. International Journal of Operations and Productions

Management. Volume 31. Issue 10. pp 1022 - 1047.

 85

Piller, F.T. (2002). Customer interaction and digitization - a structured approach to mass

customization. Moving into mass customization: Information Systems and management principle. part

II. Springer-Berlin. pp. 119 - 137.

Pine, B.J., Victor, B. & Bonyton, A.C. (1993). Making mass customization work. Harvard Business

Review. Volume 71. Issue 5. pp 108 - 119.

Pine, B.J. (1999). Mass customization: the new frontier in business competition. Harvard Business

School Press, UK.

Porter, M.E. (1985). Competitive Advantage. The Free Press.

Porter, K., Little, D., Peck, M., Rollins, R. (1999). Manufacturing Classifications: relationships with

production control systems. Integrated Manufacturing Systems, Vol. 10, No. 4, pp 189 - 199.

Reichwald, R., Piller, F.T. & Moeslein, K. (2000). Information as a critical success factor for mass

customization. Proceedings of the 2000 ASAC-IFSAM Conference, Montreal.

Rolls-Royce (2016). Rolls-Royce Motor Cars Home. Retrieved from https://www.rolls-

roycemotorcars.com/en-GB/home.html . Accessed on January 02, 2016.

Shin, Y., An, S., Cho, H., Kim, G., & Kang, K. (2008). Application of information technology for

mass customization in the housing construction industry in korea. Automation in construction, 17(7),

831-838.

Strauss, J. & Frost, R. (2001). E-marketing. 2nd Edition. Prentice-Hall, Englewood Cliffs, NJ.

Svensson, C. & Barford, A. (2002). Limits and opportunities in mass customization for “build-to-

order” SMEs. Computers in Industry. Volume 49. pp 77 - 89.

Tropicana (2016). Tropicana - 100% Pure Squeezed Sunshine. Retrieved from

http://www.tropicana.com/ . Accessed on January 02, 2016.

Turban, E. & King, D. (2003). Introduction to e-commerce. Prentice Hall, Englewood Cliffs, NJ.

Varadarajan, R. P. & Yadav, M. S. (2002). Marketing strategy and the Internet: An organizing

framework. Journal of Academy of Marketing Science. Volume 30. Issue 4. pp 296 - 312.

Verdouw, C.N., Beulens, A.J.M., Trienekens, J.H. & Verwaart, T. (2010). Towards dynamic

reference information models: readiness for ICT mass customization. Computers in Industry. Volume

61. pp 833 - 844.

 86

Vrechopoulos, A. P. (2004). Mass customization challenges in Internet retailing through information

management. International Journal of Information Management. Volume 24. pp 59 - 71.

Yang, J. and Kou, Y. (2009), Application of online simulation in production planning and scheduling,

Journal of Beijing University of Aeronautics and Astronautics, 35 (2), pp 215 - 218

Yao, S., Han, X., Yang, Y., Rong, Y., Huang, S.H., Yen, D.W. & Zhang, G. (2007). Computer aided

manufacturing planning for mass customization: part 3, information modeling. International Journal

of Advanced Manufacturing Technology. Volume 32. pp 218 - 228.

Zahed, S. & Reddy, B. K. (2004). A mass customization information framework for integration of

customer in the configuration design of a customized product. Artificial Intelligence for Engineering

Design, Analysis and Manufacturing. Volume 18. pp 71 - 85.

Section two

Complexity Drivers in Mass Customization

 87

chapter four

Complexity Issues in Mass Customized Manufacturing

Vlad Hendrik Van Landeghem, El-Houssaine Aghezzaf
Ghent University, Dept. of Industrial Systems Engineering and Product Design, Gent, Belgium
Flanders Make, Belgium

ABSTRACT

This chapter focuses on the role of complexity and its impact on the human operator within assembly

oriented mass customization manufacturing. It starts with the definition of the main complexity

concepts we need for a thorough understanding of this chapter, building on the previous chapters. It

also argues that automotive manufacturing is currently a good representative example of MCM, as the

number of variants and models continue to proliferate, and being one of the few mature industries that

practice MCM. Their size and heritage forces them to look for and adopt structural and methodical

approaches to deal with complexity.

 88

Next, the major drivers of complexity and their impacts on manufacturing systems are described and

analyzed. These were derived from operator debriefings on complexity and how they experience it

during daily work. These findings are based on a 3 year international research project (2010-2012)

involving leading OEM in automotive and vehicle manufacturing in Belgium and Sweden. One result

that is described is a fast and objective measurement to determine the complexity of a manual assembly

station, based on a statistical classification model.

The last part of the chapter is devoted to some of the methodical approaches that can be used to manage

complexity in MCM, some of these approaches are based on recent research results from our research

team at UGent. An enriched method to balance manufacturing lines, minimizing variability of station

takt times, is discussed. The main objective is minimizing overloads that prove to be most detrimental

to operator cognitive load, quality issues and assembly errors. Also technological developments to

enhance the information exchange with operators in a complex MCM setting are discussed, using

wearable devices and bi-directional information exchange.

4.1 Complexity in assembly

The trend towards mass customization in automotive industry and the complexity it induces has been

the focus of many studies in the last three decades. Specifically, the impact of increased product variety

on the performance of automotive mixed-model assembly lines, has already been studied by Fisher et

al. (1995), MacDuffie et al. (1996) and Fisher and Ittner (1999).

In particular, MacDuffie et al. (1996) investigated the effect of product variety on total labor

productivity and consumer-perceived product quality in an international study covering 70 assembly

plants from 16 countries, participating in the International Motor Vehicle Program at M.I.T. In their

paper, they analyzed complexity measures that capture different aspects of product mixes in an

assembly plants: Model Mix Complexity, Parts Complexity, Option Content and Option Variability.

Model Mix Complexity is based on the number of different platforms, body styles, and models, scaled

by the number of different body shops and assembly lines in each plant. Parts complexity is measured
 89

as an aggregate index of 6 factors: number of variants for engines, wire harnesses and paint colors;

variety of subassemblies and material flows; the total number of parts, the percentage of common parts

and the number of suppliers. Option Content reflects the overall level of installed options (from a list

of 11), and equals the percentage of vehicles built with various options, aggregated across all models in

a plant. Option Variability captures the extent to which different cars have different amounts of options

installed. A statistical analysis has shown significant but limited negative correlation between the

complexity measures and the manufacturing performance. This seminal study has however indicated

already the main factors that define complexity, and the role of variability of those across the different

products on the same line. The latter’s importance has been reported in Fisher and Ittner (1999). Section

4.2 will explore this into more details by constructing an influence model.

This work has been followed by many other investigations attempting to define, model and develop

valid and useful complexity measures for manufacturing systems. Frizelle (1996) suggests that a useful

complexity measure needs to be separable and additive as it would then simplify its computation and

allow easy analysis for managers. Deshmukh et al. (1998) provided thus a clear definition of static and

dynamic complexity. Static or structural complexity is related to the structure of the system, variety of

components and products, number of processes and machines,…. On the other hand, dynamic

complexity measures the unpredictability in the behavior of the system over a time period.

Fujimoto and Ahmed (2001) proposed a complexity index based on the ease of assembling a product.

The index takes the form of entropy to evaluate the ease of assembly of a product defined as the

uncertainty of gripping, positioning, and inserting parts in an assembly process. The workstation

entropy is the sum of variety coming from upstream and the variety added in the station itself. In

Fujimoto et al (2003) they also identify broadly some of the drivers and impacts of variety. Specifically

they identify the directions of approach as one of the drivers, which we included in our complexity

classification model in section 4.3.

ElMaraghy and Urbanic (2003, 2004) proposed a methodology to assess product and process

complexityand their interrelations in a systematic manner. A matrix methodology and an objective

 90

measure of complexity have been proposed that assess the three levels of manufacturing complexity:

product complexity, process complexity and operational complexity. Samy and Elmaraghy (2010)

define a Product assembly complexity index. Their model considers not only the complexity of

assembly introduced by factors from the Design For Assembly method but also the diversity of the

various parts used in products assembly and the total part count. The proposed model incorporates the

assembly complexity resulting from the number and diversity of the parts and fasteners used in the

product assembly using a formulation that incorporates information content and diversity.

Zhu et al. (2008) and Hu et al. (2008), inspired by Fujimoto and Ahmed (2001), proposed a measure of

manufacturing complexity at the workstation level introduced by product variety, and model its

propagation through the assembly system. The proposed complexity model quantifies human

performance in making choices, that is the uncertainty the operator is facing when making various

choices at an assembly station:

 Fixture choice: choose the right fixture according to the base part i.e., the partially completed

assemblage to be mounted on as well as the added part to be assembled.

 Tool choice: choose the right tool according to the added part to be assembled as well as the

base part to be mounted on.

 Procedure choice: choose the right procedure, e.g., part orientation, approach angle, or

temporary unload of certain parts due to geometric conflicts/subassembly stabilities.

Zhu (2009) also pointed out that complexity effects should be taken into account when determining the

Assembly Sequence, as well as the Build Sequence. The Assembly sequence is the (static) manner in

which the assembly tasks are assigned to workstations along the line, while the Build sequence is the

(dynamic) sequence into which specific vehicle configurations (with their options and variants) are

loaded unto the assembly line, typically in a daily schedule. The line balancing method we describe in

section 4.4 addresses assembly sequencing.

 91

In an attempt to understand complexity, G. Schuh et al. (2008) determine its main drivers as:

uncertainty, dynamics, multiplicity, variety, interactions and interdependencies, and a combination of

such proprieties that can render a system complex or not complex. C. Rodríguez-Toro et al. (2004)

propose a specific taxonomy where complexity is split in static and dynamic. Static complexity is

associated with the product, whereas dynamic complexity is linked to the process.

Most of these published models seem to link complexity to objective characteristics of products or

processes. However, between complexity and the company’s financial bottom line (measured through

productivity and quality) we find the operator who has to perform the tasks at the workstation. It is

intuitive that both complexity and time pressure play a factor in the mental workload of the operator.

Parasumaran et al. (2000) have provided a simplified, but useful, 4 step interaction model of how

humans deal with information from complex systems:

1) Information acquisition;

2) Information analysis;

3) Decision and action selection;

4) Action implementation.

Rasmussen (1983) provided an interesting model of how skilled operators deal with information and

what the effect is of each mode of cognition on their performance. In it he distinghuishes three behavior

levels: skill-based, rule-based and knowledge-based.

This entails the whole domain of ergonomics and the psychological side of task design. In section 4.4

we will describe how operator information systems could mediate the effects of complexity on the

cognitive load of the operator.

Very recently, ElMaraghy et al. (2012) have published a very thorough literature review of complexity

models in design and manufacturing. They state that “Designing systems for less complexity … are
 92

important issues for further research”. Our research has benefited greatly from insights gained from this

research work and we will reference it throughout this chapter.

In section 4.2 we present and discuss a first attempt to aggregate the various factors governing

complexity in assembly and populate this generic structure with tangible elements. In section 4.3 we

describe objective measurement methods to determine the complexity of workstations. In section 4.4

we present several approaches to cope and reduce the effects of complexity in MCM situations.

4.2 Modeling the DNA of complexity

The concepts and models that are introduced and discussed in section 4.2 and .4.3 are the result of a 3

year international study that was conducted in the vehicle industry of Belgium and Sweden (Zeltzer et

al, 2013;Mattson et al, 2014). It included the major vehicle OEM manufacturers of these countries. The

research focus on which this chapter is based relates to the workstations along driven assembly lines,

where manual assembly work is carried out on different models in a mixed model fashion. Different

types of assembly lines were investigated, including two for car models and 2 for engine models, a truck

model, and several subassembly lines with suppliers. This has led to a total of 76 different workstations

from which data was gathered.

4.2.1 Complexity Definition

A good definition of complexity has to be generic enough to be applicable to different manufacturing

systems and at the same time specific enough to guide and support the decisions related to whether a

system is complex or not. Although the literature review provided useful insights about manufacturing

complexity, most approaches are relatively specific. In our view there still existed a need for a clearer,

simple and generic complexity definition. After extensive exchange of ideas amongst the project
 93

partners the following definition is proposed which proved to be very useful during the workshops that

led to the causal model:

“Complexity of a workstation is the sum of all technical and ergonomical aspects and factors that make

the set of tasks to be performed within it by an operator mentally difficult, error-prone, requiring

thinking and vigilance, and inducing stress”.

This definition recognizes the fact that the inherent complexity of tasks is determined to a large extent

by the operator who executes them, hence termed subjective complexity. This means, according to the

findings in this case, that the same set of tasks can be judged differently by different operators,

production engineers, quality controllers and line managers under different circumstances. This makes

the issue of quantifying complexity in an unambiguous manner, the so-called objective complexity, a

real challenge (Mattson, 2014). One immediate consequence is that measuring the magnitude of the

subjective complexity will always involve a behavioral and psychological aspect, which is difficult to

quantify. Secondly, since complexity is a multi-faceted concept, it is almost impossible to measure it

directly as no meaningful scale exists. In order to gain more insight into the nature of complexity we

set out to construct a causal model of complexity.

4.2.2 A causal model of complexity in manufacturing

ElMaraghy et al. (2012) proposed a causal map of how manufacturing complexity cascades down from

product design to the cognitive and physical effort of the individual operator. This scheme inspired us

to a conduct a series of fact-finding workshops with vehicle manufacturing companies to validate and

refine this causal map.

To gather as much useful information as possible, it was decided that the participants in these workshops

include shop floor employees, production engineers, quality controllers and line management, i.e. all

those who deal with complexity in their daily activities. In a first phase, the project objectives were

explained to all participants. Next, the participants were asked to identify two low and two high complex
 94

workstations. The participants were then asked to use these workstations as a mental reference while

providing elements of complexity, drivers or causes of complexity and impacts or consequences of

complexity. Each participant provided his views individually by noting them on separate sticky notes.

Afterwards the answers were processed in three rounds. The relationships between these rounds were

argued and modelled into a causal model.

The results of the first round focused on aspects characterizing complexity. The results of second round

(Impacts) concentrated on revealing which consequences complexity has: areas that are affected by

complexity and the influence of complexity on the manufacturing activity and on the teams. The third

round aimed at detecting the direct Drivers of complexity, i.e. the variables that are directly linked to

the complexity elements as causal factors. In each round the relevant notes from the participants were

put up on a wall, clustered by similarity. Finally, a brainstorming session was held where the list of

ideas was discussed and finalized. The results were discussed extensively with the industrial partners,

and were found to be both highly insightful and useful.

 95

Indirect Drivers

Direct Drivers

Objectiveco
mplexity Subjective

complexity

Impacts

KPI

FIGURE 4.1 Casual model of complexity in MCM

 96

The causal links between elements obtained at the workshops were then combined in a graphical

network structure with the goal to obtain a generic complexity model for assembly workstations. The

model consists of 5 clusters of variables related with complexity characterization, complexity impacts

and KPI and complexity drivers (indirect and direct) respectively. The full model is shown in Figure

4.1.

It would be too unwieldy to describe the model and its causal links in full detail. However, we will

briefly treat each cluster.

FIGURE 4.2 Objective complexity elements

Subjective complexity is perceived by the operator, and it is a combination of mental requirements and

the time pressure to react on them (Figure 4.3). Mental requirements can be counted as “context

switches”, which means changes in the context (parts, tools, instructions, etc.). Also ambiguity coming

from similar looking but different parts adds to the mental workload.
 97

FIGURE 4.3 Subjective complexity elements

When we look at the impact of the complexity elements, a dense network emerges. The main adverse

impacts are:

 Loss of time because of variant induced workload imbalances, NVA activities linked to the

number of parts, the Border of Line organization, etc;

 Errors, that will lead to scrap, rework, line stops and in general a decrease in quality level,

which in turn invokes the need for additional control systems;

 Increased need for training of the operators, because of the larger number of work instructions,

the required breadth of knowledge of different models and part combinations, and their specific

methods and tooling;

 Mental stress of operators, inducing errors, absenteeism, accidents, and in general frustration

and loss of motivation and team spirit;

 Physical fatigue with much of the same results.

These impacts will have an adverse effect on the bottom line (KPI in the model), including direct and

indirect man-hour costs, investments, sales and unproductive capacity.

 98

Looking to the drivers’ part of the causal model, we identified 11 direct drivers of complexity (Table

4.1). We will try to use (a subset of) these to quantify workstation complexity in subsequent sections.

TABLE4.1 Direct drivers of complexity

Variable Name Values or units Description

Picking (F)ixed location (F) : Operator takes part always on the same location from bulk
technology storage.
Pick to (S)ignal (S) : Operator picks part from location indicated by a signal (light,
display)
(C)omparing
(C) :Operator must compare simple information (symbols, colors)
(M)anual (M) : Operator must read extensive information from manifest

Bulk/Sequence Kit (S)equenced kit (S) : Every part is in its package in correct assembly sequence
(K) : Parts are delivered in kits with exact set for one assembly
(K)it operation
(B) : Parts are by type in their own package
(B)ulk

# Packaging types Integer number The total number of different packaging types, a type having a specific
layout. So 2 identical boxes with different inserts are 2 different types.

#Tools per Integer number The number of tools that the operator(s) needs to handle to perform all
workstation possible assembly variants in this station, excluding automatic tools
(servants).

# Machines per Integer number Machines that perform automated tasks without operator assistance,
workstation with automatic or manual start.

# Work methods Integer number Each unique set of work methods the operator must master in this
station. A method contains several small steps.

Distance to parts Meters The furthest distance between the normal operator position (or the
center of the station) and the parts at the border of line.

# Variants same Integer number The highest number of variants belonging to one model, among all
model models of which parts are assembled in this station.

# Variants in this Integer number Total number of variant parts, combined over all models that are
workstation assembled in this workstation. So 5 types of Left hand and Right hand
mirrors x 2 models = 20 variants.

# Different parts in Integer number Total number of unique part references that are assembled in this WS,
workstation including all variants and models that typically occur in one year

# Assembly Integer number The number of different positions the operator must take to complete
directions his WS cycle, including repositioning of the upper body or the feet,
but not small repositionings of the hands
 99

These drivers are in turn influenced by a large number of elements we grouped as Indirect Drivers. The

most important ones are:

 The number of car models and the frequency of introduction of new models and features to

the plants is a major driver, caused by the market pressure from competitors and the context of

structural overcapacity. Especially the speed of introduction will lead to a host of adverse

elements that will increase complexity, such as learning curve effects (wrong or missing work

instructions, last minute schedule changes and glitches during ramp up), etc.

 The number of options that are offered because of marketing reasons

 The number of different vehicle platforms that are built on the same line, determining the

number of variants that the line (and its workstations) will have to coop with

 Many of these elements will influence how the operator will deal with them, modeled as

“SRK” level (Skills, Rules, Knowledge) according to the cognitive model of Rasmussen

(1983).

This model reveals a complex interaction of many factors. Many of them have been researched by

numerous scholars, but only a limited set was effectively quantified. So much research and fact finding

remains to be done.

4.3 Quantifying complexity in manual assembly stations

4.3.1 A brief literature review

A major insight from the above discussion is that complexity is in fact a multi-faceted concept with

objective and subjective components. Measuring complexity is therefore a challenging task. However,

as is usually done for such multi-faceted concept, attempts to propose indicators which may provide

information of the extent to which an activity or pool of activities is complex were carried out. In this

subsection we briefly discuss few of such indicators mainly using entropy (more specifically, Shannon

entropy).
 100

The concept of entropy, as introduced by Shannon (1948), measures unpredictability of information

content. In the context of mass customized manufacturing entropy refers to disorder or uncertainty. If

an operators is faced with the same model requiring the same operations he will not experience any

complexity, instead repeating these operations will help them adopt the good practices to improve the

way he can execute these operations efficiently. It is the uncertainty on the next group of tasks to be

executed by the operator that increases complexity. The entropy is therefore an appropriate indicator for

complexity.

In view of the fact that complexity to some extent has to do with the overwhelming amount of

information that the operator has to process while executing the activity, information based measures

were proposed. A frequently adopted approach to measure complexity is “entropy” and in particular the

Shannon’s information entropy (Shannon, 1948) is adopted. Frizelle and Suhov (2001) use as measures

of complexity various (long term) entropy rates that naturally emerge in the analysis of systems

involving queues and related phenomena. Sivadasan et al. (2006), propose a mathematical model for the

operational complexity of supplier–customer systems from an information-theoretic perspective. They

defined this operational complexity as the uncertainty associated with managing the time or quantity

dynamic variations across the supplier–customer s’ information and material flows.

Urbanic and ElMaraghy, (2006) proposed a complexity model based on information content, quantity,

and diversity. Information content is a relative measure of the effort needed to perform the task.

Information quantity, absolute quantity of information needed measured using entropy. Information

diversity is ratio of the specific information needed for a task to the total amount of information. Product

complexity is determined by multiplying the product’s information quantity with the sum of its content

and diversity. Complexity of each process step is determined by multiplying its information quantity

with the sum of diversity ratio and the relative complexity coefficient. The relative complexity of a

process step is calculated based on both cognitive and physical effort. The complexity of the whole

process is the sum of the product’s complexity and the sum of the complexity of all the process steps.
 101

Zhu et al. (2008) tackled the issue of measuring the variety induced manufacturing complexity in

manual mixed-model assembly lines where operators have to make choices for various assembly

activities. A complexity measure they proposed is called ‘operator choice complexity’ and is meant to

quantify human performance in making choices i.e. the more choices an operator has to make the more

he will need additional time to process the information. Here again the analytical form of the proposed

measure is an information-theoretic entropy measure of the average randomness in the choice process.

The general form of the entropy function is (Zhu et al., 2008):

𝐻(𝑋) = − ∑𝑀
𝑚=1 𝑝𝑚 𝑙𝑜𝑔2 𝑝𝑚 (4.1)

where pm = probability of element m occurring

M = number of elements that can occur

The elements can be anything, from number of parts or variants to instructions, assembly directions,

tool selection, … etc. The formula illustrates the cognitive aspect that is linked with how complexity

affects manufacturing systems through operator performance.

In section 4.4.1 we show how we used this concept of entropy not only to measure complexity of the

workstation but to rebalance the assembly line in way that this inherent complexity may also be

balanced, and hence the adverse effects mitigated.

A very comprehensive overview and comparison of nine methods to quantify complexity has been

proposed by Mattson et al. (2014). We show their main results in Table 4.2. The CXC (Complexity

Calculator) was developed by our team and will be treated more extensively in the remainder of this

section.

As part of the same study we described in section 4.2, we set out to test whether the drivers, identified

through the workshops, could be used to identify highly complex workstations, and if possible even

provide an objective way to quantify complexity. We also wanted to find the smallest subset of these
 102

drivers that is sufficient to provide meaningful results, in order to minimize the effort for data gathering.

Since no accurate information was available regarding the “real” inherent complexity of the

workstations, it was decided to ask operators and supervisors to jointly “nominate” both the most

complex and the simplest workstation within their area. This subjective label was used as benchmark

throughout the study, and each industrial partner then obtained the quantitative information about the

driving factors for these designated workstations. The models we propose and how they were

constructed was described in De Lima et al. (2014).

In this way we obtained data sets on 76 workstations from 5 different manufacturing locations (4 in

Belgium, 1 in Sweden), of which 41 were deemed of “LOW” complexity, and 35 “HIGH”. The

variables for which data was gathered are listed and explained in Table 4.1.

TABLE 4.2 Overview of 9 methods to quantify complexity (Mattson et al.),

 Appendix

Complete matrix from comparison of existing methods

The nine existing methods comparisons are presented below (due to size the table was 103
divided into two parts see Tables A1 and A2).
Table A1 Comparison of methods – first part

Type of
complexity Types of measure/-s
Method Aim
(Static or (objective/subjective)
dynamic)
Operational To monitor and Dynamic Objective: amount of
complexity manage information information required to
(Sivadasan et al., and material flows describe a state: according
2006) to flow variations, products,
reasons and variation states.
Entropic To measure the rate of Static and Objective: probability of a
measurement variety dynamic state to occur according to
(Frizelle and complexity different time measures.
Suhov, 2001) (comparison off)
Manufacturing To evaluate Dynamic Objective: quantity,
complexity index alternatives and risk complexity diversity and content
(Urbanic and with respect to information in the process
ElMaraghy, 2006) product, process or
operation task in a
design stage
Operator choice To find causes, plan Static complexity Objective: average
complexity (OCC) assembly sequences uncertainty and risk in a
(Zhu et al., 2008) and design mixed- choice for right tool,
model assembly lines fixture, parts and procedure
for variant
Knowledge and To manage software Dynamic Subjective: assessment of
technology development complexity knowledge and technology
complexity (Meyer complexity
and Curley, 1993)
Complexity To support product Static and Objective: criteria for
measurement preparation to dynamic low/high assembly
(CXB)(Falck and increase productivity complexity complexity
Rosenqvist, 2012) and decrease costs
Robustness index To evaluated risks Dynamic Subjective: robustness
(RI) and problem areas on complexity score regarding material,
a management/team method, machine and
leader level environment
Complexity To automatically Static and Objective: probability that
Calculator (CXC) assess the complexity dynamic the workstation’s
(Zeltzer et al., of stations complexity complexity is high or low
2013)
Complexity Index To find problem areas Static and Subjective: assessment of
(CXI) (Mattsson at a station level dynamic product/variants, work
et al., 2013) complexity content, layout, tools and
view of station

4.3.2 Complexity based classification of workstations

Starting with the data on 76 workstations we constructed a statistical model that can automatically decide

whether a workstation is of “LOW” or “HIGH” complexity. Such information is very useful to identify

the workstations that warrant further analysis and appropriate methods to counter the likely effects of

complexity. In statistics the logistic or logit model converts a linear combination of values of

characteristics into a probability of belonging to either value 0 (HIGH complexity) or 1 (LOW

complexity). We identified 4 characteristics of manual workstations (Table 4.3) that yield a very good

fit for our sample resulting in the equation (4.1).

 104

𝑒 18.164−3.173𝑃𝑊𝐿−2.326𝑃𝑇𝐿−2.182𝐴𝐷𝐿−0.344𝑇𝑊𝐿
𝑃𝐿𝑂𝐺𝐼𝐼𝑇_𝑆𝐴𝑀𝑃𝐿𝐸 (𝐿𝑂𝑊) = (4.2)
1+𝑒 18.164−3.173𝑃𝑊𝐿−2.326𝑃𝑇𝐿−2.182𝐴𝐷𝐿−0.344𝑇𝑊𝐿

where PWL=different parts in the workstation in Likert scale

PTL=number of packaging types in Likert scale

ADL=number of assembly directions in Likert scale

TWL=number of tools used in workstation in Likert scale

To improve goodness-of-fit we had to transcode some of the variables into a Likert scale (Table 4.4),

inspired by MacDuffie (1996). Their variable name ends with L, while those ending with R (see further)

use the measured raw value directly.

TABLE4.3 Workstation characteristics used in the complexity quantification

Variable Name Values or units Description

# Packaging Integer number The total number of different packaging types, a type having a specific
types (PT) layout. So 2 identical boxes with different inserts are 2 different types.
#Tools per Integer number The number of tools that the operator(s) needs to handle to perform all
workstation (TW) possible assembly variants in this station, excluding automatic tools
(servants).
# Work methods Integer number Each unique set of work methods the operator must master in this station. A
(WM) method contains several small steps.
# Different parts Integer number Total number of unique part references that are assembled in this WS,
in workstation including all variants and models that typically occur in one year
(PW)
# Assembly Needing The number of different positions the operator must take to complete his WS
directions (AD) repositioning of cycle, including repositioning of the upper body or the feet, but not small
tool/operator repositionings of the hands

TABLE4.4 Coding rules to transform variables into Likert Scale values

Complexity-driving variables Likert scale coding rules (raw value)

# Packaging types (PTL) 0 1 2-4 5-8 >8
0 1 2 3 4
#Tools per workstation (TWL) 0-1 2-4 5-8 >8
1 2 3 4
# Different parts in workstation 0 1-4 5-10 11-20 >20
0 1 2 3 4
# Assembly directions (ADL) 1 2-3 4-5 >5
 105

1 2 3 4

The allocation results are shown in Figure 4.4. We can clearly distinguish the sharp transition from

probability LOW to probability HIGH, yielding a distinct allocation with very few “intermediate”

workstations. When we use a cutoff level of 80% to divide HIGH from LOW, we see that out of the 76

workstations, only 10 stations in the HIGH classification were (subjectively) classified as “Low” by

their operators (the diamond shapes in the high region), and 3 as “High” in the LOW classification.

FIGURE 4.4 Classification results on 76 workstations

4.3.3 Quantifying the complexity of workstations

Using the full sample of workstations we arrived at a second logit model, that yields a more gradual

probability curve, that can be used to assign a complexity score to a given workstations. The model

looks like this (2), with 4 variables from Table 4.1:

𝑒 6.676−1.127𝑃𝑇𝐿−0.874𝑃𝑊𝐿−0.243𝐴𝐷𝑅−0.058𝑊𝑀𝑅
𝑃𝐿𝑂𝐺𝐼𝑇_𝐴𝐿𝐿 (𝐿𝑂𝑊) = 1+𝑒 6.676−1.127𝑃𝑇𝐿−0.874𝑃𝑊𝐿−0.243𝐴𝐷𝑅−0.058𝑊𝑀𝑅 (4.3)
 106

where PTL=number of packaging types in Likert scale

PWL=different parts in the workstation in Likert scale

ADR=number of assembly directions as measured directly (raw score)

WMR=number of work methods as measured directly (raw score)

The result is given in Figure 4.5. Using cutoff levels of 30% and 80% respectively we see that only 2

stations were classified differently by their operators. The overall quality of both models can be

quantified using the Receiver Operating Characteristics (ROC) theory (Fawcett, 2006). By

differentiating the cutoff level between 0 and 1 we can generate the ROC curves of both models. Figure

4.6 shows clearly that both models are of comparable quality, and equally strong. The larger the surface

between the ROC curve and the 45° line (the latter indicating expected performance of a random filter),

the more discriminating the model is in detecting the condition, in this case the complexity level of the

workstation.
 107

FIGURE 4.5 Complexity score of 76 workstations

Using both models one can identify the high complexity workstations, and quantify the effects of

changes to the work methods and layout of the stations on its complexity score.

FIGURE 4.6 ROC curve of both complexity models

4.4 Reducing the adverse impact of complexity

From the model in section 4.2 we can gather a long list of potential adverse effects of complexity. These

fall broadly into two categories: effects on the manufacturing system at large, and effects on the

performance of the operator.

Using the complexity classification model described in section 4.3.2. the operational characteristics of

two groups of workstations was determined: 19 HIGH complexity ones and 9 classified as LOW. The

average cycle time elements are shown in Figure 4.7. We can clearly conclude that complexity increases

the Balance Loss (BL) from 14 to 20% and Walking and Bending (WB) from 11 to 20%, and reduces
 108

at the same time Direct Work (DW) from 62 to 48%, or an efficiency loss of 22%! It should be noted

that the increased balance loss is partly due to the strategic choice of production engineering to allot a

higher common cycle time to absorb the workload peaks induced by the variants on certain workstations.

If this buffer time is not sufficient operators will fail to finish the operation or be forced to invade the

working space of the next workstation. This will obviously increase the stress on the operator, which

brings us to the adverse psychological effects of complexity.

FIGURE 4.7 Impact of complexity on workstation efficiency

To illustrate the effects of complexity on the operator’s performance we refer back to the notion of

choice complexity, mentioned in section 4.1. It is well known (Zhu et al, 2008) that the reaction time of

the operator to many and diverse stimuli (which we called “context switches” in the model of section

4.2) follows a distinct function as shown in Figure 4.8.

 109

FIGURE 4.8 Operator reaction time to choice complexity

So increased complexity will increase the operator reaction time in a given work sequence within a

station, further increasing the overall time pressure to the operator. A good example of this reaction time

was published by Hanson et al. (2012) showing that kitting parts (pre-ordering them in assigned

locations) as opposed to bulk supply in a mixed-model assembly line at Saab reduced the time to fetch

the parts considerably (Figure 4.9). Also the manner in which part identification was conveyed (printed

manifest versus signal lights) had an impact. Since fetch time includes information reaction time as well

as the physical grasping of the part, it illustrates how cognitive load influences operator performance.

45
Cycle time per part (Seconds)

40
35
30
25
20
15
10
5
0
Kitting same side Kitting opposite Bulk Pick to Light Bulk Printed
side

FIGURE 4.9 Effect of complexity on operator reaction time (after Hanson et al., 2012) – dark regions

are +/- 1 std.dev. around the mean

As already introduced in the beginning of this chapter, the main effect of complexity is an increased

cognitive load on the operator. In our causal model we identified many aspects of this so-called

subjective complexity. To avoid going extensively into a psychological analysis of the cause and effect

of complexity, we focus on the Rasmussen modelChyba! Nenašiel sa žiaden zdroj odkazov.. The main

effect of complexity is to present the operator with a larger proportion of level 2 (rules) and level 3

(knowledge) problems, to the detriment of the level 1 skills-based tasks for which he is trained, and on

which most of the allotted production times are based.

 110

TABLE 4.5 Complexity-induced cognitive load and mitigating actions

Level Rasmussen classification of operator Example Mitigating technology

reaction required

1 Skill-based

Repetitive assembly of different part sets on different Online tool sensoring and active
variant models using the right tools and applying the right tags for error control, increased
procedure (torque, visual control, …), incurring varying balance loss or peak shaving
workload per cycle sequencing

2 Rule-based

Assembly of rare parts from infrequent variants requiring Augmented reality wearable
different tools and/or procedures and access to instructions devices

3 Knowledge-based

Unique parts for which standard instructions are not Context-sensitive information
available or not up to date, requiring “same as but” behavior system generating on-the-fly
and supervisor assistance within normal cycle time step by step instructions,
empowered operators

4.4.1 Line balancing to reduce work overload and smoothen complexity

The assembly platform on which the various automotive models are assembled, requiring different

operating times at each workstation, is known in the literature as mixed-model assembly lines (see

Figure 4.10 below). In a mixed-model assembly line more than one model of the same general product

are intermixed and assemble on the same line. The amount of work required to assemble units can vary

from model to model, creating an uneven flow of work along the line. When the cycle time is fixed and

new models are added to the portfolio of the models on the line, work overload may occasionally or

regularly occur in some workstations. Increasing the cycle time for the whole line to moderate overload

occurrences is usually not acceptable in a competitive environment such as automotive. Instead,

rebalancing the line occasionally and sequencing the builds during each cycle to minimize the frequency

and magnitude of overloads is performed by the manager of these lines.

 111

FIGURE 4.10 Example of a mixed-model assembly line (Zeltzer, 2016)

In this section we present and discuss some optimization models and techniques to minimize overload

via intermittent line rebalancing and regular sequencing of the planned vehicle mix (builds). We then

examine how an entropy based measure of complexity can be used to reduce complexity while

rebalancing the line. Some of these models are tested on some real data sets obtained from our partners

in the automotive industry and some of their suppliers. The major results are also presented and

discussed.

For a clear presentation of the optimization models and line balancing approach presented and examined

below, we define the following relevant parameters:

We let M be the set of car models to be assembled on the mixed-model line, indexed by m. For each

model type m, let 𝑑𝑚 be the demand proportion that the model type m represents in the total model-mix

demand assembled on the line, (note that ∑𝑚∈𝑀 𝑑𝑚 = 1). Let K be the set of workstations making up

the assembly line, indexed by k. Each workstation k has a length 𝐿𝑘𝑚𝑎𝑥 𝑘 measured in time units (Figure

4.11):

FIGURE 4.11 Workstation’s major parameters

 112

We let J be the set of all tasks required to assemble the various models in M, indexed by j. Each task j

has a set 𝑃𝑟𝑒𝑑(𝑗)of direct predecessor tasks which must be performed before j. We use O to designate

the set of operators on the line. We also let Jk be the set of workstations k on which task j can be

performed and Jl be the set of operators l qualified to perform task j. We also introduce the parameter

Okmaxwhich represents the maximum number of operators that can be assigned to a workstation k. This

particular possibility is usually not taken into account in existing line balancing procedures. Finally, we

let c denotes the targetedcycle time of the line and 𝑡𝑗𝑚 the processing time of task j for the model m.

Given the above parameters, the so-called Mixed-Model Assembly Line Balancing Problem

(MMALBP), which is an optimization problem to optimally design the assembly line, consists of

finding an optimal assignment of all the tasks in J to the workstations in K that satisfies the various

constraints resulting from tasks precedence relationships together with the qualified workstations and

operators restrictions, and this while minimizing the work overload (and hence the balance loss)

resulting at each workstation when some car models are assembled.

The procedure and the underlying optimization tool to achieve such a line balance is summarized below.

4.4.1.1. Phase I. Mixed-Model Line Balancing Optimization

• Initialization Step:

 Select a “supermodel” (to reduce the problem to a single product case), a type I, II, F or E,

and then select a heuristic algorithm to determine an initial line balance.

 Determine the corresponding workstations and operators workloads and all work overloads

across all models.

• Improvement Step:
 113

 Improve the current workload balance (eliminate or reduce all overloads) using the selected

heuristic algorithm until the stopping criteria is satisfied (minimum number and magnitude

of overloads)!

• Analysis and Visualization Step:

 visualize the final line balance and compute stations and operators workloads and overloads

and all relevant statistics such as number of overloads and spreads per station and operator.

For the sake of completeness we also recall here the widely used definitions of the various basic line

balancing models: in Type I the objective is to minimize the number of workstations given the cycle

time; in Type II the objective is to minimize the cycle time for a given number of workstations; in Type

F the objective is to determine a feasible line balance, given the number of workstations and the cycle

time. Finally, in Type E the objective is to minimize both the cycle time and the number of stations.

Note at this point that the use of the “supermodel” to reduce the problem to the simple line balancing

problem is not the only alternative. Technically one can solve the problem taking the different models

explicitly into account. However, the use of a super model has an advantage in terms of computational

time and understanding by the users in the field if they do wish to use the line balancing tools as black

box.

Figure 4.12(a) and Figure 4.12(b) below show the typical summary of the results that must be generated

after the analysis and visualization step of the above procedure. This summary figure shows the number

of work overload occurrences and the spread per work station, and the models to be handle carefully.

These important performance measures will be exploited during the sequencing phase.
 114

FIGURE 4.12(a) Workstation’s major statistics: maximum spread per station

FIGURE 4.12(b) Balance loss versus overload statistics per operator, workstation for and models

The initialization step in the above procedure requires that a supermodel be defined. A supermodel is

characterized by the processing time of each task j, denoted as tˆj  f (t j ,, t j ,, t j , d1 ,, d m ,, d M )
1 m M

and is a function of the model processing times and their relative demand ratios. There are many

alternative ways to define this function. Below we propose examples of such a function:
 115


M
 Weighted Average Times: tˆj  d tm
m 1 m j

Specific model m*: tˆj  t j

m*



M
 Average times: tˆj  tm
m 1 j
M

Maximum task time: tˆj  Maxm 1,..., M t j

m


Minimum task time: tˆj  Minm 1,..., M t j

m


Most occurring task time: tˆj  MostFrequentm 1,..., M t j

m


Once a supermodel is selected, the following decision variable can now be used within an optimization

model to determine a task assignment that balances the resulting workload across the workstations:

1, If task j is assigned to station k and operator l

x ljk   (4.4)
0, Othewise

1, If operator l is assigned to station k

ylk   (4.5)
0, Othewise

The total workload of workstation k when model m is loaded is given by:

 km   jJ t mj lO xljk (4.6)

Workload of operator l when model m is loaded on workstation k is given the same way by:

 lkm   jJ t mj xljk  CLk (4.7)

where CLk is the complexity allowance on workstation k.

The average processing time at station k is determined as:

 116

Tk  
M
d
m 1 m
 
max  lkm : l  O (4.8)

Average station time assuming tasks can be divided at will:


K max
T  k 1
Tk K max (4.9)

Average station time assuming tasks can be divided at will:

Tm  
K max
k 1

max  lkm : l  O K max  (4.10)

The typical objective functions which can be minimized are:

 Minimizes maximal deviation of a station’s k or operator’s l workload for any model m from

the cycle time or average station time:


max k 1... K max ,l 1,..., Lmax ( k ), m 1... M  lkm  (Cmax , Tm or T )  (4.11)

or

  
K max
k 1
Lkmax
l 1
M
m 1

max 0, lkm  (Cmax , Tm or T )  (4.12)

where Lkmax is the maximum number of operators working in parallel which can be assigned to

workstation k.

The constraints to be satisfied are:

 Tasks assignment (1):

 kK j  lO j
xljk  1, for all j  J (4.13)

 Precedence (2):

 lO j
xljk   hK i , h  k  lOi
l
xih  0, for all j  J , and i  Pj

(4.14)
 117

 Operator Assignment (3):

• Required if each operator can be assigned only to at most one workstation:

 kW
ylk  1, for all l  O (4.15)

• Number of operators working in parallel on workstation k:

 y
lO lk
 Lkmax , for all k W (4.16)

where Lkmax is the maximum number of parallel operators on station k.

 Station’s and Operators Workload (4):

 j:lO j
xljk  J l  J k ylk  0, for all l  O and k W (4.17)

When using a Super-Model:

 tˆ xl
j :lO j j jk
 Cmax  CLk , for all l  O and k W (4.18)

To directly control maximum deviations:

 t xl
j:lO j jm jk
 Cmax  CLk   lmk , for all l  O, m  M and k W (4.19)

where Jl is the set of all tasks which can be performed by operator l, Jk is the set of all tasks

which can be performed on station k and lmk is the overload caused by model m to operator’ l

workload.

This optimization model can be used in many ways:

 In case the number of workstations and the cycle time are given, the optimization model can be

used to determine the best feasible task and operators assignments balancing the workloads and

minimizing overloads.
 118

 If tasks are already assigned, then by fixing x variables the model turns into a generalized

assignment optimization models for which optimize operators assignment, balance their

workloads and minimize overloads.

 In case of one operator per station the optimization model turns into the classical line balancing

model.

4.4.1.2 Phase II. Sequencing & Work Overload Optimization

Figure 4.13(a) and Figure 4.13(b) show that the sequence in which the models are assembled might

negatively impact the productivity of the line. If the work overload at a workstation is not resorbed at

the subsequent ones, the productivity of the line will be negatively impacted. After balancing the line to

reduce as much as possible the number and magnitude of structural work overloads, sequencing

contributes in helping to resorb some of this structural overloads.

Stations
Workstation
Overload

S2
4
O
5
O
6
O
7
O

4
O

5
O

6
O

7
O

4
O

5
O

6
O

7
O

C
M1 M2 M2

FIGURE 13(a) If the builds are not sequenced, overload occurs

Stations
No overload

S2
4
O

5
O

6
O

7
O

4
O
5
O
6
O
7
O
4
O

5
O

6
O

7
O

C
M2 M1 M2
 119

FIGURE 13(b) If the builds are sequenced, no overload occurs.

4.4.1.3 Analysis of a prototype example

To illustrate what the models approaches discussed in this section, we consider a data set consisting of:

9 workstations, 9 models, 9 operators and 109 tasks. First, we present and compare the results of the

mixed-model assembly line balancing model. Then, we analyze the impact of sequencing on the work

overload at each work station:

Figure 4.14 shows the results of the current balance. One can observe that some overloads occur due to

the fact some new model we added to the portfolio of model assemble on the line without rebalancing.

In workstation 7 for instance, model 2 produces a high underload whereas model 3 produces a high

overload.

FIGURE 4.14 Results of the current line balance

The use of the optimization proposed another assignment of tasks and consequently a different

workstation loads. Figure 4.15 shows the results of the optimal line balance obtained via the propose

optimization model. One can see there are no overloads left and even the sequencing is not necessary in

this case.
 120

FIGURE 4.15 Results of the optimized line balance

To conclude this section we now discuss how complexity can be taken into account when developing a

line balancing. After a first balance is developed, we can computing the entropy at each workstation in

the usual way:

Complexity k    jS k
p j log( p j ) , (4.20)

where pj is the probability of occurrence of task j given the models which are assembled on the

workstation and Sk is of course the set of tasks assigned to workstation k. We can then report these values

for each workstation on a graphic as done in Figure 4.16, below. On this figure the red line labeled

original balance entropy show the entropy at workstations 6 and 7 are high. These two workstation are

by the way among the once which were revealed as complex by the analysis in preceding sections 4.2

and 4.3.
 121

FIGURE4.16 Summary of Manufacturing Complexity and Line Balancing Analyses (Source: Luiza

Zeltzer, 2016)

After rebalancing, the green line in the figure gives the entropy at the workstation. These values are now

levelled and are almost the same at each workstation. One can expect now that this rebalancing has not

only reduced overload but also helped reduce the complexity experienced by the operators. The next

section elaborates on this important issue.

4.4.2 Operator information systems to reduce mental workload

Careful inspection of the Figures 4.2, 4.3 and 4.4 above reveals that, despite advances in automation and

production equipment technology, information to and from the operator will take the leading role in
 122

mitigating the effects of complexity. Whenever the context surpasses level 1 behaviour (Table 4.5)

information technology can step in. In a first attempt to trigger level 2 actions information systems can

retrieve the procedure with the best fit to the observed context and present the operator with the right

tasks. If no suitable pre-defined procedure exists the system will need operator instructions to be able to

present the operator with information, as well as record the context for possible offline treatment by

engineering.

Selecting the right amount of information so as not to overburden the operator (which would increase

his cognitive load rather than subdue it) and presenting it in a non-obtrusive manner so as not to interfere

with production tasks remains a big challenge. The fast-evolving technology of wearable devices is

emerging as a very likely solution. Already large companies such as BMW (Woollaston, 2014), but also

small ones like industrial bakeries (Vandorpe, 2014), use augmented reality goggles to support operators

in complex and error-prone tasks (Figure 4.17).

FIGURE 4.17 Augmented reality on the shop floor (Courtesy of VUZIX)

A second important reason to adopt these bi-directional operator information systems is to allow

operators to use their experience and knowledge and capture them for later use. Operators should also

be able to annotate information and flag parts that are inadequate or plainly wrong. This approach could

be a rich source of improvement targets for offline treatment by production engineering, and also enable
 123

semi-automatic generation of correct operating instructions for training purposes and to secure

improvements.

One important aspect of such information systems, which still needs to be developed, is the context

gauging. The system should at all times be able to infer from both the physical appearance of the

workstation (the objective complexity) and the cognitive situation of the operator (the subjective

complexity) what information is needed and what automated actions are possible. To make this happen

research is needed to gain better understanding of the interaction and communication between the human

assisted with technology and the production process. This will require accurate capturing of the

cognitive status of the operator (stress levels, cognitive level of information processing, etc.) as well as

his relevant physical actions. These actions can be conscious and deliberate but also more unconscious

and emotional. We can distinguish different types of physical actions such as (1) the manipulation of

the work piece (using tools or hands), (2) the control of a machine through an interface (using levers,

buttons, joysticks) and (3) the maintenance of the workplace (setup, repair, clean), etc. Different

cognitive actions such as (1) the work instructions, (2) rapid problem solving and (3) emotional states

(e.g. stress, fatigue, satisfaction) should also be studied.

4.5 Conclusions

In this chapter we have explored the main elements of complexity and how it impacts the performance

of manual assembly workers in a high variety setting. Through a causal model we have shed more light

to the different drivers of complexity, the mechanisms with which complexity interferes with

manufacturing performance and the cognitive performance of the workforce, and the different

consequences. We also proposed a quantitative approach to classifying workstations as High or Low

complexity, and subsequently a gradual measure of complexity. By studying the differences between

the two kinds of workstation we described objective degradations of workstation efficiency, as well as

heightening of cognitive load, according to a mental model. Finally we included two approaches to

mitigate complexity in manufacturing. A specific line balancing algorithm was proposed to reduce the

workload peaks in high complexity workstations (allowing a reduction in overall balance loss and
 124

complexity), and a technology driven approach to operator information systems was described to

mitigate stress related to cognitive overload.

The coming years will see both focused research into this domain of human-centered manufacturing

systems and a wave of practical experiences from early adopters, that encompass all types and sizes of

companies as was illustrated.

REFERENCES

Anil Kumar S., N. Suresh, (2009), Operations Management, New Age International Publishers.

Claeys, Arno, Hoedt, S., Soete, N., Cottyn, J., & Van Landeghem, H. (2015). Framework for evaluating

cognitive support in mixed model assembly systems. Symposium on Information Control in

Manufacturing, Proceedings (pp. 980–985). Presented at the 2015 IFAC Symposium on Information

Control in Manufacturing (INCOM 2015).

De Lima Gabriel Zeltzer, L., Limère, L., Van Landeghem, H., Aghezzaf, E.H., and Stahre, J.. (2013).

“Measuring Complexity in Mixed-model Assembly Workstations.” International Journal of Production

Research 51 (15), pp. 4630–4643.

Deshmukh, A.V., J.J. Talavage, and M.M. Barash, (1998). Complexity in Manufacturing Systems, Part

1: Analysis of Static Complexity, IIE Transactions 30, No. 7, 645–655

EIMaraghy W.H., Urbanic, R.J., (2004) “Assessment of Manufacturing Operational Complexity”, CIRP

Annals, vol. 53, issue 1, pp. 401–406.

ElMaraghy W.H., ElMaraghy H., Tomiyama T., Monostori L., (2012) “Complexity in engineering

design and manufacturing”, CIRP Annals – Manufacturing Technology, Retrieved Nov. 16 from

http://dx.doi.org/10.1016/j.cirp.2012.05.001

Fawcett T., (2006), “An introduction to ROC analysis”, Pattern Recognition Letters, 27, pp. 861-874.

Fisher, M. L. and Ittner, C. D. (1999) ,The impact of product variety on automobile assembly operations:

empirical evidence and simulation analysis‟, Management Science, Vol. 45, pp. 771-786.

Fisher, M. L., Jain, A. and MacDuffie, J. P. (1995) „Strategies for product variety: lessons from the auto

industry‟, B.Kogut & E.Bowman, Eds. Redesigning the Firm. Oxford U. Press, 116-154.
 125

Frizelle, G. (1996). "Getting the measure of complexity." Manufacturing Engineer 75(6): 268-270.

Frizelle, G. and Suhov, Y.M. (2001) ‘An entropic measurement of queueing behaviour in a class of

manufacturing operations’, Proceedings of Royal Society, London, pp.1579–1601.

Fujimoto, H., and Ahmed, A., (2001), “Entropic Evaluation of Assemblability in Concurrent Approach

to Assembly Planning,” Proceedings of the IEEE International Symposium on Assembly and Task

Planning, pp. 306–311.

Fujimoto, H., Ahmed, A., Iida, Y., Hanai, M.,(2003) “Assembly Process Design for Managing

Complexities because of product varieties”, International Journal of Flexible Manufacturing Systems,

15, pp 283-307.

Gullander P., Davidsson A., Dencker K., Fasth Å., Fässberg T., Harlin U., Stahre J., (2012) “Towards

a Production Complexity Model that Supports Operation, Re-balancing and Man-hour Planning”,

Proceedings of4th Swedish Production Symposium, SPS11, 3rd to 5th of May, 2011, Lund, Sweden

Hu, S.J., X. Zhu, H. Wang, and Y. Kkoren (2008). Product Variety and manufacturing complexity in

assembly systems and supply chain. CIRP Annals, Manuf. Techn, 57 45-48.

Kim Y.S., (1999) “A System Complexity Approach for the Integration of Product Development and

Production System Design”, MSc thesis at MIT, USA.

MacDuffie J.P., K. Sethuraman, M.L. Fisher, (1996) “Product Variety and Manufacturing Performance:

Evidence from the International Automotive Assembly Plant Study”, Management Science, vol. 42 no.

3, pp. 350-369.

Mattsson, S., Karlsson, M., Gullander, P., Van Landeghem, H., De Lima Gabriel Zeltzer, L., Limère,

V., Aghezzaf, E.-H., et al. (2014). Comparing quantifiable methods to measure complexity in assembly.

INTERNATIONAL JOURNAL MANUFACTURING RESEARCH, 9(1), 112–130.

Parasuraman, R., Sheridan, T., Wickens C., (2000), “A Model for Types and Levels of Human

Interaction with Automation”, IEEE Transactions On Systems, Man, And Cybernetics—Part a: Systems

and Humans, vol. 30, no. 3, pp. 286-297.

 126

Rasmussen, J. (1983). Skills, Rules, and Knowledge; Signals, Signs, and Symbols, and other

Distinctions in Human Performance Models. IEEE Transactions on systems, man and cybernetics.

SMC-13 (No.3,May/Jun): pp. 257-266.

Rodríguez-Toro C., G.Jared, and K. Swift, (2004) “Product-development Complexity Metrics: a

Framework for Proactive-DFA Implementation”, 8th International Design Conference, Dubrovnik,

Croatia, pp.:483 – 490.

Shannon, C. E., (1948) “A Mathematical Theory of Communication,” Bell System Technical Journal

27 (3): 379–423.

Schuh G., L. Monostori, B.Cs. Csájib, S. Döringa. (2008) “Complexity-based Modeling of

Reconfigurable Collaborations in Production Industry” , CIRP Annals, vol. 57, issue 1, pp. 445–450.

Sivadasan, S., Efstathiou, J., Calinescu, A. and Huatuco, L.H. (2006) ‘Advances on measuring the

operational complexity of supplier-customer systems’, European Journal of Operational Research, Vol.

171, No. 1, pp.208–226.

Urbanic, R.J. and Elmaraghy, W.H. (2006) ‘Modeling of manufacturing process complexity’, in

Advances in Design, ElMaraghy, H.A. and ElMaraghy, W.H. editors, Springer Series in Advanced

Manufacturing, ISBN 978-1-84628-004-7.

Vandorpe A., (2014), “Biobakkerij pioniert met smart glasses (Bio-bakery pioneers with smart glasses)”

(in Dutch), De Tijd, March 3, 2014, retrieved on Dec 1, 2015 from http://www.tijd.be/ondernemen/.

Wiendahl H.P., P. Scholtissek, (1996) “Management and Control of complexity in Manufacturing”,

CIRP Annals, vol. 43, issue 2, pp. 533–540.

Woollaston V., (2014), “The end of the mechanic?”, The Daily Mail, Jan. 21, 2014, retrieved on Dec.

1, 2016 from http://www.dailymail.co.uk/sciencetech/article-2543395/

Zeltzer Luiza, Ph.D. dissertation submitted to Ghent University, 2016.

Zhu, X., J.S. Hu, Y. Koren, and S.P. Marin (2008). Modeling of manufacturing complexity in mixed-

model assembly lines, J. of Manuf. Science and Eng, vol 130, 051013

Zhu, X.,(2009), Modeling product variety induced manufacturing complexity for assembly systems

design, Dissertation University of Michigan, 89p., Retrieved Nov 16 on

http://hdl.handle.net/2027.42/62265
 127

Chapter five
 128

Modelling of Assembly Supply Chain Structures

Vladimir Modrak, Slavomir Bednar
Technical University of Kosice,Slovakia

ABSTRACT

Assembly supply chain structures and their modelling can be effectively used when planning assembly

sequences in terms of mass customized manufacturing. Commonly for this purpose, graph theory is

applied. Generally, assembly sequence planning helps layout designers, among others, to increase

productivity and to decrease complexity. On the other hand, high flexibility of supply chains is required

to obtain short periods of order realization and high product diversity. Especially, the network

complexity reduction and finding alternative assembly supply chains is recently in the center of

managers' attention as assembly supply chain (ASC) systems are becoming increasingly complex. This

chapter provides a framework to model alternative ASC structures which are subsequently analyzed

using selected structural complexity indicators. Later, we outline initial components of mass customized

assembly and present modeling of available product configurations on a hypothetical model of labeled

graph.

5. 1 Introduction

Gradual transformation of companies toward mass customization (MC) is pulled by the growing demand

for tailor-made mass produced products and pushed by rapid development of modern supporting

technologies such as information technologies, additive manufacturing, 5th generation mobile networks,

identification technologies and others. This gradual development causes companies to take different

forms of MC. In this context, authors in [1] differ between low-level mass customization and high-level

of mass customization. In the second case, it is required that manufacturing systems are highly flexible

and manufacturing planning and control is more complex. As a consequence, requirements on process

modularity are of higher importance. Therefore, companies with a higher degree of customization might
 129

focus on reducing the complexity of the production processes, since complexity problems not only affect

the production processes, but also managerial processes [2]. Although this objective is quite clear, the

way to achieve it is not. An identification of complexity metrics is the first precondition for solving the

problem. Prior to this task, it would be helpful to have a basic understanding of what factors affect the

manufacturing complexity. A typical feature for MC is that products consist of several modules and

each module can have a certain number of variants. Combinations of these variants contribute to high

product variety that triggers high manufacturing complexity [3]. The assembly of the modules creates a

network of interconnected workstations that are frequently characterized as mixed-model assembly

systems. At each station selected components are assembled onto the partially finished product. Finally,

the end product is finalized at the last station. According to Koren et al. [4], a configuration of assembly

stations has a notable impact on the performance of manufacturing systems. Wang [5] therefore adds

that ’it is necessary to take into account the effect of system configuration when studying the variety-

induced manufacturing complexity and its impact on the performance of mixed-model assembly

systems’.

The chapter is divided into two parts. The first part proposes a framework for modeling of alternative

ASC structures and gives a brief description of possible complexity indicators with the aim to measure

the structural complexity of alternative ASCs. In the second part of the chapter, an approach to

determining all possible product configurations is outlined. The quantification of all possible product

configurations and variants will be treated in detail in chapter six.

5. 2 Modeling of Assembly Supply Chain Structures

Modeling of ASC structures is useful especially when the structures obtained as a result can be utilized

for practical guidance and applied in conjunction with a selection of optimal assembly process

structures. Generally, assembly sequence planning helps layout designers, among others, to increase

productivity, reduce costs and complexity. In particular, the network complexity reduction is recently
 130

in the center of managers' attention as ASC systems are becoming increasingly complex. On the other

hand, high flexibility of supply chains is required to obtain short delivery times.

5.2.1 Generating all possible assembly supply chain structures

Commonly for this purpose graph theory is applied. Assembly supply chains can be represented by tree

directed graphs, in which each node in the chain has at most one successor, but may have any number

of predecessors. Such supply chain structures are convergent and are divided into two basic types:

modular and non-modular. In the modular structure, the intermediate sub-assemblers are understood as

assembly modules, while the non-modular structure consists only of original suppliers and a final

assembler (root node). Steps to identify optimal ASC structures are clearly specified by Zhu et al. [6] in

these steps:

1. Generation of all possible supply chain structures;

2. Quantification of topological complexity values for each possible configuration;

3. Selection of the optimal supply chain configurations.

The same authors outlined the way forward to model possible supply chain structures depending on the

number of original suppliers i. For example, if the supply chain has from two to five original suppliers,

then we obtain 1, 2, 5 and 12 different ASC networks respectively, as can be seen in Figure 5.1.
 131

FIGURE 5.1 All possible ASC networks based on numbers of original suppliers

Analogically, it is possible to generate topological structures of ASC networks for higher numbers of

original suppliers (see reference [7]). For this purpose, it is useful to establish a framework for creating

so called topological classes of assembly supply chains for both non-modular and modular ASC

networks based on the number of initial nodes (original suppliers) i respecting the following rules:

1. ASC structures with an identical number of initial nodes are grouped into so called topological

classes categorized based on the number of initial nodes.

2. The initial nodes i in ASC structures are located on tiers tl (l = 0,…, m). The tiers are ordered

from right to left.

3. A final assembly workstation is situated in tier t0, while it is assumed to model ASCs only

with one final assembly node. In a case when a real assembly process consists of more than one final

assembly stations, then it is useful to split the assembly network into independent networks.

4. The minimal number of initial nodes i in the first tier t0 equals two.

5. Each node in the chain has at most one successor and must have at least two predecessors.
 132

6. In case of non-modular ASC structure, the number of initial nodes i in the most upstream

echelon is equal to the number of individual assembly parts or inputs.

Obtained unique topological structures are considered to be directed in three graphs so that all edges

point from left to right. Thus, they are convergent.

An example of the sets of structures for the classes with numbers of initial nodes i=6 and i=7 can be

found in Appendix A and ASC structures with numbers of initial nodes i=8 can be found in Appendices

B and C.

5.2.2 Complexity mitigation in assembly supply chain structures

In the previous paragraph 5.1.1, ASC were modeled by unlabeled graphs. However, ASCs correspond

with labeled graphs. Therefore, in order to mitigate complexity of ASC structures, it is advisable to

utilize the following steps:

1. Firstly, to transform a realistic ASC system into a labeled graph and subsequently to assign

a corresponding ASC structure to this graph. A simple example is shown in Figure 5.2.

2. Subsequently, identify alternative ASC structures to the corresponding ASC structure

considered as the original one. Such alternative ASC structures are available in all lower classes of ASC

structures than the class of given ASC structure. As an example, we can use the original ASC structure

from Figure 5.2(c).

 133

FIGURE 5.2 (a) Model of ASC network, (b) transformation into a labeled graph, and (c) transformation

into ASC structure No. 10 in Class #5

Then, alternative ASC networks can be identified as depicted in Figure 5.3. Such alternative ASC

networks are empirically less complex and at the same time they are keeping the predefined assembly

sequence unchanged. From a graph theory viewpoint, obtained alternative graphs in the lower

topological classes are then homeomorphic to the original labeled graph. From a practical point of view

such alternatives of ASC systems can be achieved based on an outsourcing method that often allows a

company's clients to reach better, faster, and more sustainable results.

FIGURE 5.3 Original ASC structure with i=5 and related alternative ASC networks
 134

3. Finally, to substitute intuitive methods for complexity comparison of benchmarked ASC

structures by appropriate and effective indicators of topological complexity.

Coming back to Figure 5.3, four alternative ASC structures are taken from ASC topological classes of

#4, #3 and #2. Then, when considering, e.g. the ASC structure taken from the topological class #4

(namely, graph No. 5), it is assumed that assembly node DE with two external suppliers is substituted

by one external supplier of the DE module. It is also evident that in alternative ASC models, the

predefined assembly sequences are unchanged. Arrangements of three other alternative ASC structures

(graphs No. 4, No. 2 and No. 1) are analogical. In the next step, it is useful to benchmark the complexity

of the original ASC structure (graph No. 10) against the four alternative ASC networks. For this purpose,

in the next section 5.3 alternative complexity indicators are described.

5.2.3 Approaches to structural complexity of assembly supply chain structures

Metrics of structural complexity for specific or general networks can help us not only to design those

systems but also better understand their topological properties. Layout design complexity metrics can

be effectively used especially when comparing two or more ASC structures. Several studies in the

literature [8-13] can be found dealing with complexity indicators for structural complexity measurement

of manufacturing systems. Since there are several types of ASC models, not all the complexity indicators

are equally effective for different groups of networks. Respecting this fact, the following three

complexity indicators are proposed.

5.2.3.1 Supply Chain Length

Nemeth and Foldesi [14] described the Supply Chain Length (SCL) indicator and its extended definition.

The SCL indicator takes besides the number of nodes, also the number of links weighted by the
 135

complexity of links into consideration. It is mainly focused on material flows. The equation formula

(5.1) of SCL is expressed by the equation:

𝑆𝐶𝐿 = 𝑐1 . ∑𝑖∈𝑃 𝑊𝑆 . 𝑉𝑖 + 𝑐2 . ∑(𝑖,𝑗)∈𝑃 𝑓(𝐷𝑖,𝑗 ). 𝐴𝑖,𝑗 , (5.1)

where: c1 - constant represents the technical and managerial level of vertices, c2 – constant represents

the technical and managerial level of edges, wS – weight corresponding the nature of node, P – path from

the origin to the destination, Vi – the vertices (nodes) in the path, Aij – the arcs (edges) in the path, Dij –

distance in logistic terms (in this study it equals 1), f(Dij) – the weight determined by the distance in

logistic terms.

Subsequently, the following SCL indictor values for all ASC structures from relevant topological classes

have been obtained (see Table 5.1).

TABLE 5.1 SCL values of ASC structures for the related topological classes

Class SCL Values for individual ASC structures

number No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11 No. 12

#5 11 13 13 13 15 15 15 15 15 17 17 17
#4 9 11 11 13 13 - - - - - - -
#3 7 9 - - - - - - - - - -
#2 5 - - - - - - - - - - -

5.2.3.2 Vertex degree index Ivd

According to Shannon’s information theory, the entropy of information H(α) in describing a message of

N system elements, distributed according to some equivalence criterion α into k groups of N1, N2,…, Nk

elements, is calculated by the formula:

𝑁𝑖 𝑁
𝐻(𝛼) = − ∑𝑘𝑖=1 𝑝𝑖 𝑙𝑜𝑔𝑝𝑖 = − ∑𝑘𝑖=1 𝑙𝑜𝑔2 𝑖, (5.2)
𝑁 𝑁

where pi specifies the probability of occurrence of the elements of the ith group.
 136

Since it is of interest to characterize entropy of information of a network according to [15], it is possible

to substitute symbols or system elements for the vertices.

In order to define the probability for a randomly chosen system element i it is possible to formulate

general weight function as pi = wi / Σwi, assuming that Σpi = 1. Considering the system elements, the

vertices, and supposing the weights assigned to each vertex to be the corresponding vertex degrees, one

easily distinguishes the null complexity of the totally disconnected graph from the high complexity of

the complete graph. Then, the probability for a randomly chosen vertex i in the complete graph of V

vertices to have a certain degree deg(v)i can be expressed by the formula:

deg⁡(𝑣)𝑖
𝑝𝑖 = ∑𝑉 . (5.3)
𝑖=1 deg⁡(𝑣)𝑖

Shannon defines information as:

𝐼 = 𝐻𝑚𝑎𝑥 − 𝐻, (5.4)

where Hmax is maximum entropy that can exist in a system with the same number of elements.

Subsequently, the information entropy of a graph with a total weight W and vertex weights wi can be

expressed in the form of the equation:

𝐻(𝑊) = 𝑊𝑙𝑜𝑔2 𝑊 − ∑𝑉𝑖=1 𝑤𝑖 𝑙𝑜𝑔2 𝑤𝑖 . (5.5)

Since the maximum entropy is when all wi =1, then

𝐻𝑚𝑎𝑥 = 𝑊𝑙𝑜𝑔2 𝑊. (5.6)

By substituting 𝑊 = ∑𝑉𝑖=1 deg⁡(𝑣)𝑖 and wi = deg(v)i , the information content of the vertex degree

distribution of a network called as Vertex degree index (Ivd) is derived by Bonchev and Buck [16] and

is expressed as follows:

𝐼𝑣𝑑 = − ∑𝑉𝑖=1 𝑑𝑒𝑔(𝑣)𝑖 𝑙𝑜𝑔2 deg⁡(𝑣)𝑖 . (5.7)

The following Table 5.2 summarizes value of Ivd indicator for all ASC structures in classes #2-5.
 137

Table 5.2 Ivd values of ASC structures for the related topological classes

Class Ivd Values for individual ASC structures

number No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11 No. 12
#5 11,61 13,61 12,75 12,75 14,75 14,75 14,75 14,26 14,26 16,26 16,26 16,26
#4 8,00 10,00 9,51 11,51 11,51 - - - - - - -
#3 4,75 6,75 - - - - - - - - - -
#2 2,00 - - - - - - - - - - -

5.2.3.3 Axiomatic design-based complexity SDC

The main definition of axiomatic design [17] states that any process can be seen in four main domains:

process, functional, customer and physical. The process consists of several steps and at the end results

in structured relations between customer needs, functional requirement (FR) and selected design

parameters (DP). These relations or dependencies between FRs and DPs within any design hierarchy

can be expressed by the relation:

FR = [A] DP, (5.8)

where each element of the matrix [A] can be expressed as A=FR/DP. Equation (5.8) can be expressed

as each FR on the product component depends on the specific DP of the product specified by customer,

so that each such dependency [A] can be understood as existing relation of FR on DP. If in the design

matrix of any process element A refers to ’0’, then FR is not in relation with DP. And vice versa for ’1’,

where there is a relation between the DP and FR.

According to this approach, we indicate each initial node of the ASC model as FR (for example FR1 to

FR10 at C10) and each sub-assembly vertex as DP (for example DP1 to DP3 depending on the specific

ASC structure at C10 shown in Figure 5.4(a) and Figure 5.4(b). This is because initial nodes practically

represent company requirements on suppliers and specify the number of initial nodes into ASC.

Subsequently, DPs are determined by these FRs as sub-structures. This way, the transformation of all

repeated ASC is possible and valuable.

 138

FIGURE 5.4 (a) ASC structure with 10 FRs and 2 DPs; (b) ASC structure with 10 FRs and 3 DPs,

both transformed into design matrix.

Analogically, we can transform each ASC structure into an axiomatic design matrix (see examples in

Figure 5.4(a) for 10 FRs and 2 DPs and Figure 5.4(b) for 10 FRs and 3 DPs).

Presented design matrices have been transformed as coupled designs. For such matrices it is

characteristic that individual elements [A] are mostly non-zero and thus the FRs cannot be satisfied

independently.

Table 5.3 SDC values of ASC structures for the related topological classes

Class SDC Values for individual ASC structures

number No.1 No. 2 No. 3 No.4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11 No. 12
#5 8,05 13,59 11,34 9,43 12,73 14,98 16,89 10,82 12,73 14,12 16,36 18,27
#4 5,55 8,84 6,93 8,32 10,23 - - - - - - -
 139

#3 3,30 4,68 - - - - - - - - - -
#2 1,38 - - - - - - - - - - -

5.2.4 Comparison of the approaches to structural complexity

The above presented indicators can be assessed in view of their applicability as follows. As seen in Table

5.1, values of SCL indicator are identical for several structures in individual ASC classes #2-5. For

example, class #5 with 12 individual ASC graphs can be divided into four SCL complexity levels. In the

case of Ivd indicator application, 12 graphs in the same ASC class #5 can be divided into six levels of

structural complexity (see Table 5.2). The two mentioned complexity indicators are therefore not the

optimal measures of structural complexity of ASCs. On the other hand, axiomatic design-based indicator

SDC considers graphs links as interactions between nodes. Then, complexity values of SDC indicator

for the same ASC structures in classes from #2 to #5 differ for each of the graphs (see Table 5.3).

Concluding the computational analysis using three structural complexity indicators, we may state that

AD-based complexity best suits for the given purpose.

Summary table of SCL, Ivd and SDC complexity values for all ASC structures in classes from i=2 to

i=8 can be found in Appendix D.

5.3 Modeling of Assembly Supply Chain Structures for Mass Customization

In section 5.2.1 it was discussed how to identify all possible assembly supply chain structures in order

to select an optimal ASC structure. It was also assumed that ASCs correspond with labeled graphs of

which initial nodes are represented by stable assembly components. In terms of mass customization,

these assembly components are commonly categorized into different types. Then, ASC structures are

composed of at least one assembly module with possible selection(s) from input components.
 140

In order to outline the categorization of assembly components in MC supply chain systems, we consider

exactly three types of initial components. They are as follows [18]: stable component; voluntary optional

component; compulsory optional component.

1. Stable components are considered to be assembled to ensure functionality of the module or final

product, respectively.

2. Voluntary optional components are those useful in some cases but not required. They can be selected

by customers and are optional in any combination, including cases when only individual components

are chosen. No component selection by customers is also an option.

3. Compulsory optional components differ from the voluntary optional ones by the number of the

components that may be chosen from all of them; so they are limited in selection. Thus, restrictions are

determined by the rules with minimum, maximum or exact requirements on a selection. The selection

rules can be in a simple way specified by a combinatorial number (𝑘𝑙)⁡, where l defines ways of picking

component combinations from a set of all k while 1≤l<k.

Let us have the simple example of an ASC using the structure No. 10 from the class #5 (see Figure

5.2(b)), in which two types of the initial components, stable and compulsory optional ones will be

considered (see Figure 5.5).

 141

FIGURE 5.5 Labeled graph incorporating compulsory optional components

Then, using simple combinatorial rules, we are able to obtain 32 design alternatives of product ABCDE.

From the customers' perspective such customized assembly offers 32 individual product configurations

composed of four changeable assembly modules with ten options for five initial parts (A to E). One can

see that if an ASC consists of multiple initial component types and multiple modules, enumeration of

all possible product configurations is not easy to proceed. Accordingly, it would be useful to establish

an effective framework for product variety quantification in terms of mass customization. Subsequently,

it would be possible to handle so called variety-induced complexity.

5.4 Conclusions

Complexity topology analysis of the ASC structures in sub-sections 5.2.3 and 5.2.4 revealed potential

tools to optimize assembly supply chain structures to be used in a mass customization environment.

Moreover, it has been found that modeling of all possible ASC is useful because any existing structure

can be simplified using the approach presented above to obtain less complex ASC alternative(s).
 142

Secondly, the three complexity metrics, namely SCL, Ivd and SDC, to capture structural properties of all

possible ASC networks have been benchmarked. One of them, namely the SDC indicator fits best for

the decision-making about the optimal supply chain as it also considers links between nodes and their

interoperability.

Subsequently, in section 5.3, it was shown that the proposed approach to model ASC networks can be

effectively used in the MC manufacturing environment. Finally, a draft of the concept for product variety

quantification has been outlined as a precondition for posterior solutions of product variety complexity

mitigation.

REFERENCES:

1. Stump, B., & Badurdeen, F. (2012). Integrating lean and other strategies for mass customization

manufacturing: a case study. Journal of Intelligent manufacturing, 23(1), 109-124.

2. Brosch, M., Beckmann, G., & Krause, D. (2011). Approach to Visualize the Supply Chain Complexity

Induced by Product Variety. In DS 68-5: Proceedings of the 18th International Conference on

Engineering Design (ICED 11), Impacting Society through Engineering Design, Vol. 5: Design for

X/Design to X, Lyngby/Copenhagen, Denmark, 15.-19.08. 2011.

3. Blecker, T., Friedrich, G., Kaluza, B., Abdelkafi, N., & Kreutler, G. (2004). Information and

management systems for product customization (Vol. 7). Springer Science & Business Media.

4. Koren, Y., Heisel, U., Jovane, F., Moriwaki, T., Pritschow, G., Ulsoy, G., & Van Brussel, H. (1999).

Reconfigurable manufacturing systems. CIRP Annals-Manufacturing Technology, 48(2), 527-540.

5 Wang, H. (2010). Product variety induced complexity and its impact on mixed-model assembly

systems and supply chains (Doctoral dissertation, General Motors).

6. Zhu, X., S. J. Hu, Y. Koren, S. P. Marin. 2008. Modeling of manufacturing complexity in mixed-

model assembly lines. Journal of Manufacturing Science and Engineering 130(5).

7. Modrak, V., Marton, D. (2013). Development of metrics and a complexity scale for the topology of

assembly supply chains. Entropy, 15(10), 4285-4299.

 143

8. ElMaraghy, H., AlGeddawy, T., Samy, S. N., & Espinoza, V. (2014). A model for assessing the layout

structural complexity of manufacturing systems. Journal of Manufacturing Systems, 33(1), 51-64.

9. Crippa, R., Bertacci, N., Larghi, L., 2006. Representing and Measuring Flow Complexity in the

Extended Enterprise: The D4G Approach, RIRL International Congress For Research in Logistics.

10. G. Frizelle, W.E. Measuring complexity as an aid to developing operational strategy International

Journal of Operations & Production Management, 15 (5) (1995), pp. 26–39

11. A.V. Deshmukh, J.J. Talavage, M.M. Barash Complexity in manufacturing systems, part 1: analysis

of static complexity. IIE Transactions, 30 (7) (1998), pp. 645–655

12. V.B. Espinoza Vega Structural Complexity of Manufacturing Systems Layout. University of

Windsor, Canada (2012) M.Sc. Thesis.

13. Wang, H., Zhu, X., Hu, S. J., & Koren, Y. (2008). Complexity Analysis of Assembly Supply Chain

Configurations. In ASME 2008 9th Biennial Conference on Engineering Systems Design and Analysis

(pp. 501-510). American Society of Mechanical Engineers.

14. P. Németh, P. Foldesi, Efficient Control of Logistic Processes Using Multi-criteria Performance

Measurement, Act. Tech. Jaur. Log. 2 (2009) 353-360.

15. Shannon CE. A mathematical theory of communication. Bell Sys Tech J 1948;27:379-423.

16. Bonchev D. and G. A. Buck, “Quantitative measures of network complexity,” in Complexity in

Chemistry, Biology and Ecology, vol. I, D. Bonchev and D. H. Rouvray, Eds. Springer, pp. 191 235,

2005.

17. Suh NP. Complexity in engineering. CIRP Annals-Manufacturing Technology, 54(2), 46-63.

18. Modrak, V., Marton, D., & Bednar, S. (2014). Modeling and Determining Product Variety for Mass-

customized Manufacturing. Procedia CIRP, 23, 258-263.

Appendix 5.1: All possible ASC structures in class #6 and class #7

 144

No.1 No.2 No.3 No.4 No.5 No.6 No.7 No.8 No.9 No.10 No.11 No.12 No.13 No.14 No.15

No.16 No.17 No.18 No.19 No.20 No.21 No.22 No.23 No.24 No.25 No.26 No.27 No.28 No.29 No.30 No.31 No.32 No.33

No.34 No.35 No.36 No.37 No.38 No.39 No.40 No.41 No.42 No.43 No.44 No.45 No.46 No.47 No.48 No.49 No.50 No.51 No.52

No.53 No.54 No.55 No.56 No.57 No.58 No.59 No.60 No.61 No.62 No.63 No.64 No.65 No.66 No.67 No.68 No.69 No.70 No.71 No.72 No.73

No.74 No.75 No.76 No.77 No.78 No.79 No.80 No.81 No.82 No.83 No.84 No.85 No.86 No.87 No.88 No.89 No.90

Appendix 5.2: All possible ASC structures in class #8 (Part 1)

 145

No.1 No.2 No.3 No.4 No.5 No.6 No.7 No.8 No.9 No.10 No.11 No.12 No.13 No.14

No.16 No.17 No.18 No.19 No.20 No.21 No.22 No.23 No.24 No.25 No.26 No.27 No.28 No.29 No.30 No.31 No.

No.34 No.35 No.36 No.37 No.38 No.39 No.40 No.41 No.42 No.43 No.44 No.45 No.46 No.47 No.48 No.49 No.50

No.53 No.54 No.55 No.56 No.57 No.58 No.59 No.60 No.61 No.62 No.63 No.64 No.65 No.66 No.67 No.68 No.69 No.70 No.71 N

No.74 No.75 No.76 No.77 No.78 No.79 No.80 No.81 No.82 No.83 No.84 No.85 No.86 No.87 No.88 No.89 No.90 No.91 No.92 No.93

No.96 No.97 No.98 No.99 No.100 No.101 No.102 No.103 No.104 No.105 No.106 No.107 No.108 No.109 No.110 No.111 No.112 No.113 No.114 N

No.118 No.119 No.120 No.121No.122 No.123 No.124 No.125 No.126 No.127 No.128 No.129 No.130 No.131 No.132 No.133 No.134 No.135 No.1
 146

Appendix 5.3: All possible ASC structures in class #8 (Part 2)

No.118 No.119 No.120 No.121No.122 No.123 No.124 No.125 No.126 No.127 No.128 No.129 No.130 No.131 No.132 No.133 No.134 No.1

No.139 No.140 No.141 No.142 No.143 No.144 No.145 No.146 No.147 No.148 No.149 No.150 No.151 No.152 No.153 No.154 No.155 No.156 No.157 N

No.162 No.163 No.164 No.165 No.166 No.167 No.168 No.169 No.170 No.171 No.172 No.173 No.174 No.175 No.176 No.177 No.178 No.179 No.180 No.

No.185 No.186 No.187 No.188 No.189 No.190 No.191 No.192 No.193 No.194 No.195 No.196 No.197 No.198 No.199 No.200 No.201 No.202 No.203 No.2

No.208 No.209 No.210 No.211 No.212 No.213 No.214 No.215 No.216 No.217 No.218 No.219 No.220 No.221 No.222 No.223 No.224 No.225 No.226 No.227 No.228 N

No.233 No.234 No.235 No.236 No.237 No.238 No.239 No.240 No. 241 No.242 No.243 No.244 No.245 No.246 No.247 No.248 No.249 No.250 No.251 No.252 No.253

No.258 No.259 No.260 No.261

 147

Appendix 5.4: Summary table of SCL, Ivd and SDC complexity values for all ASC structures in classes i=2-8
i=2 i=3 i=4 i=5 i=6 i=7 i=8
deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC deg(v)i LSC I vd SDC
2 5 2 1,39 3 7 4,75 3,3 4 9 8 5,55 5 11 11,61 8,05 6 13 15,51 10,75 7 15 19,65 13,62 8 17 24 16,64 4;4;4 21 24 27,98 4;3;3;3;3;2 27 29,02 57,90
3;2 9 6,75 4,68 4;2 11 10 8,84 5;2 13 13,61 13,59 6;2 15 17,51 17,70 7;2 17 21,65 24,37 8;2 19 26 30,26 4;4;4 21 24 23,23 4;3;3;3;3;2 27 29,02 55,99
3;3 11 9,51 6,93 4;3 13 12,75 11,34 5;3 15 16,36 12,14 6;3 17 20,26 21,67 7;3 19 24,41 27,39 4;4;4;2 23 26 36,85 4;3;3;3;3;2 27 29,02 53,74
3;3;2 13 11,51 8,32 4;3 13 12,75 9,43 5;3 15 16,36 16,30 6;3 17 20,26 15,01 7;3 19 24,41 18,02 4;4;4;2 23 26 31,27 4;3;3;3;3;2 27 29,02 53,74
3;3;2 13 11,51 10,23 4;3;2 15 14,74 12,73 5;3;2 17 18,36 17,68 6;3;2 19 22,26 32,42 7;3;2 21 26,41 41,01 4;4;4;2 23 26 41,60 4;3;3;3;3;2 27 29,02 51,23
4;3;2 15 14,74 14,98 5;3;2 17 18,36 24,34 6;3;2 19 22,26 25,76 7;3;2 21 26,41 31,64 4;4;3;3 23 25,51 38,73 4;3;3;3;3;2 27 29,02 49,33
4;3;2 15 14,74 16,89 5;3;2 17 18,36 20,18 6;3;2 19 22,26 23,05 7;3;2 21 26,41 28,77 4;4;3;3 23 25,51 36,23 4;3;3;3;3;2 27 29,02 51,23
3;3;3 15 14,26 10,8 4;4 15 16 14,05 5;4 17 19,61 19,17 6;4 19 23,51 24,68 4;4;3;3 23 25,51 34,32 4;3;3;3;3;2 27 29,02 49,33
3;3;3 15 14,26 12,73 4;4;2 17 18 17,34 5;4 17 19,61 16,92 6;4 19 23,51 19,93 4;4;3;3 23 25,51 33,98 4;3;3;3;3;2 27 29,02 48,99
3;3;3;2 17 16,26 14,12 4;4;2 17 18 22,09 5;4;2 19 21,61 29,92 6;4;2 21 25,51 38,30 4;4;3;3 23 25,51 32,07 4;3;3;3;3;2 27 29,02 47,08
3;3;3;2 17 16,26 16,36 4;3;3 17 17,51 15,43 5;4;2 19 21,61 27,67 6;4;2 21 25,51 33,55 4;4;3;3 23 25,51 29,36 4;3;3;3;3;2 27 29,02 45,17
3;3;3;2 17 16,26 18,27 4;3;3 17 17,51 13,52 5;4;2 19 21,61 22,46 6;4;2 21 25,51 27,98 4;4;3;3 23 25,51 33,52 4;3;3;3;3;2 27 29,02 48,53
4;3;3 17 17,51 19,59 5;3;3 19 21,12 27,21 6;3;3 21 25,02 35,43 4;4;3;3 23 25,51 31,61 4;3;3;3;3;2 27 29,02 46,62
4;3;3 17 17,51 17,68 5;3;3 19 21,12 23,05 6;3;3 21 25,02 28,77 4;4;3;3 23 25,51 29,36 4;3;3;3;3;2 27 29,02 44,37
4;3;3 17 17,51 15,43 5;3;3 19 21,12 20,55 6;3;3 21 25,02 26,07 4;4;3;3 23 25,51 29,36 4;3;3;3;3;2 27 29,02 44,37
4;3;3;2 19 19,51 23,48 5;3;3 19 21,12 18,30 6;3;3 21 25,02 21,32 4;4;3;3 23 25,51 27,46 4;3;3;3;3;2 27 29,02 42,46
4;3;3;2 19 19,51 21,57 5;3;3 19 21,12 16,39 6;3;3 21 25,02 19,41 4;4;3;3 23 25,51 28,77 4;3;3;3;3;2 27 29,02 48,53
4;3;3;2 19 19,51 20,98 5;3;3;2 21 23,12 37,96 6;3;3;2 23 27,02 49,05 4;4;3;3 23 25,51 26,86 4;3;3;3;3;2 27 29,02 46,62
4;3;3;2 19 19,51 19,07 5;3;3;2 21 23,12 33,81 6;3;3;2 23 27,02 42,39 4;4;3;3 23 25,51 26,86 4;3;3;3;3;2 27 29,02 44,37
4;3;3;2 19 19,51 18,73 5;3;3;2 21 23,12 31,30 6;3;3;2 23 27,02 39,69 4;4;3;3 23 25,51 24,95 4;3;3;3;3;2 27 29,02 43,78
4;3;3;2 19 19,51 27,64 5;3;3;2 21 23,12 29,05 6;3;3;2 23 27,02 34,94 4;4;3;3 23 25,51 24,61 4;3;3;3;3;2 27 29,02 41,87
4;3;3;2 19 19,51 25,73 5;3;3;2 21 23,12 27,14 6;3;3;2 23 27,02 33,03 4;4;3;3 23 25,51 24,61 4;3;3;3;3;2 27 29,02 43,78
4;3;3;2 19 19,51 23,48 5;3;3;2 21 23,12 28,60 6;3;3;2 23 27,02 36,82 4;4;3;3 23 25,51 22,70 4;3;3;3;3;2 27 29,02 41,87
3;3;3;3 19 19,02 14,91 5;3;3;2 21 23,12 24,44 6;3;3;2 23 27,02 30,16 4;4;3;3;2 25 27,51 52,35 4;3;3;3;3;2 27 29,02 41,87
3;3;3;3 19 19,02 19,07 5;3;3;2 21 23,12 23,85 6;3;3;2 23 27,02 29,36 4;4;3;3;2 25 27,51 51,83 4;3;3;3;3;2 27 29,02 39,96
3;3;3;3 19 19,02 16,82 4;4;3 19 20,75 24,96 5;5 19 23,22 22,18 4;4;3;3;2 25 27,51 49,85 4;3;3;3;3;2 27 29,02 39,62
3;3;3;3 19 19,02 20,98 4;4;3 19 20,75 22,46 5;5;2 21 25,22 27,73 4;4;3;3;2 25 27,51 47,94 4;3;3;3;3;2 27 29,02 37,71
3;3;3;3;2 21 21,02 20,11 4;4;3 19 20,75 20,55 5;5;2 21 25,22 35,80 4;4;3;3;2 25 27,51 47,60 4;3;3;3;3;2 27 29,02 45,66
3;3;3;3;2 21 21,02 20,45 4;4;3 19 20,75 20,21 5;4;3 21 24,36 32,93 4;4;3;3;2 25 27,51 45,69 4;3;3;3;3;2 27 29,02 43,75
3;3;3;3;2 21 21,02 27,12 4;4;3 19 20,75 18,30 5;4;3 21 24,36 30,68 4;4;3;3;2 25 27,51 42,99 4;3;3;3;3;2 27 29,02 41,50
3;3;3;3;2 21 21,02 22,36 4;4;3;2 21 22,75 35,71 5;4;3 21 24,36 30,23 4;4;3;3;2 25 27,51 47,15 4;3;3;3;3;2 27 29,02 41,50
3;3;3;3;2 21 21,02 24,87 4;4;3;2 21 22,75 33,21 5;4;3 21 24,36 26,07 4;4;3;3;2 25 27,51 45,24 4;3;3;3;3;2 27 29,02 39,59
3;3;3;3;2 21 21,02 29,03 4;4;3;2 21 22,75 31,30 5;4;3 21 24,36 25,48 4;4;3;3;2 25 27,51 42,99 4;3;3;3;3;2 27 29,02 39,00
4;4;3;2 21 22,75 30,96 5;4;3 21 24,36 23,57 4;4;3;3;2 25 27,51 42,99 4;3;3;3;3;2 27 29,02 37,09
4;4;3;2 21 22,75 29,05 5;4;3 21 24,36 25,48 4;4;3;3;2 25 27,51 41,08 4;3;3;3;3;2 27 29,02 39,00
4;4;3;2 21 22,75 26,35 5;4;3 21 24,36 23,57 4;4;3;3;2 25 27,51 42,39 4;3;3;3;3;2 27 29,02 37,09
4;4;3;2 21 22,75 25,76 5;4;3 21 24,36 21,32 4;4;3;3;2 25 27,51 40,48 4;3;3;3;3;2 27 29,02 36,75
4;4;3;2 21 22,75 23,85 5;4;3;2 23 26,36 46,55 4;4;3;3;2 25 27,51 40,48 4;3;3;3;3;2 27 29,02 34,84
4;3;3;3 21 22,26 30,51 5;4;3;2 23 26,36 44,30 4;4;3;3;2 25 27,51 38,57 4;3;3;3;3;2 27 29,02 32,93
4;3;3;3 21 22,26 28,60 5;4;3;2 23 26,36 43,85 4;4;3;3;2 25 27,51 38,23 4;3;3;3;3;2 27 29,02 38,21
4;3;3;3 21 22,26 26,35 5;4;3;2 23 26,36 39,69 4;4;3;3;2 25 27,51 36,33 4;3;3;3;3;2 27 29,02 36,30
4;3;3;3 21 22,26 26,35 5;4;3;2 23 26,36 39,10 4;4;3;3;2 25 27,51 40,12 4;3;3;3;3;2 27 29,02 34,05
4;3;3;3 21 22,26 24,44 5;4;3;2 23 26,36 37,19 4;4;3;3;2 25 27,51 37,61 4;3;3;3;3;2 27 29,02 34,05
4;3;3;3 21 22,26 23,85 5;4;3;2 23 26,36 39,10 4;4;3;3;2 25 27,51 35,70 4;3;3;3;3;2 27 29,02 32,14
4;3;3;3 21 22,26 21,94 5;4;3;2 23 26,36 37,19 4;4;3;3;2 25 27,51 35,36 4;3;3;3;3;2 27 29,02 38,21
4;3;3;3 21 22,26 23,85 5;4;3;2 23 26,36 34,94 4;4;3;3;2 25 27,51 33,45 4;3;3;3;3;2 27 29,02 36,30
4;3;3;3 21 22,26 21,94 5;4;3;2 23 26,36 34,32 4;4;3;3;2 25 27,51 32,66 4;3;3;3;3;2 27 29,02 34,05
4;3;3;3 21 22,26 21,60 5;4;3;2 23 26,36 32,07 4;4;3;3;2 25 27,51 36,82 4;3;3;3;3;2 27 29,02 35,70
4;3;3;3 21 22,26 19,69 5;4;3;2 23 26,36 33,52 4;4;3;3;2 25 27,51 34,91 4;3;3;3;3;2 27 29,02 33,79
4;3;3;3 21 22,26 19,69 5;4;3;2 23 26,36 29,36 4;4;3;3;2 25 27,51 32,66 4;3;3;3;3;2 27 29,02 33,79
4;3;3;3 21 22,26 17,78 5;4;3;2 23 26,36 31,02 4;4;3;3;2 25 27,51 32,66 4;3;3;3;3;2 27 29,02 31,88
4;3;3;3;2 23 24,26 41,26 5;4;3;2 23 26,36 29,11 4;4;3;3;2 25 27,51 30,75 3;3;3;3;3;3 27 28,53 45,66
4;3;3;3;2 23 24,26 39,35 5;3;3;3 23 25,87 40,98 4;4;3;3;2 25 27,51 34,32 3;3;3;3;3;3 27 28,53 43,75
4;3;3;3;2 23 24,26 37,10 5;3;3;3 23 25,87 36,82 4;4;3;3;2 25 27,51 32,41 3;3;3;3;3;3 27 28,53 41,50
4;3;3;3;2 23 24,26 37,10 5;3;3;3 23 25,87 34,32 4;4;3;3;2 25 27,51 30,50 3;3;3;3;3;3 27 28,53 39,00
4;3;3;3;2 23 24,26 35,19 5;3;3;3 23 25,87 32,07 4;3;3;3;3 25 27,02 44,27 3;3;3;3;3;3 27 28,53 37,09
4;3;3;3;2 23 24,26 34,60 5;3;3;3 23 25,87 30,16 4;3;3;3;3 25 27,02 42,36 3;3;3;3;3;3 27 28,53 36,75
4;3;3;3;2 23 24,26 32,69 5;3;3;3 23 25,87 31,61 4;3;3;3;3 25 27,02 40,12 3;3;3;3;3;3 27 28,53 36,30
4;3;3;3;2 23 24,26 34,60 5;3;3;3 23 25,87 27,46 4;3;3;3;3 25 27,02 40,12 3;3;3;3;3;3 27 28,53 34,39
4;3;3;3;2 23 24,26 32,69 5;3;3;3 23 25,87 26,86 4;3;3;3;3 25 27,02 38,21 3;3;3;3;3;3 27 28,53 32,14
4;3;3;3;2 23 24,26 32,35 5;3;3;3 23 25,87 24,95 4;3;3;3;3 25 27,02 37,61 3;3;3;3;3;3 27 28,53 31,54
4;3;3;3;2 23 24,26 30,44 5;3;3;3 23 25,87 26,86 4;3;3;3;3 25 27,02 35,70 3;3;3;3;3;3 27 28,53 29,64
4;3;3;3;2 23 24,26 28,53 5;3;3;3 23 25,87 24,95 4;3;3;3;3 25 27,02 37,61 3;3;3;3;3;3 27 28,53 29,64
4;3;3;3;2 23 24,26 31,90 5;3;3;3 23 25,87 22,70 4;3;3;3;3 25 27,02 35,70 3;3;3;3;3;3 27 28,53 27,73
4;3;3;3;2 23 24,26 29,99 5;3;3;3 23 25,87 20,79 4;3;3;3;3 25 27,02 35,36 3;3;3;3;3;3 27 28,53 27,39
4;3;3;3;2 23 24,26 27,74 5;3;3;3;2 25 27,87 54,60 4;3;3;3;3 25 27,02 33,45 3;3;3;3;3;3;2 29 30,53 59,28
4;3;3;3;2 23 24,26 27,74 5;3;3;3;2 25 27,87 50,44 4;3;3;3;3 25 27,02 31,54 3;3;3;3;3;3;2 29 30,53 57,37
4;3;3;3;2 23 24,26 25,83 5;3;3;3;2 25 27,87 47,94 4;3;3;3;3 25 27,02 34,91 3;3;3;3;3;3;2 29 30,53 55,12
4;3;3;3;2 23 24,26 27,14 5;3;3;3;2 25 27,87 45,69 4;3;3;3;3 25 27,02 33,00 3;3;3;3;3;3;2 29 30,53 52,62
4;3;3;3;2 23 24,26 25,23 5;3;3;3;2 25 27,87 43,78 4;3;3;3;3 25 27,02 30,75 3;3;3;3;3;3;2 29 30,53 50,71
4;3;3;3;2 23 24,26 27,14 5;3;3;3;2 25 27,87 45,24 4;3;3;3;3 25 27,02 30,75 3;3;3;3;3;3;2 29 30,53 50,37
4;3;3;3;2 23 24,26 25,23 5;3;3;3;2 25 27,87 41,08 4;3;3;3;3 25 27,02 28,84 3;3;3;3;3;3;2 29 30,53 49,92
3;3;3;3;3 23 23,77 31,90 5;3;3;3;2 25 27,87 40,48 4;3;3;3;3 25 27,02 34,91 3;3;3;3;3;3;2 29 30,53 48,01
3;3;3;3;3 23 23,77 29,99 5;3;3;3;2 25 27,87 40,48 4;3;3;3;3 25 27,02 33,00 3;3;3;3;3;3;2 29 30,53 45,76
3;3;3;3;3 23 23,77 27,74 5;3;3;3;2 25 27,87 38,57 4;3;3;3;3 25 27,02 30,75 3;3;3;3;3;3;2 29 30,53 45,17
3;3;3;3;3 23 23,77 25,23 5;3;3;3;2 25 27,87 36,33 4;3;3;3;3 25 27,02 30,16 3;3;3;3;3;3;2 29 30,53 43,26
3;3;3;3;3 23 23,77 23,33 5;3;3;3;2 25 27,87 34,42 4;3;3;3;3 25 27,02 28,25 3;3;3;3;3;3;2 29 30,53 47,05
3;3;3;3;3 23 23,77 22,99 5;3;3;3;2 25 27,87 42,36 4;3;3;3;3 25 27,02 30,16 3;3;3;3;3;3;2 29 30,53 45,14
3;3;3;3;3 23 23,77 21,08 5;3;3;3;2 25 27,87 38,21 4;3;3;3;3 25 27,02 28,25 3;3;3;3;3;3;2 29 30,53 42,89
3;3;3;3;3;2 25 25,77 42,65 5;3;3;3;2 25 27,87 35,70 4;3;3;3;3 25 27,02 28,25 3;3;3;3;3;3;2 29 30,53 40,39
3;3;3;3;3;2 25 25,77 40,74 5;3;3;3;2 25 27,87 33,45 4;3;3;3;3 25 27,02 26,34 3;3;3;3;3;3;2 29 30,53 38,48
3;3;3;3;3;2 25 25,77 38,49 5;3;3;3;2 25 27,87 31,54 4;3;3;3;3 25 27,02 28,25 3;3;3;3;3;3;2 29 30,53 38,14
3;3;3;3;3;2 25 25,77 35,99 5;3;3;3;2 25 27,87 34,91 4;3;3;3;3 25 27,02 26,34 3;3;3;3;3;3;2 29 30,53 39,59
3;3;3;3;3;2 25 25,77 34,08 5;3;3;3;2 25 27,87 30,75 4;3;3;3;3 25 27,02 26,00 3;3;3;3;3;3;2 29 30,53 37,68
3;3;3;3;3;2 25 25,77 33,74 5;3;3;3;2 25 27,87 32,41 4;3;3;3;3 25 27,02 26,00 3;3;3;3;3;3;2 29 30,53 35,43
3;3;3;3;3;2 25 25,77 33,28 5;3;3;3;2 25 27,87 30,50 4;3;3;3;3 25 27,02 24,09 3;3;3;3;3;3;2 29 30,53 37,09
3;3;3;3;3;2 25 25,77 31,37 4;3;3;3;3 25 27,02 22,18 3;3;3;3;3;3;2 29 30,53 35,18
3;3;3;3;3;2 25 25,77 29,12 3;3;3;3;3;3;2 29 30,53 33,27
3;3;3;3;3;2 25 25,77 28,53
3;3;3;3;3;2 25 25,77 26,62
 148

Chapter six

Variety-induced Complexity Metrics

Vladimir Modrak, Slavomir Bednar
Technical University of Kosice, Slovakia

ABSTRACT

It is frequently discussed that product variety impacts manufacturing complexity and in this context

assessment of variety-induced complexity is becoming a topical problem. Then, there is implicite need

of configuration complexity management methods to support managers in taking proper decisions. The

main scope of this chapter is description of combinatorial-based method to quantify product

configurations and variations arising in mass customized (MC) manufacturing. In this order, two

scenarios of structuring and quantifying of product components will be presented. The first one

considers two basic types of input assembly components, namely stable and optional. The second

scenario is enlarged with additional component type, which is compulsory optional component. Finally,

numbers of all possible product configurations (NPC) and newly proposed Axiomatic design-based

measure are benchmarked on the model of mass customized assembly process.As a result, both

measures can be effectively used to assist product managers and marketing advisors to independently

assess alternative product structures.

6. 1 Complexity as an important factor influencing MC strategy

Product configuration complexity problems are frequently discussed in connection with mass

customization. Moreover, product complexity management is considered as an important topic among

company managers and academics alike. At the same time, it is not easy to precisely define complexity

of product configuration options. Tihonen et al. [1] argue that the complexity of product configuration

options is directly related to the degree of modularity of the product. Probably, any definition of the
 149

product configuration complexity would be necessarily beholden to different types of products, as there

are many definitions of the term ’product configuration’ (see e.g., [2], [3], [4]). However, ability to

measure the complexity of product configuration using a reliable variety-based metric would allow

benchmarking of the concurrent or alternative product variety platforms. It is possible to identify several

pertinent facets of complexity in this domain. Calinescu et al. [5] provided comprehensive view on the

various aspects of manufacturing complexity including product structure complexity. According to Jiao

et al. [6], a product structure is defined in terms of module types, while product variants derived from

this product structure share the same module types and take on different instances of every module type.

Liu et al. [7] defines product structure by tree levels, for example: systems, sub-system, modules, in

which the components can be added, modified or deleted on each level. Fredendall and Gabriel [8]

pointed out that there are four building elements of the product structure complexity that seem to

contribute especially to product variety complexity. These are: 1. the number of assembled components

on entry to mass customized assembly system, 2. the number of manufactured items, 3. the number of

levels in a product assembly structure, and 4. the degree of part commonality. It is evident, that the

number of assembled components and the number of manufactured products definitely increases the

complexity of scheduling and even material control. But the number of components on entry to the

assembly system and the number of levels in the product structure independently contribute to the level

of product variety complexity.

Complexity of any system is affected mainly by three variables, namely: state of the system elements,

their number and relationships among them. Several definitions of manufacturing complexity have been

provided so far but the very firstdefinition is associated with the Shannon`s information theory [9]

related to the amount of information (in bits) in uncertainty of information system. From this approach,

it is evident that the fewer processes, machines and/or product configurations, the lower is the overall

complexity of the system. Zhu et al. [10] proposed a measure of complexity based on the choices that

the operator has to make at the station level. According toDesmukh et al. [11], product design

modifications can have significant impact on static complexity of the manufacturing system.
 150

Suh [12] defined complexity in relation to product design through achievement of functional and design

requirements. Kim et al. [13] introduced number of metrics for complexity on the basis of system

components, elements and their relations. These measures cover majority of system elements but cannot

be extended to other manufacturing domains, except for cell production. Frizelle and Woodcock [14]

defined two original types of complexity, static and dynamic, currently corresponding with structural

and operational complexity. Later, Frizelle and Efstathiou [15] presented the former “good complexity

and the latter ”bad complexity”. Their metrics have been further applied and even developed in terms

of mass customization by authors [16], [17].

Kampker at al. [18] cathegorized the term ’product variety’ into two types - internal and external. At

the same time they emphasize that “product architecture and technology determine the ratio between

the variety externally offered to the market and internally produced”. According to Grussenmeyer and

Blecker [19], there is a need for a novel complexity management method to support managers in taking

proper new product development decisions for the selection of suitable product variety platforms.

Our efforts in this chapter will focus on measurement and assessment of the product configuration

options from a complexity viewpoint. At the beginning, combinatorial-based methods to quantify

product configurations and variations representing product variety extent will be outlined.

Two scenarios, which differ in presence of initial component types will be assumed. The first one

(Scenario#1) considers two basic types of input components of assembly operation, namely stable (S)

and voluntary optional (VO) components. The second Scenario#2 is enlarged with additional

component type, namely compulsory optional components (CO).

6. 2 Quantification of product configurations and variations for Scenario#1

6.2.1 The conceptual model of product configurations

A quantification of product configurations for the two component types assumes the following number

of input components: unlimited number of stable components “i” starting with i=1 and unlimited
 151

number of voluntary optional components „j“ starting with j=0. In order to outline the model of product

configurations, the following simple example will be used under assumption that input components will

be assembled into a sub-module, module or product, respectively. Let us have anumber of stable

components, i.e.,i=2, anda number of voluntary optional components, i.e., j=2. This customizable

assembly operation can be categorized by the notation CLiSCLj, where CLi refers to Class of stable

components with the number of stable components “i”, and SCLj refers to Sub-class of voluntary

components with a number of voluntary components “j”. Thus, the example belongs to Class and Sub-

classCL2 SCL2. Then, four product configurations may occur (see Figure 6.1).

FIGURE6.1Case MCA illustration forCL2 SCL2

If the number of stable components isi≥2, and if zero VO are selected, we identify one product

configuration, as shown in Figure 6.1.

6.2.2 The conceptual model of product variations

In case of product configuration modeling, permutations of input components have been omitted. But

if we consider that permutations among componentsare relevant, then it is possible to generate for each

Class and Sub-class of product configuration related number of so called product variations. One may

determine mathematical expression for quantification of product variations. These variations are

pertinent if any product configuration consists of at least one voluntary optional component and the

number of stable components i ≥1.

 152

However, product variations may have more-less only theoretical importance, therefore, their relevance

will be proved in the following sub-section 6.2.4.

In order to introduce product variations and their quantification, let us use a case with three stable input

components, i=3 and two voluntary optional input components, j=2. Then, four product configurations

can be obtained (see Figure. 6.2(a)).

FIGURE6.2 (a) Four possible product configurations for product Class CL3 SCL2; (b) nine possible

product variations for the selected product configurations

Each of the four product configurations in Figure 6.2(a) may include a number of exactly related product

variations based on permutation rules. For example, product configuration, where i=3 and j=2 generates

nine product variations, as depicted in Figure. 6.2(b). In summary, the whole mass customized assembly

node CL3 SCL2 offers 16 product variations.

6.2.3 Combinatorial formulas for determination of product configurations

Following the principles of combinatorics described above, we have established formulas for

determination of product configurations. Two types of enumerations have been identified. The first of

them is for case of CL1 SCLj, and the second one is for case of CL2-∞ SCLj. For each of these types, there

are two different ways to determine number of product configurations. First of them brings only total

number of product configurations, but the second way of looking at it is through a structure of product

configurations. For example, we know that the case CL3 SCL2 generates four product configurations,
 153

but we do not know the structure and distribution of these configurations. The second way of calculation

allows us to know also structure of these configurations.

Let us further use both methods of the enumeration for the case CL1SCL0-∞, as follows:

a) a case, when a structure of product configurations is unimportant. Then, elementary formula for

quantification of product configurations using the first way, will be:

∑ 𝐶𝑜𝑛𝑓𝐶𝐿1 𝑆𝐶𝐿𝑗 = (2𝑗 ) − 1. (6.1)

In this formula, the configuration with single stable component is not considered as an assembly

operation and therefore this configuration is substracted.

b) a case, when a structure of product configurations is important. Then, for case CL1SCL0-∞, the formula

is:

𝑗 𝑛 𝑗!
∑ 𝐶𝑜𝑛𝑓𝐶𝐿1 𝑆𝐶𝐿𝑗 = ∑𝑗=1 ( ), (6.2)
𝑛𝑗 !(𝑗−𝑛𝑗 )!

where nj is a set of integers limited to a range 1 to j.

The second type of enumeration for the case CL2-∞ SCL0-∞, as follows:

a) a case,when a structure of product configurations is unimportant. Then the following formula is used:

∑ 𝐶𝑜𝑛𝑓𝐶𝐿2−∞ 𝑆𝐶𝐿𝑗 = (2𝑗 ), (6.3)

b) when a structure of product configurations is important, then the formula is as follows:

𝑗 𝑛 𝑗!
∑ 𝐶𝑜𝑛𝑓𝐶𝐿2−∞ 𝑆𝐶𝐿𝑗 = ∑𝑗=0 ( ), (6.4)
𝑛𝑗 !(𝑗−𝑛𝑗 )!

where nj is a set of integers limited to a range 0 to j.

 154

An example, how Formula (6.4) can be applied, is shown on simple case composition by using Class

and Sub-class CL4SCL3:

3! 3! 3! 3!
∑ 𝐶𝑜𝑛𝑓𝐶𝐿4 𝑆𝐶𝐿3 = ( ) + (1!(3−1)!) + (2!(3−2)!) + (3!(3−3)!)=1+3+3+1=8.
0!(3−0)!

This calculation offers a view on how j-component combinations are distributed. In our case, it shows

that we obtain one component combination consisting only from stable components, three combinations

with one of optional components, three combinations with two optional components and one component

combinatin consisting of all optional components, asgraphically shown in Figure 6.3.

FIGURE6.3 Distribution of product configurations in calculation for the product Class CL4 SCL3

In case when a distribution of product configurations is unimportant, we can apply Formula (6.3), by

which we obtain the same result:

∑ 𝐶𝑜𝑛𝑓𝐶𝐿4 𝑆𝐶𝐿03 = 23 = 8.

6.2.4 Combinatorial approach to determination of product variations

As the title of this sub-section indicates, after determination of product configurations, it is possible to

identifya related number of product variations for each product configuration. Doing so, two basic

approaches can be applied.

 155

The first way to enumerate a number of product variations (ΣVar) for any Class CLi and Sub-class SCLj

is by taking any single product configuration. Then, a formula for enumeration of product variations is

as follows:

𝑛
𝑗!𝑖 𝑗
∑ 𝑉𝑎𝑟𝐶𝐿𝑖𝑆𝐶𝐿𝑗 = ( ), (6.5)
𝑛𝑗! (𝑗−𝑛𝑗 )!

where nj is a set of integers limited to a range 0 to j.

The second way is pertinent when we need to quantify a sum of product variations for all possible

product configurations of the given Class CLi and Sub-class SCLj. Then, a formula for quantification of

product variations in Class CL1 is as follows:

𝑛
𝑗 𝑛 𝑗!𝑖 𝑗
∑ 𝑉𝑎𝑟𝐶𝐿𝑖𝑆𝐶𝐿𝑗 = ∑𝑗=1 ( ), (6.6)
𝑛𝑗! (𝑗−𝑛𝑗 )!

and an expression for quantification of product variations in Class CL2-∞ is:

𝑛
𝑗 𝑛 𝑗!𝑖 𝑗
∑ 𝑉𝑎𝑟𝐶𝐿𝑖𝑆𝐶𝐿𝑗 = ∑𝑗=0 (𝑛 ). (6.7)
𝑗! (𝑗−𝑛𝑗 )!

Validity of the Formula (6.5) for single product configuration can be proved by using an example from

Figure 6.2(b), wherei=3, j=2, and then the number of all product variations by the following way is

calculated:

2!32
∑ 𝑉𝑎𝑟𝐶𝐿3 𝑆𝐶𝐿2 = ( ) = 9.
2(2−2)!

Subsequently, the number of product variations for all relevant product configurations in Figure 6.2(a)

can be obtainedthrough the Formula (6.7):

2!30 2!31 2!32

∑ 𝑉𝑎𝑟𝐶𝐿3 𝑆𝐶𝐿2 = ( ) +( )+( ) = 16.
0!(2−0)! 1!(2−1)! 2!(2−2)!
 156

Summarizing results of the selected individual node of assembly operation with i=3, j=2, four product

configurations and 16 product variations were identified.

Accordingly, for any composition of entry components, product configurations and related number of

product variations can be quantified.

From results of numerical experiments offered in Table 6.1, it is shown that product variations grow

unrealistically while product configurations grow in a way, which is closer to real situations in terms of

mass customization.

TABLE 6.1 Related values of product configurations and variations for Classes CL2-5

CLi SCLj i j ΣConf ΣVar

CL2 SCL0 2 0 1 1
SCL1 2 1 2 3
SCL2 2 2 4 9
SCL3 2 3 8 27
SCL4 2 4 16 81
SCL5 2 5 32 243
SCL6 2 6 48 729
CL3 SCL0 3 0 1 1
SCL1 3 1 2 4
SCL2 3 2 4 16
SCL3 3 3 8 64
SCL4 3 4 16 256
SCL5 3 5 32 1 024
SCL6 3 6 48 4 096
CL4 SCL0 4 0 1 1
SCL1 4 1 2 5
SCL2 4 2 4 25
SCL3 4 3 8 125
SCL4 4 4 16 625
SCL5 4 5 32 3 125
SCL6 4 6 48 15 625
CL5 SCL0 5 0 1 1
SCL1 5 1 2 6
SCL2 5 2 4 36
SCL3 5 3 8 216
SCL4 5 4 16 1 296
SCL5 5 5 32 7 776
SCL6 5 6 48 46 656
 157

For example, if i=5 and j=6, then 48 product configurations and 46 656 product variations can be

composed from this initial component set. However, empirical experiences tell us that number of

demanded product configurations is usually lower than the number of all possible product

configurations. Therefore, a variety extent expressed through number of all possible product variations

might be more or less considered only as a theoretical construct. Further, we provide extended summary

tables for both product configurations and product variations in the following Classes and Sub-classes

CL1 SCL1-36 and CL2 SCL0-35(see Table 6.2 and Table 6.3).

TABLE 6.2 Summary table with related values of product configurations (ΣConf) and variations

(ΣVar) for product Class CL1

Class Components "S" Sub-class Components "VO" ΣConf ΣVar

SCL1 1 1 1
SCL2 2 3 3
SCL3 3 7 7
SCL4 4 15 15
SCL5 5 31 31
SCL6 6 63 63
SCL7 7 127 127
SCL8 8 255 255
SCL9 9 511 511
SCL10 10 1 023 1 023
SCL11 11 2 047 2 047
SCL12 12 4 095 4 095
SCL13 13 8 191 8 191
SCL14 14 16 383 16 383
SCL15 15 32 767 32 767
CL1 1 SCL16 16 65 535 65 535
SCL17 17 131 071 131 071
SCL18 18 262 143 262 143
SCL19 19 524 287 524 287
SCL20 20 1 048 575 1 048 575
SCL21 21 2 097 151 2 097 151
SCL22 22 4 194 303 4 194 303
SCL23 23 8 388 607 8 388 607
SCL24 24 16 777 215 16 777 215
SCL25 25 33 554 431 33 554 431
SCL26 26 67 108 863 67 108 863
SCL27 27 134 217 727 134 217 727
SCL28 28 268 435 455 268 435 455
SCL29 29 536 870 911 536 870 911
SCL30 30 1 073 741 823 1 073 741 823
SCL31 31 2 147 483 647 2 147 483 647
 158

SCL32 32 4 294 967 295 4 294 967 295

SCL33 33 8 589 934 591 8 589 934 591
SCL34 34 17 179 869 183 17 179 869 183
SCL35 35 34 359 738 367 34 359 738 367
SCL36 36 68 719 476 735 68 719 476 735

TABLE 6.3 Summary table with related values of product configurations (ΣConf) and variations

(ΣVar) for product Class CL2-∞

Class Sub-class ΣConf ΣVar

SCL1 2 3
SCL2 4 9
SCL3 8 27
SCL4 16 81
SCL5 32 243
SCL6 64 729
SCL7 128 2 187
SCL8 256 6 561
SCL9 512 19 683
SCL10 1 024 59 049
SCL11 2 048 177 147
SCL12 4 096 531 441
SCL13 8 192 1 594 323
SCL14 16 384 4 782 969
SCL15 32 768 14 348 907
SCL16 65 536 43 046 721
SCL17 131 072 129 140 163
CL2 SCL18 262 144 387 420 489
SCL19 524 288 1 162 261 467
SCL20 1 048 576 3 486 784 401
SCL21 2 097 152 10 460 353 203
SCL22 4 194 304 31 381 059 609
SCL23 8 388 608 94 143 178 827
SCL24 16 777 216 282 429 536 481
SCL25 33 554 432 847 288 609 443
SCL26 67 108 864 2 541 865 828 329
SCL27 134 217 728 7 625 597 484 987
SCL28 268 435 456 22 876 792 454 961
SCL29 536 870 912 68 630 377 364 883
SCL30 1 073 741 824 205 891 132 094 649
SCL31 2 147 483 648 617 673 396 283 947
SCL32 4 294 967 296 1 853 020 188 851 840
SCL33 8 589 934 592 5 559 060 566 555 520
SCL34 17 179 869 184 16 677 181 699 666 600
SCL35 34 359 738 368 50 031 545 098 999 700
 159

From the tables above, one can see that in case of single stable component, numbers of product

configurations and product variations are identical. If we had one generic product, for example specific

car model with 10 supplementary equipment options, then maximum number of product configurations

and variations is 1023. But in cases when number of stable components i≥2, then product variations

have strong exponential growth while product configurations grow more steadily.

6. 3 Quantification of product configurations for Scenario#2

In contrast with Scenario#1 where both, product configurations and product variations were analyzed,

this scenario will treat only product configurations.

In the Scenario#2, a quantification of product configurations is enriched with a compulsory optional

component type, denoted as CO.Then, a notation for Sub-class of voluntary components SCLkj will

include new information through index “k”, which determines the number of CO components in

theassembly component combination. Such component type frequently occurrs in many customizable

assembly operations. Requirements on the selection of components from compulsory optional

components are expressed through a variable „l“. Here, it is assumed that „l“ as integer is limited by

the rule1≤l<k.

Then, it is necessary to specify rules for selection of compulsory optional components. They are as

follows:

Rule A: Individual selectivity rule - we may define exact number of ‘l’ of components to be chosen

from all ‘k’ of CO components, or

Rule B: Maximum selectivity rule - we may define the maximum number ‘l’ of CO components to

combine within an assembly choice of all ‘k’ of CO components (note that ‘l’ is max. k-1), or finally
 160

Rule C: Minimum selectivity rule – we may choose at least ‘l’ components from available possible ‘k’

CO components.

In this Scenario, similarly as in Scenario#1, we propose two alternative groups of calculation. The first

group of methods does not consider a distribution of j-combinations and k-combinations of components

and includes two types of formulas, one for ClassCL1 and another for Classes CL2-∞.

Group 1:

In case of the Rule B and if l=k-1 and i=1 (Class CL1), then the following formula can be applied:

∑ 𝐶𝑜𝑛𝑓𝐶𝐿 𝑆𝐶𝐿𝑘 = ((2𝑗 ) − 1) ∗ ((2𝑘 ) − 1). (6.8)

𝑖 𝑗

In case of the Rule B and if l=k-1 and i=2-∞ (Class CL2-∞), then the sum of product configurations can

be obtained by the formula:

∑ 𝐶𝑜𝑛𝑓𝐶𝐿 𝑆𝐶𝐿𝑘 = (2𝑗 ) ∗ ((2𝑘 ) − 1). (6.9)

𝑖 𝑗

The second group of methods can be used when a distribution of both, voluntary optional and

compulsory optional components is a matter of interest. Then, the following two formulas can be

applied:

𝑛 𝑗! 𝑘!
𝑗
∑ 𝐶𝑜𝑛𝑓𝐶𝐿 𝑆𝐶𝐿𝑘 = ∑𝑗=1 (𝑛 ) ∗ ∑𝑘𝑙=1 (𝑙!(𝑘−𝑙)!) , (6.10)
𝑖 𝑗 𝑗 !(𝑗−𝑛𝑗 )!

and the formula for Class CL2-∞ is as follows:

𝑛 𝑗! 𝑘!
𝑗
∑ 𝐶𝑜𝑛𝑓𝐶𝐿 𝑆𝐶𝐿𝑘 = ∑𝑗=0 (𝑛 ) ∗ ∑𝑘𝑙=1 (𝑙!(𝑘−𝑙)!). (6.11)
𝑖 𝑗 𝑗 !(𝑗−𝑛𝑗 )!

Array of possible numbers of product configurations for all Classes and Sub-classes of product

configuration and for k=<2;10> is depicted in Table 6.4.

 161

TABLE 6.4 Fragment of table summarizing product configurations for the three selection rules

(situations), while j=0

Number Selection rule

ofcomponents [k]
[l] Rule A Rule B Rule C
2 1 out of 2 2 2 2
3 1 out of 3 3 3 6
2 out of 3 3 6 3
4 1 out of 4 4 4 14
2 out of 4 6 10 10
3 out of 4 4 14 4
5 1 out of 5 5 5 30
2 out of 5 10 15 25
3 out of 5 10 25 15
4 out of 5 5 30 5
6 1 out of 6 6 6 62
2 out of 6 15 21 56
3 out of 6 20 41 41
4 out of 6 15 56 21
5 out of 6 6 62 6
7 1 out of 7 7 7 126
2 out of 7 21 28 119
3 out of 7 35 63 98
4 out of 7 35 98 63
5 out of 7 21 119 28
6 out of 7 7 126 7
8 1 out of 8 8 8 254
2 out of 8 28 36 246
3 out of 8 56 92 218
4 out of 8 70 162 162
5 out of 8 56 218 92
6 out of 8 28 246 36
7 out of 8 8 254 8
9 1 out of 9 9 9 510
2 out of 9 36 45 501
3 out of 9 84 129 465
4 out of 9 126 255 381
5 out of 9 126 381 255
6 out of 9 84 465 129
7 out of 9 36 501 45
8 out of 9 9 510 9
10 1 out of 10 10 10 1022
2 out of 10 45 55 1012
3 out of 10 120 175 967
4 out of 10 210 385 847
5 out of 10 252 637 637
6 out of 10 210 847 385
7 out of 10 120 967 175
8 out of 10 45 1012 55
9 out of 10 10 1022 10
 162

Possible numbers of product configurations for extended list of compulsory optional components

k=<11; 21> are available in Appendix A of this chapter.

Selction from four CO components according to the Rule A follows Pascals binomial distribution of

choices. The concept of product configurations does not consider the following two situations:

 The outer left side of the triangle represents the selection(𝑘𝑙) = (𝑘0) = 1, and this would mean

that no compulsory optional component must be selected. In such case, the optional components

by its nature would be invalid.

 The outer right side of the triangle (for example if k=4) represents the selection (𝑘𝑙) = (44) = 1,

and this would mean that all of the compulsory optional components must be selected. In such

case, the optional components might be replaced by the category of stable components.

To present applicability of the proposed approach, the following example for Class and Sub-class CL2

SCL42 is offered. It will be applied only selection Rule A to present individual values of configurations

per selection „l out of k“ components. Then by using Formula (6.11), we obtain:

   2!     4! 
 Conf
2! 2!
CL2 SCL42
         *    (1  2  1) * 4  16 ,
 0!(2  0)!   1!(2  1)!   2!(2  2)!   1!(4  1)! 

   2!     
 Conf
2! 2! 4!
CL2 SCL42
         *    (1  2  1) * 6  24 ,
 0!(2  0)!   1!(2  1)!   2!(2  2)!   2!(4  2)! 

   2!     4! 
 Conf
2! 2!
CL2 SCL42
         *    (1  2  1) * 4  16 .
 0!(2  0)!   1!(2  1)!   2!(2  2)!   3!(4  3)! 

The product configuration generation can be also graphically demonstrated as shown in Figure 6.4(a).

Alternatively, one may obtain number of product configurations by enumeration, as results shown in

Figure 6.4(b).
 163

FFIGURE 6.4 a) Product configurations for Class CL2 SCL42depicted graphically; b) number of

product configurations obtained by enumeration

Analogically, it is possible to apply presented formulas to determine all possible product configurations

for any Sub-class and Class in arbitrary individual assembly node.

6.4 Entropy-based complexity of product variety

High frequency of topics on complexity-related problems within mass customized production in

research publications clearly evokes their significance. Especially, the product variety-basedcomplexity

is often a matter of interest for theorists and practitioners. Our focus in this section goes towards an

exploration of the variety-based complexity on the basis of axiomatic design and entropy theories.

Presented approach relates to both types of product variety - internal and external.

To elucidate the basic idea of this approach to the complexity, it is based on a transformation of mutual

relations between input components of assembly nodes and numbers of related product configurations

into a design matrix. Subsequently, the measures to enumerate product variety complexity will be

adopted.

6.4.1 Adoption of Axiomatic Design Complexity

 164

The four main domains of AD: customer, functional, physical and process ones have been originally

defined by Suh [14]. As it is known, AD theory dealt with satisfying customers' wishes in a manageable

way as their wishes are often articulated in a non-engineering expression. Thus, customer needs are

necessarily converted into so called Functional Requirements (FRs) that must be satisfied by the Design

Parameters (DPs). In this sense, each FR must be linked to at least one DP according to Axiom #1 of

AD-theory. Then one could write:

𝐹𝑅 = [𝐴]𝐷𝑃, (6.12)

where each FR refers to at least one coupling with DP. If the DP has no effect on FR, then the design

indicates no dependency and refers to “0” and vice versa for “X” indicating coupling of DP and FR.

Depending on the type of the resulting design matrix [A], three types of design matrices exist:

uncoupled, decoupled and coupled design.

Guenov [20] successfully tested Suh's hypothesis that the distribution of FR vs. DP couplings gives a

good idea of complexity. Accordingly, he derived two indicators for architectural design complexity

measurement, which are relatively simple and easy to apply. A brief description of these measures is as

follows. Let us denote ‘N’ as a number of interactions within a design matrix, and N1, N2, …, NK as

numbers of interactions per each DP of the same matrix. Then, the so called Degree of disorder 𝛺 can

be expressed by the formula:

𝑁 𝑁 𝑁 𝑁 𝑁!
𝛺 = 𝐶𝑁 1 ∗ 𝐶𝑁−𝑁
2 3
∗ 𝐶𝑁−𝑁 𝐾
… ∗ … 𝐶𝑁−𝑁1 −⋯−𝑁𝐾−1
=𝑁 , (6.13)
1 1−𝑁 2 1 !𝑁2 !…𝑁𝐾 !

where:

𝑁 𝑁!
𝐶𝑁 1 = (𝑁−𝑁 . (6.14)
1 )!𝑁1 !

Secondly, Systems Design Complexity (SDC), as denoted by authors, is expressed as follows:

SDC = ∑NP ln NP, (6.15)

 165

where ‘NP‘ is interpreted as number of interactions per single DP.

Based on our previous works, e.g. [21], where comparison of the two indicators was performed, it was

recognized that, SDC indicator is more suitable for the given application domain than the Degree of

disorder.

In order to adopt SDC indicator in terms of MC, it is firstly important to transform our concept for

determination of product configurations into the design matrix [A].It isadvised to adopt model for

generation of all possible product configurations and their bipartite graphical representation into

axiomatic design matrices under the precondition, that FRs will be represented by all possible product

configurations and DPs will be represented by initial components related to product configurations (see

Figure 6.5(a)).

FIGURE6.5(a) Case model of MCA as bipartite graph; (b) Axiomatic design matrix of the case model.

Based on that, it is possible to transform practically any model of initial components and related product

configurations into an AD matrix, as shown in Figure 6.5(b). Subsequently, one can use obtained design

matrices to apply the complexity measures described by the Formula (6.13) and/or Formula (6.15).

6.4.2 Comparison of AD-based and combinatorial product complexity indicators

For this purpose, three testing cases will be considered for individual selections from 4 to 6 compulsory

optional components. Two indicators will be benchmarked namely, number of product configurations

(NPC) and Systems Design Complexity (SDC). Obtained values are depicted in Table 6.5.
 166

TABLE 6.5 Fragment of complexity indicators NCP and SDC

i j k Selection NPC ΣNp SDC

rule
4 1 out of 4 8 20 27,7
2 out of 4 12 42 83,6
3 out of 4 8 36 65,2
5 1 out of 5 10 25 38
2 out of 5 20 70 166,1
3 out of 5 20 90 232
1-∞ 1
4 out of 5 10 50 114,3
6 1 out of 6 12 30 48,9
2 out of 6 30 105 280,8
3 out of 6 40 180 567
4 out of 6 30 165 502,1
5 out of 6 12 78 178,7

As can be seen from the Table 6.5, values of NPC indicator for individual selections from 4 to 6

compulsory optional components are in line with existing combinatorial phenomenon on the

distribution of possible combinations within a Pascal`s triangle.With regards to the values of SDC

indicatorin all three Sub-classes CL1-∞ SCL41; CL1-∞ SCL51; CL1-∞ SCL61, these do not follow the

symmetrical values and are unique to every individual selection rule.Moreover, it can be easily proved

that, e.g. choosing “five out of six” options is less complex than choosing “five out of six“ options.

Then the product variety complexity for “one out of six” selection is lower than in the other cases of

this Sub-class. The difference is a consequence of higher number of bipartite interactions, so that

∑ 𝑁𝑝(6) = 30<∑ 𝑁𝑝(6) = 78.

1 5

Thus, an important finding is that SDC indicator brings the most realistic complexity values compared

to number of possible product configurations.

6.5 The feed and transfer complexity conception

 167

Product variety complexitymeasures described in previous sub-section 6.4 have been developed with

purpose to quantify complexity for individual assembly nodes. Their application would be very limited

since real assembly processes consist of multiple layers and complicated networks. Unlike Systems

Design Complexity, configuration complexity expressed by all possible product configurations (ΣConf)

is easily applicable even for complex modular structures.

Further will be described an effective way to apply these measures in case of real assembly processes.

According to Hu et al. [22], complexity of individual assembly stations is obtained as a weighed sum

of complexities associated with every upstream assembly activities, as can be seen in Figure 6.6.

FIGURE 6.6 Complexity aggregation principle

Feed complexity exists due to the product configurations added on the previous stations and they affect

subsequent processes at stations and configuration selections.

According to the previous scheme, transfer complexitycan only flow from upstream to downstream, but

not in theopposite direction. So called feed complexitycan only beadded at a current station without any

transferring behavior. Then the total complexity is always the sum of feed/node complexity and transfer

complexity from all upstreamassembly stations. This aggregation and/or cumulative principles are

further applied for calculation of the total complexity of modular structures.

6.6 Case application of entropy-based and combinatorial product configuration complexities

To show the relevance of the above presented methodological frameworks, two selectedmeasure will

be applied and verified on a model of mass customized manufacturing of washing machines (see Figure

6.7). The assembly process consists of two independent branches. Branch #1 offers top-loading version

of washing machine and Branch #2 is dedicated to front-loading washing machine option. Customized
 168

assembly branches on the basis of two determining stable input components A1 for top-loading and A2

for front-loading machine option. The second type is voluntary component B1-3, C1-4, D1-4, E1-3, H1-2, I1-

2 and they offer so called standard options. This option is automatically selected in cases when any of

the non-standard options is selected. The last component type – compulsory optional F1-4, G1-5 are

components with obligation to choose at least one component option of all possible within appropriate

assembly module (selection Rule C).

The following Figure 6.7 presents summary values of all possible product configurations in individual

modules and the total number of product configurations.

FIGURE 6.7 Model of MCA with CC values for individual nodes and summary CC value including

legend

From the Figure 6.8 below it is clear, that if a specific node on the single level joins multiple

components/modulesand these modules are dependent, then the summary value of ∑Conf is a

multiplication of the incoming (converging) product configurations on the same level. In the case of

independent branches (e.g. Branch#1 and Branch#2 on tier t0in Figure 6.8), then the summary value of

∑Conf for such branches is a sum of incoming (converging) product configurations into the node.

In the following Figure 6.8, aggregated complexity values based on axiomatic design approach using

Formula (6.15) are presented summarily.

 169

FIGURE 6.8 Summary feed SDC values of the MCA model

Summary values of the SDC complexity for both, dependent and independent branches are obtained as

a summation of the feed (node) complexity and complexity transferred from previous-upstream stations.

As can be seen in Figure 6.8, summary SDC complexity value on the final assembly Station 5 (with

independent branches) is obtained as a sum of the two converging complexities, while feed SDC

complexity on the Station 5 is zero, due to absence of additional operations.

In summary, regarding the so called „flow“ complexity, it can be stated:

In independent branches which result with final product, final complexity in both approaches is

obtained as a sum of individual complexities.

In dependent branches resulting with semi-products, CC of the next downstream node is obtained as a

multiplication of converging complexities which is moreover multiplied by the optional component(s)

according to combinatorial rules.

 170

6.7 Conclusion

This chapter considered number of approaches to assess product variety complexity. One of the

approaches, namely combinatorial-based approach to quantify number of product configurations (NPC)

in mass customization is quite straightforward method do so, thus product development teams can

effectively measure the variety of concurrent product platforms. Counting number of product variations,

as the second method described in sub-section 6.2.2, seems not to be very efficient, as product variations

represent only imaginary products by connecting all voluntary components in different positions of

stable components.

The measure adopting an axiomatic design entropy, called SDC, proved that this indicator works better

for the given purpose.This has been shown the case application in section 6.6.

Assessment of the product variety complexity of modular models by SDC indicator in combination with

aggregation principle results in flat values of complexity. Vice-versa, complexity values obtained by

NPC indicator are generated asmultiplication of all converging complexities and bring exponential

values.

Finally, both measures (NPC and SDC) can be combined to assist product managers and marketing

advisors to independently assess alternative product platforms and to evaluate their customization

characteristics such as product variety complexity in tasks to decide about optimal customizable product

platforms.

REFERENCES

1. Tiihonen, J., Soininen, T., Männistö, T., & Sulonen, R. (1996). State-of-the-practice in product

configuration—a survey of 10 cases in the Finnish industry. In Knowledge intensive CAD (pp. 95-114).

Springer US.

2. Klein R., Buchheit M., Nutt, W. (1994) Configuration as Model Construction: The Constructive

Problem Solving Approach. In Proceedings of Artificial Intelligence in Design 1994, 201–18.

 171

3. Mittal S., Frayman F. (1989) Towards a Generic Model of Configuration Tasks. In IJCAI-89:

proceedings of the Eleventh IJCAI, 20-25.

4. Najman O., Stein B. (1992) A Theoretical Framework for Configurations. In Proceedings of

Industrial and engineering applications of artificial intelligence and expert systems : 5th international

conference, IEA/AIE -92, Belli F., Radermacher F. J. (eds.), 441–50.

5. Calinescu, A., Efstathiou, J., Schirn, J., & Bermejo, J. (1998). Applying and assessing two

methods for measuring complexity in manufacturing. Journal of the Operational Research Society, 723-

733.

6. Jiao, J., Ma, Q., & Tseng, M. M. (2003). Towards high value-added products and services: mass

customization and beyond. Technovation, 23(10), 809-821.

7. Liu, X. F., Kane, G., & Bambroo, M. (2006). An intelligent early warning system for software

quality improvement and project management. Journal of Systems and Software, 79(11), 1552-1564.

8. Fredendall D., Gabriel TJ., Manufacturing complexity: a quantitative measure, POMS

Conference – proceedings, April 4-7, 2003, Savannah, GA

9. Shannon CE. A mathematical theory of communication. Bell Syst Tech J 1948;27(3):379-423.

10. Zhu X, Koren SJ, Marin SP. Modeling of manufacturing complexity in mixed-model assembly

lines. J Manuf Scie-T Asme 2008;130(5):313-34.

11. Desmukh AV, Talavage JJ, Barash MM. Complexity in manufacturing systems, part 1:

Analysis of static complexity. IEEE Trans Eng Manag 1998;30(7):645-55.

12. Suh NP. Complexity in engineering. Ann Manuf Tech 2005;54(2):46-63.

13. Kim, Y.-S. 1999. A System Complexity Approach for the Integration of Product Development

and Production System Design. Master of Science, Massachussetts Institute of Technology.

14. Frizelle, G., & Woodcock, E. (1995). Measuring complexity as an aid to developing operational

strategy. International Journal of Operations & Production Management, 15, 26-39.

15. Frizelle, G., & Efstathiou, J. (2002). Seminar Notes On Measuring Complex Systems. Resource

document. The London School of Economics and Political Science. Retrieved from http:// www. psych.

lse.ac.uk complexity/ PDFiles/ Seminars/ Gerjan April02lastversion.pdf. Accessed 12 May 2014.

 172

16. Krus, P.: Design Space Configuration for Minimizing Design Information Entropy, in:

Proceedings of the ICoRD’15 - Research into Design Across Boundaries: Theory, Research

Methodology, Aesthetics, Human Factors and Education, 07-09.01.2015, Springer India, pp. 51-60.

17. Modrak, V., Krus, P., Bednar, S.: Approaches to Product Variety Management Assuming

Configuration Conflict Problem. FME Transaction. 2015; 43:271-78.

18. Kampker, A., Burggräf, P., Swist, M., & Nowacki, C. (2014). Assessment and Configuration

of a Product Production System. In Enabling Manufacturing Competitiveness and Economic

Sustainability (pp. 147-152). Springer International Publishing.

19. Grussenmeyer, R., & Blecker, T. (2013). Requirements for the design of a complexity

management method in new product development of integral and modular products. International

Journal of Engineering, Science and Technology, 5(2), 132-149.

20. Guenov MD. Complexity and Cost Effectiveness Measures for Systems Design. Manufacturing

Complexity Network Conference 2002, Cambridge, UK, ISBN 1-902546-24-5.

21. Modrak V, Bednar S. Using Axiomatic Design and Entropy to Measure Complexity in Mass

Customization. Procedia CIRP 2015;34:87-92.

22. Hu SJ, Zhu X, Wang H, Koren Y. Product variety and manufacturing complexity in assembly

systems and supply chains. Annals of CIRP 2008;57:45-8.

Appendix 6.1: Summary table of avaialble product configurations when k=11-21

Selection rules
Components [k] [l] [l out of k]
Rule A Rule B Rule C
1 1 out of 11 11 11 2 047
2 2 out of 11 55 66 2 036
11 3 3 out of 11 165 231 1 981
4 4 out of 11 330 561 1 816
5 5 out of 11 462 10 1 486
 173

6 6 out of 11 462 1 485 1 024

7 7 out of 11 330 1 815 562
8 8 out of 11 165 1 980 232
9 9 out of 11 55 2 035 67
10 10 out of 11 11 2 046 12
1 1 out of 12 12 12 4 095
2 2 out of 12 66 78 4 083
3 3 out of 12 220 298 4 017
4 4 out of 12 495 793 3 797
5 5 out of 12 792 1 585 3 302
12 6 6 out of 12 924 2 509 2 510
7 7 out of 12 792 3 301 1 586
8 8 out of 12 495 3 796 794
9 9 out of 12 220 4 016 299
10 10 out of 12 66 4 082 79
11 11 out of 12 12 4 094 13
1 1 out of 13 13 13 8 191
2 2 out of 13 78 91 8 178
3 3 out of 13 286 377 8 100
4 4 out of 13 715 1 092 7 814
5 5 out of 13 1 287 2 379 7 099
6 6 out of 13 1 716 4 095 5 812
13
7 7 out of 13 1 716 5 811 4 096
8 8 out of 13 1 287 7 098 2 380
9 9 out of 13 715 7 813 1 093
10 10 out of 13 286 8 099 378
11 11 out of 13 78 8 177 92
12 12 out of 13 13 8 190 14
1 1 out of 14 14 14 16 383
2 2 out of 14 91 105 16 369
3 3 out of 14 364 469 16 278
4 4 out of 14 1 001 1 470 15 914
5 5 out of 14 2 002 3 472 14 913
6 6 out of 14 3 003 6 475 12 911
14 7 7 out of 14 3 432 9 907 9 908
8 8 out of 14 3 003 12 910 6 476
9 9 out of 14 2 002 14 912 3 473
10 10 out of 14 1 001 15 913 1 471
11 11 out of 14 364 16 277 470
12 12 out of 14 91 16 368 106
13 13 out of 14 14 16 382 15
1 1 out of 15 15 15 32 767
15 2 2 out of 15 105 120 32 752
3 3 out of 15 455 575 32 646
 174

4 4 out of 15 1 365 1 940 32 192

5 5 out of 15 3 003 4 943 30 827
6 6 out of 15 5 005 9 948 27 824
7 7 out of 15 6 435 16 383 22 819
8 8 out of 15 6 435 22 818 16 384
9 9 out of 15 5 005 27 823 9 949
10 10 out of 15 3 003 30 826 4 944
11 11 out of 15 1 365 32 191 1 941
12 12 out of 15 455 32 646 576
13 13 out of 15 105 32 751 121
14 14 out of 15 15 32 766 16
1 1 out of 16 16 16 65 535
2 2 out of 16 120 136 65 519
3 3 out of 16 560 696 65 399
4 4 out of 16 1 820 2 516 64 839
5 5 out of 16 4 368 6 884 63 019
6 6 out of 16 8 008 14 892 58 651
7 7 out of 16 11 440 26 332 50 643
16 8 8 out of 16 12 870 39 202 39 203
9 9 out of 16 11 440 50 642 26 333
10 10 out of 16 8 008 58 650 14 893
11 11 out of 16 4 368 63 018 6 885
12 12 out of 16 1 820 64 838 2 517
13 13 out of 16 560 65 398 697
14 14 out of 16 120 65 518 137
15 15 out of 16 16 65 534 17
1 1 out of 17 17 17 131 071
2 2 out of 17 136 153 131 054
3 3 out of 17 680 833 130 918
4 4 out of 17 2 380 3 213 130 238
5 5 out of 17 6 188 9 401 127 858
6 6 out of 17 12 376 21 777 121 670
7 7 out of 17 19 448 41 225 109 294
8 8 out of 17 24 310 65 535 89 846
17
9 9 out of 17 24 310 89 845 65 536
10 10 out of 17 19 448 109 293 41 226
11 11 out of 17 12 376 121 669 21 778
12 12 out of 17 6 188 127 857 9 402
13 13 out of 17 2 380 130 237 3 214
14 14 out of 17 680 130 917 834
15 15 out of 17 136 131 053 154
16 16 out of 17 17 131 070 18
1 1 out of 18 18 18 262 143
18
2 2 out of 18 153 171 262 125
 175

3 3 out of 18 816 987 261 972

4 4 out of 18 3 060 4 047 261 156
5 5 out of 18 8 568 12 615 258 096
6 6 out of 18 18 564 31 179 249 528
7 7 out of 18 31 824 63 003 230 964
8 8 out of 18 43 758 106 761 199 140
9 9 out of 18 48 620 155 381 155 382
10 10 out of 18 43 758 199 139 106 762
11 11 out of 18 31 824 230 963 63 003
12 12 out of 18 18 564 249 527 31 180
13 13 out of 18 8 568 258 095 12 616
14 14 out of 18 3 060 261 155 4 048
15 15 out of 18 816 261 971 988
16 16 out of 18 153 262 124 172
17 17 out of 18 18 262 142 19
1 1 out of 19 19 19 524 287
2 2 out of 19 171 190 524 268
3 3 out of 19 969 1 159 524 097
4 4 out of 19 3 876 5 035 523 128
5 5 out of 19 11 628 16 663 519 252
6 6 out of 19 27 132 43 795 507 624
7 7 out of 19 50 388 94 183 480 492
8 8 out of 19 75 582 169 765 430 104
9 9 out of 19 92 378 262 143 354 522
19
10 10 out of 19 92 378 354 521 262 144
11 11 out of 19 75 582 430 103 169 766
12 12 out of 19 50 388 480 491 94 184
13 13 out of 19 27 132 507 623 43 796
14 14 out of 19 11 628 519 251 16 664
15 15 out of 19 3 876 523 127 5 036
16 16 out of 19 969 524 096 1 160
17 17 out of 19 171 524 267 191
18 18 out of 19 19 524 286 20
1 1 out of 20 20 20 1 048 575
2 2 out of 20 190 210 1 048 555
3 3 out of 20 1 140 1 350 1 048 365
4 4 out of 20 4 845 6 195 1 047 225
5 5 out of 20 15 504 21 699 1 042 380
20 6 6 out of 20 38 760 60 459 1 026 876
7 7 out of 20 77 520 137 979 988 116
8 8 out of 20 125 970 263 949 910 596
9 9 out of 20 167 960 431 909 784 626
10 10 out of 20 184 756 616 665 616 666
11 11 out of 20 167 960 784 625 431 910
 176

12 12 out of 20 125 970 910 595 263 950

13 13 out of 20 77 520 988 115 137 980
14 14 out of 20 38 760 1 026 875 60 460
15 15 out of 20 15 504 1 042 379 21 700
16 16 out of 20 4 845 1 047 224 6 196
17 17 out of 20 1 140 1 048 364 1 351
18 18 out of 20 190 1 048 554 211
19 19 out of 20 20 1 048 574 21
1 1 out of 21 21 21 2 097 151
2 2 out of 21 210 231 2 097 130
3 3 out of 21 1 330 1 561 2 096 920
4 4 out of 21 5 985 7 546 2 095 590
5 5 out of 21 20 349 27 895 2 089 605
6 6 out of 21 54 264 82 159 2 069 256
7 7 out of 21 116 280 198 439 2 014 992
8 8 out of 21 203 490 401 929 1 898 712
9 9 out of 21 293 930 695 859 1 695 222
10 10 out of 21 352 716 1 048 575 1 401 292
21
11 11 out of 21 352 716 1 401 291 1 048 576
12 12 out of 21 293 930 1 695 221 695 860
13 13 out of 21 203 490 1 898 711 401 930
14 14 out of 21 116 280 2 014 991 198 440
15 15 out of 21 54 264 2 069 255 82 160
16 16 out of 21 20 349 2 089 604 27 896
17 17 out of 21 5 985 2 095 589 7 547
18 18 out of 21 1 330 2 096 919 1 562
19 19 out of 21 210 2 097 129 232
20 20 out of 21 21 2 097 150 22

Chapter seven

Product variety management assuming product

configuration conflicts
Vladimir Modrak, Slavomir Bednar
Technical University of Kosice, Slovakia
 177

ABSTRACT

An important part of product variety management is finding optimum variety extent. Usually, product

variety extent is, among other restrictions, limited by, e.g. production capabilities and/ or production

capacities.In this chapter, it is intended to show that product variety extent can be influenced also by

configurations conflicts.Making decisions about product configurations is part of standard product

architecture development activities. Producers need to deliver variants to address diversity to consumer

needs. However, when configuration conflicts occur, the question is how they will be perceived by

customers. Here, we propose analytical method to be used in decisions about elimination or retainment

of configuration conflicts. This method will be verified through realistic case where alternative product

design platforms are compared. The newly developed method can be employed to assist product

managers to independently assess competitive product variety platforms against each other.

7.1 Introduction

Once managers are in the early stage of product architecture design, they might decide about the most

suitable product component/module structure. Normally, marketing managers strive to maximize the

variety offer with aim to satisfy a wide range of customers knowing also that some incompatible

components can occur in possible product configurations. The problem is that they are not aware of the

number of infeasible product configurations. Moreover, it is not easy to identify them using calculation

methods. On every fall, relatively high number of such infeasible product configurations, as a rule,

negatively affects customer perception and buying behavior. On the other hand, if these infeasible

product configurations would be totally eliminated, it would have negative impact on the extent of

customization. Normally, extent of customization for products is perceived in a sense that the bigger

the product variety, the better, and vice-versa [1]. Therefore, required extent of customization pulled by

customers should not be ignored to remain competitive.

 178

However, high product variety is resulting in less flexibility and higher costs in the manufacturing

systems, and becoming serious problem. Then, product designers have to consider both the problems

and find optimal balance between them.

The main scope of this chapter is to explore the possibilities to solve this issue by changing a rate

between infeasible product configurations and all possible product configurations when restrictions are

omitted. In a simple way, numbers of product configurations are closely related to variety-induced

complexity. However, numbers of product configurations, both, viable as well as unviable are not

optimal indicators of variety-induced complexity used to solve this problem and to express the rate.

Therefore, instead of the numbers of product configurations, an entropy-based complexity metrics will

be used as a tool for decision-making in product variety management. Finally, in this chapter, decision-

making algorithm to solve issues related to optimal selection of product component platform will be

presented.

7.2 Entropy-based complexity metrics for product variety extent

7.2.1 Theoretical background

The aim of this sub-section is to analyserelation between infeasible product configurations and all

possible product configurations. The very first notion of complexity was outlined in the work of

Shannon [2], where information theory was originally introduced. Few years later, information became

a key complexity element for the description and analysis of the systems entropy. Definition of the

discrete-case entropy (Hd) has been defined by the probability Pi of the n-state occurrence, as follows

[2]:

𝐻𝑑 = ∑𝑛𝑖=1 𝑃𝑖 𝑙𝑜𝑔2 𝑃𝑖 . (7.1)

Differential information entropy of the probability density function p(x) for continuous signals (Hc) has

been expressed as:

∞
𝐻𝑐 = ∫−∞ 𝑝(𝑥)𝑙𝑜𝑔2 (𝑝(𝑥))𝑑𝑥. (7.2)
 179

Krus [3] adopted the design information entropy for multidimensional case (Hx) in the following form:

𝐻𝑥 = ∫−∞ 𝑝(𝑥)𝑙𝑜𝑔2 (𝑝(𝑥)𝑆)𝑑𝑥, (7.3)

where D is the design space where the particular design x is defined. S is the size of the design space

expressed as:

𝑆 = ∫𝐷 𝑥𝑑𝑥. (7.4)

In the case of general multivariable design space, its information entropyH can be expressed as follows:

𝑆
𝐻 = 𝑙𝑜𝑔2 𝑠 , (7.5)

where s is the region of uncertainty for the final design of validated system architecture.

According to Krus [3], each particular design x with regards to its design space has information entropy

Hx:

𝐻𝑥 = 𝑙𝑜𝑔2 𝑛𝑠 , (7.6)

where, ns is number of unique design alternatives (representing so called complete design space) that

are results of a combination of product options and Hx is denoted as Entropy of complete design space.

There are many real cases, in which some variants are impractical due to presence of constraint(s).

Then, Information entropy of constrained design space Hc can be enumerated as:

𝐻𝑐 = 𝑙𝑜𝑔2 𝑛𝑣 , (7.7)

where nvis a number of viable design alternatives.

Since higher number of all possible design variants has more positive impact on consumers than smaller,

constrained design space, the Entropy of constrained design space in terms of mass customization

environment should be maximized. In this sense, Entropy of constrained design space can be considered

as positive entropy.
 180

In this context, the same author proposed to express a quality of a modular design through the rest of

the design space that is outside the constrained design space by the term “waste” information entropy

of design space and to quantify it using the formula:

𝐻𝑊 = 𝐻𝑋 − 𝐻𝐶 . (7.8)

In line with the logic used for characteristics of Entropy of constrained design space, Waste entropy can

be considered as negative.

Once background of the Waste entropy is outlined, we may proceed towards its application.

In order to catch the effect of product design optimization by using the concept of negative entropy, we

firstly need to generate concurrent product design architectures that will be mutually benchmarked.

One way to do so is through a gradual execution of selected components from original product design

architecture. Subsequently, mutual relation between so called positive and negative entropy can be

purposefuly analyzed. For this objective, it is necessary to enumerate numbers of all possible product

configurations when restrictions are omitted, and all possible product configurations with component

restrictions. Such procedure is shown in the following section.

7.2.2Methodical approach to enumeration of product configurations and waste entropy

To show a practicability of this approach, a realistic case is provided to motivate practitioners to solve

possible similar problems. For this reason, an assembly model of personal computer adopted from Yang

and Dong [4] has been used to identify product configurations as seen in Figure 7.1.

PC

CD-unit HD MB CPU OS

6 options
SCSI IDE 486 586 OS1 OS2

[1:2] [0:2]
SCSI-disk SCSI-cont. MB1 MB2 MB3 PI P II
5 options

Mandatory Optional Alternative Or

FIGURE 7.1 Case of product structure with rules for constraint satisfaction problem
 181

In case of higher degree of customization, there may be reasons for restrictions or/and obligations

between two or more components. They may be of functional, design, connectivity or other nature.In

order to solve possible consequences of rectrictions and obligations, required relations or

incompatibilities between components might be specified for any product architecture. In our

experimental assembly case, four types of configuration rules may arise [4]:

• Require rule,

• Incompatible rule,

• Port-connection rule, and

• Resource balancing rule.

Model depicted in Figure 7.1 is a representation of MC assembly of personal computer consisting of

five basic modules: CD-drive (one option), HD-unit (six individual customer options), Motherboard

(MB) (three options as MB_1, MB_2 and MB_3), CPU (586_I, 586_II, 486) and Server Operating

system (OS) (OS_1 and OS_2). The case model has various customizable options depending on the

customer choice but with predefined restrictions in the form of rules related to incompatibility of

components defined as follows:

R#1 – Component CPU3 must not be in the same configuration with component MB1.

R#2 – Component MB_2 must not be in the same configuration with components CPU1 and CPU2.

R#3 – Component CPU3 must not be in the same configuration with component MB3.

R#4 – Component OS1 must not be in the same configuration with component MB1 and MB3.

R#5 – Component OS2 must not be in the same configuration with component MB2 and MB3.

R#6 – Components MB2 and MB3must not be in the same configuration with components HD4,

HD5 and HD6.

R#7 – Component OS2 must not be in the same configuration with components HD2and HD4.

7.2.2.1 Enumeration of product configurations with and without restrictions

 182

At the beginning, it is useful to transform the computer structure with constraints shown in Figure 7.1

into a simplified assembly graph depicted in Figure 7.2. Any such structure usually consists of number

of assembly stations – nodes. These can be identified within a multi-level network. In our case, two-

level network is sufficient to model final assembly operation of Personal computer. Additionally,

specific number of component alternatives can be identified at each node of tier t1.

FIGURE 7.2 Assembly graph of personal computer without component restrictions

HD unit is represented by six individual alternatives. Number of all possible component combinations

is seven but one of them is omitted, namely combination consisting of two SCSI-Controllers with single

SCSI disk, as the second controller in such hypothetical product is consideredas redundant.

On the bottom tier t0, all possible product configurations without restrictions can be identified for the

original product design platform (D0):

∑ 𝐶𝑜𝑛𝑓𝐷0 = 1 ∗ 6 ∗ 3 ∗ 3 ∗ 2 = 108.

Subsequently, it is necessary to determine the total number of configurations when restriction rules are

considered. For this purpose, incidence matrix with component restrictions R#1-7 is constructed (see

Table 7.1).

TABLE 7.1 Incidence Matrix for Component Restrictions R#1-7

Group 4 -
Group 1 - HD Group 2 - CPU Group 3 - MB
OS
HD1 HD2 HD3 HD4 HD5 HD6 CPU1 CPU2 CPU3 MB1 MB2 MB3 OS1 OS2
 183

HD1

HD2 ❼
Group 1 - HD HD3

HD4 ❻ ❻ ❼
HD5 ❻ ❻
HD6 ❻ ❻
CPU1 ❷
Group 2 - CPU

CPU2 ❷
CPU3 ❶ ❸
MB1 ❹
Group 3 - MB

MB2 ❺
MB3 ❹ ❺
Group 4 - OS

OS1

OS2

To enumerate number of restricted product configurations, the following procedure is proposed. In the

first step, let us select e.g. group of HD units. Then, we select arbitrary configuration from the group,

for example HD2, which is one of the six HD unit options. Afterwards, we may construct an incidence

sub-matrix for the HD2 option and group of CPU components. As there is no restriction, HD2 as option

can be combined with any CPU component (see Figure 7.3). Then, we need to create a three dimensional

matrix of relations between configurations HD2, group of CPU components and a group of Motherboard

components (Step 2). Four restrictions are identified and accordingly CPU components can be

combined with compatible MBs (see Figure 7.3). Finally, four dimensional matrix relations are

constructed. Then, it is possible to exactly determine number of restricted product configurations, where

HD2 is exclusively involved. Moreover, this procedure allows generating product component structure

of all identified product configurations, as can be seen in Figure 7.3.

 184

Step 1
HD2

PI P II 486

Step 2

HD2

PI P II 486

MB1 MB3 MB1 MB3 MB2

Step 3

HD2

PI P II 486

MB1 MB3 MB1 MB3 MB2

OS1

FIGURE 7.3 Proposed approach to transform incidence matrix into product configurations model

The sub-procedure depicted in the Figure 7.3 has to be repeated for the rest of components from the

group 1. Then, the sum of configurations for individual components of the group 1 determines all

possible restricted product configurations. The sum is 21, as can be seen from Figure 7.4.
 185

HD HD1 HD2 HD3

5 conf. 5 conf. 5 conf.

CPU PI P II 486 PI P II 486 PI P II 486

MB MB1 MB3 MB1 MB3 MB2 MB1 MB3 MB1 MB3 MB2 MB1 MB3 MB1 MB3 MB2

OS OS2 OS2 OS1 OS1 OS2 OS2 OS1

HD4 HD5 HD6

HD 2 conf. 2 conf. 2 conf.

CPU PI P II PI P II PI P II

MB MB1 MB1 MB1 MB1 MB1 MB1

OS2 OS2 OS2 OS2

OS

FIGURE 7.4 Model of 21 possible product configurations respecting the configuration rules

7.2.2.2Proposed procedure to reduce waste entropy

As mentioned in the introduction, the goal of the configuration conflict solutions in terms of mass

customization is to reduce number of infeasible configurations. One possible way to reach this goal is

changing the rate between infeasible product configurations and all possible product configurations

when restrictions are omitted [5]. This rate can be changed through execution of restricted components

from an original product design platform D0.

For this reason, new product design platform D1 can be obtained when e.g. one of motherboards, namely

Motherboad_2 is selected for execution. Configurations where MB_2 is present, are not counted and

therefore the total number of model configurations equals 72 without accepting the rules and restrictions

R#1-7 and the number was reached by the following multiplications:

∑ 𝐶𝑜𝑛𝑓𝐷1 = 1 ∗ 6 ∗ 2 ∗ 3 ∗ 2 = 72.

Then, applying the procedure proposed in the Figure 7.3, the number of available restricted product

configurations will decrease to 18, as enumerated by the following formula:

 186

∑ 𝑅𝑒𝑠_𝐶𝑜𝑛𝑓𝐷1 = 5 + 5 + 5 + 2 + 2 + 2 = 18.

To obtain another alternative product design structure D2 for benchmarking purposes, another

component CPU_3 has been eliminated. Then, the number of total model configurations is calculated

as follows:

∑ 𝐶𝑜𝑛𝑓𝐷2 = 1 ∗ 6 ∗ 2 ∗ 2 ∗ 2 = 48.

The number of restricted product configurations will remain at 18. Obtained numbers of configurations

with and without restrictions are summarily depicted in Table 7.2.

TABLE 7.2 Computational results of numbers of product configurations

Product Number of product configurations

design Without restrictions With restrictions
platforms ns (complete design space) nv (constrained design space)
D0 108 21
D1 72 18
D2 48 18

Subsequently, waste entropy and waste entropy rate for each of the design platforms D0-2 is calculated

in Table 7.3.

TABLE 7.3 Computational results of waste entropy for different numbers of product configurations

Platform D0 Platform D1 Platform D2

Indicator Before reduction of After withdrawal of After withdrawal of

components MB_2 MB_2, CPU486
Entropy of design Hx=log2ns [bit] Hx=log2ns [bit] Hx=log2ns [bit]
space ns=108 ns=72 ns=48
Hx=6,75 bits Hx=6,2 bits Hx=5,58 bits
Entropy of HC=log2nv [bit] HC=log2nv [bit] HC=log2nv [bit]
constrained design nv=21 nv=18 nv=18
space HC=4,39 bits HC=4,17 bits HC=4,17 bits

Waste entropy Hw=Hx-Hc=2,36 bits Hw=Hx-Hc=2,03 bits Hw=Hx-Hc=1,4 bit

Waste entropy Hw/Hx=34,9% Hw/Hx=32,7% Hw/Hx=25,3%
ratio
 187

Table 7.2 shows how waste entropy ratiohas changed by reducing the number of restricted components.

Both, the reductions from D0 to D1 and from D1 to D2 seem to be favorable in order to reduce waste

entropy. However, the entropy rate, according to previous experiments, decreased markedly at interest

of the number of feasible configurations. In such cases, decision-makers may have a dilemma on what

design platform is optimal from the customer’s perspective. Therefore, the following decision-making

tool to avoid the problem is proposed.

7.3 Decision-making algorithm

In this section, we describe a decision-making procedure to select optimal platform of product variants

by using mutual relations between waste entropy Hwand constrained design space Hc.

We start by taking so called draft design platform D0, representing an existing product design platform

generating both, feasible and unfeasible product configurations for customers, where ns0 presents a

number of unique product design configurations as a result of a combinations of product components

and nv0 is a number of feasible product design configurations.

Let us further assume that we remove single component from the platform D0, which is in conflict with

other component(s). Then, D0 can be transformed into a new state with 𝑛𝑠1 for all unique product design

configurations and 𝑛𝑣1 for feasible product configurations, denoted as platform D1.

If continued in such reduction of components, the design platform D1 is modified into D2. Obviously,

it is possible to continue inthe reduction of system components depending on specific conditions.

To compare exactly two arbitrary design platforms against each other, e.g. D0 and D1, the following two

measures can be proposed:

𝐻𝑤1
∆𝐻𝑤0,1 = | − 1|; (7.9)
𝐻𝑤0

𝐻𝑐1
∆𝐻𝑐0,1 = | − 1|; (7.10)
𝐻𝑐0
 188

where, ∆Hw0,1 is the difference between the waste entropies of the design platforms D1 and D0; and

∆Hc0,1 is the difference between the entropies of constrained design space of platforms D1 and D0.

Then, if ∆Hw0,1> ∆Hc0,1 => design platform D1 is more preferable for MC than D0. To compare between

three alternative design platforms, the following sub-procedure can be used. Let us suppose that design

platforms D1 and D2 are more preferable for MC than D0, based on criteria:

∆Hw0,1> ∆Hc0,1,

∆Hw0,2> ∆Hc0,1.

Then, one can select more preferable design platform between D1 and D2 using the three criteria:

I.If ∆Hw0,1 - ∆Hc0,1> ∆Hw0,2 - ∆Hc0,2 => design platform D1, is more suitable than D2.

II.If ∆Hw0,1 - ∆Hc0,1< ∆Hw0,2 - ∆Hc0,2 => design platform D2, is more suitable than D1.

III.If ∆Hw0,1 - ∆Hc0,1 = ∆Hw0,2 - ∆Hc0,2 => both design platforms D1, and D2 are equally preferable for

buyers.

Proposed procedure for selection of optimal design platform is graphically depicted in Figure 7.5 in the

form of decision-making algorithm.

 189

FIGURE 7.5 Procedure for the selection of optimal design platform

7.4 Practical case application

In order to proof the relevance of the proposed decision-making tool to select the most optimal product

design platform, the following realistic case of customized bycicle components is used (see Table 7.4).

The case application in this section is represented by restrictions between the two inter-operating

modules: front drivetrain,and crankset that can be found in every bicycle model. Restrictions in these

models will be denoted as “X” and represent an incompatibility of the two modules. The starting

platform D0 consists of 12 groups - nine for gears, and three for chain stay angle (CSA). Each of the

nine groups has specific number of alternative components to be combined with front drivetrain (FD),

e.g. gear 42-32-24T can be combined with six Front Cranksets (FC): M980, M780, M670, M610, M552,

M522. With this assignment of components, one could construct an incidence symmetric matrix with
 190

size n=57 by the same principle as was constructed matrix in section 7.2.2.1. In such case, the

symmetrical matrix would be needlessly complex. Therefore, it is advantageous to keep it in its original

form as a non-symmetric matrix consisting of 38 rows and 19 columns.

For the design platform D0, complete design space is determined by 𝑛𝑠0 = 722 product configurations,

and restricted design space expressed by 𝑛𝑣0 = 239 product configurations.

By using this matrix, it is possible to gradually remove selected component entries with restrictions

from the product platform D0 to obtain concurrent platforms.

In order to benchmark possible concurrent product platforms at once, gears 48-36-26T including 5 crank

sets (M610, T780, M670, T781, T671, T611, T551 and T521) have been selected for an execution into

the platform D1. This group of components was selected for the execution based on the criterion of the

highest density of restrictions. Subsequently, we obtain compatibility table, as seen in Annex 7.1, where

gears 48-36-26T are omitted.The number of rows in this table was reduced from 38 to 30.

Obtained platform D1 consists of 570drivetrain configurations(𝑛𝑠1 ), and 215restricted (viable) product

configurations⁡(𝑛𝑣1 ).

Afterwards, for determination of the platform D2, we proceed towards reduction of the gear type 44-32-

24T (including three crank sets T611, T551 and T521), as can be seen in Annex 7.2. The number of

rows in this table was reduced from 30 to 27.

Then, obtained platform D2is determinedby 𝑛𝑠2 = 513 drivetrain configurations and 𝑛𝑣2 =

206restricted (viable) product configurations.

Table 7.4 Module ompatibility platform D0 for gears and front drivetrain

Platform D0 Front drivetrain (FD)

For triple For double
CSA 66° - 69° 63° - 66° 66° - 69°
M981
M781
M671

M611
M981
M781
M671
M611

M986
M786
M676
M616
M986
M786
M676
M616
T781
T671
T611
-A-D
-A-D
-A

-A

-D

-D

-D
-D
-D
-D
-3
-3
-3
 191

crankset
Gears
Front

(FC)
M980 X X X X X X X X X X X
M780 X X X X X X X X X X X
42-32-24T

M670 X X X X X X X X X X X
M610 X X X X X X X X X X X
M552 X X X X X X X X X X X
M522 X X X X X X X X X X X
M782 X X X X X X X X X X X X X
M672 X X X X X X X X X X X X X
40-30-22T

M622 X X X X X X X X X X X X X
M612 X X X X X X X X X X X X X
M523 X X X X X X X X X X X X X
M610 X X X X X X X X X X X X X X X X
T780 X X X X X X X X X X X X X X X X
M670 X X X X X X X X X X X X X X X X
48-36-26T

T781 X X X X X X X X X X X X X X X X
T671 X X X X X X X X X X X X X X X X
T611 X X X X X X X X X X X X X X X X
T551 X X X X X X X X X X X X X X X X
T521 X X X X X X X X X X X X X X X X
T611 X X X X X X X X X X X X X X X X
44-32-24T

T551 X X X X X X X X X X X X X X X X
T521 X X X X X X X X X X X X X X X X
42-30T 44-30T

M985 X X X X X X X X X X X

M985 X X X X X X X X X X X

M985 X X X X X X X X X X X
M785 X X X X X X X X X X X
40-28T

M675 X X X X X X X X X X X
M625 X X X X X X X X X X X
M615 X X X X X X X X X X X
38-26T

M980 X X X X X X X X X X X
M785 X X X X X X X X X X X
 192

M675 X X X X X X X X X X X
M625 X X X X X X X X X X X
M615 X X X X X X X X X X X
M785 X X X X X X X X X X X
38-24T

M675 X X X X X X X X X X X

M625 X X X X X X X X X X X

M615 X X X X X X X X X X X

In order to provide the next alternative product platform D3 for the benchmarking study, two FDs M981

and M981-D have been eliminated due to the high number of restrictions related to the two components.

Then, we obtained platform D3 defined by 𝑛𝑠3 = 459 drivetrain configurations and 𝑛𝑣3 = 194restricted

product configurations (see Annex 7.3), while the number of columns decreased from 19 to 17.

Obtained numbers of drivetrain configurations withrelated values of waste entropy Hwand constrained

design space Hc (with and without restrictions) are summarily depicted in Table 7.5.

In the next step, the decision-making algorithm to determine suitable extent of product variety for

different platforms can be applied. Since algorithm in Figure 7.5 is dedicated for maximum three

alternative design platforms, an extension of this algorithm for maximum four design platforms has

been constructed (see Figure 7.6). After using the algorithm,as the most suitable alternative with respect

to the amount of waste entropy, Platform D3 is indicated.

 193

INSERT
ns0, ns1, ns2, ns3
nv0, nv1, nv2, nv3

Hx0=log2ns0 Hw0=Hx0 - Hc0

Hx1=log2ns1 Hw1=Hx1 - Hc1
Hx2=log2ns2 Hw2=Hx2 - Hc2
Hx3=log2ns3 Hw3=Hx3 - Hc3
Hc0=log2nv0
Hc1=log2nv1
Hc2=log2nv2
Hc3=log2nv3

Yes is No
ΔHw0,1>ΔHc0,1

is Yes
ΔHw0,2>ΔHc0,2 is Yes
ΔHw0,2>ΔHc0,2

No is Yes
ΔHw0,3>ΔHc0,3 No is Yes
is Yes ΔHw0,2-ΔHc0,2=ΔHw0,3-ΔHc0,3
ΔHw0,1-ΔHc0,1=ΔHw0,3-ΔHc0,3 is Yes
No ΔHw0,3>ΔHc0,3
is Yes
No ΔHw0,1-ΔHc0,1=ΔHw0,2-ΔHc0,2 No
is Yes
ΔHw0,1-ΔHc0,1=ΔHw0,2-ΔHc0,2 No
is No is Yes
Yes is ΔHw0,2-ΔHc0,2>ΔHw0,3-ΔHc0,3
ΔHw0,1-ΔHc0,1>ΔHw0,3-ΔHc0,3 ΔHw0,1-ΔHc0,1=ΔHw0,2- Yes
No ΔHc0,2=ΔHw0,3-ΔHc0,3
is D3(nv3,ns3)
No Yes
ΔHw0,1-ΔHc0,1>ΔHw0,2-ΔHc0,2 No
is Yes No
D1(nv1,ns1) ΔHw0,1-ΔHc0,1>ΔHw0,2-ΔHc0,2
D3(nv3,ns3) No D3(nv3,ns3) D2(nv2,ns2)
is Yes
ΔHw0,1-ΔHc0,1=ΔHw0,2-
D1(nv1,ns1) No ΔHc0,2>ΔHw0,3-ΔHc0,3 D2(nv2,ns2)
D3(nv3,ns3) is Yes
ΔHw0,2-ΔHc0,2=ΔHw0,3-ΔHc0,3 D3(nv3,ns3)
D1(nv1,ns1) No
D2(nv2,ns2)

No D1(nv1,ns1)
D1(nv1,ns1) D3(nv3,ns3)
D2(nv2,ns2)
D2(nv2,ns2)
is Yes
ΔHw0,2-ΔHc0,2>ΔHw0,3-ΔHc0,3 D1(nv1,ns1)
D2(nv2,ns2)
D3(nv3,ns3)
No
D0(nv0,ns0)

D3(nv3,ns3) D2(nv2,ns2)

D2(nv2,ns2)
D3(nv3,ns3)

is Yes
ΔHw0,1-ΔHc0,1=ΔHw0,3-ΔHc0,3

No

is
Yes
ΔHw0,1-ΔHc0,1>ΔHw0,3-ΔHc0,3

No

D3(nv3,ns3) D1(nv1,ns1)

D1(nv1,ns1)
D3(nv3,ns3)

STOP

FIGURE 7.6 Procedure for the selection of optimal design platform from the three available
 194

TABLE 7.5 Computational results of numbers of product configurations

Indicator Platform D0 Platform D1 Platform D2 Platform D3

Entropy of Hx=log2ns [bit] Hx=log2ns [bit] Hx=log2ns [bit] Hx=log2ns [bit]
design space ns=722 ns=570 ns=513 ns=459
Hx=9.50 bits Hx=9.15 bits Hx=9.00 bits Hx=8.84 bits

Entropy of HC=log2nv [bit] HC=log2nv [bit] HC=log2nv [bit] HC=log2nv [bit]

constrained nv=239 nv=215 nv=206 nv=194
design space HC=7.90 bits HC=7.75 bits HC=7.69 bits HC=7.60 bits

Waste Hw=Hx-Hc=1.59 bits Hw=Hx-Hc=1.41 bits Hw=Hx-Hc=1.32 bit Hw=Hx-Hc=1.24 bit

entropy

7.5 Concluding remarks

The proposed novel method can be employed to assist product managers to independently assess

competitive product variety platforms against each other and to evaluate their customization

characteristics quantitatively. As it was shown and proved on multiple cases, the above described

approach based on the empirical criteria leads to decision for optimal product platform. In other words,

a suitable product platform is understood as structure of product components (compulsory, voluntary

and stable) with minimum waste entropy and at the same time keeping so called positive entropy of

constrained design space (enumerated from feasible options) sufficiently high, i.e. options that still

satisfy a wide range of requirements.

Development of this method has been motivated by previous experiences that restricted options in terms

of MC are not perceived positively by individual users. In this context, authors, e.g. [6, 7, 8, 9] argue

that infeasible configurations might be hidden by using algorithm-based product configurators. Quelch

and Jocz [10] underlined the role of information technology to ensurethat buyers will not choose

incompatible options.However, it is evident that one of types of configurator engines is developed

especially for options that include also infeasible component combinations [11]. Thus, the problem

treated in this chapter opens new research perspectives as each different sector of mass customization
 195

requires a specific, effective approach to solve configuration conflict problems related to product

structures with restricted configuration options.

REFERENCES:

1. Bonev, M., Hvam, L., Clarkson, J., & Maier, A. (2015). Formal computer-aided product family

architecture design for mass customization. Computers in Industry, 74, 58-70.

2. Shannon CE. A mathematical theory of communication. Bell Syst Tech J 1948;27(3):379-423.

3. Krus, P. (2015). Design Space Configuration for Minimizing Design Information Entropy. In

ICoRD’15 Research into Design Across Boundaries Volume 1 (pp. 51-60). Springer India.

4. Yang, D. and Dong, M.: A constraint satisfaction approach to resolving product configuration

conflicts, Adv Eng Inform, Vol. 26, No. 3, pp. 592-602, 2012.

5. Modrak V, Krus P, Bednar S. Approaches to Product Variety Management Assuming Configuration

Conflict Problem. FME Transaction 2015;43:271-78.

6. Pitiot, Paul, Michel Aldanondo, and Elise Vareilles. "Concurrent product configuration and process

planning: Some optimization experimental results." Computers in Industry 65.4 (2014): 610-621.

7. P.T. Helo, Q.L. Xu, S.J. Kyllonen, R.J. Jiao, Integrated vehicle configuration system connecting the

domains of mass customization, Computers in Industry 61 (1) (2010) 44–52.

8. D. Mailharro, A classification and constraint-based framework for configuration, Artificial

Intelligence for Engineering Design, Analysis and Manufacturing 12 (4) (1998) 383–397.

9. M. Aldanondo, E. Vareilles, M. Djefel, Towards an association of product configuration with

production planning, International Journal of Mass Customisation 3 (4) (2010) 316–332.

10. Quelch, J. A., & Jocz, K. E. (2013). Greater good: How good marketing makes for better democracy.

Harvard Business Press.

11. Orsvärn, K., & Axling, T. (1999, April). The Tacton view of configuration tasks and engines. In

Workshop on Configuration, Sixteenth National Conference on Artificial Intelligence (AAAI-99).

 196

Annex 7.1 Reduced module compatibility platform D1

Platform D1 Front drivetrain (FD)

For triple For double
CSA 66° - 69° 63° - 66° 66° - 69°

M781-A-D
M671-A-D
M781-A
M671-A

M981-D

M611-D

M986-D
M786-D
M676-D
M616-D
T781-3
T671-3
T611-3
M981

M611

M986
M786
M676
M616
crankset
Gears
Front

(FC)

M980 X X X X X X X X X X X
M780 X X X X X X X X X X X
42-32-24T

M670 X X X X X X X X X X X
M610 X X X X X X X X X X X
M552 X X X X X X X X X X X
M522 X X X X X X X X X X X
M782 X X X X X X X X X X X X X
M672 X X X X X X X X X X X X X
40-30-22T

M622 X X X X X X X X X X X X X
M612 X X X X X X X X X X X X X
M523 X X X X X X X X X X X X X
T611 X X X X X X X X X X X X X X X X
42-30T 44-30T 44-32-24T

T551 X X X X X X X X X X X X X X X X
T521 X X X X X X X X X X X X X X X X

M985 X X X X X X X X X X X

M985 X X X X X X X X X X X

M985 X X X X X X X X X X X
M785 X X X X X X X X X X X
40-28T

M675 X X X X X X X X X X X
M625 X X X X X X X X X X X
M615 X X X X X X X X X X X
M980 X X X X X X X X X X X
M785 X X X X X X X X X X X
38-26T

M675 X X X X X X X X X X X
M625 X X X X X X X X X X X
M615 X X X X X X X X X X X
24T
38-

M785 X X X X X X X X X X X
 197

M675 X X X X X X X X X X X

M625 X X X X X X X X X X X

M615 X X X X X X X X X X X

Annex 7.2 Reduced module compatibility platform D2

Platform D2 Front drivetrain (FD)

For triple For double
CSA 66° - 69° 63° - 66° 66° - 69°

M781-A-D
M671-A-D
M781-A
M671-A

M981-D

M611-D

M986-D
M786-D
M676-D
M616-D
T781-3
T671-3
T611-3
M981

M611

M986
M786
M676
M616
crankset
Gears
Front

(FC)

M980 X X X X X X X X X X X
M780 X X X X X X X X X X X
42-32-24T

M670 X X X X X X X X X X X
M610 X X X X X X X X X X X
M552 X X X X X X X X X X X
M522 X X X X X X X X X X X
M782 X X X X X X X X X X X X X
M672 X X X X X X X X X X X X X
40-30-22T

M622 X X X X X X X X X X X X X
M612 X X X X X X X X X X X X X
M523 X X X X X X X X X X X X X
42-30T 44-30T

M985 X X X X X X X X X X X

M985 X X X X X X X X X X X

M985 X X X X X X X X X X X
M785 X X X X X X X X X X X
40-28T

M675 X X X X X X X X X X X
M625 X X X X X X X X X X X
M615 X X X X X X X X X X X
M980 X X X X X X X X X X X
38-26T

M785 X X X X X X X X X X X
M675 X X X X X X X X X X X
M625 X X X X X X X X X X X
 198

M615 X X X X X X X X X X X
M785 X X X X X X X X X X X
38-24T

M675 X X X X X X X X X X X

M625 X X X X X X X X X X X

M615 X X X X X X X X X X X

Annex 7.3 Reduced module compatibility platform D3

Platform D3 Front drivetrain (FD)

For triple For double
CSA 66° - 69° 63° - 66° 66° - 69°
M781-A-D
M671-A-D
M781-A
M671-A

M611-D

M986-D
M786-D
M676-D
M616-D
T781-3
T671-3
T611-3
M611

M986
M786
M676
M616
crankset
Gears
Front

(FC)

M980 X X X X X X X X X X X
M780 X X X X X X X X X X X
42-32-24T

M670 X X X X X X X X X X X
M610 X X X X X X X X X X X
M552 X X X X X X X X X X X
M522 X X X X X X X X X X X
M782 X X X X X X X X X X X
M672 X X X X X X X X X X X
40-30-22T

M622 X X X X X X X X X X X
M612 X X X X X X X X X X X
M523 X X X X X X X X X X X
42-30T 44-30T

M985 X X X X X X X X X

M985 X X X X X X X X X

M985 X X X X X X X X X
M785 X X X X X X X X X
40-28T

M675 X X X X X X X X X
M625 X X X X X X X X X
M615 X X X X X X X X X
38-26T

M980 X X X X X X X X X
M785 X X X X X X X X X
 199

M675 X X X X X X X X X
M625 X X X X X X X X X
M615 X X X X X X X X X
M785 X X X X X X X X X
38-24T

M675 X X X X X X X X X

M625 X X X X X X X X X

M615 X X X X X X X X X
 200

Section three

Management and Sustainability of Mass

Customization
 201

Chapter eight

Product Configuration for Order Acquisition and Fulfilment

Dario Antonelli, Giulia Bruno
Polytechnic University of Turin, Italy

ABSTRACT

A key issue for an effective application of ‘Mass Customization’ (MC) design principles is associated

with the order acquisition and the fulfillment process that becomes increasingly more complex with the

proliferation of distinct product variants. Nowadays the process cannot do without the assistance of a

Product Configurator System. The evolution of Product Configurator toward a tool able to support the

design of new product variants is a challenging problem, not fully solved in a large variety of production

sectors.

An underestimated issue in the implementation of a Product Configurator derives from the lack of a

vision encompassing the whole Product Lifecycle already during the early phase of product definition.

As a matter of fact the design of a new product variant could not leave out of consideration the technical

and economic feasibility of the product manufacturing, its impact on the Capacity Planning and the

Material Resource Planning of the whole industrial plant.

Present chapterdescribes the problem by means of the interactions with a meta-model of a generic

product family. The meta-model has the form of an ontology and would enable an interactive software,

namely the ProductConfigurator, to assist the customer during the definition of a customized product.

The meta-model should give the rules in order to build every possible product tree (the reciprocate of

the BOM) for every product variant in the family.

The PLM approach extends the product description to the process constraints and production

requirements in order to solve simultaneously product configuration and order fulfillment problems.
 202

8.1 Introduction

Configurators are software applications intended to help both the customer and the dealer in the stage

of order acquisition and fulfilment (Forza and Salvador 2002). They are largely applied in several

industrial sectors, like computer, automotive and clothing companies. Their objective is to help the

customers to configure a personalized version of the product and conversely to ensure a fast, complete

and reliable transmission of the order to the production. As a matter of fact, the growth of customizable

features of the products forced a far larger growth in the number and variety of product data which have

to be communicated from the customer to the company before the execution of the order. Therefore it

is helpful to use the support of an automatic system dealing with the number and the variety of product

features and their related data. The Product Configurator translates the customer choices in a coherent

bill of material (BOM). It controls the availability of every optional component of the product and

detects conflicting components, i.e. components which cannot be used / assembled together in the final

product.

Configurable Products can be definedas pre-designed products that (Tiihonenet al. 1996):

 Are adapted according to the requirements of the customer for each order.

 Consist of (almost) only pre-designed components.

 Have a pre-designed product structure.

 Are adapted by a routine, systematic product configuration process.

Each configurable product has to be described by a product configuration model that encompasses the

product trees of every individual product.A computer system supporting product configuration tasks, a

product configurator,must be able to utilize product configuration models.

A main distinction can be made between the application of the Product Configurator to help the

customer in selecting among several product alternatives, already existing in the company catalogue, or

the application as a support tool to design and develop a new variant on the basis of the customer

specifications (Figure 8.1). The second alternative, that can be called generative approach, is far more

challenging and difficult to implement in industrial sectors characterized by continuous innovation or

high product complexity. Nevertheless, it would be particularly useful in assisting the order acquisition
 203

for a large variety of industrial products that have better to be tailored on customer specifics, as

commercial vehicles, data servers, clothes, furniture (kitchens, libraries) or machine tools.

In the generative case, the Product Configurator has two tasks: firstly to support the customer in the

setup of a feasible product variant, by enforcing the application of the design constraints, secondly to

fulfil the order by assembly all the parts in a new product variant in order to generate the corresponding

Bill of Materials (BOM).

Customer requirements
New variant
Existing variant
Product Configurator
Product Configurator (feasibility analysis)

Product’s BOM Product

Detailed design
engineering

Order execution

FIGURE 8.1 Comparison of the two approaches to Product Configuration

There are two approachesto the design of a product family: top down and bottom up (Farrel 2003). In

the top down approach the company develops single products after having decided a set of common

features which are used in all the family products. In the bottom up approach, different products are

redesigned in order to include them in a family, increasing the number of common features among them.

The first approach gets the maximum advantage from the product platform (Muffatto and Moreno 1999,

Meyer and Dhaval 2002).

To achieve the scope of a full generative Product Configurator delivering reliable output without

significant human intervention there are two competing approaches: increase the intelligence of the

configuration software or simplify the complexity of the product by modularity.

The first technique refers to writing down an expert system made of a knowledge base containing all

the rules necessary to correctly assembly different items in order to generate a new product from a
 204

library of sub-assemblies. An inferential engine would inquiry the knowledge base to comply with the

customer requests and to check their feasibility. The first expert system XCON,introduced in Digital

Research, became a typical example of the successful application of expert systems Technology.XCON

was a good illustration of the complexity of knowledge maintenance. In 1989, its knowledge base had

more than 31,000 components and approximately 17,500 rules (Sabin 1998).

The second technique refers to the simplification of the assembly interfaces to a point which won’t

require any ‘intelligence’ to assembly the different items and the procedure will be implemented on

existing Configurator software. To do this it is necessary to redesign all the existing families of products

and to accept the inevitable shortcoming of a more expensive and less optimized product.

Nevertheless there is a third way which gained the majority of the choices since the beginning of the

century: to embed intelligence inside the product description. If the product tree contains not only the

hierarchy of the assembly but also the constraints to be followed and the producing techniques, it is

possible to analyse a new product request and to generate the corresponding specifications of the process

plan by a smart application of the production rules embedded in the product tree. This approach requires

an extra effort to the product designer because he has not only to design a number of parts belonging to

the product family but also to make explicit the constraints and incompatibilities that link each other

the various items that compose a product. The advantage is that it is not required a redesign of the

existing products, and it is possible to adapt the existing Product Configurator to a more structured set

of building rules.

This approach has seen a consistent evolution. It started from the building of product meta-trees, that

are a set of rules to build the tree of all the products belonging to the same family (Olsen et al. 1997).

Then it was proposed to adopt an object oriented representation of the product in order to exploit the

advantages of object oriented philosophy: inheritance, easy reusability of models, functional description

of the product components (Hedin et al. 1998). This approach naturally leaded to a representation of

the product tree in terms of UML Class diagram or, more generally, by modelling the classes and data

types in a representation independent by their implementation.

The final arrival point was the representation of the ontology of the product(Stark 2005).
 205

In order to clarify the rationale behind this evolving pattern it is useful to make a step backward and

describe in detail the meaning of BOM and its constitutive rules.

8.2 Product tree and BOM

Configuring a product implies the selection of items from a set of sub-components and

theirorganizationand structuring in order to build a functional assembly. The final product has to satisfy

a number of requirements that derive from the customer, from the enterprise’s dealer or from the

conceptual definition of a new product. Put it simply, configuration is an activity whose input is

determined by the customer requests specifying the peculiar variant of the product he is aiming to buy.

The output is the generation of a document containing the description of the product mapping the

specifications of the customer. This document is the Bill of Materials (BOM). A BOM is organized

hierarchically with a root indicating the product and a number of branches representing the

subassemblies that compose the product. From every branch, other sub-branches can depart, allowing

for many different detail levels. The description is made in terms of assembly, subassembly or

individual item. BOM has a graphic representation in the product tree:a graph describing ordered

networks of interconnected nodes. Visually it reproduces an "inverted tree".

For sake of discussion, let’s consider the example of Figure 8.2, that reportsthe product tree of a personal

computer.
 206

FIGURE 8.2 Product tree of a Computer

(https://en.wikipedia.org/wiki/Product_breakdown_structure)

If all of the nodes in the graph are sorted starting from the end node (which corresponds to the assembled

product), from top to bottom layer, it derives a representation of the product in terms of its

components.For each component at a given level, it is possible to associate simpler derivedcomponents,

at a lower level. Level 1 is referred to the final product; the level 2 is referred to the sub-assembled

components and so on down, until reaching a partbought as is from external providers or produced

internally from raw material. The BOM corresponding to the product tree of Figure 8.2 is reported in

Table 8.1.

TABLE 8.1 BOM corresponding to the product tree of Figure 8.2

Bill of Materials - product Computer

 207

level code description quantity

1 1-0-0 Computer 1

2 1-1-0 Mainunit 1

2 1-2-0 Monitor LED 28” Full HD 1

2 1-3-0 Mouse wirelessoptical 1

2 1-4-0 Keyboardwireless 1

3 1-1-1 Housing ATX 1

3 1-1-2 Hard disk SATA 1TB 2

3 1-1-3 Video card 2GB GDDR5 1

3 1-1-4 Motherboard 1

4 1-1-4-1 Motherboard plate ATX 1

4 1-1-4-2 CPU 64 bit 4GHz 1

4 1-1-4-3 RAM card 4GBytes 2

4 ... ... ...

The extension of the product description to the product family description requires a generalization step

that can be named product meta-tree (Antonelli and Villa, 2006).

Let us consider a family of products with a still undefined number of components. Each product

configuration is represented by the BOM. Every item inside the BOM is an object defined by a number

of geometric and functional attributes (states) and by the techniques used to produce/acquire/assembly

it (behaviors). Representation of the product family requires the design of a super-class that enclose all

items of the family. The super-class is equivalent to a product meta-tree, that is a product tree with only

the branch structure and the levels but without instancing the items corresponding to one specific

product. The connections among different items on different levels are described by means of logic

operators: AND, OR, XOR. It is called meta-tree because it represents all the possible product trees of
 208

a product family by giving the connection rules. In Figure 8.3 a sample family tree is represented

together with the corresponding functional tree which express the functions assigned to every

component.

C(0,1) F(0,1)

OR OR

C(1,3) F(1,3)
AND AND

C(1,1) C(1,2) F(1,1) F(1,2)

FIGURE 8.3 Exampleofcomponenttree and functiontree

As far as now it has not been defined a single product but a set of functional requisites matched with

corresponding physical components. The pairs of meta-trees give a thorough description of a family of

products in terms of the functionalities which can be performed by different product instances and the

contribution of every component to the product operations, as well as in terms of the assembly

configuration of each product. A product family is described in terms of a function vector, i.e. the list

of functions related respectively to the product family F(0,1) and to each instance j on a given detail

level I, i.e. F(i,j).

Therefore the function tree can be modeled by a connection matrix, i.e. an incidence matrix which

specifies both the existence of a connection between two nodes in the tree and the type of connection.

The connection matrix MF is generated by the application of the following rules:

 MF{F(i,j),F(n,m)} = 1 if F(i,j) is a sub-item of F(n,m);

 MF{F(i,j),F(n,m)} = AND if F(i,j) has to be joined with F(n,m);

 MF{F(i,j),F(n,m)} = OR if F(i,j) is alternative with F(n,m) ;

 MF{F(i,j),F(n,m)} = XOR if F(i,j) is excluding alternative with F(n,m).

 209

The connection matrix is asymmetric and it is composed of a lower triangular matrix with the product

structure and a upper triangular matrix with the connections.

Next step is an object oriented representation of the super-class of products by assigning to every

component one or more functions and its geometric and technical attributes (Figure 8.4). This is done

in STEP, the standard representation of product data using the EXPRESS formal language (STEP).

Now, let assume the customer to propose a new variant in terms of new function requirements and/or

new components, selected in the component library. The configuration of a new product variant is

decomposed in the three subsequent issues (Figure 8.4).

The first issue is the making of a product and function tree corresponding to the desired product variant.

This is equivalent to instancing the class of product trees corresponding to the product family. The

solution of the first problem is obtained by a thorough application of Object Orientedmethods: mapping

the functions of the new project onto the functions of the product family, masking class functions non

required by the customer, checking the compatibility of the corresponding interfaces. The first step is

made possible by the rule of inheritance inside an object class. As a matter of fact, if the customer ask

for a new function which substitute an existing one, all the functions which are in a lower hierarchical

level will be inherited by the new function.

The second issue is to define a draft process planning method in order to verify the feasibility. This

should be obtained by the transformation of the product tree in a corresponding manufacturing

sequence. The structure of the product tree is directly converted in a manufacturing tree by applying

concepts of “pattern recognition”.

The third issue is to define a schedule of the production process of the new product variant in order to

estimate its production lead time and the manufacturing cost. For every operation the production lead

times and costs are given, therefore it is possible to define the completion time of the entire product and

the productive cost.

 210

New productrequest I issue

New component New function

Componenttree Functiontree

II issue
Familydatabase
Matching New components

III issue

Processesdataba
Matching New operations
se

Technical /
EconomicFeasibility

FIGURE 8.4 Decomposition of the problem in three main issues

FIGURE 8.5 Class diagram of a component and application to the mixer example

As an example let’s consider a case study made by an electric mixer to be employed in the preparation

of different kinds of food. The high level component tree is provided in Figure 8.5 right, where a circle
 211

substitutes the AND a rhombus the OR, functions are in italics. The mixer is obtained by assembling

together the individual items with a choice between two socket variants: thread for slow but secure

connection, plug for a fast connect / disconnect operation. It is worth remembering that functions are

inherited among objects along the relationship chain. As an example, the cover inherit the supporting

function from its parent, the base of the mixer. Let’s now suppose that the customer need a universal

adapter for the same mixer. It is a common component, available in the company database, but it has

not been provided in this case. A conventional Product Configurator would discard the request, while

the generative one can generate a new product tree for this variant replacing the discarded function. The

component will inherit all the functions owned by the parent items, therefore its individual connection

matrix will be mapped onto the family connection matrix of Table 8.2 (the difference being in the fact

substitution of AND and OR with ‘1’ or ‘0’).

TABLE 8. 2 Individualconnection matrix forthe mixer withthe universal adapter

From \ Proc Cont Cut Sup See Join Work Ins Fast Link
To

Process

Contain 1 A A

Cut 1 A

Support 1

See 1 A

Join 1

Work 1 A

Insulate 1

Fasten

Std. link 1
 212

8.3 Approaches to model productconfigurators

There are many intersecting study fields involved in the conception of a Product Configurator. On the

side of the product design, the guidelines of mass customization have to be followed: modularity (Hsuan

1999), commonality (Agard and Tollenaere, 2002), postponement (Markham et al. 2001). All of these

concepts are known and overall accepted and have been employed together with object oriented design,

e.g. in the studies of (Grady and Liang 1998, Huang and Kusiak1998, Jiao 2003).

8.3.1 Mass customization

The modular conception aims at designing product elements with functional over-capabilities (they are

able to satisfy a redundant set of requirements), in such a way to be interchangeable: modules. The

commonality is the generalized use of the same module inside many different products. Modularity

depends on two product characteristics: similarity among physic and functional architectures,

minimization of the interaction among the components. Some researchers, such as He et al. (1998) or

Jiao and Tseng(1999), propose methods aimed at building modular product families. In particular, He

proposes a matrix decomposition of the product tree in order to highlight the interchangeable or

independent elements. Jiao proposes a method to develop architecture of product families by applying

three different points of view, functional, technical and physical, to the classification of product variant.

8.3.2 Object oriented design

In the object oriented (OO) approach, an object is a bundle of related variables and methods (Eden,

2002). Objects are a model of the real-world objects, with a state and a behavior. A class is a blueprint

or prototype that defines the variables and the methods common to all objects of a certain kind. Every

object is obtained by instancing a class. Classes can be defined in terms of other classes. Each lower

level class (subclass) inherits the state from the upper level class (superclass).
 213

8.3.3 UML class diagram

The class structures are graphically represented using the Unified Modeling Language (UML) class

diagram (Fowler andScott 2000). UML is widely adopted as a specification and modeling formalisms

in various fields. It evolved as a strong convergence amongst different, competing but semantically very

similar, other specification languages and formalisms that were independently developed by software

engineers. It is now standardized by the Object Management Group (OMG). Currently, UML is a de-

jure and de-facto standard for describing software artifacts, from the initial conception to the

specification, design, implementation, testing and deployment phases. UML (http://www.uml.org/) is

based on a set of several graphical diagram types (Figure 8.6), each with its specific semantics, that

allow expressing static or dynamic system behavior. The graphical notation is extensible through the

stereotyping mechanism (i.e., define a new semantics to an existing symbol by attaching a <<label>>

to it), and a corresponding textual representation, for information interchange between different tools

and automatic information processing by user defined procedures, is also standardized by the OMG

(XML Metadata Interchange (XMI)). Finally, a formally specified constraint language is defined for

UML models, which help us expressing domain-specific constraints and complex interdependencies.

FIGURE 8.6 Set of UML 2.1 standard diagrams

 214

The evolution of UML as a specification formalism specifically oriented to software artefacts is evident

from the strong support of object-oriented constructs (classes, properties, methods, inheritance,

interfaces …), particularly suited to the high degree of flexibility possessed by software systems.

However, thanks to the completeness of the formalism, to its widespread popularity, and to several tools

supporting it, UML was also successfully applied to non-software domains. For example, the

automotive industry developed Automotive UML (von der Beeck et al. 2003) to model the complex

interaction of active subsystem in a modern vehicle. Also, real-time systems can be modelled thanks to

an extension called Embedded UML (Martin et al. 2001), or manufacturing systems can be described

and simulated with UML representations (Barbarisi et al. 2004).

Figure 8.7 shows an example of an UML class diagram. Each class is drawn as a rectangle with three

sectionsinside. The name section at the top holds the object name. The attribute section in the middle

holds a list component’s features. The operation section at the bottom holds a list of processes to be

applied to that object. Classes can be connected through different kinds of relationships (generic

associations, aggregation, composition, generalization …) represented as lines with different kinds of

endpoint terminators (arrows, diamonds, …).

FIGURE 8.7 An example of the class diagram notation.

On the side of the product description methods, the information conveyed through the standard BOM

is not sufficient to drive the generation of new product alternatives. Product components can be

accurately described (Demartini 1998) by using the three layers PDES/STEP methodology (ISO

10303), including the reference model, the format object class and schema definition language
 215

EXPRESS, the file structure. Incidentally it can be noted that EXPRESS follows an object-oriented

paradigm.The adoption of UML diagrams inside STEP standard is also being discussed in

ISO/TC184/SC4. While STEP standard is the preferable choice for the communication of a detailed

product description,nonetheless its detail level is too high to allow an efficient implementation fora

generative Product Configurator. Therefore the solution is to force somewhat the structure of STEP

inside a low detailed description of the product.

8.3.4 Ontology representations

As described by Ardito et al. 2011, the idea of exploiting ontology-based product configurators brings

the advantage of giving customers more freedom in creating products that best fittheir desires. The

authors analysed a case study of a company producing pieces of furniture directly ordered by customers,

who look at the company catalogues and provide a sketch of each piece of furniture they want, which

may be composed by parts chosen from different items in the catalogues, and assembled together. In

this way, customers have much more freedom in designing their furniture, thanks to the use of an

ontology that models the possible composition of different parts in a whole piece.

The ontology was used for two scopes: (i) connect the objects from different suppliers’ catalogues, so

that customers can consider all of them as a unique catalogue, and (ii) describe the components of each

piece of furniture and their properties (e.g., colours, size, decorations, shapes, and materials). The

integration of heterogeneous information sources implies the design of a data integration system aimed

at dealing with data residing in several sources and at hiding to the user the source of the data s/he is

accessing and its structure. Moreover, the ontology provides the rules and constraints to be applied to

assemble various components in order to generate only those pieces of furniture that are considered by

the ontology.

8.3.5 Proposed approach

The proposed work extends the previous approaches that are only focused in the early stage of the

product lifecycle (i.e., product design), by proposing an ontology covering the whole product lifecycle,
 216

in order to assess the influence of the choices done during the product configurations on the later stages

of the product lifecycle, i.e., manufacturing, maintenance and recycling.

8.4 Product lifecycle

Product lifecycle (PLC) is the term used to describe the stages a product goes through its life. The

lifecycle is an intrinsic attribute of a physical product.

This section defines the concept of product lifecycle, provides a summary of the ontology that have

been developed in this domain and define a reference ontology to be used as a common model to be at

the basis of a product configurator.

8.4.1 Product lifecycle definition

Although a generally accepted product lifecycle definition, the product lifecycle is still not a very clear

and exact concept, because it depends from the perspective it is analysed. For example, from marketing

perspective, the product lifecycle covers the phases of introduction, growth, maturity, and decline

(Anderson and Zeithaml 1984). In this context, the evolution of the product is represented from the

manufacturer perspective (Stark 2005), thus the following main phases are considered.

 Design The first stage of the product lifecycle is the definition of its requirements based on

customer, company, market and regulatory bodies’ viewpoints. From this specification of the

products major technical parameters can be defined. This phase also involves the functional

analysis, the requirements allocation, the definition of product components and assembly

processes and the development specification.

 Manufacturing The manufacturing(or production) is the physical construction of the product

and product components,including the acceptance testing, the operational testing and the

evaluation assessment. Once the design of the product’s components is complete, the method

of manufacturing is defined. This includes CAD tasks such as tool design, creation of CNC

Machining instructions for the product’s parts as well as tools to manufacture those parts, using

integrated or separate CAM computer-aided manufacturing software. Once the manufacturing

method has been identified CPM comes into play. This involves computer-aided production
 217

engineering or production planning tools for carrying out factory, plant and facility layout and

production simulation. Once components are manufactured their geometrical form and size can

be checked against the original CAD data with the use of computer-aided inspection equipment

and software.

 MaintenanceThe maintenance regards the product operation in the user environment. It

includes the management of service information, i.e. providing customers and service engineers

with support information for repair and maintenance. This involves using tools such as

Maintenance, Repair and Operations Management (MRO) software. It also includes performing

routine actions which keep the device in working order or prevent trouble from arising.

 Recycling The recycling is the process to convert old products (waste) into new products to

prevent waste of potentially useful materials, reduce the consumption of fresh raw materials,

reduce energy usage, and reduce pollution.

8.4.2 Ontologies for product knowledge management

Ontologies have already been proposed for knowledge management in a number of papers (Jurisica and

Mylopoulos 2010). Since Product Lifecycle Management (PLM) is a complex organization of activities

for product and process engineering management, starting from conceptual design, detailed design,

engineering, production, and usage, including disposal and recycling, the main reasons for the

development of ontology models for PLM are the need for a clear understanding of the product lifecycle

phases (McKenzie-Veal at al. 2010) and the need of systems interoperability (Abdul-Ghafour 2012).

These reasons motivate the ontology-based data and knowledge representations to support the

collaborations among the actors operating along the product lifecycle, resulting in expected lower

efforts and shorter time to market.

In order to support the system integration, approaches dedicated to studying and developing semantic

data models with concepts, relation and their respective properties, have been introduced (Fiorentini et

al. 2007, Kwak and Yong 2008). Ontologies have also been developed for STEP (Wang et al. 2010). A

product design ontology that formalizes the functionality of shape processing methods in the design

workflow is defined by Catalano at al. (2009), while an ontology to manage both the product and the
 218

PLM is proposed by Matsokis and Kiritsis (2010)to automatically handling multiple data from multiple

physical products and to support system interoperability and data integration.

The application of ontologies in product development is also fundamental to allow the knowledge

sharing (Lutters et al. 2001,Imran and Young 2013). Lutters et al. (2001) developed an information

management system based on an ontological approach on design and engineering processes, while

Imran and Young (2013) showed the benefits of applying ontologies to support knowledge sharing in

PLM with a focus on manufacturing processes. By using a product ontology, Panetto et al. (2012)

introduced an approach to support interoperability in Product Data Management (PDM).

In Sim and Duffy (2003) an ontology for engineering design activities was described, even it was not

implemented in software. An approach for conflict mitigation in collaborative engineering using

ontologies implemented in a software module was presented by Slimani et al. (2006). Another approach,

also supported by software, was developed by Lu et al. (2009) to allow the management of engineering

drawings, in the context of ship-building. The work presented by Eckstein (2010) describes the SWOP

project, which focused on the application of ontologies and problem solving methods to develop

optimized combinations for products and production systems in order to solve customer-specific

problems.

8.4.3 Reference ontology for product lifecycle

According to Imran e Young 2013, the process to define an ontology starts from the identification of

the set of relevant concepts, then proceeds with the organization of concepts in a formal model

representing the ontology structure, and finally performs the implementation of the model in OWL.

This procedure was followed to define the reference PLC ontology.

The set of high level concepts needed to represent the product lifecycle knowledge are the following

(see Antonelli et al. 2014 and Bruno et al. 2014 for more details).

 Production item: either a product or a product component. Each product can be made of several

components, and the same component can be used by different products. Each component can be

in turn be composed by other components.

 219

 Characteristic: a material, a functional characteristic or a physical characteristic (e.g., height,

length, width, weight, etc) which refers to a production item.

 Operation: a manufacturing operation executed by one or more resources to produce a product.

 Resource: an entity that is involved in the execution of an activity. It can be of two kinds, Person

or Machine.

 Role: the role of a person denoting his/her skills in the company.

 Failure: a failure occurred to a product.

 Recycling procedure: the procedure to follow to recycle a component.

These concepts were organized in a formal model, shown in Figure 8.8, according to the UML class

diagram formalism (Fowler and Scott 2000). The rectangles identify the concepts involved in each

lifecycle phase.

The Production item class is linked to the Customer class to store the customer who ordered each

product. A production item is also associated to the Characteristic class to store the characteristic of the

item. A production item can be a Product or a Component. A product can be made of several

components. Each component can be in turn be composed by several other components.

A Product is associated with the operations needed to produce it. To keep trace of the resources involved

in the operations, the Resource class was linked to the Operation class. Also the other specialization of

the Resource class, i.e., the Machine class, is linked to the Operation class to store the machines used

in each activity. For each Person, it is know the Role he/she has in the company, and also the roles that

can execute each activity are stored.

For each Product is recorded the list of Failures reported by the users, and for each Component it is

known the Recycling procedure, if available.

Figure 8.9 shows the implementation of thePLC reference ontology in Protégé.

Once the model to represent a general product lifecycle is designed, it can be further specified depending

on the industrial domain of application. This process is explained in the following section.
 220

FIGURE 8.8 UML class diagram of the PLC reference ontology.

 221

FIGURE 8.9 Implementation of thePLC reference ontology in Protégé.

8.5 Use case

The use case regards a company (named RTT) which is a manufacturer of precision microwave, coaxial

cable assemblies, harnesses and looms in the electronics and telecommunication industry, mechanical

assembly, filters for mobile telephony, complete assembly of racks, junction box and electronic

equipment, especially for aerial, naval means, for civil and military use.

The use case revolves around a telecommunications filter unit, a variant of the RF filter family of units,

a technical drawing of which is depicted in Figure 8.10. A telecommunications filter is a product used

to send radio signals, such as those used in broadcasting radio, television and wireless communications.

The filter aspect of the unit is to filter out particular frequencies, transmitted or received via tuning of

the product.
 222

RF filter assembly and tuning can be complex processes. While production tends to follow serial lines

and the standard assembly steps for any given filter are provided it must be noted that many variants

within the family of filters and mechanical and electronic assemblies are concurrently on the line thus

increasing complexity of sequencing and scheduling along with tracking along the assembly line (Bruno

et al. 2015).

Having a generative product configurator to support RTT in addressing the requests of the customers

will be highly valuable. Furthermore, knowing in advance the impact of the chooses made by the clients

during the product design on the subsequent phases of the product lifecycle will help in organizing the

company activities.

FIGURE 8.10 A RF filter unit produced by the RTTcompany

The definition of filter lifecycle is managed closely with the client. The clients exploit their own

experience in the design and engineering phases, while relying on the execution capabilities of the

company for the filter assembly and tuning.

The design of the filter is done by an Italian enterprise, which then sends the order to assembly a specific

filter, identified by its code (e.g. E15R01). During the design phase, the files needed to perform the

assembly of the filter, i.e., the BOM, the CAD and the Assembly sequence for the filter and its
 223

components, together with the other files needed to tune and test the filters, are generated. By analyzing

these files, it is possible to derive the information to create the ontology classes of the Design and

Manufacturing phases of the lifecycle.

After the testing of filters, it is possible to report the failures causing filter malfunctions, which are

stored in text files. This information is used to derive the ontology concepts for the Maintenance phase.

By the analysing the documents generated through the filter lifecycle, the general concepts were

specified into the concepts related to the specific filter lifecycle.

Particularly, the Product component class represents the hierarchy of components that can be included

in a filter (e.g., Panel, Plate, Capacitor, Inductor, Resistor, Resonator, Fixing elements, etc..), the

Functional characteristic represents the characteristics required to produce a filter, divided in Electrical

specification (e.g., Power supply, Frequency range), Mechanical specification (e.g., DC connector,

Forward and Reverse RC connector) and Environmental specification (e.g., Operating temperature,

Relative humidity).

The Failure concept was specified in the three main types of failures occurring to a filter: Assembly

failure, Interference and Damaging.

The specification of these concepts is graphically represented in Figure 8.11, and the screenshot of the

Protégé implementation is shown in Figure 8.12.

 224

FIGURE 8.11 UML class diagram containing the specification of the basic concepts for RTT.
 225

FIGURE 8.12 Implementation of theRTT ontology in Protégé.

By exploiting the structure of information reported in the ontology, on the one hand the clients of RTT

will be able to directly design a TCL filter with the desired characteristics, and on the other hand RTT

will be able to track the effect of the filter characteristics required by the clients on the production and

the failures of the filters.

8.6 Conclusions

There is a large number of industrial sectors whose manufacture benefits from mass customization

strategy which are unable to support it with an automatic Product Configurator.These sectors are
 226

constrained by the requirement of a set of product alternatives far too large to be enumerable and

determined in advance.

The aim of this chapter is to describe the evolution of product configurator design in the last decades,

by highlighting the new trends in exploiting ontologies to define a common vocabulary and integrate

data from different sources.

The first evolution of Product Configurator has been the substitution of lists of product variants with a

meta-model of a generic product family usable at the industrial level. The meta-model enablesthe

Product Configurator to assist the customer during the definition of the product. The meta-model gives

the rules in order to build every possible product tree for every product variant in the family. Since the

end of the last century, concepts taken from Object Oriented Approachhave been applied to design the

meta-tree of the product family.

A further step has been the adoption of Ontology to define not only the part variants that constitute the

product but also the different kinds of relationships among the parts.

Presently, the product configurator has been integrated inside the PLM logic because there is the need

of adding information deriving from later stages of product lifecycle. Information can thus beexploited

in order to evaluate the impact of the choices made during the product design on the following stages

of product manufacturing and product maintenance.

An example of implementation of Ontology based representation of the product tree is shown in the

industrial scenario of the RF filter production.

REFERENCES

Abdul-Ghafour, S. (2012) Integration of product models by ontology development. In: IEEE IntConf

on Information Reuse and Integration, pp. 548 - 555.

Agard, B. and Tollenaere, M. (2002): Conception d’assemblages pour la customisation de masse,

Labora-toire GILCO, Grenoble, France, Mécanique & Industries, 3. pp 113–119.

Anderson, C. andZeithaml, C. (1984) Stage of the product life cycle, business strategy, and business

per-formance.Academy of Management Journal, 27 (1), pp. 5-24.

 227

Antonelli, D. and Villa, A. (2006)Object-oriented design of a generative product configurator.In: Intern.

Conference on the Modern Information Technology in the Innovation Processes of the Industrial

Enterprises (MITIP2006),pp. 529-534.

Antonelli, D., Bruno, G., Schwichtenberg, A. and Villa, A. (2014) Full exploitation of Product Lifecycle

Management by integrating static and dynamic viewpoints, Advances in Production Management

Systems, Competitive Manufacturing for Innovative Products and Services, Springer Berlin Heidelberg,

pp. 176-183.

Ardito, C., Barricelli, B.R., Buono, P., Costabile, M.F, Piccinino, A., Valtolina, S., Zhu, L. (2011). An

Ontology-Based Approach to Product Customization, Lecture Notes in Computer Science, 6654, pp.

92-106,

Barbarisi, O., Del Vecchio, C., Glielmo, L., Vasca, F. (2004):Why adopting UML to model hybrid

manufacturing systems? International Conference on Systems, Man and Cybernetics.

Bruno, G., Antonelli, D., Korf, R., Lentes, J. and Zimmermann, N. (2014) Exploitation of a semantic

platform to store and reuse PLM knowledge, Advances in Production Management Systems, IFIP

Advances in Information and Communication Technology, Springer Berlin Heidelberg, 438, pp. 59–

66.

Bruno, G., Antonelli, D., Villa, A. (2015). A reference ontology to support product lifecycle

management, Procedia CIRP, Elsevier, 33, pp. 41-46.

Catalano, C.E., Camossi, E., Ferrandes, R., Cheutet, V. andSevilmis, N. (2009) A product design

ontology for enhancing shape processing in design workflows. Journal of Intelligent Manufacturing,

20, pp. 553-567.

Demartini, C., Rivoira, S., Valenzano, A. (1998): Product data exchange using STEP, Proceedings of

the Tenth International IFIP WG 5.2/5.3.

Eckstein H. (2010) SWOP Project - Semantic Web-based Open engineering Platform,

http://www.swop-project.eu/swop-approach/final-public-report.

Eden, A. (2002): A theory of Object-Oriented design, Information Systems Frontier, 4(4), pp. 379-391.
 228

Farrell, R. S., Simpson, T. W. (2003): Product platform design to improve commonality in custom

products, J. Intelligent Manufacturing, 14/6: 541-556.

Fiorentini, X., Gambino, I., Liang, V.C., Foufou, S., Rachuri, S., Bock, C. and Mani, M. (2007) Towards

an Ontology for Open Assembly model. In: Int. Conf. Product Lifecycle Management, pp. 445-456.

Forza, C. andSalvador, F. (2002): Managing for variety in the order acquisition and fulfilment process:

The contribution of product configuration systems, International Journal of Production Economics, 76,

pp. 87–98.

Fowler, M. andScott, K. (2000) UML distilled, Addison-Wesley.

Grady, P. and Liang, W.-Y. (1998): An object oriented approach to design with modules, Computer

Integrated Manufacturing Systems, 11(4), pp. 267-283.

Imran, M. andYoung, B. (2013) The application of common logic based formal ontologies to assembly

knowledge sharing. Journal of Intelligent Manufacturing, pp. 139-158.

Huang, C.C. and Kusiak, A(1998): Modularity in design of products and systems, IEEE Transactions

on Systems, Man and Cybernetics, Part A : Systems and Humans 28: 66–77.

He, D., Kusiak, A. and Tseng, T. (1998): Delayed product differentiation: a design and manufacturing

perspective, Computer-Aided Design, 30(2), pp. 105-113.

Hedin, G., Ohlsson, L.and McKenna, J. (1998) Product configuration using object oriented grammars.

System Configuration Management. Springer Berlin Heidelberg, pp. 107-126.

Hsuan, J. (1999)Impacts of supplier-buyer relationships on modularization in new product

development. European J. Purchasing & Supply Management, 5, pp.197-209.

Jiao, J., Ma, Q. and Tseng, M.M.(2003): Towards high value-added products and services: mass

customization and beyond, Technovation, 23, pp. 809–821.

Jiao, J. and Tseng, M.M. (1999) A methodology of developing product family architecture for mass

customization, Journal of Intelligent Manufacturing, 10, pp.3–20.

Jurisica, I., Mylopoulos, J. and Yu, E. (2004) Ontologies for Knowledge Management: An Information

Systems Perspective. Knowledge and Information Systems, 6, pp. 380–401.

 229

Kwak, J.A.and Yong, H.S. (2008) An Approach to Ontology-Based Semantic Integration for PLM

Object. In: IEEE Int.Work. on Semantic Computing and Applications, pp. 19-26.

Markham T., Frohlich A. and Westbrook R. (2001): Arcs of integration: an international study of supply

chain strategies, Journal of Operations Management, 19, pp. 185–200.

Martin, G., Lavagno, L. and Louis-Guerin, J. (2001): Embedded UML: a merger of real-time UML and

co-design. In: Proceedings of the Ninth international Symposium on Hardware/Software Codesign.

Matsokis, A., KiritsisandD. (2010) An ontology-based approach for Product Lifecycle Management.

Computers in Industry, 61, pp. 787–797.

McKenzie-Veal, D., Hartman, N.W. andSpringer, J. (2010) Implementing Ontology-based Information

Sharing in Product Lifecycle Management. Academia.edu

Meyer, M. H. and Dhaval, D. (2002) Managing platform architectures and manufacturing processes for

nonassembled products, Journal of Product Innovation Management, 19(4), pp. 277-293.

Muffatto and Moreno (1999): Introducing a platform strategy in product development, Int. J. Production

Economics, 60: 145-153.

Lu, T., Guan, F., Gu, N. and Wang, F. (2008) Semantic classification and query of engineering

drawings in the shipbuilding industry. International Journal Production Research, 46(9), pp. 2471-

2483.

Lutters, D., Mentink, R.J., Van Houten, F. andKals, H. (2001) Workflow management based on

information management,CIRP Annals, 50 (1), pp. 309-312.

Olsen, K.A., Sætre, P. and Thorstenson, A. (1997) A procedure-oriented generic bill of

materials. Computers & industrial engineering 32(1), pp. 29-45.

Panetto, H., Dassisti, M. andTursi, A. (2012) ONTO-PDM: product-driven ONTOlogy for Product Data

Man-agement interoperability within manufacturing process environment. Advanced Engineering

Informatics archive, 26(2), pp. 334-348.

Sabin, D. and Rainer, W. (1998) Product configuration frameworks-a survey. IEEE Intelligent

Systems, 4, pp. 42-49.

 230

Sim, S.K. andDuffy, A.H.B. (2003) Towards an ontology of generic engineering design activities. In:

Res Eng Design, 14, pp. 200-223.

Slimani, K., Silva, C., Ferreira, D., Médini, L. andGhodous, P. (2006) Conflict mitigation in

collaborative design. International Journal of Production Research, 44(9), pp. 1681-1702.

Stark, J. (2005) Product lifecycle management: 21st century paradigm for product realization. Springer-

Verlag, London.

Tiihonen, J., Soininen, T., Männistö, T. and Sulonen, R. (1996) State of the practice in product

configuration–a survey of 10 cases in the finnish industry. Knowledge intensive CAD, pp. 95-114.

von der Beeck, M., Braun, P., Rappl, M., and Schröder, C. (2003): Automotive UML: a (meta) model-

based approach for systems development. In UML For Real: Design of Embedded Real-Time Systems,

L. Lavagno, G. Martin, and B. Selic, Eds. Kluwer Academic Publishers, Norwell, MA, 271-299.

Wang, Q., Peng, W. andYu, X. (2010) Ontology based geometry recognition for STEP. In: IEEE Inter-

national Symposium on Industrial Electronics, pp. 1686-1691.

 231

Chapter nine

Shoe Configurators: A Comparative Analysis of

Capabilitiesand Benefits
Enrico Sandrin1, Cipriano Forza1,Zoran Anisic2, Nikola Suzic1,2, Chiara Grosso1, Thomas Aicher3,
Alessio Trentin1,
1
Department of Management and Engineering, University of Padova, Italy
2
Department of Industrial Engineering and Management, Faculty of Technical Sciences, University of
Novi Sad, Serbia
3
Faculty of Economics and Management, Free University of Bozen-Bolzano, Italy

ABSTRACT

Mass customizers (MCs) increasinglysell their products on the web through web-based sales

configurators (WBSCs). This selling approach has proven to be beneficial to both MCs and their

customers because, on one hand, it facilitates the customization process and, on the other hand, it

provides a real-time preview of the customized product. However, selling through WBSCs is

challenging. Different WBSCs havedifferent capabilities and, consequently, customers perceive

different levels of benefits from both the configured products and the customization experience. The

present work performs an analysis of the state-of-the-art of WBSCs for shoes and comparesthem with

other fashion WBSCs in order to help companies and researchers to adoptordevelopinnovative

approaches toenhancing WBSCs.

9.1 Introduction

The vast majority of mass customizers (MCs) rely on web-based sales configurators (WBSCs) to sell

their products online (Fogliatto et al., 2012). This selling approach has proven to be beneficial to both

MCs (Heiskala et al., 2007; Forza and Salvador, 2008) and their customers (Grosso et al, 2014; Trentin

et al. 2014; Franke et al., 2010). However, selling through WBSCs is challenging, not only because it

is a new way of selling for many companies, but also because WBSCs are witnessing a number of
 232

continuous technical innovations. This challenge can be observed across different sectors, but it is

particularly evident in the fashion industry and, more precisely, in the footwear industry.The trend to

mass customize shoes is in line with the exponential increase in product variety of this product category.

For example, in the USA there has been an increase from 5 sport shoe models in 1970 to 285 models in

1998 to 3,371 models in 2012 (Aichner and Coletti, 2013).

The present work performs an analysis of the state-of-the-art of WBSCsfor shoes in order to help

companies and researchers to adopt or develop innovative solutionstosupport online product

customization. More specifically, it analyses WBSCs both forshoes and for other fashion-related

products. The analysis identifies improvement opportunities and best practices for shoe configurators

to augment the value perceived by customers from shoe personalization.

Evaluating product configurators is a complex task. In order to ensure that the presentresearch is

rigorous,the assessment is based on the web-based sales configurator capabilities proposed by Trentin

et al. (2013). The capabilities under consideration (user-friendly product-space description, focused

navigation, flexible navigation, benefit-cost communication, easy comparison) have been proposed to

reduce the difficulties a customer faces when he/she customizes a product (Trentin et al., 2013).

However, sales configurators can act not only as tools to reduce customer difficulties but also as means

to increase the benefits customers derive from customization. In that respect,the benefits identified by

Merle et al. (2010), that isutilitarian, uniqueness, self-expressiveness, hedonic, and creative

achievementbenefits are taken into account.

Based on the measures proposed and tested by Merle et al. (2010), Trentin et al. (2013, 2014), and

Grosso et al. (2014),both shoe configurators and configurators of other fashion products are evaluated.

The 68 web-based configurators were evaluated by a total of 98 users. Each configuratorwas evaluated,

on average, by fivedifferent users. The 333 different configuration experiences have been analyzed to

identify the strengths and weaknesses of WBSCs for shoes. Subsequently, the shoe configurators with

the highest average evaluations have been further analyzed to identify the solutions that those

configurators use to achieve each capability. Finally, a number of non-shoe configurators with high

capabilities have been analyzed to identify solutions that shoe configurators could adopt to improve

those capabilities on which they are lower.

 233

9.2Product configurator capabilities

Drawing upon prior research concerning sales configurators and the customer decision process, Trentin

et al. (2013) defined five capabilities that sales configurators should deploy in order to alleviate the risk

that offering more product variety and customization in an attempt to increase sales paradoxically

results in a loss of sales: (1)user-friendly product-space description, (2) focused navigation, (3) flexible

navigation, (4) benefit-cost communication, and (5) easy comparison capabilities.

9.2.1User-friendlyproduct-spacedescriptioncapability

User-friendly product-space description capability is the ability of a sales configurator to adapt the

description of a company's product-space to the individual characteristics of a potential customer as

well as to the situational characteristics of his/her use of the sales configurator (Trentin et al.,

2013;2014).The essence of this capability is captured, for example, by the following statements which

were used, with few modifications, as measures in Trentin et al. (2014):

 The system gives an adequate presentation of the choice options for when the user is in a hurry,

as well as when the user has enough time to go into the details

 The product features are adequately presented for the user who just wants to find out about

them, as well as for the user who wants to go into specific details

 The choice options are adequately presented for both the expert and inexpert user of the product

An example of a WBSC deploying this capabilityis on Volkswagen’s website

(www.volkswagen.co.uk/configurator).This configurator allows potential customers to customize Polo

utility cars and,for each available choice, enables users to opt for either a brief description or a detailed

one with more technical information, which is available by selecting the “More Info” button. In

addition, choices affecting the esthetics of the car are described using both text and product images,

which changeautomatically as the potential customer selects different options.

9.2.2 Focused navigation capability

 234

Focused navigation capability is the ability of a sales configurator to quickly focus a potential

customer's search on those solutions of a company's product-space that are most relevant to the

customer, such as those that are most likely to satisfy his/her idiosyncratic needs (Trentin et al., 2013,

2014). The essence of this capability is captured, for example, by the following statements which were

used as measures in Trentin et al. (2014, p. 702):

 The system made me immediately understand which way to go to find what I needed

 The system enabled me to quickly eliminate from further consideration everything that was not

interesting to me at all

 The system immediately led me to what was more interesting to me

 This system quickly leads the user to those solutions that best meet his/her requirements

An example of a WBSC with this capability is Lenovo’s laptop customization website

(www.lenovo.com). This configurator has a search engine, called laptop finder,which enables potential

customers to narrow their search to laptops with a specific color, processor, and/or any other

customizable feature of the product.

9.2.3 Flexible navigation capability

Flexible navigation capability is the ability of a sales configurator to let its users easily and quickly

modify a product configuration they have previously created or the one they are currently creating

(Trentin et al., 2013, 2014). The essence of this capability is captured, for example, by the following

statements which were used, with few modifications, as measures in Trentin et al. (2014):

 The system enables the user to change some of the choices the user has previously made during

the configuration process without having to start it over again

 With this system, it takes very little effort to modify the choices the userhaspreviously made

during the configuration process

 Once the user has completed the configuration process, this system enables the user to quickly

change any choice made during that process

 235

An example of a WBSC deploying thiscapability is on Converse’s website (www.converse.com/us).

This configurator allows potential customers to customize their sports shoes through a multi-step

configuration process, where each step corresponds to one customizable feature of the product. This

process is depicted by a progress bar with a box for each step. Potential customers can modify any

previously selected feature simply by clicking on the related box, without losing any other choice that

they have madebefore.

9.2.4 Benefit-cost communication capability

Benefit-cost communication capability is the ability of a sales configurator to effectively communicate

the consequences of the configuration choices made by a potential customer, both in terms of what

he/she would get and in terms of what he/she would give (Trentin et al., 2013, 2014). The essence of

this capability is captured, for example, by the following statements which were used as measures in

Trentin et al. (2014, p. 702):

 Thanks to this system, I understood how the various choice options influence the value that this

product has for me

 Thanks to this system, I realized the advantages and drawbacks of each of the options I had to

choose from

 This system made me exactly understand what value the product I was configuring had for me

An example of a WBSC with this capability is Dell’s laptop customization website (www.dell.com). At

each step of the configuration process, this site gives potential customers the possibility to click on the

“Help Me Choose” button, which opens up a page with a list of recommendations outliningthe

advantages of every single option. Furthermore, the site communicates the price variations that selecting

each of the available options would cause with respect to the price of the current configuration.

9.2.5 Easy comparison capability

Easy comparison capability is the ability of a sales configurator to support its users in comparing

product configurations they have previously created (Trentin et al., 2013, 2014).The essence of this
 236

capability is captured, for example, by the following statements which were used, with few

modifications, as measures in Trentin et al. (2014):

 The system enables easy comparison of product configurations previously created by the user

 The system lets the userto easily understand what previously created configurations have in

common

 The system enables side-by-side comparison of the details of previously saved configurations

 The system lets the user to easily understand the differences between previously created

configurations

An example of a WBSC that deploys this capability is on Nike’s website(www.nike.com).This

configurator enables potential customers to save configured shoes in their “Wish List” area and to

subsequentlyaccess them at any time. Once logged in, the site user finds all of his/her previously saved

shoes portrayed on the same webpage, so they can be easily compared.

9.3 Consumer-perceived benefits

As recently acknowledged in literature, mass customization involves not only improving compatibility

between product customization and a firm’s operational performance (Pine, 1993; Tu et al., 2001;

McCarthy, 2004), but also augmenting the customer’s perceived benefits with regard to both the

customized products and the customization experience (Schreier, 2006; Merle et al., 2010; Fogliatto et

al, 2012). Prior research on the consumer perceived value of a mass-customized product (e.g.,Merle et

al., 2010) explains that in addition to the well-researched utilitarian benefit, there are two benefitsa

consumer could derive from the possession of a mass-customized product, namely uniqueness and self-

expressiveness benefits. Mass-customization research has also identified two other benefits, namely a

creative-achievement benefit and a hedonic benefit, which a consumer derives not from the possession

of a mass-customized product, but from the experience of self-customizing such a product using a sales

configurator (e.g.,Merle et al., 2010).

9.3.1 Utilitarian benefit

 237

The utilitarian benefit is a benefit derived from the closeness of fit between objective product

characteristics (i.e., physical, aesthetic, functional characteristics) and individual preferences about the

product’s functional/instrumental characteristics (Merle et al., 2010). Therefore, the utilitarian benefit

deriving from a mass-customized product fulfillstheconsumer's functional or aesthetic needs related toa

self-designed product (Addis and Holbrook, 2001). To measure this benefit,Grosso et al. (2014, p. 86),

building on Merle et al. (2010), usedthe following three statements:

 This product is exactly what I had hoped for

 I was able tocreateaproduct that was the most adapted to what I was looking for

 I was able tocreate the product I really wanted to have

9.3.2 Uniqueness benefit

The uniqueness benefit of possessing a mass-customized product is defined as the benefit that a

consumer derives from the opportunity to assert his/her personal uniqueness by possessing acustomized

product (Merle et al., 2010). The uniqueness benefit is related to the symbolic meanings a person

attributes to objects as a result of social construction. Brewer’s optimal distinctiveness theory posits

that people have opposing motives: to fit in and to stand out from social groups (Brewer, 1991). Whereas

threats to one’s inclusionary status produce increased attempts to fit in and conform, threats to one’s

individuality produce attempts to demonstrate how different one is from the rest of the group.

Consequently, the uniqueness benefit deriving from a mass-customized product meets the individual

need to assert his/her own personality by differentiating him/her self from others. To measure this

benefit,Grosso et al. (2014, p. 86), building on Merle et al. (2010), usedthe following three statements:

 With this product, I will not look like everybody else

 With this program, I was able todesign a product that others will not have

 With this product, I have asmall element of differentiation compared to others

9.3.3 Self-expressiveness benefit

 238

The self-expressiveness benefit is the benefit that originates from the opportunity to possess a product

that is a reflection of the consumer’s self(Merle et al., 2010). This is in accordance with the self-

consistency motive underlying self-concept, where the term self-consistency denotes the tendency for

an individual to behave consistently with his/her view of him/herself. Like uniqueness, the self-

expressiveness benefit is related to the symbolic meanings a person attributes to objects as a result of

social construction. Possessions are often an extension of the self. As Belk (1988) stated, "people seek,

express, confirm, and ascertain a sense of being through what they have" (p. 146). The above statement

implicitly relates identity with consumption. Consumers deliberately acquire things and engage in

consumption practices to achieve a pre-conceived notion of their selves. Thus, a mass-customized

product will fulfillindividualneedsforself-consistency by offeringpossession of a product that is a

reflection of the consumer’s self. To measure this benefit, Grosso et al. (2014, p. 87), building on Merle

et al. (2010), usedthe following three statements:

 I was able tocreate a product that is just like me

 This product reflects exactly who I am

 This product is in my own image

9.3.4 Creative-achievement benefit

The creative-achievement benefit derives from the capacity of the mass-customization experience to

arouse pride of authorship. In general, pride is a positive emotion of self-reward that follows the

assessment of one’s competence in a situation that is, in some measure, challenging, such as an exam

or climbing a mountain. Pride of authorship, in particular, is the feeling of pride that an individual

experiences whenever he/she creates, or at least has a sense of being the creator of, an artifact that

constitutes positive feedback on his/her own competencies. To measure this benefit,Trentin et al. (2014,

p.703), building on Merle et al. (2010), usedthe following three statements:

 I felt really creative while configuring this product

 The company gave me a lot of freedom while creating this product

 By personalizing this product, I had the impression of creating something

 239

9.3.5 Hedonic benefit

Unlike the creative-achievement benefit, the hedonic benefit stems only from the characteristics of the

mass-customization experience and, therefore, may be enjoyed even though a potential customer does

not complete the configuration task. Hedonic benefit derives from the capacity of the mass-

customization experience to be intrinsically rewarding. There may be various reasons why a generic

activity can be an end in itself, thus implying the actor’s positive affect (enjoyment, contentment,

satisfaction, etc.). To measure this benefit,Trentin et al. (2014, p. 703), building on Merle et al. (2010),

usedthe following three statements:

 I found it fun to customize this product

 Customizing this product was a real pleasure

 Customizing this product was like a game

9.4 Method

To assess the state-of-the-art of web-based shoeconfigurators, a set of 333 customization experiences

from 98 users on a sample of 68 WBSCs was used. These experiences allowed assessment of both the

WBSC capabilities and the benefits presented in sections 9.2 and 9.3, respectively. This analytical

approach has been successfully adopted by Merle et al. (2010), Trentin et al. (2013, 2014), and Grosso

et al. (2014).

The sample of 68 WBSCsused in the present study includes both shoe configurators and other fashion-

related product configurators. Each configurator has been evaluated, on average, through

fivedifferentexperiences,each one performed by a different user. In each experience, one of the 98 users

browsed an assigned sales configurator website and configured a product according to his/her own

preferences. At the end of each experience,the user filled out a questionnaire developed to measure the

capabilities and benefits presented in the previous sections. In this questionnaire,the measures proposed

and tested by Merle et al. (2010), Trentin et al. (2013, 2014), and Grosso et al. (2014) were used. These

measures were already presented, with few modifications, in sections 9.2 and 9.3. Table 9.1 shows the
 240

product categories involved and, for each category, the number of different web-sales configurators and

the number of mass-customization experiences (observations).

TABLE 9.1 Sample of web-based sales configurators (WBSCs)

No. of web No. of

Product category
configurators observations
Men's and women’s shoes 5 32
Sneakers, Sports shoes 7 46
Total shoes 12 78
Suits and elegant men’s clothing, Women’s clothing, Wedding dresses 4 24
Wedding rings, Watches, Women’sjewelry 14 49
Bags, Messenger bags, Women’s handbags, Backpacks 12 54
Shirts 9 53
T-Shirts, Jeans, Shorts, etc. 19 75
Total other fashion products 57* 255
Total 68* 333
* This number is lower than the sum per category becauseone configurator sells products belonging to
more than one product category

9.5 Benchmarking: Capabilities

For each capability,the calculation was done of the average score within the 12 shoe configurators

(product categories of “men’s and women’s shoes” and “sneakers, sports shoes”) and the average score

within the 57 other fashionproduct configurators (all the other product categories except shoes). The

results are reported in the radar chart in Figure 9.1. This chart allows comparison of the various

capabilities of shoe configurators to identify the more and less advanced capabilities. It also allows

comparison of shoe and non-shoe fashion product configurators to assess whether shoe configurators

are more or less advanced in comparison to other fashion product configurators. The scale used in Figure

9.1 is the same as the response scale used in the questionnaire. When responding to the questionnaire,

respondents specified their level of agreement or disagreement on a 1 to 7 symmetric agree-disagree

scale, where 1 means “completely agree,” 4 means “neither agree nor disagree,” and 7 means

“completely disagree,” for the set of statements used to measure each capability. The final score foreach

capability is the simple mean of the capabilitystatement scores.

In absolute terms, shoe configurators feature the following levels of the five capabilities:

 High on flexible navigation, and focused navigation

 241

 Intermediate-high on user-friendly product-space description

 Intermediate-low on benefit-cost communication

 Low on easy comparison

Compared with non-shoe fashion product websites, the scores for WBSCs for shoeswere as follows:

 Higher for flexible navigation, focused navigation, user-friendly product-space descriptions,

and easy comparison

 Slightly lower for benefit-cost communication

USER-FRIENDLY
PRODUCT-SPACE
DESCRIPTION
7
6
5
4
FOCUSED
EASY COMPARISON 3
NAVIGATION
2
1

BENEFIT-COST FLEXIBLE
COMMUNICATION NAVIGATION

Shoes Fashion not shoes

FIGURE 9.1 Benchmarking of web-based sales configurator (WBSC)capabilities

9.6 Benchmarking: Benefits

For each benefit,the calculation was done ofthe average score within the 12 shoe configurators (product

categories of “men’s and women’s shoes” and “sneakers, sports shoes”) and the average score within

the 57 other fashionproduct configurators (all the other product categories except shoes). The results

are reported in the radar chart in Figure 9.2. This chart compares shoe WBSCs and other fashion WBSCs

on product-related and customization experience-related benefits. Prior research provides empirical

evidence that WBSC capabilities influence at least some of these benefits (Grosso et al., 2014; Trentin
 242

et al., 2014). Therefore, even though other factors, such as the characteristics of the company's solution

space, are likely to play a role in determining such benefits, it is meaningful to compare shoe WBSCs

and other fashion WBSCs on these benefitsaswell.The scale used in Figure 9.2 is the same as the

response scale used in the questionnaire. When responding to the questionnaire, respondents specified

their level of agreement or disagreement on a 1 to 5 symmetric agree-disagree scale, where 1 means

“completely agree,” 3 means “neither agree nor disagree,” and 5 means “completely disagree,” for the

set of statements used to measure each benefit. The final score for eachbenefitis the simple mean of the

scores for the benefits statements on the questionnaire.

UTILITARIAN
5

3
HEDONIC UNIQUENESS
2

CREATIVE- SELF-
ACHIEVEMENT EXPRESSIVENESS
Shoes Fashion not shoes

FIGURE 9.2 Benchmarking of WBSCs on product-related and customization experience-related

benefits

In absolute terms,the evaluation of shoe WBSCson product-related and customization experience-

related benefits provides the following levels of the five benefits:

 Highlevels on hedonic, creative-achievement, and utilitarian benefits

 Intermediate-high levels on self-expressiveness and uniqueness benefits

Compared with non-shoe fashion product websites, web-based shoe configurators scored as follows:

 A little higher on hedonic, creative-achievement, and utilitarian benefits

 243

 Similar on self-expressiveness and uniqueness benefits

9.7 Best practices from shoe configurators

This section presents, for each capability, some of the solutions that the shoe configurators with the

highest average evaluations adopt to achieve that capability.

9.7.1 User-friendly product-space description capability

From the analyses conducted, some of the solutions that are associated with higher user-friendly

product-space description capability are those implemented by Shoes of Prey and Adidas.

Shoes of Prey (www.shoesofprey.com)

The Shoes of Prey website provides a large numberof color and material options that allow a user to

personalize many details of the shoe. Foreach element, anexplanatoryimageisprovided (for example,

the material, the style, or the type of heel). In addition, by holdingthe mouse cursor over a certain option,

theuser is provided with a drop-down window with atextualdescription of the element that is

represented. The elements are fairly intuitive, so an imageis more than enough to understand what is

intended. However, a description of the material is still useful because small product pictures cannot

effectively provide this kind of information, leaving the user to guess. The site also provides a detailed

explanation that guides the user in navigating through the configurator and, in particular, in choosing

shoesizes. In addition, there is a step-by-step tutorial that guides the user the first time he/shestarts the

configuration of aproduct, explaining the main features of the process.

 244

FIGURE 9.3 The user-friendly product-space description capability in the Shoes of Prey WBSC

Adidas (www.adidas.com)

The Adidas site guides the customer step by step in the customization of a shoe, which can go very

deep. In particular, once the user choosesa basic model from which to start the configuration (which is

divided into “men's shoes,”“women,” or “kids”), the website describes the shoe performance in detail,

explains in what contexts the shoe ensures better performance, etc.

A detailed description of alternatives from which to choose is not provided because the choices are very

intuitive, relating mostly to the color of the customizable part, for which a representation of the available

color variations is sufficient.

Moreover, this configurator even provides a detailed table with shoesizes in the specific units of

measure fordifferent countries to facilitate the user’s decisionregarding what size to buy.

9.7.2 Focused navigation capability

Some of the solutions that are associated with higher focused navigation capability are those

implemented by Shoes of Prey and Reebok.

 245

Shoes of Prey (www.shoesofprey.com)

The Shoes of Prey women’s shoes configurator shows, from the beginning of its use, the various ways

users can navigate. Various ways tonavigateare specified clearly atthe top of the page and, from there,

theuser can select what he/shewants to do or see. In addition, the configurator menus offer the abilityto

immediately choose a basic shoe model to start the configuration process, in case the customer is already

clear abouthis/herneeds. Many alternatives are offered, including ankle bootshigh-heeled shoes,

medium-heeled shoes, low-heeled shoes, ballet-style shoes, sandals, open-toedshoes, closed-toed shoes,

etc.

Reebok (www.reebok.com)

The Reebok sneakers configurator, as well as the Adidas one, typically interfaces with a diverse

clientele and supportsusers in making their choices with mechanisms that allow simplification of the

search space of the offered product. For example, the search is simplified through the activation of

several customizable filters, making it possible to remove shoe models that do not reflect the search

criteria.Then the customer is directed towards models that might better reflect his/her needs.

Anexampleofthesetoolsisshown in Figure 9.4. Theconfiguratorallowstheuser to narrowtheinitial set

ofavailableshoes to configure by selectinga set ofoptions in the menu on theleft and

sortingtheavailableshoes by variouscriteria, suchas “top sellers” or “price.”By meansofthe menu on

theleft, theusercanselectshoecharacteristicsthroughvarious filter optionssuchas “gender” (forexample,

iftheshoeisformen, forwomen, or forkids) and thetype and useforwhichtheuserisconfiguringtheshoe

(forexample, iftheshoeisclassic, CrossFit, or forrunning). Subsequently,itispossible to

clickonthepreferred model fromthe set of shoe models that meet the selected criteria, as the starting

point for the configuration process.

 246

Moreover, the configurator directs the customization steps toward the final configuration, presenting

one attribute of the product at a time to the customer.

FIGURE 9.4 The focused navigation capability in the Reebok WBSC

9.7.3 Flexible navigation capability

Some of the solutions that are associated with higher flexible navigation capability are those

implemented by Shoes of Prey and Adidas.

 247

Shoes of Prey (www.shoesofprey.com)

Shoes of Prey is a very flexible website, since it is easy to change the choices the user makes during the

configuration of the product, and it is equally easy to change a configuration made earlier. Each choice

is independent of the others, and it is up to the user to decide how to proceed. For example,the user can

decide the order in which he/shedefines individual attributes and can change them if they are not in tune

with each other or ifthey are not to the user’s liking. The site permitsthe maximum freedom of choice

in configuration, although some materials cannot be used for certain parts of the shoe. The user can also

save his/herproduct configurations to resume them later, without prior registration on the website.

Adidas (www.adidas.com)

Adidas configurator, like the majority of shoe configurators, such as Reebok, Nike, or Converse, allows

the user to easily change the current configuration, returning to previous options through a navigation

menu or by selecting a feature directly on the displayed shoe (both highlighted in Figure 9.5), thus

changing the previous choices by directly selecting another option. There are no constraints between

options belonging to different characteristics. With the Adidas configurator, every potential consumer

is able to go back and fairly easily change his/hercustomization choices, without triggering changes in

the parameters that they do notwant to change.

 248

Flexible navigation Flexible navigation

through menu through model interaction

FIGURE 9.5 The flexible navigation capability in the Adidas WBSC

9.7.4 Benefit-cost communication capability

Some of the solutions that are associated with higher benefit-cost communication capability are those

implemented by Shoes of Prey and New Balance.

Shoes of Prey (www.shoesofprey.com)

The Shoes of Prey WBSC clearly and unequivocally presents the features related to each choice and

allows verification of all the costsfor each choice in addition to the base price. During the configuration

process, the total price is always displayed at the top right of the page and it changes asthe various

options are selected. The user can see the window that contains the price details by moving the cursor

over the price, as shown in Figure 9.6. An explanatory image is provided for each option and, if the
 249

image is not sufficiently clear, the user can holdthe mouse cursor over the option and textualinformation

appears on a drop-down menu.However, the additional costs associated with each alternative are not

shown before or during their selection.Instead, the user has to check the price from time to time during

the configuration process.

Other shoe configurators, such as Nike and Reebok, show the price composition details, specifying the

extra costs associated witheachselected feature near the virtual image of the customized shoe.

FIGURE 9.6 The benefit-cost communication capability in the Shoes of Prey WBSC

New Balance (www.newbalance.com)

New Balance’s configurator adopts a different approach for benefit-cost communication compared to

Shoes of Prey and other companies, such as Nike, Adidas, and Reebok. The latterfour configurators

inform the user about the final price and the extra costsassociatedwitheach option. Conversely, the price

of New Balance’s customized shoe is fixed regardless of the type of customization the customer

chooses,and it is communicated from the beginning of the configuration process. Therefore, the user

knows the price ofthe selected model of shoes independently from the type of customization. Some

users have perceived this characteristic as a good method to effectivelycommunicate benefits and costs,
 250

given that during the configuration process there were no consequences of their choices in terms of final

price.

Note that configuration is based mostly on choices concerning the color of the different parts of the

shoe, which is very intuitive and does not require additional clarification besides the image of the

selected color. Nevertheless, there is a section at the bottom of the page thatshows the brand history and

adescription of the particular model selected, with anexplanation of the characteristics and materials of

which product is composed.

9.7.5Easy comparison capability

Some of the solutions that are associated with higher easy comparison capability are those implemented

by Shoes of Prey and Nike.

Shoes of Prey (www.shoesofprey.com)

Shoes of Prey gives the user an opportunity to compare the configurations that have been created

previously by simply placing them in the cart, where they are saved and will remain even if the user

leaves the site and returns to it again weeks later. From there, theuser can compare products through

images of the respective configurations, the prices, and the materials of the product.

Alternatively, the user can create an account to store all of his/hercreations and to build apersonal

collection. Furthermore, the usercan add the products that he/shelikes the most to a “wish list,” which

he/shecan access and which gives an overview of all the products he/shehas selected. Moreover,the user

can compare product characteristics such as the model of shoe, the heel height, the type of texture, the

material etc.

Nike (www.nike.com)
 251

The Nike site offers the possibility ofcomparing the results of all the user’s configurations on a single

page, which, in this case, is constituted by the virtual shopping cart. Acomparison of the products is

made by placing different types of information about the different customizations of the productside by

side. The configurator shows a visual representation of the final results of the configuration process, a

textual summary of the configuration choices, and the final price of each product, allowing visual,

textual, and price comparisons (Figure 9.7).

Visual Textual Price

Comparison Comparison Comparison

FIGURE 9.7 The easy comparison capability on the Nike WBSC

 252

9.8. Best practices from non-shoe configurators

This section presents some of the solutions adopted by the non-shoe configurators with the highest

average evaluations to achieve the user-friendly product description capability, the benefit-cost

communication capability, and the easy comparison capability—the three capabilities in which web-

based shoeconfigurators are less advanced.

9.8.1 User-friendly product description capability

Audi (www.audi.com)

A solution that the developers of shoe WBSCs could consider to obtain the user-friendly product

description capability is that adopted by the Audiwebsite, which usesboth a 3D model of the car at the

center of the screen (which accuratelypresents the product) and detailed descriptions. For example, the

site provides adescription of the car’s characteristics (rim material, type of brakes, headlights, etc.), and

any constraints in the combinations of attributes. Moreover, explanatory photos and videos for the

various options that the user can choose from are available, so that even the novice user can gain an

idea of the featureand, thus, configure his/hercar in a conscious way. The 3D model of the interior and

exterior of the car changes in real-time as the customer selects various options.Furthermore, during the

configuration of the particular model, a detailed list of the features included in each version is provided.

Moja Mix (www.mojamix.com)

Another site that the developers of shoe WBSCs could consider to obtain this capability is Moja Mix,

which offers usersthe possibility to personalize breakfast cereals based on the customer’sown tastes and

nutritional needs. The Moja Mix website allows customers to select their ideal cereal mix composed of

a cereal base, dried fruit (cranberries, currants, mango, etc.), nuts and seeds (walnuts, hazelnuts,

almonds, etc.) and extras (chocolate chips, chocolate coffee beans, etc.).During the configuration

process, the user can choose to simply look at the name of each option and the associated image.If

he/she wishes to get more information,however, the user can click on the “more”button highlighted in

red and, thus, learn aboutthe product’s properties in terms of the nutrientintake associated with the

particular choice. These explanations are provided in language comprehensible even to the most

inexperienced user, putting the optimal diet choicesin understandable terms.Finally, by clicking on the
 253

“Nutrition Facts” button, the userwho is more experienced in the nutrition field can find a nutritional

table explaining in detail the supply of nutrients per serving.

9.8.2 Benefit-cost communication capability

Hemdwerk (www.hemdwerk.com)

TheHemdwerksite,which sells men's shirts,presentsaninteresting solution to achieve the benefit-cost

communicationcapability. A 3D model within the configurator effectivelycommunicates the

outcomesof the customization choices. The model provides a 3D representation of the shirton the right

side of the window, and it changes as the user selects various options. This model helps the user gain a

clear idea of how the shirt will look like,allowinghim/her to see the effects of his/her choices.

The configurator also lets the user know the price of the configured product, which is updated based on

the sum of the prices corresponding to the selected components/features. There are additional charges

for some options. For example, the addition of astandard pocket button does not involve additional

costs, while the addition of aright pocket incurs a small charge. For options that include a specific price,

the price is shown clearly beside the option. In this way, when setting up the product, the customer is

immediately aware of what features will most increasethe final price.

NFL Shop (www.nflshop.com)

In a similar way to Hemdwerk, the NFL Shop, a website where users can purchase customizable

sportswear, shows good communication of the prices associated with each of the choices available to

the customer. The user is free to modify the product that he/she has selected and a model of the product

shown at the top of the site is updated to reflect any changes. Also, the options that can be chosen for

the clothing productare at the bottom of the screen, and include various team logos, the range of colors

available, and more. All this is accompanied by the options’ price, which is written on each available

option.

An overall summary of the user’s choices isprogressivelyupdatedto the right of the product model,

showing all the cost items that make up the final pricein detail, so the user can control the price based

onthe product options.

Moja Mix (www.mojamix.com)

 254

Another site that the developers of shoe WBSCs could consider to achieve thebenefit-cost

communication capabilityisMoja Mix.

The communication of the ingredients characteristicsis showed for each available component of the

custom mix,both in terms of the cost of each component and in terms of properties such as its nutritional

content. This configurator provides significant support to the potential customer,communicating every

available alternative through macro-categories that includeboth additional costs and information about

the beneficial properties of certain foods. For example, the configurator explains the nutritional intake,

the detailsof the products, and whattheyarerecommended for, guiding the user in order to align the

choice to his/herneeds.

In addition to the mix of cereals that the user is composing,product nutrition information is summarized

in atable to assistcustomers who are less experienced in the field of nutrition. A running list of the

chosen ingredients along with the final cost is provided as the customer progresses through the menu.All

of this makes the experience nice because the user is able to fully understand the product itself and the

properties of the producthe/sheis creating, and he/sheis aware, moment by moment, of the final cost

that may incur.

9.8.3Easy comparison capability

Ferrari (www.ferrari.com)

An interesting way to implement the easy comparison capability has beenproposed by Ferrari, whose

website permits the user todo parallel configurations on selected cars. The user can very simply compare

two different configurations, which differ in the color of the body.This comparison is done directly on

the same screen by cutting the car image into two parts. With each of the two image parts being a

different color,the user can choose options to continue the configuration process.

Therefore, the user has the opportunity to compare different product options throughout the course of

the configuration and not only at theend, as happens on many other sites (such as those that have already

been analyzed). The user can immediately discard the combinations of attributes that do not reflect

his/her tastes and does not waste time completing separateconfigurations.

Timbuk2 (www.timbuk2.com)
 255

Another interesting solution for the comparison of different product alternatives is adopted on

theTimbuk2 website, where the user can choose different products simultaneously and compare them

with each other on a single page after selecting the product features that interest him/herthe most and

configuring more product variants. The comparison is made based on the technical characteristics and

features that differentiate each variant.

Furthermore, reviews from customers who have already bought and used the company's products are

available. In these customer reviews, the user can find answers about the quality and reliability of these

products, if he/she does not havethem already, and can clarify indecisions that may emerge during the

configuration process.

9.9 Conclusions

The analysis of web-based shoe configurators showed that, on average, shoe configurators are not

inferior to other fashion product configurators. However, a number of them could improve some

specific capabilities. In general, the capabilities that offer the greatest degreeforimprovement are “easy

comparison” and,toa lesser extent, “benefit-cost communication” and “user-friendly product-space

description.” The specific examples of best practices that have been identified in this research can help

improveshoe configurators and ultimately increase a potential customer’s purchase intention.Forthe less

advanced capabilities, examples of best practices from WBSCs of both fashion and non-fashion

products are provided. These improvements in WBSC capabilities could contribute to enhancing self-

expressiveness and uniqueness benefits, which currently obtain the lowest score among the benefits

perceived by customers as a result of their shoe configuration experiences.

Acknowledgements

We acknowledge the financial support of (a) the Italian Ministry of Education, University and Research

under the Italian Flagship Program “La Fabbrica del Futuro,” project MADE4FOOT (FdF-SP2-T1.1)

and (b) the University of Padova, project ID CPDA129273.

 256

REFERENCES

Addis, M. and Holbrook, M. B. (2001). On the conceptual link between mass customisation and

experiential consumption: An explosion of subjectivity. Journal of Consumer Behaviour, 1(1), 50–66.

Aichner, T. and Coletti, P. (2013). Customers' online shopping preferences in mass customization.

Journal of Direct, Data and Digital Marketing Practice, 15(1), 20–35.

Belk, R. W. (1988). Possessions and the extended self. Journal of Consumer Research, 15(2), 139–168.

Brewer, M. B. (1991). The social self: On being the same and different at the same time. Personality

and Social Psychology Bulletin, 17(5), 475–482.

Fogliatto, F. S., da Silveira, G. J. and Borenstein, D. (2012). The mass customization decade: An

updated review of the literature. International Journal of Production Economics, 138(1), 14–25.

Forza, C. and Salvador, F. (2008). Application support to product variety management. International

Journal of Production Research, 46(3), 817–836.

Franke, N., Schreier, M. and Kaiser, U. (2010). The "I designed it myself" effect in mass customization.

Management Science, 56(1), 125–140.

Grosso, C., Trentin, A. and Forza, C. (2014). Towards an understanding of how the capabilities

deployed by a Web-based sales configurator can increase the benefits of possessing a mass-customized

product, in Felfernig,A.,Forza, C. and Haag, A. (Eds.), Proceedings of the 16th International

Configuration Workshop, September 25-26, 2014,Novi Sad; Serbia, CEUR-WSVol. 1220: 81–88.

Heiskala, M., Tiihonen, J., Paloheimo, K.-S. and Soininen, T. (2007). Mass customization with

configurable products and configurators: A review of benefits and challenges,inBlecker,T. and

Friedrich, G. (Eds.), Mass customization information systems in business,Hershey, PA: IGI Global, 1–

32.

McCarthy, I.P.(2004). Special issue editorial: the what, why and how of mass customization.

Production Planning and Control, 15(4), 347–351.

Merle, A.,Chandon, J.-L., Roux, E. and Alizon, F. (2010). Perceived value of the mass-customized

product and mass customization experience for individual consumers. Production and Operations

Management, 19(5), 503–514.

 257

Pine II, B.J.,Mass Customization: The new frontier in business competition,MA: Harvard Business

School Press, Boston, 1993.

Schreier, M. (2006). The value increment of mass-customized products: An empirical assessment.

Journal of Consumer Behaviour, 5(4), 317–327.

Trentin, A., Perin, E. and Forza, C. (2011). Overcoming the customization-responsiveness squeeze by

using product configurators: Beyond anecdotal evidence. Computers in Industry, 62(3), 260–268.

Trentin, A., Perin, E. and Forza, C. (2012). Product configurator impact on product quality.

International Journal of Production Economics, 135(2), 850–859.

Trentin, A., Perin, E. and Forza, C. (2013). Sales configurator capabilities to avoid the product variety

paradox: Construct development and validation. Computers in Industry, 64(4), 436–447.

Trentin, A., Perin, E. and Forza, C. (2014). Increasing the consumer-perceived benefits of a mass-

customization experience through sales-configurator capabilities. Computers in Industry, 65(4), 693–

705.

Tu, Q., Vonderembse,M.A. and Ragu-Nathan,T.S.(2001).The impact of time-based manufacturing

practices on mass customization and value to customer. Journal of Operations Management, 19(2),201–

217.

Chapter ten

Empirical investigation on implications of configurator

applications for mass customization
LindaL. Zhang, Petri T. Helo
IESEG School of Management (LEM-CNRS), Lille-Paris, France
Networked Value Systems, University of Vaasa, Vaasa, Finland
 258

ABSTRACT

In pursuing mass customization, many companies have applied product configurators for configuring

a right amount of product variety. While studies have been reported to shed light on how product

configurator applications achieve time reduction and quality improvement in fulfilling mass

customization, investigations addressing the implications of product configurator applications for

companies’ business activities have not been seen. However, understanding the implications is very

important for companies because with such understanding, they can better plan actions and make

changes to embrace the application of product configurators, thus further reaping the benefits of mass

customization. Based on survey, this study investigates the implications of product configurator

applications. The results indicate (i) how product configurator applications affect companies’ business

activities, (ii) the difficulties in designing, developing, and using product configurators, and (iii) the

potential barriers preventing companies from effectively applying product configurators in the future.

The results show several improvement areas for companies to investigate in the future so as toachieve,

to the largest extend, the benefits of implementing product configurators.

10.1 Introduction

Since the publiation of Pine’s book (Pine, 1993), many companies have adopted mass customization as

a new manufacturing strategy to replace mass production, in hope of delivering diverse customized

products while utilizing the available manufacturing capabilities (Zhang, 2007). In pursuing mass

customization, companies have adopted product configurators, which are information systems, for

configuring a right amount of product variety (also called customer wanted variety in the literature)

(Zhang, 2014). In view of the benefits that they bring to companies, such asbetter managing product

variety, shortening the sales-delivery process, simplifying order acquisition and fulfillment activities

(Aldanondo et al., 2000; Haug et al., 2011; Forza & Salvador, 2002), product configurators have been

receiving continuous interests and investigations from both academia and industrial alike.
 259

Resulting from the countless efforts, many articles have been published to present solutions to diverse

configurator related issues. Many of these articles address configuration knowledge representation and

modeling,configuration modelling and solving, and methods and approaches for product configurator

design (e.g., Chen & Wang, 2009; Chu et al., 2005; Falkner et al., 2011; Felfernig, 2007; Felfernig et

al., 2001; Ong et al., 2006). The feature common to these articles is that the solutions are proposed from

a theoretical point of view and are validated using ad hoc industrial examples. In comparison, a

relatively small number of articles deals with the practical issues related to product configurator

applications. Among these articles, some empirically investigate how product configurators achieve

lead-time reduction and quality improvement (Trentin et al., 2011 &2012;Haug et al., 2011); some use

single cases to show (i) how product configurator development can be facilitated (Haug et al., 2010;

Hvam et al., 2003; Hvam & Ladeby, 2007), (ii) how product configurators contribute to variety

management (Forza & Salvador, 2002), and (iii) the suitable product configurator development

strategies (Haug et al., 2012).

In spite of the above efforts, to the best of our knowledge, limited investigations have been made to

study the implications of product configurator applications for companies’ business activities. Due to

the lack of studies, several important issues related to configurator applications are unclear. How

product configurator applications affect companies’ business activities? What are the objectives of

implemneting product configurators? Whether the configurators are designed, developed, and used in

line with the objectives? What are the difficulties in designing, developing, and using product

configurators in practice? What are the barriers potentially preventing companies from effectively

applying configurators in the future? It is very important to have a clear understanding of all such issues.

This is because such an understanding can help companies better design, develop, and apply product

configurators, which in turn can help reap the benefits of product configurator applications to the largest

extent.

In view of the aobve lack of investigations and the unclear practical issues pertaining to product

configurator applications, this study empirically addresses the implications of product configurator

applications for companies‘ business activities. Among all the issues mentioned above, it investigates

(i) how product configurator applications affect companies’ business activities, (ii) the difficulties in
 260

designing, developing, and using product configurators, and (iii) the potential barriers influencing the

effective application of product configurators in the future. Survey is used to collect data for clarifying

these three issues. With respect to how product configurator applications affect companies’ business

activities, the data results and analysis show the major tasks and users of product configurators, the

functional units reorganized conrresponidng to the adoption of product configurators, the changes to

business processes, to companies’ legacy systems, and to the number of employees, and performance

improvements thanks to the adoption of product configurators. Besides the above mentioned

difficulties, the potential barriers influencing the effective application of product configurators include

two new ones: the unclear customer requirements and the unsafe feeling of employees.Based on the

results, this study further highlight three areas for companies to investigate in the future, in hope of

realizing the optimal benefits from implementing product configurators by carrying out suitable

decisions and activities. They include IT capacity and capability enhancement, organizational redesign,

and Top-down support and company-wide engagement,

This chapter is organized as follows. The data collection and analysis methods are presented in section

10.2. Results and corresponding analysis are provided in section 10.3. Built upon the insights in section

10.3, we further provide in section 10.4 the possible changes and improvements that companies may

undertake for improving product configurator applications. Section 10.5 concludes the chapter by

discussing limitations and potential future research.

10.2 Research Methods: Data Collection and Analysis

As the purpose of this study is to investigate the implications of product configurator applications for

companies’ business activities, an empirical study was performed. In doing so, a questionnaire was

developed for collecting data. The collected data was computed and analyzed for revealing the

application implications and for identifying the related opportunities as well.

In designing the questionnaire, we included questions related to matters such asfunctions that product

configurators perform, the business process and IT system changes caused by product configurator

applications, the difficulties in implementing product configurator projects, the potential barriers

preventing the effective application of product configurators in the future, and the performance
 261

improvements resulting from product configurator applications. Besides these questions, there are also

general demographic questions such as position titles, company size, and industry type. (See the major

questions used in the survey in Appendix A.) By considering the explorative nature of this study,

nominal scales were used to present alternative choices for each question. Each respondent can select

more than one alternative for each question based on his or her own product configurator application

practice. In this regard, the respondents are not mutually exclusive with respect to alternatives in each

question.The alternatives were determined based on the literature and our experiences of working with

companies. To avoid missing potential alternatives, an “other” option was included for each question

for respondents to give details, in addition to selecting the given alternatives.

To verify the initial questionnaire with respect to the sufficiency and appropriateness of questions, we

pretested it in 5 Finnish companies, with which we have collaborations. In the pretest, 5 company

representatives filled up the questionnaire and provided comments. In addition, we made phone contacts

for clarifications of the comments and additional remarks. Based on the feedback, some questions were

revised, leading to the finalized questionnaire.

Based on the finalized questionnaire, the survey was conducted in collaboration with EMpanel Online

consulting company. After preselection, the questionnaire was sent to 305 companies, which had

balanced distributions with respect to the company size, industry, and time duration (in year) of product

configurator applications. We received 61 completed questionnaires and thus the response rate was

20%. The respondents were mainly IT managers or managers with sales IT responsibilities, and the

companies representedthe computer, telecommunication system, and industrial machinery industries.

In analyzing the data, we computed the total occurrence of each alternative, which was selected by the

respondents, and the corresponding percentage. In this regard, we analyzed the distribution of the

selected alternatives for each question. As the respondents selected more than one alternative for most

of the questions, each respondent was counted more than once in the computation of the percentage of

selected alternatives for each question.

10.3 Results and Analysis

 262

In accordance with the questionnaire, the collected data is analyzed with respect to (i) how the

application of product configurators affect companies’ business activities, (ii) the difficulties in

designing, developing, and using product configurators, and (iii) the barriers potentially influencing the

effective application of product configurators in the future.

10.3.1 Product Configurator Applications AffectingCompanies’ Business Activities

In studying how companies’ business activities are affected by the application of product configurators,

it is essential to understand the major tasks that product configurators perform. This is because these

tasks contribute to companies’ activities for designing, producing, and delivering products. It is equally

important to understand the major users of product configurators. In accordance with the configurators’

tasks and users, there might be changes to companies’ business processes, functional units, IT systems,

the number of employees, and operational performance.

Major tasks of product configurators. The literature suggests, either directly or indirectly, that a product

configurator carries out diverse tasks. The survey result confirms this. As shown in Figure 10.1, the

tasks that configurators perform can be classified into three groups: sales order processing, product

documentation, and production documentation. Configurators in 62% of the respondents perform sales

order processing, which includes quotation preparation, sales order specification, and product

specification. Whileconfigurators in 26% of the respondents carry out product documentation: BOM

and drawing generation, in 11% of the respondents, they performproduction documentation, such as

routing generation, process plan generation. (Note:The word: respondent means the responding

company from this section onward.) While published articles provide anecdotal evidences, there is no

study presenting such a complete distribution of the major tasks that product configurators perform.

Based on our analysis method (see section 10.2), the respondents are not mutully exclusive in these

percentages. This is the same for the rest results.

 263

Major tasks of product configurators

Production
documentation
12%

Sales order
processing
Product 62%
documentation
26%

FIGURE 10.1 Major tasks of product configurators and corresponding percentages

Main users of product configurators. The main users of product configurators include end customers,

sales staff, production planners, product designers, as shown in Figure 10.2. Also provided in Figure

10.2 are the percentages of respondents that selected the corresponding users. Sale staff includes both

sale agents and internal sales staff; product designers include product designers and product developers.

Being consistent with the major tasks that configurators perform, product configurators are used by end

customers in 17% of the respondents, sales staff in 44% of the respondents, designers in 25% of the

respondents, and production planners in 14% of the respondents. Similarly, there is no study presenting

such a complete distribution of the main users of product configurators.

Main users of product configurators

Planners
14% End customers
17%

Product
designers
25%
Sales staff
44%
 264

FIGURE 10.2 Main users of product configurators and corresponding percentages

Functional units reorganized. Figure 10.3 shows the result with respect to reorganized functional units

caused by product configurator applications. The percentages in the figure represent the percentges of

respondents that selected each of the given alternatives, including (i) no change made to functional

units, (ii) sales unit is affected and reorganized, (iii) product design and development unit is affected

and reorganized, and (iv) production unit is affected and reorganized. As product configurators take

over tasks, which are performed previously in different functional units, their applications may bring

many changes to the organization of these units. According to the survey result, both the sales and

product design units in33% of the respondents are affected, thus reorganized; the production unit in

18% of the respondents is reorganized (Figure 10.3). It is interesting to see that product configurator

applications do not bring any changes to the functional units in 16% of the respondents. Our experiences

of working with a number of companies show that this finding is rational. This is especially true in the

situations, where the end customers are the main users of product configurators. In this regard, 16% is

not a surprising figure by considering the result above: end customers are the users in 17% of the

respondents. However, this connection needs to be verified in future studies.

Functional units reorganized

Production unit No changes

18% 16%

Sales unit
33%

Prod. design &

development
unit
33%

FIGURE 10.3 Function unites reorganized and corresponding percentages

 265

Business process changes. As product configurators automatically perform many activities, which are

carried out manually in the past, their applications might incur business process changes as well. Our

survey result confirms this. In the survey, many respondents indicate that their business processes have

been changed. These changes include (i) the original manual quotation preparation is done

automatically by configurators in 24% of the respondents, (ii) specifiying product functions is done

automatically by configurators, (iii) product technical details, e.g., BOMs, drawings, are generated

automatically by configurators, (iv) manufacturing documents are generated automatically. The

percentages of the respondents where the above changes occur are 24%, 27%, 21%, and 15%,

respectively, as shown in Figure 10.4. While most respondents experience business process changes,

13% of the respondents indicate that product configurator applications do not bring changes to their

business processes (see Figure 10.4). This seems consistent with what we have found above: 16% of

respondents point out that product configurator applications do not cause changes to the functional

units. When there are no changes to companies’ functional units, there may not be the changes to the

business processes.

Business processes changed

Production No changes
documents 12%
generation
23%
Quotation
preparation
22%

Product technical
details generation
19%
Product function
specification
24%

FIGURE 10.4 Business process changes and corresponding percentages

Changes to companies’ legacy systems. In performing tasks, product configurators interact with

companies’ other IT systems for receiving inputs and/or sending outputs. Consequently, product

configurator applications may cause changes to companies’ legacy systems. The result in Figure 10.5
 266

confirms this. As shown, the changes include (i) design systems are modified to be linked with

configurators, (ii) production planning and control systems are modified to be linked with product

configurators, (iii) material requirements planning systems are modified to be linked with product

configurators, and (iv) accounting systems are modified to be linked with product configurators. The

percentages of the respondents that experience each of these changes are 32%, 26%, 20%, and 11%,

respectively, as shown in Figure 10.5. At last, 11% of the respondents indicate that there are no changes

to their existing legacy systems (see Figure 10.5). When product configurators are built-in modules of

ERP systems, it might be possible that companies do not need to modify their legacy systems. Further

studies are needed to investigate this issue.

Legacy systems changed

Accounting No changes
systems 11%
11%

Material req.
planning systems
20%

Design systems
32%

Production
planning &
control systems
26%

FIGURE 10.5 Changes to legacy systems and corresponding percentages

Changes to the number of employees. As product configurators perform automatically many activities,

which are performed manually in the past, intuitively, product configurator applications should reduce

the number of full time employees. However, it is surprising to see that 67% of the respondents hire

full time employees, whereas only 5% indicate that they lay off employees, as shown in Figure 10.6.

A share of 28% of the respondents claims that there are no changes to the number of their employees.

As indicated in the following results (subsections 10.3.2 and 10.3.3), companies do not have sufficient

good IT system designers and developers. In this regard, the application of product configurators may

lead to the recruitment of new employees.

 267

Changes to the number of employees

Decreased
employees from
layoff
5%
No changes
28%

Increased
employees from
hiring
67%

FIGURE 10.6 Changes to the number of employees and corresponding percentages

Performance improvements. The available literature reports anectodally diverse performance

improvements resulting from the application of product configurators. This study finds similar results

(see Table 10.1), thus supporting the literature. In Table 10.1, the first column indicates the performance

measures considered, whereas the first row provides the alternative improvement ranges (in

percentage). The values (in percentage) in the cells present the percentages of respondents that achieved

the improvements falling in the corresponding ranges for each performance measure. As shown, for all

the respondents, the improvements include (i) increased sales volume, (ii) increased correct sales orders,

(iii) reduced production rework, (iv) increased customer orders accepted, (v) reduced order processing

time, and (vi) reduced sales delivery time.

TABLE 10.1 Performance improvements acheived by the respondents

Alternative improvement
ranges Higher than
0-30 30-50 50-80
Performance 80
measures
Increased sales volume 18 43 33 6

Increased correct sales orders 8 46 27 19

Reduced production rework 30 37 25 8

 268

Increased customer orders

20 31 31 18
accepted

Reduced order processing time 28 25 25 22

Reduced sales delivery lead time 33 26 21 20

Note: all values are in percentage.

As shown in Table 10.1, for four performance measures, including increased sales volume, increased

correct sales orders, reduced production rework, and increased customer orders accepted, most

respondents achieve an improvement ranging from 30% to 50%, as indicated by their corresponding

percentages: 43%, 46%, 37%, and 31%. As seen from the table, it is difficult for most respondent to

achieve higher improvement for these measures. For example, there are only 6% of respondents that

increased their sales volume higher than 80%. For reduced order processing time and reduced sales

delivery lead time, the majority - 28% and 33% of the respondents, respectively - achieve improvements

by no higher than 30%. As there are more interactions among different functional units in the entire

cycle of delivering sales orders than in order processing, improvements in sales delivery lead time might

be lower than in order processing time. However, it is interesting to see that respondents achieve similar

improvements in order processing time and sales delivery lead time. Further investigations should be

carried out to dig deeper this issue.

10.3.2 Product Configurator Design, Development, and Usage Difficulties

The resultsshow that most respondents experience difficulties in designing, developing, and using

product configurators. 50% of the respondents indicate that it is rather difficult for them to design

product configurators. The two major reasons are (i) the lack of IT system designers (in 50% of the

respondents) and (ii) a problem where IT system designers and product designers cannot communicate

well (in 45% of the respondents). With our experiences of working with companies and based on the

literature, these results are understandable. Manufacturing companies normally hire IT engineers for

maintaining systems in support of their core business activities: design and production. In this regard,

the IT engineers may not possess sufficient skills and capabilities for designing product configurators.
 269

The early literature points out that due to the differences in communication languages, configurator

designers and product experts have difficulties in making effective communications (Haug et al., 2010).

The results found in this study are consistent with the literature.

Similarly, most respondents have difficulties in developing product configurators. The biggest

challenge for companies to develop product configurators is the high complexity of product

configurators, as supported by 52% of the respondents. The other two main difficulties include (i) the

lack of good IT system developers (in 24% of the respondents) and (ii) the continuous evolution of

products and the resulting high product complexity (in 24% of the respondents). While product

complexities do not appear as a main difficulty in designing configurators, they do cause difficulties in

developing product configurators. This is because in accordance with product complexities, the product

configurator design is complex too. It is understandable that complex product configurators are difficult

to develop.

In using product configurators, companies also have difficulties. These are caused by (i) un-user

friendly interfaces (in 44% of the respondents), (ii) the inefficient communications for getting required

inputs (in 31% of the respondents), (iii) the high complexity of product configurators (in 12.5% of the

respondents), and (iv) the lack of sufficient training (in 12.5% of the respondents). In processing

customer orders, configuring products, and generating product/manufacturing documents, configurators

require diverse inputs. These inputs originate from customers, sales staff, designers, etc. In many cases,

the input providers are from different offices or even different companies. This location dispersion may

hinder the effective communications for required inputs. Additionally, even in the situation where the

input providers are in the same location, due to, e.g., other tasks that they need to deal with, the input

providers may not be able to supply required inputs on-time. In our view, the other three difficulties are

all interconnected with one another. Firstly, complex product configurators may have many interrelated

modules and procedures. Secondy, more training is required to understand and use these modules and

procedures. However, companies are busy with dealing with daily operations activities and may not

give enough training time to users. As shown in practice, e.g., the fail of SAP project in Avon

(http://blogs.wsj.com/cio/2013/12/11/avons-failed-sap-implementation-reflects-rise-of-usability/),

caused by design difficulties, complex IT systems tend to have un-user friendly interfaces. In this
 270

regard, the un-user friendly interfaces may be, at least, partially related to product configurator

complexities.

10.3.3 Potential Barriers influencing Product Configurators’ Effective Applications

One question in the questionnaire asked respondents for the barriers, which may potentially prevent

them from effectively applying product configurators in the future. The results here are consistent with

these discussed earlier. The earlier results show that companies have difficulties in designing and

developing product configurators because of the lack of technical IT staff. Similarly, the lack of IT staff

also appears to be a major barrier for companies to effectively use product configurators in the future.

The fact that products keep evolving is one of the three difficulties in using product configurators is

also acknowledged as one barrier for future application. Two additional barriers, including the unclear

customer requirements and the unsafe feeling of employees, are brought up. Due to the linguistic

origins, customer requirements are often imprecise and ambiguous (Jiao & Zhang, 2005), In addition,

they often conflict with one another or are subject to changes (Kristianto et al., 2015). As product

configurators need articulated customer requirements, the ambiguous and conflicting requirements will

negatively affect the effective application of product configurators. As a matter of fact, during the initial

questionnaire pretest, some of the 5 Finnish company representatives indicated this. As product

configurators execute activities, which are carried out earlier by the employees, the affected employees

perceive product configurators as a menace to their positions (Forza & Salvador, 2002). In this regard,

the unsafe feeling of employees may become an obstacle for the effective application of product

configurators in the future.

In summary, the largest barrier is the continuous evolution of products, as pointed out by 75% of the

respondents. The other barriers include (i) the lack of technical IT staff for maintaining the configurators

(seen by 47% of the respondents), (ii) the unclear customer requirements (perceived by 47% of the

respondents), and (iii) the unsafe feeling from the employees because of the possibilities of losing jobs

(agreed by 34% of the respondents).

10.4 Discussions
 271

Along with the benefits achieved, product configurator applications bring many additional requirements

and changes to companies’ existing way of doing business, as shown in this study. While the changes

and requirements may not be perceived beneficial, they open up opportunities for companies to improve

the new ways of doing business, which involves product configurators. Based on data analysis and

results, this study highlights three areas for investigation:(1) IT capacity and capability enhancement,

(2) organization redesign, and(3) top down support and company-wide engagement.

IT capacity and capability enhancement. Product configurators are basically IT systems. The optimal

design and development of these systems will bring many advantages to companies, such as the

configuration of optimal products, the cut-down of configuration time, the reduction of configuration

errors, the easy application, the reduction of training time, etc (Haug et al., 2012). Such design and

development demands sufficient system designers and developers with high skills and experiences.

However, the results indicate that many companies do not have sufficient good designers and

developers. In this regard, it will be beneficial to companies, especially these that design and develop

configurators in house, for having sufficient well-trained system designers and developers. These

designers and developers bring companies additional IT capacities and capabilities. Developing such

IT capacities and capabilities can be also justified by other issues. The fact that products keep evolving

necessitates continuous maintenance and upgrading to be performed (subsection 10.3.2). Caused by its

complexity, configurator maintenance and upgrading is not easy tasks and difficult to perform In

addition, if they are not well performed on-time, companies may delay product configuration,

production, and delivery. This may, in turn, cause companies to lose customers. In this regard,

sufficient, well-trained system designers and developers can also contribute to continuous maintenance

and upgrading of configurators and product models.

Organizational redesign. Product configurator applications bring many changes to companies existing

activities, processes, and functional units (subsections 10.3.1). While simply reorganizing the affected

units, as what the practice does (see Functional units reorganized in subsections 10.3.1), may to certain

degree facilitate product configurator applications, it is insufficient for companies to realize the full

benefits of product configurators (Salvador & Forza, 2004). In fact, the communication difficulties

(subsection 10.3.2) lend themselves to this point. In accordance with the tasks and functions that product
 272

configurators perform, companies should reorganize their business processes and structures by

reallocating the responsibilities of each individual employee and functional unit. The reorganization

should be performed such that each employee has a clear vision for his activities, tasks, and

responsibilities. The same applies for the functional units. Besides, information exchange protocol and

procedures need to be (re)designed such that communication difficulties in applying product

configurators can be eliminated. At last, as one of the potential barriers for effective configurator

application in the future lies in unclear customer requirements, some efforts in organization redesign

may be directed to the suitable tools, techniques, systems, etc and the related issues for obtaining clear

customer requirements or proper change management process.

Top-down support and company-wide engagement. As with the implementation of any new

technologies, the implementation of product configurators needs continuous support and commitment

from all levels, especially the top management level, in a company. The support and commitment is

very important for completing the necessary organization changes (see above) and for successfully

implementing product configurator projects. The literature shows that the lack of long-term

commitment is one of the main reasons for the failure of many technology implementation projects

(Bergey et al., 1999). As the employees including the middle management level have a tendency to

resist changes (Paper & Chang, 2005), regular encouragement and incentives from the top management

level are required to remove employees’ hostile attitude towards the application of product

configurators. Once the employees positively look at product configurator applications, they are willing

to accept and implement organization changes. Perceived by companies, the employees’ unsafe feeling

for losing jobs due to process automation is one of the important barriers potentially preventing

companies from effectively applying configurators in the future (subsection 10.3.3). To encourage the

employees and remove their unsafe feelings, the top management level should create more training

activities. With these training activities, the employees may master additional skills. They may also

involve employees in company’s important meetings, share with employees company’s daily or weekly

news and development, etc. All these supports may help employees regain their confidence and develop

correct attitudes towards configurator applications.

 273

10.5 Conclusion

In view of the contribution of product configurators to pursuing mass customization, this study

investigated the implications of product configurator applications for companies’ business activities.

The belief is that it is beneficial for companies to understand the difficulties and challenges before

embarking upon a product configurator project. As shown in the results, product configurator

applications bring many changes and difficulties along with performance improvements. The changes

together with the difficulties highlight a number of areas to be investigated if companies want to achieve

the optimal benefits from using product configurators. These improvement areas include (i)

development of IT capacity and capability for addressing product configurator design, development and

maintenance, (ii) reorganization of company’s structure, e.g., individuals’ and functional units’

responsibilities and tasks, information exchange protocol, for eliminating communication difficulties in

configurator design, development, and application, and (iii) top level support and company-wide

engagement which is fundamental to the achievement of the above two.

In conducting the survey, we used nominal scales by considering the explorative nature of this study.

While the nominal scale permits an easy understandable questionnaire, it makes analysis less exact than

a Likert scale. In another words, it may not be able to identify the causal relationships among the

interesting elements involved in the implications of configurator applications. In this regard, this study

highlights an interesting future research topic. An extended quantitative method involving data analysis

based on Likert scales might be conducted to reveal these causal relationships. Moreover, since product

configurator applications incur business process changes and require organization redesign, business

process reengineering, where product configurators are applied, might deserve future research efforts

as well.

REFERENCES

Aldanondo, M., Rouge, S. & Reron, M.(2000). Expert configurator for concurrent engineering:

Cameleon software and model.Journal of Intelligent Manufacturing, 11, 127-134.

Bergey, J., Smith, D., Tiley, S., Weiderman, N. & Woods, S. (1999). Why reengineering projects fail,

Carnegie Mellon Software Engineering Institute-Product Line Practice Initiative, 1, 1-30.

 274

Chen, Z. & Wang, L.(2009). Adaptable product configuration system based on neural network,

International Journal of Production Research, 47(18), 5037-5066.

Chu, C.H., Lin, C.W., Li, Y.W. & Yang, J.Y. (2005). Online product configuration in E-commerce with

3D viewing technology, International Journal of Electronic Business Management, 3(3), 225-234.

Falkner, A., Haselbock, A., Schenner, G. & Schreiner, H.(2011). Modeling and solving technical

product configuration problems, Artificial Intelligence for Engineering Design, Analysis and

Manufacturing, 25, 115-129.

Felfernig, A. (2007).Standardized configuration knowledge representations as technological foundation

for mass customization, IEEE Transactions on Engineering Management, 54(1), 41-56.

Felfernig, A., Friedrich, G.& Jannach, D. (2001).Conceptual modeling for configuration of mass-

customizable products, Artificial Intelligence in Engineering, 15, 165-176.

Forza, C. & Salvador, F. (2002). Managing for variety in the order acquisition and fulfillment process:

the contribution of product configuration systems, International Journal of Production Economics, 76,

87-98.

Haug, A., Hvam, L. & Mortensen, N.H. (2010). A layout technique for class diagrams to be used in

product configuration projects, Computers in Industry, 61, 409-418.

Haug, A., Hvam, L. & Mortensen, N.H. (2011). The impact of product configurators on lead times in

engineering-oriented companies, Artificial Intelligence for Engineering Design, Analysis and

Manufacturing, 25, 197-206.

Haug, A., Hvam, L. & Mortensen, N.H. (2012) Definition and evaluation of product configurator

development strategies, Computers in Industry, 63, 471-481.

http://blogs.wsj.com/cio/2013/12/11/avons-failed-sap-implementation-reflects-rise-of-usability/

Hvam, L. and Ladeby, K., 2007, An approach for the development of visual configuration systems,

Computers & Industrial Engineering, 53, 401-419.

Hvam, L., Riis, J., and Hansen, B.L., 2003, CRC cards for product modeling, Computers in Industry,

50, 57–70.

Jiao, J. and Zhang, Y., 2005, Product portfolio identification based on association rule mining,

Computer-Aided Design, 37(2), 149-172.

 275

Kristianto, Y., Helo, P., & Jiao, R. J. (2015). A system level product configurator for engineer-to-order

supply chains. Computers in Industry, 72, 82-91.

Ong, S.K., Lin, Q. & Nee, A.Y.C. (2006). Web-based configuration design system for product

customization, International Journal of Production Research, 44(2), 351-382.

Paper, D. and Chang, R., 2005, The state of business process reengineering: a search for success factors,

Total Quality Management, 16(1), 121-133.

Pine, B.J. (1993).Mass Customization: The New Frontier in Business Competition, Harvard Business

School Press, Boston.

Trentin, A., Perin, E.& Forza, C. (2011). Overcoming the customization-responsiveness squeeze by

using product configurators: beyond anecdotal evidence, Computers in Industry, 62, 260-268.

Trentin, A., Perin, E. & Forza, C.(2012). Product configurator impact on product quality, International

Journal of Production Economics, 135, 850-859.

Zhang, L. (2007). Process platform-based production configuration for mass customization. PhD

Dissertation, Division of Systems and Engineering Management, Nanyang Technological University,

Singapore.

Zhang, L. (2014). Product configuration: A review of the-state-of-the-art and future research,

International Journal of Production Research, 52(21), 6381-6398.

 276

Appendix 10.1: Questions in the survey

General demographic questions:

1: Are you knowledgeable about the design or development or application of the product configurator.

2. What is your position title.

3. What is your industry.

4. What is your company size.

4. How long have the configurator been in use.

5. How many types of products are the configurator used for.

Product configurator applications:

1. Which tasks are performed by the configurator.

2. Who are the users.

3. Which functional units were reorganized.

4. Which business processes were changed.

5. Which changes were made to other computer systems.

6. Did you layoff/hire full time employees.

Performance of product configurator applications:

1. The percentage of increased sales volume.

2. The percentage of increased correct sales orders.

3. The percentage of reduced production rework.

4. The percentage of increased customer orders accepted.

5. The percentage of reduced order processing time.

6. The percentage of reduced sales delivery time.

Difficulties in implementing product configurator project:

 277

1. The difficulties in designing the product configurator.

2. The difficulties in developing/maintaining the configurator.

3. The difficulties in using the product configurator.

Barriers preventing companies from continuously applying product configurator:

1. The continuous evolution of products.

2. The lack of technical staff for maintaining the configurator.

3. Customer requirements are rather unclear.

4. The unsafe feeling from the employees because of the possibility of losing jobs.
 278

Chapter eleven

Sustainability Assessments for Mass Customization Supply

Chains
Carlo Brondi, Rosanna Fornasiero, Davide Collatina
Institute of Industrial Technologies and Automation, Italy

ABSTRACT

The evaluation of the environmental impact of production networks has been under debate during the

last years. At present, there is a shift of production paradigm from mass production to customization

and personalization. The aim of this chapter is to evaluate sustainability of supply chains applying a

model based on the integration of LCA (Life Cycle Assessment) with discrete simulation to compare

different customization policies in a networked context. In the developed model, the supply chain

environmental impact is assessed through an innovative modular LCA where different level of

customization have been analysed and then the paper compares the scenarios based on variation of

drivers like lead-time to the customer, quality in terms of scraps and level of sustainability of the

suppliers. The model is validated by collecting data from a fashion based case study taking into

consideration the environmental impact for a certain batch production. The preliminary results highlight

that specific decisional areas under the control of supply managers (e.g. supplier selection and

manufacturing defects) can significantly affect the environmental impact of the whole supply chain.

11.1 Introduction
 279

The quantitative assessment of the environmental sustainability is a recurrent area of interest for

evaluating production phases, transportation and suppliers within the literature. The sustainability

assessment also concerns the modern production paradigms like the knowledge intensive services to

customer (Gallouj et al., 2015). In particular, Petersen T.D. et al. (2011) addresses the issue whether the

concepts mass customization and sustainability are fundamentally compatible. The updated mass

customization paradigm calls both for personalized outputs and cost/eco-efficiency tracking in order

for companies to maintain their competitiveness and create value (Mourtzis and Doukas, 2014; Ueda et

al.,2009). The development of customized production and their related services seems implicitly to call

for new collaborative supply chains (Romeo, Cavalieri and Resta, 2014) as well as for reliable models

for sustainability characterization going beyond the qualitative assessment (Kohtala, 2014). Preliminary

involvement of consumer within product-service systems including a referenced sustainability tracking

seems also to produce positive effects in fashion sectors (Amstrong et al., 2105). At the current stage,

different studies have proposed alternatives for the design of sustainable supply chain and eco-efficient

product in order to be compliant with the mass customization paradigm (Govindan et al., 2014; Osorio

et al., 2014; Lee and Huang, 2011; Piplani, Pujavan and Ray, 2007) but literature review emphasizes

the lack of verification criteria in presence of diverging possible effects due to customization policies

(Kohtala, 2014). Positive mentioned environmental effect accounts the reduction of pre-consumer

waste, the lower transport emission, the minor product replacement, the greater potential for re-

manufacturing, the intermediary reduction and the use-phase extension for customized products.

Negative possible effects instead account the augmented difficulties in product reuse, the need of

energy/resource intensive transformation process, the exposition to uncontrolled emission in local

environments and the possible failure in the replacement of traditional mass production.

From the industry perspective, despite the large agreement on the importance of sustainability aspects

for the long-term competitive advantages, often companies need strong triggers in order to put into

action initiatives for integrating these dimensions in their strategies. On one side legal regulations,

response to stakeholders, customer demand, reputation loss, environmental and social groups’ pressure

are often listed as triggers for companies to implement sustainability. On the other side, some barriers

to implement actions for Sustainable Supply Chain (SSC) are (Piplani, Pujavan and Ray, 2007):
 280

1. costs to implement measurement systems for sustainability;

2. hitch to define a value for the output with respect to environmental outcomes;

3. perception that data to be collected from different actors in the network are not manageable and

some having low impact on the global outcome;

4. difficulties to take no popular and high-priced decisions for network.

In order to overcome these limits, researchers and managers are trying to answer to the following

questions: How should a supply chain accomplish the trade-off between economic and non-economic

objectives while making managerial decisions? What activities are necessary to implement a SSC?

Which types of incentives are necessary to induce people to pursue sustainability’s objectives? (Noci,

1997).

When dealing with networked companies, the availability of data on time, quality, service etc. of

suppliers along the network is at state of the art while for the environmental impact analysis there are

still problems with sharing of data which are still considered confidential (e.g. energy consumption,

release on water and heating) and once the data are shared to make them homogeneous. Specific sectoral

inventory data are useful to calculate the Global Warming Potential (GWP) of a company and of a

network.

11.2 State of the art

In the current business environment, purchasing process has become critical activity for adding value

to products and a vital determinant to ensure the competitiveness of a company. This process becomes

more complicated when environmental issues are considered because green purchasing must consider

the supplier’s environmental responsibility, depending on product chain assets, in addition to the

traditional factors such as the supplier’s costs, quality, lead-time and flexibility. The management of

suppliers based on strict environmental compliance seems to be not sufficient in view of a more

proactive or strategic approach. Noci (1997) designed a green vendor rating system for the assessment

of a supplier’s environmental performance based on four environmental categories, namely, ‘green’

competencies, current environmental efficiency, suppliers’ green image and net life cycle cost. Main

limit in attributing a unique environmental performance index to a company seems to be linked to the
 281

management of reliable quantitative scientific set of values which can be considered constant in

different comparison. While literature related to supplier evaluation is plentiful, the works on green

supplier evaluation or supplier evaluation that consider environmental factors are rather limited

(Handfield et al., 2002; Humphreys, McIvor, & Chan, 2003). Two general aspects seem then to emerge

as relevant in the sustainability assessment of mass customization: from one side it seems important to

identify the sustainability features for a proper assessment while on the other hand the environmental

assessment of scalable product chains requires specific modelling issues. These aspect are separately

faced in the next two paragraphs.

11.2.1 Sustainable supply chains in the mass customization

The high variability of the customer demand and the legislative pressure in the EU Countries on

environmental aspects, push academic and industrial communities to tackle the question related on how

to implement sustainable production systems. In order to accomplish this objective, a strong integration

among the units of the Supply Chain (SC) is necessary and can help to maintain and to build a durable

competitive advantage with respect to competitors. For this reason in the last years many approaches

have been proposed on international journals to support the implementation of Sustainable Supply

Chain (SSC) (Seuring and Muller, 2008; Dyllick and Hockerts, 2002). The result of this academic and

corporate interest has been the achievement of important goals for the sustainable success of firms in

terms of Integrated Supply Chain, Green Supply Chain, Ecology Industry, and Long Terms Competitive

Advantages.

Despite in recent years, there is a large agreement on the importance of sustainability aspects for the

long term competitive advantages, often companies need strong triggers in order to put into action

initiatives for integrating these dimensions. Legal regulations, response to stakeholders, customers

demand, reputation loss, environmental and social groups pressure are often listed as triggers for

companies to implement sustainability.

Zhu, Sarkis and Lai (2008) identifies five Green Supply Chain Management (GSCM) practices

implementation factors: internal environmental management (IEM), green purchasing (GP),

cooperation with customers including environmental requirements (CC), eco-design practices (ECO)
 282

and investment recovery (IR). The authors present the implications in the GSCM for closing the loop

of supply chain. Zhu, Sarkis and Lai (2008b).

It is clear that the adoption of green practices impacts on environmental results, for example in terms

of pollution reduction (Klassen and Whybark, 1999), but at the same time companies need to take over

other environmental dimensions without forgetting to pursue profit objectives. In literature we can find

some references to the positive role that environmental management plays in order to achieve

operational performance (and it is established that operational performance are strictly and positively

linked to financial performance) linking the lean and the “green” approach to management.

Hart (1997) and Florida (1996) suggest that environmental management can also provide cost savings,

by increasing efficiency in production processes and improving the firm’s performance, by facilitating

the creation of resources and capabilities as well as the ability to innovate (Porter and Van der Linde,

1995; Russo and Fouts, 1997; Reinhardt, 1999). Moreover, Rusinko (2007) suggests a positive impact

of pollution prevention on cost savings and competitive advantage. Christmann (2000)draws on the

resource-based view of the firm and finds a moderating effect of innovation and implementation on the

relationship between environmental practice and cost advantage.

On the other hand, literature also raises a trade-off issue between environmental initiatives and

operational performance (Clark, 1994; Walley and Whitehead, 1994), but in more recent works it was

evaluated the impact of the cost of compliance with environmental goals (Yu, Ting and Wu, 2009). For

this reason a “lean & green” perspective is adopted in the development of the performance measurement

system, in order to monitor and control the trade-offs resulting from the implementation of the

environmental management.

The current debate on customization paradigm poses a number of further issues for the sustainability

paradigm. A customer-driven manufacturing could in fact addresses the reduction of the environmental

impact since the closest link between manufacturer and customer can imply a reduction of the

environmental load due to operation and distribution (e.g. electricity, heating and transport). The

reduction of item stock and the increase in the value of traditional products (Bruno et al., 2013; Bernard

et al., 2011; Zhang et al., 2011) seem to contribute to reduce environmental impact particularly in the

product distribution to customer, in its use and in the eventual recovery phase. A proper product
 283

modularization and proper efficiency policies in the factory management can be the best way to increase

efficiency as well as to counterbalance negative effects of customization. Another open issue concerns

the transportation reduction which depends on the supply chain configuration. The down size of

transport network could, in fact, conflict with the reduction of the efficiency of economies of scale.

According to some author the relative environmental contributions from the stages of supply,

manufacture and waste production is affected by a strong sectorial characterization (Su and Chen and

Yang, 2015).

Different authors proposed simulation and optimization techniques to manage such divergent aspects.

Mourtzis et al. (2013) and Mourtzis et al. (2014) proposed a toolbox to deal with the supply chain

conflicting drivers in a network set-up, metaheuristic and Artificial Intelligence methods integrate

assessment of carbon footprint limited to standard transport processes. A complete life cycle approach

for supply chain carbon footprint modelling was proposed by Trappey et al. (2012) by using the I-O

matrix in a real supply chain case based on three areas of investigations: materials, production and

logistic for the supply chain. Despite the completeness of the approach, assessment is developed in

presence of the same service model and by using sectoral data to assess the cluster carbon footprint. An

effort for the integration of life cycle assessment and business models with a mass-customization

perspective is faced by Boër, C.R. et al. (2013) where a complete reference set of environmental

indicators for the modular subdivision of the whole product life cycle has been applied at level of single

component, life cycle phase and supplier. The detailed set of equations can, however, be difficultly

implemented to common multi-tier networks in which other LCI indicators are available and distinction

between part manufacturing and assembly phase is often not clear.

Then as primary issue a proper characterization of real effects seem necessary (Su, Chen and Yang,

2015). In this respect, the environmental impact characterization due to current industrial practices is

also affected by serious operative limits. Particularly, the definition of company’s environmental

performance is generally considered based uniquely on the type of transformation process or, instead,

is referred to standard operating conditions which are focused on a single product type. Real industrial

practices are instead based on ever changing production item batches.

 284

11.2.2 Environmental impact assessment of networks for customization strategy

As mentioned in the introduction, it is important to underline that the customization processes involves

divergent environmental effects. On one hand, the simultaneous presence of these effects can lead to

higher impact from customization processes compared to processes for mass-produced items.

Traditional process can in fact benefit from scale economies. On the other hand, the comparison

between a customized product and a standard product should require the same functional unit, which is

the performed service towards the final consumer. In such view, the product customization result to be

an additional service towards the consumer that changes the traditional functional unit of the mass-

customized products. According to these premises, the environmental impact minimization can be

relevant for identification of best implementation scenarios rather than the single comparison of

customized product against similar traditional products.

LCA can represent a proper methodology to assess product environmental sustainability cause the

intrinsic perspective on the whole product life cycle (Sun, Chen and Yang, 2015; Hugo and

Pistikopolous, 2005; New et al., 2010; Bojarski et al., 2009; Brondi et al.,2014). Nevertheless, the

proper adoption of traditional LCA to the customized environment should overcome the following

barriers:

- Alignment between life cycle perspective and business perspective: inventory schemes for

physical flows within SMEs can require a business-compliant approach that can significantly

differ from LCI schemes. Internal operations can be committed to external suppliers so that the

mass and energy tracking is interrupted. Furthermore, the capability to provide reliable data

from company could not overcome a limited extent of the product life cycle (i.e. from first

supply level up to final product distribution). Corporate environmental policies usually have to

face on how many product-chain levels should be included within the data inventory process.

Ideally, the entire value chain should be analysed, but resources and data availability can pose

a serious constrains for the assessment models (Brondi et al., 2012; Unep, 2015; ISO/TS, 2014).

In a factory perspective, common “knowledge horizon” by product manager can cover the

background phases up to a certain supplier and the foreground phases up to the gate for the

customer (see Figure 11.1).

 285

FIGURE 11.1 Knowledge horizons in modelling environmental impact of the product chain

 Adaptation of inadequate data to new LCA studies (Hagelaar and Van der Vorst, 2002). Life-

cycle analysts can frequently abuse of literature and general-purpose database in place of supply

chain data in case the assessment involves a limited view on the product chain.

 Misalignment between company environmental assessments and product design. The designer

and the life-cycle analyst can require radically different procedures in order to modify the final

solution (Brondi et al., 2012).Furthermore, the design of a product require a set of parameters

that can result insufficient for an environmental impact assessment.

 Limited extent in reuse of previous LCA studies.(Klöpffer, 2012). Literature suggests that a

study review starts with the draft Goal and Scope chapter. In fact, each LCA is performed under

specific assumptions and purposes, e.g. an LCA for comparative assessment requires different

rules than an LCA for internal assessment. Changes in the functional unit, in the single

processes or in the system boundaries can then revoke the study validity for other purposes.

 Uncertainty in the life cycle determination. Existing product benchmarks commonly provide

results with reference to the entire life cycle of a single product. The proper determination of

life cycles requires the statistical tracking of a certain stock of products. Such stock involves

different life cycles. The combined variance of specific environmental drivers result then to be

fundamental. As an example economy of scale, transport network, stock variance and

environmental profiles from different suppliers can influence the variance analysis.
 286

 Misalignment between consequential and attributional methodologies. LCA study that aims to

optimize a supply chain should compare different configurations of technologies and materials.

The resulting comparative study (consequential methodologies) should require the assessment

of additional marginal effects that can be difficultly modelled (i.e. marginal demand for a

certain choice, avoided impacts etc.). On the other hand, the not-comparative studies

(attributional methodologies) focus on the life cycle for a specific product. In particular, the

attributional methodologies make use of allocation factors requiring an impact sub-division

according to eventual co-products and services.

11.3 A proposal for integrated model of supply chain assessment

A modular parametric approach can introduce a flexible and precise way to assess the relative

contributions of scalable supply chains within a product chain.

Such approach structures the available data (e.g. information on energy and material input, quantitative

emissions to water, soil and atmosphere, transport data from suppliers to focal company) in terms of

input and output impacts for each product chain node.

Further simulation of supply chain trade-offs, which account also other quantitative indicators, assign

performance indicators to each supplier. Other reference indicators for such assessment are the delivery

time, the quality of the product, the flexibility, the inventory strategies and the environmental profile.

2. Assessment of the product 3. Modular LCA Approach

environmental profile (Information Segmentation)

Impact assessment of independent information modulus

linked to key drivers in customization process

4. Simulation of the
1. Collection of KPIs from each supplier
supply chain trade-offs

Data for average

Assessment of different product chain configurations
production ,
Energy and
Emissions to
distances to final
according with previous analysis and economic performances
water, soil and
material input on assembly
atmosphere on
average sampling Waste
average sampling
management
system etc. 5. Product Environmental Profiles for
the simulated configuration
 287

FIGURE 11.2 Integrated model for supply chain evaluation

As reported in the grey boxes of Figure 11.2, firstly, product-chain modularization provides the set of

quantitative data, then, dynamic simulation integrates such information and provides quantitative values

for not available data. By such approach, the simulation can perform assessment for several products

and supply chain configurations. A final analysis of customized production models allow to assess the

sustainability due to different manufacturing scenarios within the make-to-order paradigm.

11.3.1 Modular LCA approach for supply chain modelling

The modularization of the impact assessment starts from a comparative LCA. As first step, LCA

execution is compliant with the LCA guidelines (DIN EN ISO 14040:2006/14044:2006). LCA consists

of four phases: a) definition of goal and scope b) inventory analysis phase (LCI), c) impact assessment

phase (LCIA), and d) interpretation phase.

The modular Life Cycle approach includes the definition of examined system, functional unit, system

boundaries, allocation procedure, data quality requirements and any other assumptions.

- Goal and Scope: differently from traditional LCA scopes, which depend on a specific product and

the intended use of the study, the modular approach aims to identify single information modules for

each recurrent macro flow within the product chain. Macro flows are aggregated flows (i.e. specific

products or services) commonly exchanged within the supply chain. According to the extent of the

optimization both cradle-to-gate and cradle-to-grave perspectives can be adopted.

- The Life Cycle Inventory phase (LCI phase) is an inventory of input/output data with regard to the

examined system involving collection, calculation and allocation of the necessary data. The

modular life cycle requires tracking product-chain data according to recurring flows for a wide

range of possible products.

- The Life Cycle Impact Assessment phase (LCIA) provides additional information to help assess a

product system’s LCI results in order to understand their environmental significance. The approach

focuses on the environmental impact significance on relative contribution of each flow. Then results

represent each individual impact in a comparative perspective.

 288

- Life Cycle Interpretation discusses the results of the LCI or the LCIA, as a basis for conclusions,

recommendations and decision-making in accordance with the goal and scope definition. Such

phase compare single results in addition with other weighting factors as times and cost drivers. The

supply-chain assessment model integrates the LCIA phase in order to identify best supply chain

configuration.

The equation 11.1 formalizes the modular approach in quantitative terms. Such quantification takes into

account a series of modularization for the traditional impact assessment.

Modularization of the single life cycle phases: The impact assessment of product life cycle results as

sum of environmental profiles of different phases (i.e. manufacturing phase or supplier operations). In

particular, the indexing of such modules identifies the supply level from a player to another (i.e. from

s-1 to s). The LCIA of the customized product results as sum of incremental contribution due to different

supply chain stages.

Modularization of the customized product: A certain number of physical components forms the

customized product with reference to a specific supply chain configuration (i.e. a material type from a

specific supplier). The environmental impact assessment of the customized product results as sum of

environmental profiles of such different components.

Explication of the manufacturing key drivers: With reference to the specific customized product, the

approach identifies and cluster relevant drivers with a significant variance and potential environmental

impact. Common drivers due to product customization are the material composition of a product, the

product weight, the product chain transport, the warehouse stock for each supplier and the material

waste for single operation. The environmental impact assessment of the customized product result as

sum of environmental profiles depending on such key drivers.

In a supply chain perspective, each node of the network represents a single company while an input-

output modelling defines the flow inventory of the single company (Hart, 1997).

More in detail, firstly, Life Cycle (LC) analyst assesses impact of background flows (auxiliary flows

and processing materials) and impact of foreground flows (final products and emissions to nature) for

a certain company. Such approach requires tracking and collecting the flows crossing the physical

factory boundaries with reference to the final product. Common flows are input energy vectors (e.g. the
 289

electricity and natural gas used for the operation of the production plant), input primary resources, (e.g.

water supply), output emission to air and water (e.g. VOC, PTS and wastewater), output solid waste

with related treatment (e.g. wastes from packaging and finishing activities, paints and coatings).After

the identification of such recurrent flows, LC analyst calculates the respective environmental impact

for a reference unit in compliancy with the LCA general rules. The results constitute a set of independent

information modules for a company.

Subsequently, in a further calculation, the LC analyst gathers together and calibrates the information

modules according to the overall mass and energy balance related to the product chain activities. While

impact of elementary flows (e.g. basic chemicals and energy vectors) requires standard values from

international database, specific materials and components require specific LCA study or, alternatively,

the adoption of the same approach to the foregoing supplier.

The final evaluation of the impact assessment infers from the combination of company inventory and

Life Cycle studies for specific industry flows and elementary flows.

The following equations formalize the modular decomposition of the LCA approach in order to express

flexibly the environmental profiles of a customizable item:

The Equation (11.1) assesses the impact categories of the customized product as a sum of independent

previously calculated vectors. The final array expresses a cradle-to-gate assessment of a specific product

from the raw material extraction up to the factory gate. The calculation method appears to be compliant

with the supplier perspective. The same approach assesses different manufactured product within the

same factory through a flexible supply chain and distribution methodologies.

𝒆𝒑𝒏 = ⁡ 𝑬𝑷𝐶𝑜𝑚𝑝𝑎𝑛𝑦 (𝑝𝑛 , 𝑧) + 𝑬𝑷𝑠𝑢𝑝𝑝𝑙𝑦 (𝑝𝑛 , 𝑧 − 1), (11.1)

𝑚(𝑝𝑛 ) 𝐽
𝑬𝑷𝐶𝑜𝑚𝑝𝑎𝑛𝑦 (𝑝𝑛, 𝑧) = ∙ [[∑𝐼𝑖=1(𝒖𝒊 ∙ ∑𝐾 𝐹
𝑘=1 𝑞𝑘 (𝑝𝑛 ))]𝐼𝑛𝑝𝑢𝑡 + ⁡⁡[∑𝑗=1(𝒖𝒋 ∙ ∑𝑓=1 𝑞𝑓 (𝑝𝑛 ))] ]⁡, (11.2)
𝑀∙𝑁 𝑂𝑢𝑡𝑝𝑢𝑡

𝑆 𝑇 𝑅

𝑬𝑠𝑢𝑝𝑝𝑙𝑦 (𝑝𝑛 , 𝑧 − 1) = ⁡ ∑ 𝑞𝑠 ∙ 𝒆𝒑𝒔 ⁡⁡⁡ + ⁡⁡ ∑ 𝒖𝒕 ∙ (∑ 𝑙 𝑡 ∙ 𝑚𝑟𝑡 (𝑝𝑛 ) ∙ 𝑑𝑟𝑡 (𝑝𝑛 )) ⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡(11.3)

𝑠=1 𝑡=1 r=1
 290

𝒆𝒑𝒏 = ⁡ 𝑬𝑷𝑡𝑎𝑟𝑒 (𝑝𝑛 ) + ∑ 𝑋𝑑 (𝑝𝑛 , 𝑐) ∙ 𝒖𝒅 ⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡(11.4)

𝑑=1

TABLE 11.1 Notations for the Equations 11.1-11.4

Variable Explanation

epn Environmental Profile for the Customized Product (EPCP): vectorial array of data

representing Environmental Impact for the production of one customized product

pnby a specific company.

EPCompany Environmental contribution to the environmental profile of the product pnby the

company internal processes. It represents the environmental impact due to

consumption and emission of company activities.

EPSupply Environmental contribution to the environmental profile of the product pnby the

company supply. It represents the Cradle-to-gate environmental impact of

supplied items for the production of a reference quantity of customized products

pn in the reference period. The bill-of-material of the customized product can help

to list such items. The supply refers to a specific level (i.e. the direct supply to the

company).

m(pn) Weight (or alternatively the value) of the customized product pnin the reference

period

M Total weight (or alternatively the value) of the total production of the company in

the reference period

N Total number of the total customized products pn manufactured by the company in

the reference period

ui Vectorial array representing the environmental impact for an incoming unitary

flow of energy /mass. The inventoried mass and flows are not included in the final
 291

productpn. Such vectorial array refers to homogenous flow type both in terms of

physical features (e.g. the same energy input type) and in terms of product chain

features (e.g. the same supplier).

uj Vectorial array representing the environmental impact for an outgoing unit of

energy /mass flow type. The inventoried mass and flows are wastes changing with

the kind of production (pn). Such vectorial array refers to homogenous flow type

both in terms of physical features (e.g. the same waste type) and in terms of product

chain features (e.g. the same dismissal procedure).

qk Inventoried quantity for a specific incoming flow type.

qf Inventoried quantity for specific outgoing flow type.

I Total number of incoming flow types.

J Total number of output flow types.

K Number of total supplies for the incoming auxiliary flows iby the examined

company in the reference period.

F Number of total disposal for the outgoing flows j from the examined company in

the reference period.

eps Environmental Profile for the Supplied Items (EPSI): vectorial array of data

representing Environmental Impact for the supplied item psthat compose the final

customized product pn.

S Total number of supplied items for the production of the customized product pn.

qs Quantity of supplied items psthat are required for a single unit of the customized

productpn.

ut Vectorial array of environmental impact for a specific transport type t. The vector

is assessed for 1 ton*km and for a set of predetermined impact categories.

T Total number of transport types.

l Load factor for a single round trip travel

 292

m Mass of the supplied items, to or from the company, which is transported by the

transport type t.

d Distance covered by the transport t in a roundtrip travel.

R Total number of roundtrip travels between suppliers and company for the

production of the product pn

D Number of customization drivers changing during customization process

X Value of the customization driver in a specific manufacturing scenario c (e.g. the

quantity of a specific waste)

us Vectorial array representing the environmental impact for a unit of specific

operational driver which changes during the customization process

EPtare Total contribution to the Environmental profile of the product pndue to operational

drivers which remains unchanged during customization process

ps Items and services provided from specific suppliers at a specific tier level

pn Customized product manufactured in the examined company

 293

FIGURE 11.3 Modular representation of a product chain for a specific customized product

The approach can be applied both in presence of previous cradle-to-gate LCA studies for specific

product components and to address the further analysis of the intermediate suppliers up to raw material

level. Furthermore, the same approach can be applied in the consumer perspective by a simple extension

of the product chain to the level z+1.

Finally, the equation 11.4 modularizes the same impact categories according to different operational

drivers in order to introduce an explicit dependency of the LCA calculation from the production

management choices.

11.3.2 The simulation of the supply chain trade-offs

The output of the modular LCA is used for the second stage of the model. Discrete event simulation is

used as a tool which enables to evaluate alternative production network configuration and operating

procedures in a convenient way when optimization models are not practical (Bernard et al., 2011).The

model is developed to compare different scenarios with the initial configuration of the SC model. This

part of the model gives companies the possibility to create different configuration scenarios and to make

what-if analysis to evaluate the trade-offs due to customization between different performance
 294

dimensions which are otherwise difficult to compare like delivery time and sustainability. For example,

considering different customization policies, the need to shorten the delivery time to each customer can

increase the number of travels and increasing therefore pollution. The model studies how to optimize

the number of travels in the upstream supply chain without compromising delivery time to customer

and without compromising sustainability. The model evaluates also the impact of applying different

aggregation of orders to suppliers as a way to reduce their lead time and also environmental impact.

The modelling of supply networkswas often used as a way to check the balance of inventory, especially

to compare standard production methods with just-in-time approach. In literature three different

approaches can be found: organizational, analytical and simulation (Zhang et al., 2011). The first one

relies on process modelling based on systems theory, but the models developed with this approach are

not dynamic and they do not take into account the system's behaviour through time. The second one

relies on mathematical formalization of the supply chains. These models, however, require

approximations, usually restrictive, that can also be limited for consider time.

Simulation refers to a broad collection of methods and applications to mimic the behaviour of real

systems. Simulation models enable to evaluate alternative system designs and operating procedures in

a convenient way when optimization of models is not practical due to dimension of the problem in terms

of complexity. Moreover simulation is a support when testing alternatives on a real production system

is usually too expensive and time consuming.

The model created for the specific case of comparing different customization strategies is based on the

following starting points:

 Supply chain is based on a hierarchical relationship with the focal company: suppliers

deliver to the focal company their materials and components upon specific requests.

 Production orders are pulled by the customer orders, therefore MTO (Make-to-order)

strategy is applied.

 It is assumed that there is one warehouse where all the materials and components are

sent by the suppliers and are ready to be used according to the customer orders request.
 295

 Customer orders are received by the focal company dispatched to suppliers with fixed

date’s policy and taking into consideration minimal safety stock.

 The customer orders are queued according to the request date from the customer with

the first in first out strategy.

 Performance of the suppliers is used to evaluate the overall performance of the supply

network and is based on delivery time, quality (scraps), flexibility, etc. For each of

these indicators, also the variance is taken into consideration based on the real

performance collected from the ERP of the focal company.

 Contractors are also part of the network structure, i.e. company working in parallel with

the focal company when there is a capacity problem at the focal company.

 Environmental profile is assigned to the three phases identified in the application of the

Modular LCA: suppliers, transports and production at focal company/contractors.

This model is modular and can be used and customised for different companies according to their

specific data. Suppliers can be added according to the dimension of the specific network and

performance adapted to the needs of the specific case.

11.3.2.1 The formulation of the model

The simulation allows to verify the performance of different scenarios for each network configuration

defined to analyse the effect of improving the performance in the case of traditional or personalized

products by considering also the possibility to change the number of suppliers and considering how

much the overall performance change when the performance of suppliers are improved.

Defining supplier i, where i = 1, …, n and order j (where j = 1, ..., M), the performance of each supplier

is evaluated based on the following indicators:

 T(i) = delivery time for supplier (i) evaluated as the average time to delivery an order. This

performance is particularly relevant in customization because of the necessity of providing the

customized product to customers in short time meaning high flexibility.

 296

 Q(i) = quality for supplier (i) evaluated as the average percentage of defected pieces on each

delivered order. This performance is particularly relevant in customization context because

defected pieces are hardly tolerated by consumers willing to pay even higher premium price for

customized products; defect products create delays in delivering products due to the required

rework.

For what concerns the production orders that the focal company assigns to suppliers, their demand

occurrence follows a normal distribution N(µ, ), where µ is mean of demand, and where  is standard

deviation. µ,  depends on the type of order (small or large order) and  of small order is in average

higher than in large order representing the high variability of the small orders.

Simulation is replicated to create different supply chain configurations. Then each configuration is

evaluated based on the following supply chain performance indicators:

 OLT = Order Lead-Time. Time from the reception of order from the customer (i.e. focal company’s

retailer) that starts supply chain production processes till the delivery to the customer of products

(that is the end of supply chain processes).

 IV = Inventory Volume at the product factory. The volume of inventories of the components that is

transferred from suppliers and is used at the product factory.

The creation of comparative supply chain configurations scenarios (i.e. scenarios 1, 2,..n) is based on

the variation of the suppliers’ performance starting from the scenario zero. In the simulation model, the

production costs are not considered because it is assumed they are not a discriminant in the choice of

customization since it is demonstrated that customers are willing to pay a premium price for customized

products (Alptkinoğlu and Corbett, 2008). Table 11.2 shows the to-be supply network configurations

scenarios created with the simulation.

11.3.3 The manufacturing scenarios

The definition of different manufacturing scenarios allows to directly assessing environmental

implication of customization policies. This means to identify recurring customization in industry

 297

practices, the degree of variability of the product and the degree of variability of the related supply

chain.

The customization scenarios aim to fix the drivers variance for a certain product batch in presence of

progressive increase in the product variance towards the final consumer. The table reports the general

assumptions in the mass production and mass customization.

TABLE 11.2 Manufacturing scenarios

Manufacturing
Description Design changes Supply chain changes
scenarios
Supply matches a certain Suppliers are located in various
Mass quantity with a minimal The Bill of Material is fixed and countries (according to current
production flexibility within a year. The design changes are limited to location of the real case)
(current modelling of material supply standard size and colour according to cost and quality
situation) requires an average load per changes. parameters.
travel.
Accessories and components
The Bill-of-Material can change can be supplied in a local
Production follows the style network (average distance from
in terms of components size and
preferences of the customer focal company limited to a
component type.
within a certain degree of certain mileage) while suppliers
The number of materials for the
freedom in the choice. Such of raw materials are kept
Mass components and the number of
higher grade of selection unchanged. Possible variation
customization suppliers increase by allowing
includes a variability in of stored stock, extra
the customer to change the
design features, ergonomic consumptions and waste and
material type within certain
features, material features and transport depends on the
components for aesthetical,
aesthetical features. consumer order sequence.
technical or ergonomic options.

11.3.3.1 Customization drivers

Customization strategies can vary according to the combined variation of technical, market and

organizational drivers. The following list reports relevant drivers according to previous literature studies

and to an analysis of several companies dealing with customization.

Operational drivers

 Number of models within the same production batch. Starting from a specific type of product, the

number of models available can vary per thousand of shoes produced. The change involves a limited

improvement of ergonomic features, the variation in the type of material for product components

(in case of shoes can be increase in the types of leather for the upper, variations to the outer sole).

The more extensively the customization is applied, the more the bill of material of the pattern while

maintaining a consistency in terms of the components has a variance related to each component.
 298

 Processing materials and scrap rate. The increase in the variability of the final product can affect

the efficiency of the traditional manufacturing process. In particular, the requirement of material

per pair product should include the gross material requirement. As an example, in the comparison

with mass production, the customization can increase the waste production in the upper production

in a concurrent manufacturing of different shapes within the same production batch.

 Defectiveness rate. Product defects depend heavily on technological and managerial processes that

exist within a single company. Although it is difficult to quantify the change in the defectiveness

levels with the product variability, it is possible to assume that such defectiveness increase cause to

the increased complexity in manufacturing options. Defective products can affect environmental

impact due the additional resources consumption for single product and with the increase waste

contribution.

 Transportation. The increase in the number of transport for a manufactured product seems to

depend on the supply chain management and size of production batches. In general, the increase in

materials types from different suppliers can imply a decrease in transport efficiency and in the loads

optimization. Such effect can be registered both in entrance to the factory gate (more limited

supplies) and in output to the consumer distribution (smaller lots at the points of sale).

 Auxiliary material consumption. Consumption of materials and auxiliary resources (consumables

not integrated into the final product) in general has a limited dependence from the variability of the

product. In fact, the consumption of auxiliary materials depends on the increase of the variety of

the product only within a limited amount. Instead, the technology for production process

significantly infers the consumption and emissions for each type of model. However, a growing

complexity of the production processes may entail a limited increase of these consumptions.

Economical drivers

 Unsold items. The unsold items depend on the failure to predict the market demand. Despite the

economic and environmental damage related to the overproduction, the price elasticity of demand

for goods could reset the stocks of unsold items. In the case of customization it is possible to assume

that the increase of choice for the consumer can satisfy better the demand and reduce the unsold

items.
 299

 Average product life cycle. Some economic studies (Brodie et al., 2013) suggest that increased

demand satisfaction has a limiting effect on the replacement of an asset. There is a lack of empirical

link between the increasing customization of a product and the reduction of its replacement.

However, it is possible to assume (within the further assumption the satisfaction remain the same

during the product use) that customized products fit better with customer needs and may increase

their time of use, reducing new consumption in a certain period.

 Order size from selling points. The increasing of market segmentation and increased customization

of the product may increase frequency of supplying to retailers, shops, and multistore. There is no

reason to keep high stocks of customized items that can in fact be risky in case of fluctuations in

demand.

 Time to service. In a scenario of stable technologies, lack of optimization within the product chain

is highly dependent on the required time-to-service and the demand trend. Segmented markets with

a high variability may in fact require rapid production organization with the implications on the

related resources demand and environmental emissions.

Organizational drivers

 Make-to-order supply chain. A chain of suppliers which is organized according to the lean “make-

to-order" paradigm with a reduced stock at the final assembler and a frequent supply depending on

the customized product demand. This chain type requires an efficient organization and a restrained

time-to-market. Transportation can remain frequent and not optimized even if assembler and

suppliers are synchronized.

 Factory flexibility. A flexible factory is able to meet a variable demand for customized products

and a proper time-to-market. In order to perform such operations, the factory include many

production departments and an adequate internal materials stock.

11.4 Sustainability assessment for a customization case in a fashion company

11.4.1 Application of the Modular LCA

The application of modular LCA to a footwear case enabled a comparative assessment of environmental

burdens due to customization policies in a fashion company. LCA involved a cradle-to-gate perspective
 300

on the footwear company, and the analysis takes into account the product supply chain from raw

materials acquisition up to the product manufacturing and industrial waste disposal. The product use

and its dismissal were not included in such model. The modelling of the factory waste included also the

waste treatment processes after the initial deposit. Differently, the analysis did not take into account the

waste flows sent to economic recovery (i.e. material recycling, energy recovery, composting etc.). In

such a case, system boundaries are limited up to the facility gate where the recycling or recovery

processes take place (e.g., transportation to the facility shall be included).

A) In a first stage, a classical LCA assess common recurrent flows for an Italian footwear company.

The combination of such recurrent flows provided the total environmental profile for the factory in

a reference period. Such impact results from the combination of the industry flows (i.e. average

energy for each shoe pair) and the processing materials (i.e. specific content of material per

footwear type).

The use of data from the international database (i.e. Ecoinvent, Gabi etc.) supported the LCA

modelling particularly for elementary flows. The formalization of environmental impact through

impact categories follow is compliant with the CML 2011 standard and EPD system. The impact

categories to assess inventory flows are the global warming potential (GWP), the acidification

potential (AP), the eutrophication potential (EP), the ozone depletion potential (ODP) and

photochemical ozone creation potential (POCP).

B) In a second stage, a modular LCA assess the environmental impact variance due to customization

of a production batch in specific conditions. A number of company drivers address such variance

assessment. In terms of technology options, we assumed that the production of customized footwear

required the same resources and technologies than the current technologies.

11.4.2 Life Cycle Inventory (LCI) for the case study

A methodology of data collection compliant with the modular approach allowed acquiring datasheets

and inventory data from the examined company. At factory level, the data account mass and energy

recurrent flows for an average year production of 477.569 footwear pairs.

 301

Specific energy supply configurations referred to energy mix of the energy utility serving the company

(i.e. Kwh supplied by a specific Utility).In addition, the modelling of waste treatments complies with

the European Waste Catalogue (EWC) for industrial waste (i.e. recycled paint and varnish containing

organic solvents).

FIGURE 11.4 Inventory data for the definition of in- and outgoing flows

Data inventory for different footwear models allowed identifying a general impact for average footwear.

Supply scenarios integrated the number of deliveries within a year, the distance among suppliers and

factory, the mean of transport and the load capacity for each supply type. Stages from resource

extraction up to the creation of process materials, involves suppliers to large distances. For example,

the production of leather requires a breeding outside Europe, a transport in European tanneries, and then

the manufacture of the materials. Stages from the acquisition of process materials to the shoe

manufacturing involves manufacturing at local level, so the producers of components and the footwear

manufacturing company are placed in a local district over 100 km. We finally assume that same final

footwear consumer stays in a local area (minor than 200 km).

 302

FIGURE 11.5 Model application to the study case

Different drivers are considered according to equation 11.4 in order to better define the scenarios to be

analysed.

FIGURE 11.6 Operational drivers for the product customization

 303

The description of product physical features, supply chain configuration and manufacturing features are

reported in the Table 11.3. The selection of these drivers defines a basic scenario in which each

operational driver has a base value.

TABLE 11.3 Product features

The sole is composed of Ethylene-vinyl acetate and for a weight of

Sole
520 g
The upper is composed of leather upper for a total area of 1,5 m² per
Upper
pair
Product
features Leather Upper, Polyurethane Sole, Other materials - Polyester, Nylon
Materialcomposition 6, Elastan, Spandex, Conventional Cotton, Rayon viscose, Ethylene-
vinyl acetate
Pairweight 920 g

11.4.3 Environmental impact due to customization activities

In the assumptions, the production batch remains constant (10,000 pairs) while changes in the product

design are introduced within the same batch. It is assumed that each driver varies according to a range

that has been defined for this work in agreement with literature data and empirical evidence. Each

drivers have been varied as reported in Table 11.4 and multiple impacts have been analyzed to

understand how positive and negative influence of different drivers can impact on the overall

environmental performance.

TABLE 11.4 Description of customization drivers and the related variance range

Driver Description Initial Value Maximum

Value
Scrap rate can vary according to the efficiency of the cutting
process for the upper. Such efficiency can be reduced by the 20% respect 60% respect
alternation of different model within the same factory. Single to the bill of to the bill of
E1 Scrap Rate
supply could be oversized respect to the batch size. The base material material
value bases on "Waste generated in the leather products values values
industry", UNIDO, 2000.
Transport depends on the number of delivery to the factory,
distance between producer, suppliers, and freight load factors.
According with the study case, the components supplier
E2 Transports remains in an area within a radius of 50 Km. Then number of 50 km 1000 km
supply increases. In the base scenario, an average truck type
performs the supply in a single transport for the whole batch
(9200 kg)
 304

In the analysed case study, the defectiveness rate can increase

with the manufacturing complexity. Each defective pair require 0,1% respect 10% respect
Defectiveness
E3 a new production to maintain the batch dimensions. In the to the batch to the batch
rate base scenario (mass production), the defectiveness amount to size size
10 pairs per 10000 shoe pairs.
This driver introduces a positive effect. Life cycle extension
2 years
reduces product replacement that creates a minor provision
Life Cycle duration 3 years
E4 of further resources from eco-sphere in a certain period time.
extension (expected use duration
According to assumptions, the shoe pair can last than two
phase)
years while its effective average use amount to two years.
The overall batch size is 10.000 pairs. Uppers of the shoe are
22 materials
in different materials within the same production batch and
type for the
for the same model. Every time a new material is added, the
One material upper (5
Material batch is split in smaller sub-batches. For ex. two material types
E5 type for the leather types
variance means that 5000 pairs are produced with one material and
upper (leather) and 17
other 5000 pairs are produced with the other material type, 3
synthetic
material types means 3.333 pairs per each material type and
types)
so on.

In particular, the following environmental indicators are evaluated:

 Global warming potential (GWP100) is a relative measure of how much heat a greenhouse gas

traps in the atmosphere. It compares the amount of heat trapped by a certain mass of the gas in

question to the amount of heat trapped by a similar mass of carbon dioxide. A GWP is

calculated over a specific time interval, 100 years. GWP is expressed as a kg of carbon dioxide

equivalents (whose GWP is standardized to 1).

 Ozone depletion potential (ODP) describes the decline in the total amount of ozone in Earth's

stratosphere. ODP is expressed as the sum of ozone-depleting potential in kg CFC 11-

equivalents, 20 years).

 Acidification potential (AP) measures acid gases that are released into the air or resulting from

the reaction of non-acid components of the emissions. The acidification potential is expressed

in kg SO2- equivalents.

 Photochemical Ozone Creation (POCP) measure the emissions of gases that contribute to the

creation of ground level ozone. The POCP is expressed as kg ethene-equivalents.

 Eutrophication potential (EP) measure the ecosystem's response to the addition of artificial or

natural nutrients, mainly phosphates, through detergents, fertilizers, or sewage, to an aquatic

system. Emission of substances to water contributing to oxygen depletion is then expressed as

kg PO4 3-equivalents.
 305

Appendices 11.1. to 11.5 report respectively the environmental impact due to customization activities

for the five environmental indicators, GWP100, AP, EP, ODP and POCP.

It is important to emphasize that the results were calculated under the assumption that the technology

framework remain the same assuming that increase in the level of customization is linked to an increase

of the product model variability.

In the charts in the annexes, the variation of all the operational drivers mentioned in the Table 11.4 (life

cycle extension, scraps, defectiveness, transport, material variance) is normalized to a scale 1-100 to

make them comparable. the starting situation is represented with a blue bar and all the bars above it

represent the cases when the variation of the 5 drivers is such to cause an increase in the environmental

impact, while the bars below it represent the decrease in the environmental impact.

The results for the specific case study suggest the following conclusions:

 The customization process can have both a positive and negative environmental impacts, and

when it is linked to the increase in the possibility of using new materials may in fact include

more eco-efficient than traditional materials and thus minor environmental impact.

 The most significant drivers to control the environmental impact are the choice of the material

type, the rate of defective parts and scrap rate.

 The positive effect on environmental impact brought by lifecycle extension of the use phase

of the product, avoiding the use of new resources consumption, balance the increase in other

drivers. Similarly, a more informed consumer choice on the test material could affect final

impact on the final product.

 Impact categories are differently affected by the product variance; in the analysed case, AP

and ODP indicators double their impact according to the change of the operational parameters.

 The allocation rules can also significantly affect the background impacts. Standardization

plays a determinant role in determine best supplier options for foreground sectors. A clear

alignment between allocation rules and system boundary selection with respect to background

suppliers (material producers) seems necessary in order to reduce the potential high variability

in LCA results.
 306

11.4.4 Results of the simulation of customization scenarios

The simulation was based on the data collected from the ERP of the company and on the results of the

modular LCA. The preliminary analysis of data extracted from ERP shows that the suppliers of the shoe

company are asked to produce both large and small orders according to the needs of the company with

a wide range of order dimension both in terms of number of rows, number of pieces per row, number

of different items. According to the order dimension suppliers have different performance in terms of

delivery time, product quality, etc. (see an example in Table 11.1). Before to apply the simulation

model, a Pareto analysis allowed to categorize suppliers to identify the most strategic ones in terms of

total delivered amount. In the analyzed case study, it emerges that some supplier performance are linked

to the order dimension like average delivery time while other are independent from it like average

scraps. These performance indicators are taken into consideration in the simulation model and are used

to create the scenarios for large and small orders.

According to the defined model, the customization strategies have been applied to choose the most

suitable suppliers for each scenario and a commercial simulator like Simio® was used to compare

different scenarios based on suppliers performance. The initial scenario was based on data collected

from the footwear company and it represents a simplified model of its network where most of the

suppliers are considered. The model is based on the following assumptions:

A contractor works in parallel with the shoe producer to manufacture the orders which can’t be

assembled by the shoe producer for capacity limits.

Some product models can be produced only by the shoe producer, others only by the contractor and

other by both of them.

In case of a product that can be processed both by the shoe producer and by the contractor, it is sent to

the one with the shortest queue.

The shoe producer manages the materials necessary for the contractor and forwards them to him when

necessary for the production.

The warehouse and the distribution centre are located at the shoe producer's site.
 307

The working time of the contractor includes an extra time both for delivery time of materials to

contractor and for shipping time of the final products to the distribution centre. The advantage of

producing at the contractor is given by the fact that there is the possibility to shorten the queue of the

company.

The application of the modular LCA in the previous chapter showed that out of the three identified

macro-categories (supplying process, transportation, production), the supplying process has a large

impact on the overall sustainability of the network and for this reason the scenarios are built mainly to

evaluate how their performance can impact on the sustainability in particular considering the most

important environmental indicator, the GWP.

As it was described in the previous chapter, the scenarios defined in the LCA evaluation are used to

link the level of customization (in terms of number of product variations) with drivers like

transportation, scraps, defectiveness, etc. The model is based on keeping the same type of raw material

(leather) provided by the same suppliers or by similar suppliers to evaluate the environmental impact

of their operative performance beside the environmental impact of new materials. Therefore starting

from the standard production scenario, the other scenarios are analysed according to possible changes

in the operative performance of the suppliers given the materials they can provide. A set of different

what-if scenarios based on the variation in the suppliers' performances has been defined in order to

evaluate how the change in the supplier performance can impact on overall supply chain performance.

In particular, it has been analysed how improvements in their delivery time (from10 to 35% of suppliers'

lead time) and in products quality (from 10 to 35% in scraps) can affect the overall performance of the

supply chain. Based on the data collected from the company and the established model, the performance

of the supply network is dynamically evaluated considering the value of the initial inventory, the

average inventory during the analysed period, the average and maximum lead time to fulfil customers'

orders (see Table 11.5).

TABLE 11.5 Simulation scenarios

Description of the
uctio

uctio
varia

varia

varia
Prod

Prod

Prod
prod
Mas

tion

tion

tion
uct

uct

20
…

scenarios
n

n
s

3
 308

SQ(i) = Scenarios with

Improvement in supplier
quality (scraps)
SQ1 -10%
SQ2 -15%
SQ3 -20%
SQ4 -25%
SQ5 -30%
SQ6 -35%
SL(i) = Improvement in
supplier lead time
SL1 -10%
SL2 -15%
SL3 -20%
SL4 -25%
SL5 -30%
SL3 -35%

Preliminary results show that the variation in the suppliers lead time has a different impact according

to the applied level of customization (which means number of product variants). For simplicity, Figure

11.7 does not show the details of all the cases from 1 to 20 product variants but consider only 4 cases

of increasing the level of customization, from 5 to 10 to 15 and 20 product variants. The Figure 11.7(a)

shows how the improvement in supplier lead time performance can bring an improvement in the

customer order time which is more than the improvement caused by the supplier quality on the product

defectiveness represented in Figure 11.7(b).

Figure 11.7b
Figure11.7a

FIGURE 11.7 Impact of improvement in suppliers’ performance

 309

In fact, as for the impact of changes in products delivered by suppliers in terms of quality (less scraps),

with the data of the specific company it turns out that a reduction in scraps give a reduction in final

product defectiveness. The level of scraps represents a limited share of production (5%) and for this

reason the impact is more limited than in the case of lead time changes. Improvement on scraps level

of product components means less reworking and less mistakes which from the suppliers goes until the

final customer and less defectiveness during production. Generally speaking improvements in the

suppliers’ performance bring different degree of improvement in the overall Supply Chain performance

and many variables need to be taken into consideration. In this study, some of them have been

considered and analysed, but further studies will be necessary to complete the flow.

11.5 Conclusions and Recommendations

The complexity of evaluating the environmental impact of supply chain modelling seems to require

novel methodologies to properly identify the key decisional areas. In this chapter, a new approach has

been presented based on the integration of LCA data with discrete simulation and has been tested in a

specific case by collecting data from a footwear company and considering different customization

strategies.

 The functional unit for manufacturing product is commonly based on a single product. By adopting

a factory perspective, it seems necessary to shift the focus on production batches rather than on a

single product since the Life cycle Inventory. Such shift can in fact include new inventory categories

which can represent more precisely the real hidden flows for a customized production. Examples

can be the modelling of the distribution platform or the use pattern for a certain product. When the

inventory is based on a batch such evaluation can include new variables by resulting more in line

with mass and energy balance at supply chain level.

 The chapter analyses also the variance due to different product chain configurations in altering the

environmental impact based on simulation of multiple scenarios considering different degree of

variability of the operational drivers. In the preliminary results, it is highlighted that specific

decisional areas under the control of product managers are also key drivers in environmental impact
 310

creation. Further studies in other sectors could better contextualize the environmental implications.

In particular, aspects such as economies of scale, warehouse management, and the use of alternative

technologies could significantly affect this analysis.

FIGURE 11.8 Scenarios for implementation of customization within the product chain

The outcome of the model suggests that a proper implementation of customization practices can result

in an environmental benefit. In general, it is possible to identify four subsequent scenarios for the

implementation of sustainability issues in the customization practices (see Figure 11.8):

 In the first scenario, the weak implementation, customization focus on limited aspects of the

product such as design or some functional parameters without any framework to support

consumption reduction or emissions. In this case, it becomes likely that the customized product

has a higher environmental impact.

 In the second scenario, the efficient customization, dedicated tools to minimize customization

cost and consumptions can be implemented in a single factory perspective. In particular,

emphasis is given to methods for the effects quantification and the data management from the

manufacturer.

 In the third scenario, the knowledge-based customization, personalization pushes on multiple

aspects concerning the use phase of the product and background phases so that data concerning

the whole product life-cycle can be analysed by the producer. This type of implementation
 311

make clear the effects induced on the product chain and acts proactively to reduce these effect

in a single player perspective.

 In the fourth and final scenario, the sustainable customization, data regarding the sustainability

of the product are exchanged within the product chain with a standard protocol. The diffusion

process involves the whole chain starting from raw material producers up to the final

consumer. Furthermore, distributed methods and tools for the quantification of the social and

environmental effects related to the choice of customization concurrently support the product

chain players (e.g. consumer, producer, material developer) at each stage. The diffusion of this

type of information introduces emergent properties and feedback within the system. Such

framework,jointly with the increased buyer decisional power can directly link the product

environmental profile associated with customization preferences.

Modelling based on simulation was used because it offers a realistic observation of the supply chain

behaviour and allow an analysis of the supply chain dynamics. It provides an observation of the

behaviour of the network over time, to understand the organizational decision-making process, analyse

the interdependencies between the actors of the chain and analyse the consistency between the

coordination modes and the decisional policies. Moreover, simulation can also be coupled with an

optimization approach, to validate the relevance and the consequences of its results.

Future developments in the model will be based on making available for companies reliable libraries

on environmental impact and on refining the simulation model to ease what-if analysis. Further analysis

of the trade-offs between the operative and sustainability performance is also necessary. In a

perspective view, the authors will further develop and customize the framework for other specific

industrial case studies, with the definition of transversal methods and tools for the sustainability

performance analysis. Relationships among critical processes, improvement actions and sustainability

dimensions as well as suitable indicators will be deepened and updated.

REFERENCES
 312

Armstrong C. M., Niinimäki K., Kujala S., Karell E., Lang C. 2015. “Sustainable product-service

systems for clothing: exploring consumer perceptions of consumption alternatives in Finland” Journal

of Cleaner Production 97: 30-39.

Bernard A., Daaboul J., Laroche F., Da Cunha C. 2011. “Mass Customisation as a Competitive Factor

for Sustainability.” Proceedings of the 4th International Conference on Changeable, Agile,

Reconfigurable and Virtual Production, Montreal, Canada 2011.

Bojarski D., Laínez J. M., Espuna A., Puigjaner L. 2009. “Incorporating environmental impacts and

regulations in a holistic supply chains modeling: an LCA approach.” Computers and Chemical

Engineering 33: 1747–1759.

Brodie R.J., Ilic A., Juric B., Hollebeek L. 2013. "Consumer engagement in a virtual brand community:

An exploratory analysis" Journal of Business Research 66: 105-114.

Brondi C., Fornasiero R., Vale M., Vidali L., Brugnoli F. 2012. “Modular Framework for Reliable

LCA-Based Indicators Supporting Supplier Selection within Complex Supply Chains.” APMS (1): 200-

207.

Brondi C., Fragassi F., Pasetti T., Fornasiero R. 2014. “Evaluating sustainability trade-offs along

Supply Chain.” 20th ICE Conference – IEEE TMC Europe Conference 23-25 June 2014, Bergamo.

Bruno T. D., Nielsen K., Taps S. B., Jorgensen K. A. 2013. “Sustainability Evaluation of Mass

Customization.” IFIP Advances in Information and Communication Technology 414: 175-182.

Christmann P. 2000. “Effects of Best Practice of Environmental Management on Cost Ad-vantage: The

Role of Complementary Assets.” The Academy of Management Journal 43(4): 663-680.

Clark R.A. 1994. “The Challenge of going green.” Harvard Business Review 72(4): 37-38.

Dyllick T., Hockerts K. 2002. “Beyond the business case for corporate sustainability.” Business

Strategy and the Environment 11(2):130-141.

Florida R. 1996. “Lean And Green: The Move To Environmentally Conscious Manufacturing.”

California Management Review 39(1): 80.

Gallouj F., Weber K. M., Stare M., Rubalcaba L. 2015. “The futures of the service economy in Europe:

A foresight analysis” Technological Forecasting & Social Change 94: 80–96.

General Programme instructions for the International EPD System 2.0 - Annex A
 313

Govindan K., Azevedo S. G., Carvalho H., Cruz-Machado V. 2014. “Impact of supply chain

management practices on sustainability”. Journal of Cleaner Production 85: 212–225.

Gu P., Hashemian M., Nee A.Y.C. 2004. “Adaptable Design” CIRP Annals Manufacturing

Technology53(2): 539-557.

Hagelaar G. J. L. F., van der Vorst J. G. A. J. 2002. “Environmental supply chain management: using

life cycle assessment to structure supply chains.” International Food and Agribusiness, Management

Review 4: 399-412.

Handfield R., Walton S. V., Sroufe R. 2002. “Applying environmental criteria to supplier as-sessment:

A study in the application of the Analytical Hierarchy Process.” European Journal of Operational

Research 141:70-87.

Hart S. 1997. “Beyond Greening: Strategies for a Sustainable World” Harvard Business Review 75: 66-

76.

Hugo A. and Pistikopoulos E.N. 2005. “Environmentally conscious long-range planning and design of

supply chain networks.” Journal of Cleaner Production 13: 1428-1448.

Humphreys P., McIvor R., Chan F. 2003. “Using case-based reasoning to evaluate supplier

environmental management performance.” Expert Systems with Applications 25: 141–153.

ISO/TS 14072:2014 - Environmental management -- Life cycle assessment -- Requirements and

guidelines for organizational life cycle assessment

Klassen R.D. and Whybark C. D. 1999. “The impact of environmental technologies on manufacturing

performance.” Decision Sciences 30(3): 599–615.

Kohtala C. 2014. “Addressing sustainability in research on distributed production: an integrated

literature review”. Journal of Cleaner Production. In press.

Lee Y. , Huang F. 2011. “Recommender system architecture for adaptive green marketing” Expert

Systems with Applications 38: 9696–9703

Mourtzis D., Doukas M., Psarommatis F. 2013. “Design and operation of manufacturing networks for

mass customisation” CIRP Annals - Manufacturing Technology 62: 467–470.

Mourtzis D., Doukas M., Psarommatis F. 2014. “A toolbox for the design, planning and operation of

manufacturing networks in a mass customisation environment.” Journal of Manufacturing Systems

 314

Mourtzis, D., &Doukas, M. 2014. “Design and planning of manufacturing networks for mass

customisation and personalisation: challenges and outlook”. Procedia CIRP, 19: 1-13.

Noci G. 1997. “Designing green vendor rating systems for the assessment of a supplier’s environmental

performance.” European Journal of Purchasing and Supply Management 2: 103–114.

Nwe E. S., Adhityab A., Halimb I., Srinivasan R. 2010. “Green Supply Chain Design and Operation by

Integrating LCA and Dynamic Simulation.” 20th European Symposium on Computer Aided Process

Engineering – ESCAPE20.

Osorio J. , Romero D., Betancur M., Molina A. 2014. “Design for Sustainable Mass-Customization:

Design Guidelines for Sustainable Mass-Customized Products.” Proceeding of the 20th ICE Conference

on Engineering, Technology and Innovation.

Petersen, T. D., Jørgensen, K. A., Nielsen, K., & Taps, S. B. 2011. “Is Mass Customization

Sustainable?” In MCPC 2011 World Conference on Mass Customization and Personalization.

Piplani R., Pujavan N. and Ray S. 2007. “Sustainable Supply Chain Management.” International

Journal of Production Economics 111(2): 193-4.

Porter M.E. and Van der Linde C. 1995. “Toward a new conception of the Environment

Competitiveness Relationship.” Journal of Economic Perspectives 9(3): 97-118.

Reinhardt F. L. 1999. “Bringing the environment down to earth.” Harvard Business Review 77(4): 149-

157.

Romero D., Cavalieri S., Resta B. 2014. “Green Virtual Enterprise Broker: Enabling Build-to-Order

Supply Chains for Sustainable Customer-Driven Small Series Production” APMS 2014, Part II, IFIP

AICT 439, 431–441.

Rusinko C. 2007. “Green manufacturing: An evaluation of environmentally sustainable manufacturing

practices and their impact on competitive outcomes.” IEEE Transaction on Engineering Management

54(3): 445-454.

Russo M.V., Fouts P.A. 1997. “A resource-based perspective on corporate environmental performance

and profitability.” Academy Management Journal 40(3): 534-59.

Seuring S. and Müller M. 2008. “From a literature review to a conceptual frame work for sustaina-ble

supply chain management.” Journal of Cleaner Production 16: 1699-1710.

 315

Su M., Chen C. , Yang Z. 2015. “Urban energy structure optimization at the sector scale: considering

environmental impact based on life cycle assessment”. Journal of Cleaner Production, online (2015)1-

11.

The International EPD® System - Product Category Rules. 2011. “CPC Class 2912 Finished Bovine

Leather.” Version 1.0 Dated 28-09-2011.

The International EPD® System. 2013. “General Programme Instructions for the international EPD

system.” Dated 04-06-2013.

Trappey A.J.C., Trappey C.V., Hsiao C., Ou J.J.R. and Chang C. 2012 “System dynamics modelling of

product carbon footprint life cycles for collaborative green supply chains.” International Journal of

Computer Integrated Manufacturing 25(10): 934–945.

Ueda K, Takenaka T, Váncza J, Monostori L. 2009. “Value creation and decision-making in sustainable

society” CIRP Annals 58(2): 681-700.

UNEP (2015): Guidance on Organizational Life Cycle Assessment

Walley N. and Whitehead B. 1994. “It’s not easy being green.” Harvard Business Review 72(3): 46-5.

Yu V., Ting H., Wu Y.J. 2009. “Assessing the greenness effort for European firms. A resource

efficiency perspective.” Management Decision 47(7):1065-1079.

Zhang Y., Luximon A., Ma X., Guo X., Zhang M. 2011 “Mass Customization Methodology for

Footwear Design” Digital Human Modeling, pp. 367-375. Springer Berlin Heidelberg

Zhu Q., Sarkis J., Lai K. L. 2008 (a). “Confirmation of a measurement model for green supply chain

management practices implementation.” International Journal of Production Economics 111: 261-273.

Zhu Q., Sarkis J., Lai K. L. 2008 (b). “Green supply chain management implications for closing the

loop.” Transportation Research Part E 44: 1-18.

Appendix 11.1: Global warming potential (GWP100)

 316

365 624

326 037

291 170

264 965

244 527

225 105

216 873

209 386

200 405

196 702

186 789
Global Warming Potential (GWP)

180 452

174 502

166 359
kg CO2eq

164 649

162 654

156 676

150 304

147 945

142 940

137 377

132 911

116 794

107 318

102 098

98 677

94 462

90 054

82 324

0% 100% 200% 300% 400% 500%

Percentage change from base scenario

LC extension Scrap rate Defectiveness rate Transports Materials Base scenario

 317

Appendix 11.2: Acidification potential (AP)

Appendix 11.3: Eutrophication potential (EP)

84 229

82 578

78 941

76 346

63 155

61 934

59 206

57 259

42 103

41 289

39 471

38 173
Acidification Potential (AP)

15 980

15 254
kg SO2eq

14 542

14 309

11 985

11 440

10 906

7 990

7 627

7 271

3 697

2 773

2 524

1 933

1 775

1 295

1 168

0% 100% 200% 300% 400% 500%

Percentage change from base scenario

LC extension Scrap rate Defectiveness rate Transports Materials Base scenario

 318

5 257

4 616

3 979

3 761

3 462

2 984

2 626

2 308

1 990

1 383

1 230

1 144
Eutrophication Potential (EP)

1 077
kg PO4---eq

1 028

980

944

938

896

854

808

765

704

660

615

572

540

512

471

427

0% 100% 200% 300% 400% 500%

Percentage change from base scenario

LC extension Scrap rate Defectiveness rate Transports Materials Base scenario

 319

Appendix 11.4: Ozone depletion potential (ODP)

4,7E+00

4,1E+00

3,6E+00

3,4E+00

3,1E+00

2,7E+00

2,4E+00

2,1E+00

1,8E+00

5,7E-02

5,1E-02
Ozone Depletion Potentia (ODP)

4,5E-02

4,1E-02
kg CFC-11eq

3,8E-02

3,3E-02

2,8E-02

2,6E-02

2,4E-02

2,3E-02

2,2E-02

2,1E-02

1,9E-02

2,0E-02

1,8E-02

1,6E-02

1,3E-02

1,2E-02

1,1E-02

9,6E-03

0% 100% 200% 300% 400% 500%

Percentage change from base scenario

LC extension Scrap rate Defectiveness rate Transports Materials Base scenario

 320

Appendix 11.5: Photochemical Ozone Creation (POCP)

232

210

195

173

161

157

151

150

147

143
Photochemical Ozone Creation Potentials (POCP)

142

136

130

124
kg C2H4eq

118

113

111

108

107

103

101

98

82

78

75

72

72

68

65

0% 100% 200% 300% 400% 500%

Percentage change from base scenario

LC extension Scrap rate Defectiveness rate Transports Materials Base scenario

 321

Chapter twelve

Sustainability Issues in Mass Customized Manufacturing

Arun Nambiar,
California State University Fresno, USA

ABSTRACT

Stiff global competition and volatile customer demand have become the mainstay of today’s industrial

and service sectors. With more and more countries opening up their economies, companies are striving

hard to stay competitive and are constantly seeking new venues that provide them the necessary

advantage over their competitors. The mass customization paradigm provides companies that

opportunity to create a niche for themselves by providing customized solutions to customer needs. A

environmentally conscious customer combined with increased federal and state regulations have forced

companies to focus on the environmental impact of their operations while trying to maximize the value

of their products and services to keep their costs down. Sustainable development combined with mass

customization affords new synergies that can potentially accrue benefits in the form of new product or

service designs, efficient supply chains and increased profits. Companies can leverage information

systems and other new technology to obtain maximum value from their resources thus lowering their

environmental impact while enabling them to stay competitive.

12.1 Introduction

Mass customization is a manufacturing paradigm that aims to provide customized products at mass

production prices. Though this concept was originally introduced in the 1960s, it has been gaining

significant traction in the recent years due to stiff market competition and ever-changing customer

preferences. Companies are shifting from a one-size-fits-all approach to providing highly customized

products and services designed to satisfy the needs of most customers.

 322

While this is a good idea and definitely a market winner if done right, it engenders a plethora of

problems for the manufacturing and production side of the businesses not to mention the logistical

challenges. Manufacturing and production facilities have traditionally been optimized to run the same

set of operations over and over again in an highly efficient manner. With this shift to highly customized

products, the production mix has changed from one of low-variety high-volume to a high-variety, low-

volume one. This creates a new optimization scenario where parts required, and production operations

needed for this high-variety product mix needs to be efficiently managed to minimize waste and

maximize throughput. Modern manufacturing paradigms (Nambiar 2010) such as quick response

manufacturing, lean principles (Womack, Jones & Roos 1999), and agile manufacturing (Nambiar,

2009a) when combined with mass customization (Davis 1989, Pine 1999) purport to create a win-win

scenario for companies by helping them produce highly customized products and services quickly and

efficiently.

In the recent years though, there has been increased emphasis on “going green”. This is essentially a

moniker in common parlance for all things related to sustainable development. Sustainable development

has been defined in the Brundtland Report (UN 1987) as “the development that meets the needs of the

present without compromising the ability of future generations to meet their own needs”. Thus, it can

be seen that focus of sustainable development is to utilize minimum resources, produce minimum

wastes, and cause minimal pollution. The three main pillars of sustainable development focus on

economical, sociological and environmental aspects of development respectively. That is to say, the

products and services resulting from a sustainable endeavour must be economically viable for both the

customer and the business, sociologically acceptable and environment-friendly. With the advances in

information technology, customers are better informed about the environmental impact of their

purchases and the manufacturing practices of the products they purchase. This engenders an informed

customer base that is shifting towards environment-friendly products and shunning products and

companies with a poor track record of sustainable practices. As a result, companies are hopping on to

this green bandwagon both as a marketing tool to gain more customers in today’s highly uncertain and

volatile market rife with global competition and as part of their corporate social responsibility (CSR).
 323

Marketing gimmicks aside, sustainable practices do have significant benefits. Through these practices,

fewer raw materials may be used in production or existing materials may be reused or recycled to obtain

maximum benefit. This is of particular interest since countries like Brazil, Russia, India, China and

South Africa (often known as BRICS) are poised to grow significantly in the coming decades thus

creating an increased demand for goods and services which in turn places undue strain on the already

limited resources. This has been quite evident in this past decade with China’s voracious appetite for

natural resources to sustain its manufacturing facilities and unprecedented growth necessitating forays

into many resource-rich countries especially in the African continent. Although China’s economy has

been cooling in the past couple years which was to be expected, other parts of the world are growing as

well. The combined aspirations of large swathes of population in developing countries for lifestyles

similar to that of the developed nations creates unprecedented demands on limited resources and puts

undue stress on the planet through the resulting developmental activities. Hence, instituting sustainable

practices both for material conservation and environment preservation can really go a long way in

ensuring a sustainable planet for future generations to come.

Though the general premise behind this initiative is laudable and its benefits purportedly manifold, there

are significant challenges to be overcome in order for sustainable practices to become the norm and to

obtain maximum benefit through the collective efforts of everyone involved. The first important

challenge is changing the mindsets and encouraging people to take a more holistic view thus enabling

them to see the larger impact of their actions. This is a drastic change in perspectives from the “not in

my backyard” or “not my problem” attitudes that tend to be more prevalent. Often times, companies

take a narrow approach to sustainability by focusing on the environmental impact alone which brings

to fore the next challenge. It is imperative to focus on economical, and sociological aspects of

sustainable development along with the focus on environment. This is important because in order of

products and services that are a result of sustainable practices to become popular, they have to be

affordable and acceptable to the consumers. Companies also tend to focus on the proverbial low-

hanging fruit which provides immediate results and instant gratification. However, a holistic approach

that focuses on the entire value chain and critically examines all operations through the sustainability

lens in order to identify areas of innovation and improvement will provide greater returns on investment
 324

in the long term. Information systems can serve as a key enabler in overcoming some of these challenges

by integrating various partners of the value chain into a central system. This brings to light the fourth

challenge of getting these disparate systems to communicate with each other and seamlessly integrate

into a single system. Researchers world over are focusing on ways to overcome these challenges and

issues involved in a successful implementation of sustainable practices.

With mass customization, the aim is to produce products and services that the customers want at prices

they can afford. This takes care of two aspects of sustainable development - economically viable and

sociologically acceptable from the customer’s perspective. In order for the business to leverage this to

gain market share and increase profit margins, it is imperative that these efforts be economically viable

from the manufacturing, production and logistics perspectives as well. Moreover, being

environmentally friendly helps from a marketing standpoint. Thus, it can be seen that sustainability

initiatives can be combined with mass customization efforts to reap maximum benefits from providing

highly customized products and services at affordable prices in an environment-friendly manner.

This chapter explores the sustainability issues involved in mass customized manufacturing. The chapter

identifies the key areas of concern, and examines best practices to handle sustainable development in

a mass customization environment.

12.2Background

Porter’s theory of competitive advantage originally proposed in the eighties (Porter 1985) suggests that

companies can stay competitive by adopting one or a combination of two strategies:

● Product differentiation - By providing customers products that meet their expectations

and are different from those of competition, companies can build a reputation that will attract

more customers.

● Cost leadership - Companies can compete by offering products at lower prices while

remaining profitable through the ensuing increased sales volume.

Modern manufacturing paradigms seek to address these tenets either individually or in a combined

fashion. For example, the primary objective of lean principles is to reduce wastes which in turn reduces

costs thus providing the cost leadership advantage. Quick response manufacturing and agile
 325

manufacturing seek to reduce the design-to-market times in an effort to provide customers with products

that meet their requirements and expectations before competition thus providing the product

differentiation advantage. Mass customization also seeks to provide companies with the competitive

advantage through product differentiation. However, in today’s globalized economy, it is impractical

to demand a cost premium for customized products and hence companies have to strive to not only

provide customized products at lower cost but also seek out other venues to reduce their operating costs

even further.

In this mad race to lower operating costs, companies are off-loading routine activities to subcontractors

and vendors while focusing on their core competencies. This allows companies to direct valuable

resources towards what makes them unique and different from competition. Mass customization is a

concept originally introduced by Davis (Davis, 1987) and later developed by Joseph B Pine that seeks

to provide companies the much needed competitive advantage through product differentiation and cost

leadership. Pine defines (Pine, 1999) mass customization as “providing tremendous variety and

individual customization, at prices comparable to standard goods and services ...with enough variety

and customization that nearly everyone gets exactly what they want”. The underlying premise for this

paradigm is that there is high demand for customized goods and services that are tailored specifically

to meet individual customer needs and expectations. There are four basic types (Gilmore & Pine, 1997)

of customization identified as the four faces of mass customization. These vary depending on the level

of involvement of the customer in the design of the product and the level of customization possible for

the product. These four faces include adaptive, cosmetic, transparent and collaborative approaches.

These focus primarily on the end-product and how it is modified or used based on individual customer

needs. However, with the expanding world markets and stiff global competition, it becomes imperative

that companies adopt a combination of these approaches rather than one or the other as shown in Figure

12.1.
 326

FIGURE 12.1 Faces of Mass Customization (Modified from Gilmore & Pine, 1997)

The four approaches have been modified from the original representation in Gilmore & Pine (1997) to

suggest that a company may choose its position anywhere within the continuum similar to the

production continuum proposed by Lampel & Mintzberg (1996). Moreover, with increasing operations

crossing geographic boundaries, distribution and delivery are also important venues for customization

by ensuring customers receive what they want where they want it and when they want it. Companies

also tend to differentiate from competition by providing support services such as installation, repair and

maintenance, recycling, and replacement. Hence, customization can be achieved at eight different levels

(Da Silveira, Borenstein & Fogliatto,2001) viz. standardization, usage, package and distribution,

services, specific customization, assembly, fabrication and finally design. These eight levels can be

correlated to the four faces mentioned earlier. For example, packaging and distribution customization

could be more of a cosmetic change or design customization could be as a result of a more collaborative

effort between the company and the customer. Needlessly to say, irrespective of the slight differences

in the focus area for customization, the ultimate objective of the company is to provide customers with

products and services that meet their expectations in the hopes that this would translate into increased

market share. However, stiff global competition leads to price wars where companies try to outcompete

each other by providing goods and services at the lowest cost often undercutting their profit margins.

This makes it crucial to invest in mechanisms and methodologies that allow companies to lower their

operating costs. Another approach might be to seek differentiation through increased emphasis on
 327

environmental impact of their operations given that the environmental debate about global warming

and greenhouse gas emissions has been on the forefront in the recent years.

Resource-based theory (Hart 1995) builds on Porter’s theory of competitive advantage by suggesting

that the firm’s core competency arises from the value being added by its resources and hence resources

are critical to the company’s success. These resources include:

● physical - machines, facilities, workforce

● financial - capital to be invested

● tacit - knowledge and experience acquired over time

Hart (Hart 1995) further underscores the importance of sustainable development by proposing a three-

pronged natural resource-based approach towards achieving competitive advantage. This includes:

● preventing pollution by minimizing greenhouse gas emissions, effluents thus lowering

costs through increased efficiencies, lower material costs and regulatory compliance

● developing new products that require fewer raw materials and can be reused or

remanufactured to lower the overall cost of the product throughout its life from cradle to grave.

● engaging in sustainable practices that minimize the long-term impact of the company’s

operations both on the environment and the society in which it operates through investment in

research to develop technologies that achieve that goal

This is in keeping with the tenets of sustainable development which has been eloquently captured by

Theodore Roosevelt (Roosevelt 2016) as “our duty to the whole including the unborn generations, bids

us to restrain an unprincipled present-day minority from wasting the heritage of these unborn

generations”. In a subsequent follow-up (Hart & Dowell 2011) to the original propositions in Hart

(1995), the authors find that there is a dearth of research in the area of sustainable development and

how it provides companies the necessary competitive advantage to make it more viable. This is despite

prevalence of numerous standards such as ISO 14000 (ISO 2016b) and ISO 14040 (ISO 2016c) which

purport codify the procedures for companies to adopt in order to implement sustainable practices. These

standards have not caught on as much as other ISO standards related to quality such as ISO 9000 (ISO

2016a). However, despite the lack of strong causal relationship (Brunoslash et al 2013) between mass

customization and sustainability, it can be seen that sustainable practices could lead to new products,
 328

improved efficiencies, lower operating costs and increased customer loyalty. There are numerous

benefits to be accrued by combining sustainable initiatives and mass customization initiatives due to

the inherent similarities.

12.3 Sustainability and Mass Customization

As seen in the discussion above, companies stand to gain a lot by applying mass customization

principles in concert with sustainable practices. Both these paradigms are geared towards helping

organization to ultimately lower costs and both often result in new products and processes. It has been

shown (Medini, Da Cunha, Bernard 2012) that demand management is one crucial area when attempting

to incorporate sustainability considerations into customization. Some of the key aspects of both these

paradigms have been examined below with a special emphasis on the synergies to be achieved by

combining both approaches.

12.3.1 Product Design

The design of the product or service is the most crucial element in sustainable development and mass

customization. Products have to be designed to meet the customer’s needs and requirements. At the

same time, designers have to take into consideration sustainability aspects such as impact on

environment, and ability to reuse or recycle. Through appropriate choice of materials and production

processes, the tenets of sustainability can be incorporated into the design of the product. Modular design

which is critical for the success of mass customization (Kumar 2004) also helps with sustainability since

modules could be reused or disassembled for remanufacture. Concepts such as design for engineering

(Glavic & Lukman 2007; Bevilacqua, Ciarapica & Giacchetta 2007) and Environmentally Conscious

Quality Function Deployment (Kaebernick, Kara & Sun 2003) incorporate consideration for sustainable

practices into the regular design process. Some of the popular design approaches include:

● Design for Environment (DfE) - ensures that the environmental impact of the product

both while being manufactured or assembled and being used is taken into consideration during

the design phase.

 329

● Design for Manufacturability (DfM) - focuses on ensuring that the design process

includes consideration for all aspects of the product including manufacturing, assembly,

shipping and service.

● Design for Disassembly (DfD) - emphasizes the ease of taking apart a product after its

end-of-life and reusing its parts. This is particularly relevant and has gained popularity with the

growing emphasis on sustainability.

● Design for Recyclability (DfR) - focuses on alternate uses of parts or products at the

end of their life.

● Design for Reuse - aligns the design process to focus on reusing parts in their original

form after their initial use has ended.

A holistic product design model (Howarth & Hadfield, 2006; Osorio et al 2014) that takes into account

input from all the stakeholders in a combined approach is indispensable. These stakeholders include the

customer, regulatory agencies, community, suppliers and production. In the mass customization

framework, products are often referred to as being co-designed by the designer and the customer. By

incorporating sustainability principles, the design is often termed as eco-design. Thus, through a

combination of co-design and eco-design, companies develop products and services that meet customer

expectations in all aspects while achieving regulatory compliance and meeting their social

responsibilities.

12.3.2 Life Cycle Management

Much of the attention in life cycle management has been on handling the different growth phases (Levitt

1965) of a product such as development or introduction, growth, maturity and finally decline. The focus

is on the external influences on the product demand growth and how the enterprise can handle the

varying demands at a strategic level. However, another related aspect is the process life cycle (Hayes

& Wheelwright 1979) where the focus is on the production process used to fabricate or assemble the

product. As the product evolves through the aforementioned stages, the production process used also

evolve from a purely job-shop where a large variety of products are produced in low volumes to an

automated assembly line where large volumes of a single family of products are assembled.
 330

FIGURE 12.2 Product and Process Characteristic (Adapted from Hayes &Wheelwright, 1979)

Looking at this continuum through the prism of mass customization, it can be seen that this paradigm

is designed to use medium to high volume production processes to manufacture or assemble a large

variety of products as shown in Figure 12.2 adapted from Hayes & Wheelwright (1979). Thus, it

becomes imperative that sustainable manufacturing practices be used through the life cycle of the

customized product. On one end of the spectrum, highly automated systems typically tend to be

designed for efficient use of resources such as raw materials thus minimizing wastes. However, these

fully automated systems can tend to be energy intensive and hence there is a need to examine

efficiencies of these systems and explore alternative sources of energy for running these systems. At

the other end of the spectrum, processes are not as automated and subsequently not as energy intensive.

However, these systems are not designed to be efficient in terms of their use of raw materials primarily

due to the economies of scale. Hence, there is a need to apply lean principles (Womack & Jones, 2003)

to continuously improve the operations in an effort to minimize wastes. Mathematical models have been

developed (Hu & Bidanda 2009) to take into account these various factors influencing the design of the

product throughout its life cycle. Irrespective of the production process chosen, flexibility (Nielson &

Brunoslash 2013a) is an essential capability required in order to be able to quickly respond to changes

in demand. A holistic approach to life cycle management that takes into consideration the impact of

product design, production processes, product usage and subsequent recycling or reuse helps companies

extract maximum value from its products.

 331

12.3.3 Sustainable Production

Production is an important aspect of any manufacturing enterprise and it assumes all the more

importance in a sustainable environment since it is often an energy intensive process too. There have

been different production approaches to implementing mass customization. The most successful

approach is the modular approach where the features of the product are separated into modules and can

be quickly plugged together in different ways to provide a customized product. This approach requires

standardization of parts across multiple products or product families thus providing the economies of

scale required to produce these individual components in large volumes. Concepts from lean principles

such as value-stream mapping, continuous improvement and pull systems can employed to help make

the production process more sustainable. Some of the benefits of applying lean principles to the

production process include:

● Value Identification - Though the concept of value chain has been around for quite some

time (Porter 1985), the popularity of lean principles and its underlying Toyota Production

System has brought renewed focus on value as perceived by the customer and the stages through

the product chain where value is created or added which is also known as value chain. Value-

stream mapping (Womack & Jones 2003) is a technique within the lean principles that help

companies identify the value of the product to its customers and examine the activities that

contribute to that value. Since a lot of the success of mass customization hinges on the

company’s ability to provide more value to its customers than the typical mass-produced

products, knowledge about the value of its products and its value chain can be capitalized to

improve its product offerings. This also helps with the sustainable development since a focus

on value and creating value would automatically result in an efficient use of the available

resources.

● Waste reduction - As per lean principles (Womack & Jones 2003), every activity within

the enterprise can be classified into three broad categories viz. value-added activities, necessary

non-value-added activities and unnecessary non-value-added activities. Clearly, by definition,

 332

the most desirable is the value-added group of activities. By critically examining the entire

value stream in detail, the non-value-added activities can be identified and eliminated.

● Inventory minimization - Eliminating wastes and redundancies in the production

process allows the process to function just in time with minimal if not zero inventory. This is

particularly important in a mass customization framework because of the inherent vagaries in

the product demand. With modular design and just in time assembly inventory of finished goods

can be kept to the bare minimum. This allows companies to respond to changing demands in

an agile manner (Nambiar 2009b).

● Order-driven production - One of the important tenets of lean principles is the kanban

system or pull system. This is essentially an order-driven process where downstream operations

“pull” parts or products from their upstream counterparts who in turn pull from their upstream

operations. This ensures that there is no unnecessary inventory build-up. This is again of utmost

importance in a mass customization environment since companies would not want to be saddled

with inventory of customized products with no demand.

12.3.4 Sustainable Packaging

One of the eight generic levels of mass customization identified by Da Silveira, Borenstein & Fogliatto

(2001) is packaging. This correlates to the “cosmetic” face among the four faces of mass customization

as put forth by Gilmore and Pine (1997). In this form of customization, the focus is on how the product

is packaged for each segment of customer. In many cases, the product inside is the same irrespective of

the external packaging. For example, candies may be wrapped in orange and black during Halloween

or in red and green during Christmas. The product may also be available in different quantities. For

example, the same breakfast cereal may be available in to-go containers for one-time use or in family-

size value boxes. This is one of the easily achieved forms of customization since there is very little

change in the product and hence more widely adopted. However, with the growing emphasis on

sustainability, companies are looking towards minimizing the use of materials for packaging their

products. This focus on sustainable packaging has multiple advantages such as:
 333

● Reduced cost - In many cases, examining the packaging can help identify better and

cheaper alternatives that are more sustainable thus lowering costs.

● Faster shipping - As a result of the efficient packaging, more products can be fit into a

standard shipping container thus facilitating more products to be shipped at once and at the

same price.

Companies have already begun finding innovative ways to implement sustainable packaging. For

example, Amazon’s frustration-free packaging (Amazon 2016) ships many products in their original

box thus saving on costs while providing customers a pleasant experience opening packages. Apple Co.

is purported to have a separate packaging room (Lashinsky 2012) where researchers spend countless

hours trying to get the packaging right. Some of the mechanisms of implementing sustainable packaging

include:

● Minimalist packaging - Minimalist design can be achieved by reducing the packaging

to the bare minimum necessary to ensure that the product arrives in good condition and it

provides customers a good experience unravelling to reveal the product. This ensures that very

little raw material is used to begin with.

● Recycled material - Another approach is to use recycled material in product packaging.

This may be through the use of recycled paper or plastic.

● Recyclable - A third aspect is ensuring that all the packaging material is recyclable.

Some companies go beyond simply ensuring that the material is recyclable by offering to collect

the packaging material to be recycled.

Companies use one or more of these approaches in an effort to ensure minimum impact on the

environment while providing customized products to meet customer demand.

12.3.5 Sustainable Supply Chain Management

Improving customer satisfaction and streamlining material flow rank high among the critical factors

motivating companies towards an efficient supply chain management (Tummala, Phillips & Johnson,

2007). In a mass customization environment, the problem of material flow is compounded many times
 334

over due to the sheer variety of products. Thus, an efficient supply chain management is crucial for the

success of mass customization initiatives. Increasing pressures from environment-conscious customers

and regulating bodies combined with the ensuing competitive advantage, companies are exploring ways

to make their supply chain more sustainable. This necessitates a closed-loop supply chain (Nielson &

Brunoslash 2013b) that takes into account sustainable issues such as energy consumption and wastes

along with other aspects of a typical supply chain such as cost, time and quality.

Sustainable or Green supply chain management has been defined (Srivastava 2007) as “integrating

environmental thinking into supply chain management including product design, material sourcing and

selection, manufacturing processes, delivery of the final product to the consumers as well as end-of-life

management of the product after its useful life”. An integral part of the green supply chain management

is the reverse supply chain which includes five main activities (Guide & van Wassenhove, 2002) geared

towards the reverse flow of material:

● acquiring the product through an efficient collection system spread across all supply

chain partners,

● transporting the collected products to a central facility,

● inspectingand sorting of the products and their constituent parts,

● reusing and/or remanufacturing parts and finally

● distributing recycled or remanufactured products.

The large number of products in a mass customized industry has adverse effects (Shui-Mu & Su 2013)

on the reverse supply chain since it makes the aforementioned steps in the process all the more arduous.

However, in order to stay competitive in today’s market characterized by stiff competition and shrinking

profit margins, it is imperative that companies strive to recover as much value from its products once

they reach their end of life. In order to achieve this, it is imperative to create a strong corporate culture

of cooperation and collaboration among supply chain partners (Trappey & Wognum 2012, Liu 2013).

It is also important to have the necessary support system to achieve this collaboration in the form of a

robust information system and an efficient communication network. There is also an increased emphasis

on the need for the supply chain to be more agile to be able to respond to uncertainties (Tachizawa &
 335

Thomsen 2007) in product demand in terms of variety and quantity and the uncertainties in deliveries

given the global nature of the market.

12.3.6 Sustainable Logistics

The Council of Supply Chain Management Professionals (Ballou 2007) define logistics as “..that part

of supply chain management that plans, implements, and controls the efficient forward and reverse flow

and storage of goods, services, and related information between the point of origin and point of

consumption in order to meet customer requirements”. Logistics assumes increased importance in a

mass customization environment because of the sheer complexity of the operations with increased

product variety. Hence, an efficient logistics system is indispensable (Gooley 1998) in a mass

customization framework. It is also important for the logistics system to be flexible (Hamid 2015) in

today’s truly globalized and highly volatile markets. The system should be able to quickly respond to

the uncertainties inherent in a mass customization framework. This response also needs to be in an

efficient manner so that companies can still stay competitive. This underscores the need for sustainable

logistics with increased focus on the use of “clean” vehicles with zero or low emissions. This is also

especially true because transportation accounts for a significant percentage of the greenhouse gas

emissions released into the atmosphere.

Another area of growing importance is reverse logistics. Reverse logistics has been defined (Rogers &

Tibben-Lembke 1999) as “the process of planning, implementing and controlling the efficient cost-

effective flow of raw materials, in-process inventory, finished goods, and related information from the

point of consumption to the point of origin for the purpose of recapturing or creating value, or for proper

disposal”. This is an integral part of the reverse supply chain that is essential for sustainable use of

limited resources. Reverse logistics typically involves a network of players each performing a vital

function of the process. The vital functions in reverse logistics (Srivastava 2015) are:

● Collection - Once a product has been expended, it is important to urge consumers to

return the product to the manufacturer in order to maximize the value of the product. This

process of returning goods to the manufacturer needs to be easy and economical in order to

motivate consumers. Many establishments have begun having collection points at various
 336

locations for consumers to return their no longer needed products. The geographic location of

these collection centers is critical in order to make it effective.

● Inspection - The collected products have to be inspected to evaluate their current state.

This is important because consumers often return products in various stages of disarray which

determines the recovery process.

● Sorting - The products are then sorted into various categories based on material

recoverable, ease of recovery, etc.

● Processing - The products often undergo some sort of processing to recover parts or

raw materials to be recycled or reused. In some cases, it might be as simple as taking apart the

product and cleaning up the parts for reuse while in other cases the parts may have to be

processed into raw materials for alternative uses.

Supporting these various functions is the underlying logistical support and distribution network that is

so critical to successful implementations of reverse logistics. It can be seen that there are a lot of

common features between the forward supply chain and reverse supply chain the primary being the

distribution network. One important difference though is that often times speed is not an issue in reverse

logistics whereas it is of utmost important in the forward supply chain. Despite the critical difference,

there are numerous advantages to be accrued through the synergy of forward and reverse supply chain

networks. For example, as the trucks make deliveries in the forward supply chain, they can also be

collecting recyclable and reusable products from the various collection centers. Reverse logistics is

particularly important in the electronics industry especially with the rapid proliferation of mobile

phones and other computing devices. It has been shown (Srivastava 2015) that implementing an

efficient reverse logistics mechanism and in turn a reverse supply chain system can provide a

competitive advantage through increased return on investment and improved customer image.

12.3.7 Information Systems

Successful implementation of sustainable and mass customization practices require an integrated

information system (Mahajan, Srinivasan & Wind, 2002; Frutos & Borenstein, 2004; Dean, Tu & Xu,

2008; Ngiatedema, 2012) that connects all the value chain partners while providing an effective
 337

communication network for sharing information (Bai & Gu 2010). It can be seen that supply chain

integration is crucial for mass customization due to the sheer diversity of products and the volatility of

demand. In order to be able to quickly respond to changing customer demands, it is imperative to be

able to communicate the customer needs through the entire value chain. This becomes even more critical

with numerous parent companies adopting drop shipping where customer orders are often fulfilled

directly by the supplier. With increased emphasis on reverse supply chain as a result of the growing

impetus for recycling and remanufacturing parts, it is essential that supply chain partners work closely

with each other. Information systems serve as the key enabler in providing the infrastructural support

needed for increased collaboration and cooperation between value chain partners. There are two

approaches to building the necessary information systems infrastructure:

● Centralized - A centralized networked system with real-time monitoring will allow for

informed and instantaneous decision-making regarding product mixes, production volumes,

product distribution and collection.

● Decentralized - In this scenario, each player in the value chain implements their own

information system with the caveat that these systems should be able to communicate and

interact with other systems in the value chain. This can be accomplished by adopting commonly

accepted design patterns and ensuring uniform standards for data encoding.

In either case, it is essential that the information system is capable of being scaled up and down to reflect

the growth patterns of the organization. It is also important for the system to be flexible and easily

adaptable to the changing market conditions. Ensuring data security and integrity is also an essential

feature of the information system. The challenges involved in building such an information system

scaffolding include:

● Diverse needs - The requirements of individual players are so diverse and disparate that

it is difficult for a single unified system to address all of the needs.

● Legacy systems - Many organization already have an existing technology infrastructure

that they have been using for many years and continue to use them because all their data

conforms to that system. It is often difficult for these legacy systems to interact with newer

systems without intermediate data exchange mechanisms

 338

● Piecemeal implementations - For many years, information systems and technology was

more of an afterthought and hence piecemeal implementations have been put in place by

organizations to address specific needs as they arise thus making seamless integration an

onerous task.

● Multitude platforms and devices - The proliferation of mobile devices of all forms and

sizes combined with the burgeoning range of platforms catalyzes into cyclopean combinations

which engenders compatibility issues.

12.4 Future Research Direction

Even though sustainability and mass customization concepts have been around for more than a few

decades, there has been a heightened sense of urgency and need to focus on these aspects in today’s

world due to numerous reasons such as stiff global competition, and stringent environmental regulations

to name a few. Both these concepts have significant overlap in their reliance on robust information

systems, agile supply chains and efficient product design. In this chapter, some of the benefits and

challenges involved have been identified. There needs to be more research into the benefits of

combining the two concepts based on actual evidence. Though concepts such as design for excellence

(DfX) and its offshoots such as design for environment (DfE) and design for manufacturability (DfM)

have been applied in isolation for specific objectives such as improving production efficiency or

reducing environmental impact, there needs to be greater integration between these systems to facilitate

efficient and effective product design. Reverse logistics is a vital and significant component of

sustainable supply chain management and the synergies afforded by combining forward and reverse

logistics need to be explored and exploited. Since a lot of the success of these principles rely on

information sharing and exchange, it is important to have well-defined industry wide data formats that

allow individual disparate systems to communicate with each other across the multitude of platforms

and provide a seamless user experience across the various devices ranging from handhelds to desktops.

As with any new concept or paradigm, change management is a significant issue especially if it requires

a complete overhaul of the existing operations. In order to reduce some of the implementation throes,
 339

it might be also be useful to better communicate and share some of the best practices from companies

that have been successful since sustainability is about collective good for future generations.

12.5 Conclusion

Globalization has provided customers world over access to a wide variety of products while providing

enterprises access to new sources of raw materials and skilled workforce. This has made customers

more demanding thus engendering stiff competition among companies to gain market share. Mass

customization is gaining popularity in this milieu since it purports to provide customized products and

services at mass production prices. However, in this mad rush to drive down costs in order to stay

competitive, many enterprises have become highly distributed with operations in many parts of the

world and highly diversified with a plethora of products. This has created a supply chain nightmare.

With global warming and drastic climate changes around the world, there is increased attention on the

environmental impact of human development. Sustainable development has assumed paramount

importance and companies are striving hard to reduce the environmental footprint of their operations.

A synergy of sustainable development and mass customization could lead to efficient product or service

design, agile and flexible supply chain and robust information systems thus providing companies the

much needed competitive advantage to survive in this era of globalization.

REFERENCES

Abraham, M. (2003). Random thoughts on sustainability. Environmental Progress. Volume 25. No. 1.
pp 9 - 10.

Amazon (2016). About Amazon certified frustration-free packaging. Retrieved from

https://www.amazon.com/gp/help/customer/display.html?nodeId=200285450. Accessed on January
07, 2016.

Bai, J. & Gu, C. (2010). Research on model of information sharing in mass customization supply
chain. Proceedings of the 2010 International Conference of Logistics Engineering and Management:
Logistics for Sustained Economic Development - Infrastructure, Information, Integration, Chengdu,
China. pp 4224 - 4229.

Ballou, R. H. (2007). The evolution and future of logistics and supply chain management. European
Business Review. Volume 19. No. 4. pp 332 - 348.
 340

Bevilacqua, M., Ciarapica, F.E. & Giacchetta, G. (2007). Development of a sustainable product
lifecycle in manufacturing firms: a case study. International Journal of Production Research. Volume
45. No. 18-19. pp 4073 - 4098.

Brunoslash, T.D., Nielsen, K., Taps, S.B. & Joslashrgensen, K.A. (2013). Sustainability evaluation of
mass customization. Proceedings of the International Conference on Advances in Production
Management Systems, State College, PA. Volume 414. pp 175 - 182.

Da Silveira, G., Borenstein, D. & Fogliatto, F. (2001). Mass customization: literature review and
research directions. International Journal of Production Economics. Volume 72. No. 1. pp 1- 13.

Davis, S. (1987) From future perfect: mass customizing. Planning Review. Volume 17. No. 2. pp 16 -
21.

Dean, P.R., Tu, Y.L.,Xue, D. (2008) A framework for generating product production information for
mass customization. International Journal of Advanced Manufacturing Technology. Volume 38. Issue
11-12. pp 1244 - 1259.

Frutos, J.D. & Borenstein, D. (2004). A framework to support customer-company interaction in mass
customization environments. Computers in Industry. Volume 54. pp 115 - 135.

Fullana i Palmer, P., Puig, R., Bala, A., Baquero, G., Riba, J., & Raugei, M. (2011). From Life Cycle
Assessment to Life Cycle Management. Journal Of Industrial Ecology, 15(3), 458-475.

Gilmore, J. H. & Pine, B.J. (1997). The four faces of mass customization. Harvard Business Review.
Volume 75. Issue 1. pp 91 - 91.

Glavic, P. & Lukman, R. (2007). Review of sustainability terms and their definitions. Journal of
Cleaner Production. Volume 15. pp 1875 - 1885.

Gooley, T. (1998). Mass customization: how logistics makes it happen. Computers and Industrial
Engineering. Volume 37. Issue 4. pp 49 - 54.

Guide, V.D.R. van Wassenhove, L.N. (2002). The Reverse Supply Chain. Harvard Business Review.
Volume 18. Issue 2. pp 25 - 26.

Hamid, J. (2015). Logistics flexibility - a systematic review. International Journal of Productivity and
Performance Management. Volume 64. Issue 7. pp 947 - 970.

Hart, S.L. (1995). A natural resource-based view of the firm. Academy of Management Review. Volume
20. Issue 4. pp 986 - 1014.

Hart, S.L. & Dowell, G. (2011). A natural-resource-based view of the firm - fifteen years after. Journal
of Management. Volume 37. Issue 5. pp 1464 - 1469.

Hayes, R.H. & Wheelwright, S.C. (1979). Linking manufacturing process and product life cycles.
Harvard Business Review. Volume 57. pp 133 - 140.

Howarth, G. & Hadfield, M. (2006). A sustainable product design model. Materials and Design.
Volume 27. pp 1128 - 1133.

Hu, G. & Bidanda, B. (2009). Modeling sustainable product lifecycle decision support systems.
International Journal of Production Economics. Volume 122. pp 366 - 375.
 341

ISO (2016a). ISO 9000 - Quality management systems. Retrieved from

http://www.iso.org/iso/home/store/catalogue_tc/catalogue_detail.htm?csnumber=45481. Accessed on
January 08, 2016.

ISO (2016b). ISO 14000 - Environmental Management. Retrieved from

http://www.iso.org/iso/home/standards/management-standards/iso14000.htm. Accessed on January 03,
2016.

ISO (2016c). ISO 14040:2006 - Environmental Management: Life Cycle Assessment - principles and
framework. Retrieved from http://www.iso.org/iso/catalogue_detail?csnumber=37456. Accessed on
January 03, 2016.

Kaebernick, H., Kara, S. & Sun, M. (2003). Sustainable product development and manufacturing by
considering environmental requirements. Robotics and Computer Integrated Manufacturing. Volume
19. pp 461 - 468.

Kumar, A. (2004). Mass customization: metrics and modularity. International Journal of Flexible
Manufacturing Systems. Volume 16. Issue 4. pp 287 - 311.

Lampel, J. & Mintzberg, H. (1996). Customizing customization. Sloan Management Review. Volume
38. Issue 1. pp 21 - 30.

Lashinksy, A. (2012). Inside Apple: How America’s most admired and secretive company really works.
John Murray Publishers, London, UK.

Levitt, T. (1965). Exploit the product life cycle. Harvard Business Review. Volume 43. pp 81- 94.

Liu, L. (2013). Supply chain management under mass customization. Proceedings of the 2nd
International Conference on Energy, Environment, and Sustainable Development, Jilin, China.
Volume 616 - 618. pp 2044 - 2047.

Mahajan, V., Srinivasan, R. & Wind, J. (2002). The Dot.com retail failures of 2000: Were there any
winners? Journal of the Academy of Marketing Science. Volume 30. Issue 4. pp 474 - 486.

Medini, K., Da Cunha, C. & Bernard, A. (2012). Sustainable mass customized enterprise: Key concepts,
enablers and assessment techniques. Proceedings of the 14th IFAC Symposium on Information Control
Problems in Manufacturing, Bucharest, Romania. Volume 14. pp 522 - 527.

Nambiar, A.N. (2009a), Mass Customization - Where do we go from here?, Proceedings of the 2009
World Congress on Engineering, London, U.K, July 1 - 3, Vol I.

Nambiar, A.N. (2009b), Agile Manufacturing: A Taxonomic Framework for Research, Proceedings
of the 2009 International Conferences on Computers and Industrial Engineering, July 6 - 9, pp 684 -
689. Nambiar, A.N. (2010), Modern Manufacturing Paradigms - A Comparison, Proceedings of the
2010 International Multiconference of Engineers and Computer Scientists, Hong Kong, March 17 -
19, vol III.

Nielsen, K. & Brunoslash, T.D. (2013a). Assessment of process robustness for mass customization.
Proceedings of the International Conference on Advances in Production Management Systems, State
College, PA. Volume 414. pp 191 - 198.
 342

Nielsen, K. & Brunoslash, T.D. (2013b). Closed loop supply chains for sustainable mass
customization. Proceedings of the International Conference on Advances in Production Management
Systems, State College, PA. Volume 414. pp 425 - 432.

Ngniatedema, T. (2012). A mass customization information systems architecture framework. The

Journal of Computer Information Systems, Volume 52. Issue 3. pp 60-70.

Osorio, J., Romero, D., Betancur, M. & Molina, A. (2014). Design for sustainable mass-customization:
design guidelines for sustainable mass-customized products. Proceedings of the International
Conference on Engineering, Technology, and Innovation: Engineering Responsible Innovation in
Products and Services, Bergamo, Italy.

Pham, D.T., Pham, P.T.N. and Thomas, A. (2008), Integrated production machines and systems -
beyond lean manufacturing. Journal of Manufacturing Technology Management , vol. 19, no. 6, pp.
695–711.

Pine, B.J. (1993). Mass customizing products and services. Planning Review. Volume 21. Issue 4. pp
6 - 13.

Pine, B.J. (1999). Mass customization: the new frontier in business competition. Harvard Business
School Press, UK.

Porter, M.E. (1985). Competitive Advantage. The Free Press.

Porter, K., Little, D., Peck, M., Rollins, R. (1999). Manufacturing Classifications: relationships with
production control systems. Integrated Manufacturing Systems, Vol. 10, No. 4, pp 189 - 199.

Rogers, D.S. & Tibben-Lembke, R.S. (1999). Going backwards: reverse logistics trends and practices.
RLEC Press, Pittsburgh, PA.

Roosevelt, T. (2016). Theodore Roosevelt Association. Retrieved from

http://www.theodoreroosevelt.org/ Accessed on January 09 2016.

Rusinko, C.A. (2008). Towards more sustainable management systems: through life cycle management
and integration., Journal of Cleaner Production, vol. 16, pp. 1071–1080.

Shui-Mu, H. & Su, J.C.P. (2013). Impact of product proliferation on the reverse supply chain. Omega.
Volume 41. pp 626 - 639.

Srivastava, S. K. (2007). Green supply chain management - a state-of-the-art literature review.

International Journal of Management Reviews. Volume 9. Issue 1. pp 53 - 80.

Srivastava, S. K. (2015). Network design for reverse logistics. Omega. Volume 36. Issue 4. pp 535 -
548.

Tachizawa, E.M.& Thomsen, C. G. (2007). Drivers and sources of supply flexibility: an exploratory
study. International Journal of Operations & Production Management. Volume 27. No. 10. pp 1115 -
1136.

Trappey, A.J.C. & Wognum, P.M. (2012). Network and supply chain system integration for mass
customization and sustainable behavior. Advanced Engineering Informatics. Volume 26. Issue 1. pp 3
- 4.
 343

Tummala, V.M.R. , Phillips, C.L.M. & Johnson, M. (2006). Assessing supply chain management
success factors: a case study. Supply Chain Management: An International Journal , Volume. 11. No.
2, pp.179-192.

Womack, J.P. and Jones, D.T. (2003). Lean Thinking: Banish Waste and Create Wealth in Your
Corporation, Simon & Schuster, New York, NY.

Womack, J.P., Jones, D.T., and Roos, D. (1999), The Machine that Changed the World, Rawson
Associates, New York, NY.

UN (1987), Report on the World Commission on Environment and Development: Our Common Future,
http://www.un-documents.net/our-common-future.pdf, Accessed on January 03, 2016.

View publication stats