Interdisciplinary Contributions to Theory for Collaborative Networks 9781781904664, 9781781904657

Citation preview

jmtm cover (i).qxd

05/10/2012

12:58

Page 1

ISSN 1741-038X

Volume 23 Number 8 2012

Journal of

Manufacturing Technology Management Interdisciplinary contributions to theory for collaborative networks Guest Editors: Rob Dekkers and Hermann Kühnle

www.emeraldinsight.com

Journal of Manufacturing Technology Management

ISSN 1741-038X Volume 23 Number 8 2012

Interdisciplinary contributions to theory for collaborative networks Guest Editors Rob Dekkers and Hermann Ku¨hnle

Access this journal online _________________________

959

Editorial board ___________________________________

960

CONTENTS

GUEST EDITORIAL Some thoughts on interdisciplinarity in collaborative networks’ research and manufacturing sciences Hermann Ku¨hnle and Rob Dekkers ________________________________

961

The inter-disciplinary modelling of supply chains in the context of collaborative multi-structural cyber-physical networks Dmitry Ivanov and Boris Sokolov _________________________________

976

Entropy assessment of supply chain disruption Olatunde Amoo Durowoju, Hing Kai Chan and Xiaojun Wang_________________________________________________

998

A model to determine complexity in supply networks Markus Gerschberger, Corinna Engelhardt-Nowitzki, Sebastian Kummer and Franz Staberhofer__________________________ 1015

Alignment prediction in collaborative networks Roberto da Piedade Francisco, Ame´rico Azevedo and Anto´nio Almeida_______________________________________________ 1038

Habitual domain exploration in inter-firm networks: a framework for understanding network behaviour Lei Ma, Yongjiang Shi and Wenwen Zhao __________________________ 1057

Towards the explanation of goal-oriented and opportunity-based networks of organizations Jens Eschenba¨cher and Novica Zarvic´ ______________________________ 1071

This journal is a member of and subscribes to the principles of the Committee on Publication Ethics

CONTENTS continued

Appraising interdisciplinary contributions to theory for collaborative (manufacturing) networks: still a long way to go? Rob Dekkers and Hermann Ku¨hnle ________________________________ 1090

Lessons learned from the lifecycle management of collaborative enterprises networks: the case of Swiss Microtech Naoufel Cheikhrouhou, Michel Pouly, Charles Huber and Jean Beeler ___________________________________________________ 1129

www.emeraldinsight.com/jmtm.htm As a subscriber to this journal, you can benefit from instant, electronic access to this title via Emerald Management eJournals. Your access includes a variety of features that increase the value of your journal subscription.

Structured abstracts Emerald structured abstracts provide consistent, clear and informative summaries of the content of the articles, allowing faster evaluation of papers.

How to access this journal electronically

Additional complementary services available

Our liberal institution-wide licence allows everyone within your institution to access your journal electronically, making your subscription more cost-effective. Our web site has been designed to provide you with a comprehensive, simple system that needs only minimum administration. Access is available via IP authentication or username and password.

When you register your journal subscription online you will gain access to additional resources for Authors and Librarians, offering key information and support to subscribers. In addition, our dedicated Research, Teaching and Learning Zones provide specialist ‘‘How to guides’’, case studies, book reviews, management interviews and key readings.

To benefit from electronic access to this journal, please contact [email protected] A set of login details will then be provided to you. Should you wish to access via IP, please provide these details in your e-mail. Once registration is completed, your institution will have instant access to all articles through the journal’s Table of Contents page at www.emeraldinsight.com/ 1741-038X.htm More information about the journal is also available at www.emeraldinsight.com/jmtm.htm

E-mail alert services These services allow you to be kept up to date with the latest additions to the journal via e-mail, as soon as new material enters the database. Further information about the services available can be found at www.emeraldinsight.com/alerts

Emerald online training services Visit www.emeraldinsight.com/help and take an Emerald online tour to help you get the most from your subscription.

Emerald Research Connections An online meeting place for the world-wide research community, offering an opportunity for researchers to present their own work and find others to participate in future projects, or simply share ideas. Register yourself or search our database of researchers at www.emeraldinsight.com/connections

Key features of Emerald electronic journals

Choice of access

Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. within the institution only. This only applies to articles of which Emerald owns copyright. For further details visit www.emeraldinsight.com/copyright Online publishing and archiving As well as current volumes of the journal, you can also gain access to past volumes on the internet via Emerald Management eJournals. You can browse or search these databases for relevant articles. Key readings This feature provides abstracts of related articles chosen by the journal editor, selected to provide readers with current awareness of interesting articles from other publications in the field. Non-article content Material in our journals such as product information, industry trends, company news, conferences, etc. is available online and can be accessed by users. Reference linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article. E-mail an article Allows users to e-mail links only to relevant and interesting articles to another computer for later use, reference or printing purposes.

Electronic access to this journal is available via a number of channels. Our web site www.emeraldinsight.com is the recommended means of electronic access, as it provides fully searchable and value added access to the complete content of the journal. However, you can also access and search the article content of this journal through the following journal delivery services: EBSCOHost Electronic Journals Service ejournals.ebsco.com Informatics J-Gate www.j-gate.informindia.co.in Ingenta www.ingenta.com Minerva Electronic Online Services www.minerva.at OCLC FirstSearch www.oclc.org/firstsearch SilverLinker www.ovid.com SwetsWise www.swetswise.com

Emerald Customer Support For customer support and technical help contact: E-mail [email protected] Web http://info.emeraldinsight.com/products/ subs/customercharter.htm Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785201

JMTM 23,8

960

EDITORIAL REVIEW BOARD

Azmawani Abd Rahman University Putra Malaysia, Malaysia

Doug Love Aston University, UK

Lynne Baxter University of York, UK

Douglas Macbeth University of Southampton, UK

Nourredine Boubekri University of North Texas, USA

Bart MacCarthy University of Nottingham, UK

Felix Chan The Hong Kong Polytechnic University, Hong Kong

Shunji Mohri Hokkaido University, Japan

Marly Monteiro de Carvalho Universidade de Sa˜o Paulo, Brazil

Andrew Neely University of Cambridge, UK

Ian Gibson National University of Singapore, Singapore

Adegoke Oke Arizona State University, USA

Angappa Gunasekaran University of Massachusetts Dartmouth, USA Jinsheng He Tianjin University, People’s Republic of China Abdel-Aziz Hegazy Helwan University, Egypt Robert Hollier University of Manchester, UK Tarek Khalil Nile University, Egypt Ashok Kochhar Aston University, UK

Kulwant Pawar University of Nottingham, UK Roy Snaddon Polytechnic of Namibia, Namibia Amrik Sohal Monash University, Australia Harm-Jan Steenhuis Eastern Washington University, USA Mile´ Terziovski University of South Australia, Australia Juite Wang National Chung Hsing University, Taiwan

Siau Ching Lenny Koh University of Sheffield, UK Hermann Ku¨hnle Otto-von-Guericke-Universita¨t Magdeburg, Germany

Journal of Manufacturing Technology Management Vol. 23 No. 8, 2012 p. 960 # Emerald Group Publishing Limited 1741-038X

The current issue and full text archive of this journal is available at www.emeraldinsight.com/1741-038X.htm

GUEST EDITORIAL

Some thoughts on interdisciplinarity in collaborative networks’ research and manufacturing sciences

Some thoughts on interdisciplinarity

961

Hermann Ku¨hnle IAF, Otto-von-Guericke University Magdeburg, Magdeburg, Germany, and

Rob Dekkers UWS Business School, University of the West of Scotland, Paisley, UK Abstract Purpose – Scientific progress in a field is mostly discussed within disciplines. Far less attention is paid to outside or between disciplines’ work. To speed up research progresses for collaborative networks (CN) in manufacturing, a base for further grounded theory establishment is propagated, recalling some of the most relevant chapters of philosophy of science. The focus is put onto the roles of disciplines and their scholars involved in interdisciplinary contexts, in order to further motivate as well as to hint at a number of catalysing forces and fruitful impacts of outside disciplines’ work. Design/methodology/approach – The intentions of this Special Issue are mirrored to important and well-accepted findings in the philosophy of science. All papers that are included in this journal issue are positioned within a general framework of scientific disciplines and theory building understanding. Findings – Interdisciplinary work is speeding up theory building and innovation in CNs in general and in all applications for manufacturing in particular. In order to encourage publications of project work and solutions that do not neatly fit into the scientific disciplines set up, it is pointed out that exactly these papers have the potential to unveil unattended and valuable insights. This kind of outline often confirms both gut feelings of managers, as well as vague hypotheses of researchers and scientists. Research limitations/implications – The paper shows that more attention might be paid to outside contributions and to mechanisms to increase their impact on theory building in manufacturing science. Originality/value – For the field of CN, the paper represents a first and unique attempt to enhance scientific progress by emphasising theory contributions from other disciplines. The approach contributes to theoretically as well as methodically supporting the fast growing number of practical solutions beyond state of art. Keywords Manufacturing industries, Sciences, Research methods, Philosophy of science, Theory building, Scientific disciplines Paper type General review

In search for competitive excellence in manufacturing, collaborative networks (CNs) have received much attention during recent years. Understanding and anticipating on network characteristics in “product design and engineering”[1] and manufacturing creates potentially competitive advantage for the firms that participate in those CNs. However, concepts for CNs, typologies and enabling software have generated mostly isolated solutions to problems so far. Large portions of the acquired knowledge about CN

Journal of Manufacturing Technology Management Vol. 23 No. 8, 2012 pp. 961-975 q Emerald Group Publishing Limited 1741-038X DOI 10.1108/17410381211276826

JMTM 23,8

962

are cast into rather singular models or solution-oriented procedures uniquely based on case experiences and anecdotal verifications that need further validation to establish their scope beyond instances (see Wacker, 1998 for how theory is formed in operations management). These phenomena will even become more manifest since major trends in manufacturing are distributed organisational and geographically dispersed structures (Kuehnle, 2007), more loosely coupled entities for industrial networks (Dekkers and Bennett, 2009) and fragmentation of manufacturing and control processes (Kuehnle, 2007). Hence, we posit that incoherent approaches originating from these different aspects have led only to heterogeneous and inconsistent fragments for consolidated knowledge about CNs or for underpinning theory. At the very same time, there is an ongoing debate concerning the nature of theory necessary for manufacturing networks, especially footing on CNs. The once clearly defined domain of manufacturing science has to recognise that only increasingly contributions drawn from other disciplines might ensure more commanding generic concepts, models or theoretical approaches for CNs. Among the eligible disciplines that might make a worthy contribution we enumerate theory from complex adaptive systems, decision sciences, evolutionary biology, game theory, organisational theory, sociology and topology, alongside more traditional approaches from data exchange and network management. Within the narrower scope of management sciences, network management has established itself as a field for CNs. And concepts from data exchange are seen as relevant to structured communication protocols. But within all these disciplines and fields attempts have been made to investigate and to describe phenomena of collaboration in networks and related changes that are taking place in industrial entities. However, to-date there has not been published an edited, collective account of the different perspectives that exist among the various academic and industrial research communities towards the new science that might emerge, even though Camarinha-Matos and Afsarmanesh (2005) have called desperately for furthering insight; this special issue aims to remedy this gap. 1. Collaborative (manufacturing) networks – stuck between disciplines? Rien ne va plus – Anything goes.

But even though the contributions of the special issue aim to fill this void, one imperative question beyond the state-of-the-art for researchers as well as managers working with and within company networks emerges: what direction will manufacturing and management science take within this setting of networks? That requires looking first at where manufacturing science came from. Since the 1990s, manufacturing science and management have been advancing at fast pace, echoing Buffa’s (1980) call. After the propositions of Skinner (see Dekkers and Bennett, 2009 for placing it in the context of industrial networks) and Drucker that manufacturing matters, its management has become more conceptual and has grown into the product of principles and a number of practices, together seen as new approaches; lean production is a case in point. Many of these new approaches already went beyond the limits of the systems’ thinking as its restrictions appeared critical during the 1980s, even though recognised later (Forrester, 1994). Hence, in the early 1990s a plea for empirical research based on quantitative analysis emerged (Flynn et al., 1990; Swamidass, 1991), inspired by a social science perspective to arrive a theory; a re-iteration of this position came regularly about (Bertrand and Fransoo, 2002; Forza, 2002) and proved popular for advancing operations

management science (Filippini, 1997; Rungtusanatham et al., 2003). That was followed by a call for the antidote, the case study research methodology (McCutcheon and Meredith, 1993; Meredith, 1998), which picked up strongly in the beginning of the 2000s (Stuart et al., 2002) in combination with action research (Coughlan and Cogghlan, 2002; Westbrook, 1995). These developments led also to taking best practice as source for new approaches (Voss et al., 2002) but this type of research has strong limitations (Davies and Kochhar, 2002). Hence, this diversification of research methods (supported by Boyer’s and Swink’s (2008) position) resulted in either generalisations yielding limited insight (quantitative research) or solutions for specific circumstances with little attention for contingencies (as noted by Sousa and Voss, 2008). With the increasing number of publications over the past 30 years or so, the domain of manufacturing management has become entrenched in research philosophies and conceptual approaches, with no direct alternative to the aspirations of (general) systems theories as metatheory. In the meanwhile, it has become accepted more commonly that other disciplines, such as network theories or complexity thinking (Wiendahl and Scholtissek, 1994), are seen to be much more adequate to address recent challenges for manufacturing and its management than the widespread general systems thinking approach. Moreover, shifts in perceptions about manufacturing are regularly inducing paradigmatic debates often pointing at social, resources, technological dimensions, etc. that should be included stronger and hence demanding the widening of the scope. On the other hand neither established production and manufacturing technology nor management sciences do seriously deny that their respective bodies of knowledge clearly hit limitations in applicability for an ever wider range of phenomena and loudly encourage further incorporating inter-disciplinary approaches and novel dimensions achieved by other disciplines. Prominent examples are the best practice concept of lean production being well embedded in Theory of Constraints ( Jacob et al., 2009) and agile manufacturing (Christopher, 2010), relying on reconfiguration principles found in nature. Consequently, considerable work in the domain of (dispersed) manufacturing networks has already been undertaken at intersections with other disciplines and fields of knowledge for providing a solid scientific base or at least for a more theoretic foundation. However, attempts to develop more advanced theories for collaborative (manufacturing) networks, too, face deep-rooted challenges for interdisciplinary work. First, there is the extension of the validity of construct resulting from the application to aspects from the complex nature of CNs. Such extensions of validity face opposition from traditionalists whose research instruments and approaches are confined to the prevailing notions and application; examples are the non-foreseeable market movements that only immediate restructuring can cope with, but not the more common modus operandi of high-frequency adapted traditional planning. Second, disciplines already active in research work on CNs, such as social sciences and network’s software agents, tend to claim all manufacturing networks as more or less trivial cases within their sciences and thus strongly marginalise outcomes of interdisciplinary research. Consequently, interdisciplinary research appears less attractive for promoting researchers (Rhoten and Parker, 2004), because its standing is lowering their benefits from the results and decreasing their motivation for further theorising (an example is Campbell’s (2005) cry for help). In addition, interdisciplinary researchers face personal difficulties, as researchers rooted in more than one field may experience disciplinary critiques as the pressure on researchers in most disciplines to keep strictly engaging

Some thoughts on interdisciplinarity

963

JMTM 23,8

964

in “normal puzzle-solving science” is very strong (Alvesson and Sko¨ldberg, 2000; Pickett et al., 1999). Following such academic pressure to remain within traditionally established disciplines is financially and professionally rewarding and suppresses interdisciplinary outcomes. Interdisciplinary work therefore merits support, a quest that this special issue aims to support, and individual researchers should be commended for their courage to pursue such an avenue. The practice of problem solving however shows that interdisciplinary researchers do neither see themselves as breaching disciplinary walls nor crossing disciplinary boundaries, but as conducting negotiations across different groups of disciplines (Aboelela et al., 2006). Above all, they experience largely a high level of acceptance of their work by managers and other stakeholders. For CNs in manufacturing, therefore, interdisciplinary approaches have resulted frequently in quite a number of literally “unnoticed” excellently executed strategies with brilliant implementations, for example, the owner-managed German “Hidden Champions” prove (Simon, 2012; Venohr and Meyer, 2007); evidently these solutions do not attempt gaining broader interest in manufacturing and management science. Some come up accidentally and much later, as successful global capacity loading practices for global car assembly in automotive industry or flexible supply networks in local/regional small- and medium-sized enterprise contexts. Their influences on manufacturing and management science, therefore, remain limited and possible impact theory-building are far from being fully exploited. For being able to grasp and to promote better results from interdisciplinary research in CNs and manufacturing science, and for building fecund and parsimonious theories, the roles of disciplines in science and research in general have to be clarified. 2. Disciplines, interdisciplinarity and paradigms Philosophy of science is about as useful to scientists as ornithology is to birds (Richard Feynman, 1918-88, Physician).

In that context, science may be named any intersubjectively verifiable examination of facts, including their systematic descriptions and – if possible – their explications (Carnap, 1966). With an identified object of interest as a starting point, any science traditionally strives for understanding and principles in line with specificities of the associated branch of knowledge, also referred to as the accounting scientific discipline within the relevant classification of sciences (Popper, 1959). Well-established scientific disciplines have considerable impact on research. The content of theory to be proven seems to strongly depend on presumptions, experiential evidence and ad hoc explanations that constitute scientific progress, however always tightly held together by a dominant paradigm that may as well be referred to as the identity of the accounting discipline. In this perspective, we speak of a pure discipline or of mono-disciplinarity if a certain domain is scientifically permeated with a consistent paradigmatic and theory-rich concept. However, the environment of sciences rarely corresponds with the internal differentiation of disciplines of science to start with. Therefore, typically any progress in science is partially interdisciplinary and (applied) scientific research is indeed one of the triggers for collaboration between disciplines (Luhmann, 1990). Moreover, scientific work, notably in manufacturing and management contexts, is accompanied by reiterations to stay closer to practice (for example, co-production of knowledge as mentioned by Tranfield et al. (2004) and Hartley and Benington (2000)); the calls for the

case study methodology and action research appearing in the first section of this paper testify to that. In this perspective, science is more and more confronted with new so-called complex problems resulting from “everyday challenges”. Consequently, more holistic requirements result in especially frequent demands for adaptation of disciplines’ borders and lending from other domains of knowledge. From this viewpoint, interdisciplinarity may also be interpreted as a reaction to external challenges in manufacturing, triggering efforts for establishing novel methods and concepts that promise to be more adequate to solve research items or practical problems than canonical pure disciplinary approaches appear to offer. At this point, some clarification is needed to narrow down to the scope of interdisciplinarity as used in this special issue about CNs and in manufacturing and management research, with the intention to address the problem areas that cross the border of individual disciplines. This issue will not emphasise transdisciplinarity, where several fields of knowledge areas are providing pure disciplinary accesses and may be networked with each other by intermediate concepts (Aboelela et al., 2006). Neither intends it to go deeper into multidisciplinarity, with several disciplines from different knowledge areas simply co-working within a common context (Aboelela et al., 2006). Noticeable impact on disciplines and disciplines’ boundaries strongly demand for “melting” together disciplines and their followers, a mechanism for which interdisciplinary research seems more promising. At the current stage of thoughts and ideas, transdisciplinarity and multidisciplinarity seem to be far less promising for CNs in manufacturing research than interdisciplinary contributions. A permanent point of discussions is the internal differentiation of science with its consequences in terms of institutional “atomisation of the disciplines and subject areas” (Mittelstraß, 1987). From this point of view, interdisciplinarity clearly responds to progressive specialisation and knowledge fragmentation as it is widely criticised. Without interdisciplinarity, one might argue that progressing knowledge fragmentation limits innovative capabilities, usefulness of scientific knowledge or even relevance of science in general. Moreover, there are debates within this context, seriously putting into question scientific specialisations and differentiations of disciplinary knowledge as barriers to science and knowledge in total, which can only be overcome by means of interdisciplinarity. In order to leave behind such barriers to management science and manufacturing research regularly draw from external theories dependent on the problem domain or the most suitable models to be engaged with, for example, integer programming and control theory. Additionally, in manufacturing and management sciences the disciplines and categories of theories used (e.g. general systems theory, operations research) are often overlapping, since not all work falls neatly into a single school of thought or topic area. This observation may be a helpful starting point for interdisciplinary discussions, as – vice versa – many individuals’ works could have been listed in more than one category; categories of theories may simply appear as temporary and heuristically useful for sorting out major approaches. In no way, they should represent any barriers for interdisciplinary work of fundamental nature. On the contrary, interdisciplinary methods are designed to answer questions differently and to study both phenomena for which we have sufficient hypotheses and phenomena about which too little is known to even formulate hypotheses within pure disciplines’ frameworks. In conclusion, interdisciplinary perspectives enable to make truly original and useful contributions to knowledge, as well as to critically review both the fields from which they

Some thoughts on interdisciplinarity

965

JMTM 23,8

966

draw and the domain at which they aim. Based on such conjectures, interdisciplinary researchers ally with colleagues in traditional disciplines who are also increasingly becoming open to external inputs. It is not surprising, that one of Kuhn’s groundbreaking observations was that anomalies leading to the toppling of a reigning theory or paradigm almost invariably were observed by researchers whose backgrounds were in other disciplines than those in which opinion leaders in the field traditionally had been trained (Kuhn, 1962). Researchers from different disciplines generally use different methods and have different interests toward their object of study. Therefore, it is not surprising that many of the most breakthroughs in the study of management, organisations and markets have come from scholars who stood astride two or more academic disciplines. The benefits of interdisciplinarity for manufacturing research are clearly obvious. Whether championed, vilified, tolerated, or marginalised in manufacturing and management science, interdisciplinarity has stepped in the core of its research to stay. As the departure from pure disciplinary studies, interdisciplinarity exhibits the development of theory innovation by being both an embattled site of controversy and a battle cry (Hutcheon, 1997). This is true in general, but in particular for theories for collaborative (manufacturing) networks. 3. Theory-building and disciplines There is no particular method guaranteeing success. Scientists do not solve problems just because they swing a methodological magic wand (Paul Feyerabend; Die Wissenschaft in einer freien Gesellschaft, 1978).

This call for interdisciplinary contributions has become prominent since increasing numbers of enterprises are faced with the huge so far unseen challenges of manufacturing efficiently in CNs and distributed structures while operating beyond the consolidated state-of-the-art. For support by theoretical insights in recent developments and up-coming concepts, the collection of contributions in this volume attempt to provide a thoroughly evaluated selection of concepts and theory approaches that ought to give considerable underpinning in response to these actual challenges. Some of these challenges may have been already outlined elsewhere; and in numerous cases, workable solutions have been predominant and of uncontested practical impact for “daily” management. Powerful theories, however, can offer additional valuable lenses, which enable managers and stakeholders to frame issues, to compensate for the unreliability of intuition and common sense, to ascertain belly feelings and to clarify many causal relationships that have impact on firms’ objectives as well as resource allocation tactics. In this sense, theories have already contributed and certainly strongly will further contribute a lot to enhance modes and practice of management. For instance, the theory for bottleneck management, elsewhere referred to as the Theory of Constraints, has shed considerable light on methods for optimisation of inventory and flow of materials in factories. Also, chaos theory (a.k.a. theory of complexity) has completely revised the concepts for team structures and for strategy formation in companies. These achievements have rapidly shifted paradigms in manufacturing sciences; the resulting lean and complexity thinking is a key constituent of manufacturing science nowadays. More than for any other field, for manufacturing and management science, the dictum: “Nothing is more practical than good theory is” holds true referring to both substantial progresses in theory as well as for the credibility of managers and practitioners.

The numerous alternative definitions for the term “theory”, each of them highlighting specific aspects and emphasising distinct points of view, have all in common that a theory is represented by a set of laws linked by related derivations. For example, the Popperian as well as the Carnapian Schools see theories as sets of statements: scientific theories are general theses and statements that are, as any representation, symbols and systems of characters (Popper, 1982). Similar thinking is provided by Sutton and Staw (1995), who regard “theory” as a set of logically interconnected arguments that tell a story about why certain acts, events, structures and thoughts occur. So, theories do not just ascertain practical insights, theories are considered the bases of all science and establishing powerful theories is crucial to any scientific progress, but they are also subject to discourse (Foucault, 1969). Returning to our line of reasoning, the development of appropriate theories brought considerable progress for manufacturing sciences. A case in point is the broadening of technological-driven transformations to the total organisational design of manufacturing companies by establishing the Tayloristic thinking, that could later be embedded in the General Systems Theory (von Bertalanffy, 1950, 1973). New ways of modelling, by interpreting technical transformation as inputs and outputs, allowed deeper insight into the logic of manufacturing organisations and its implications to integration of aspects, decomposition for analysis and appropriate control mechanisms. The resulting thoughts actually are indispensable constituents of all current manufacturing systems’ theories. This way of working scientifically is generally referred to as theory-building. More precisely: theory-building is considered any process aiming to produce new theory about empirical phenomena (Weick, 1995). Building of theory may occur through steps, as induction or deduction, comparative analysis or theoretical sampling within a discipline and formation of more general formal theories (Glaser and Strauss, 1967; Suddaby, 2006). Consequently, a scientist’s work consists of establishing theses or systems of theses, which are systematically challenged by the researcher. Possible outcomes will be falsifications, verifications of theories or portions of it as contributions to theory-building. In his seminal work, Kuhn (1962) observed that confusion and contradiction typically are the norm during theory-building, often characterised by a plethora of categorisation schemes. One possible implication of this view may be even that theories can be scientific at one period in time and unscientific at another, depending on their progressiveness (Thagard, 1988). Using this logic, also far-reaching and apparently exotic approaches should always be further encouraged if explicative elements are seen, like during specific case studies. For approaches as holonic manufacturing or soft artefact, only much later fundamental scientific qualities came to light, in synthesis with complexity theories and life-cycle approaches. In such instance, principles of abstraction (Dekkers, 2013; Timpf, 1999) – classification, aggregation and generalisation – will support the extension of principles and solutions to becoming underpinning theories. The domain of collaborative (manufacturing) networks seems to be momentarily in such a phase of theory-building phase. This is caused by a number of phenomena that are not explicable within established organisation theory (e.g. bullwhip effect, instability in turbulent markets). Furthermore, we observe various research approaches offering frameworks, taxonomies, guidelines, etc. in addition to a number of solution descriptions drawn from cases and projects as well as very abstract approaches built upon allegories and metaphors, for example, the footprint or the holon. For collaborative (manufacturing)

Some thoughts on interdisciplinarity

967

JMTM 23,8

968

networks, evidently intersections of scholarly fields and disciplines offer important opportunities to theorise in ways that challenge, reframe and redefine core issues in an emerging discipline and provide new areas for original ideas, fostering theory-building and testing. They also provide opportunities to challenge and revise accepted assumptions as well as established questions and traditions in the original fields from which theories have been drawn. For example, multi-agent systems theory totally revised logic and reasoning for decision making in CNs as well as cooperation and teamwork, including distributed problem solving, coalition formation and coordination. But with the list of examples not being exhaustive and the opportunities for grabs, we expect the number of interdisciplinary works to increase substantially over the years to come. A more extended example for theory-building is a conjecture that has been proposed for establishing complex system sciences and theories (Kuehnle, 2012). It is set up as a collection of shells around a core of theories, enclosed by a shell of laws, principles and rules or generic elements, respectively. These shells are embraced by another shell of models that may be either newly established or frequently applied within the context. The three shells are viewed as embedded in the real world context, which is the manufacturing world, the practitioners view, the successful implementation by proving and verifying practical needs, effectiveness and adaptations. The conjecture may be seen as outcome of introducing topology to the setting of manufacturing networks, in particular the theory of manifolds. Envisioned like this, manufacturing network nodes do not represent just simple units but elements that encapsulate rich structures, able to unfold numerous attributes and properties into the attached realm of models. Manufacturing networks may then be interpreted as specific Hausdorff spaces. The topological structure of Hausdorff spaces allows separating the points representing the production network nodes and thus supports all mappings perfectly. This structure appears rich enough too, to capture a vast majority of configurations occurring in manufacturing networks. It may be accomplished by “attaching” respective models of attributes, relations and aspects as tangent spaces assigned to the manufacturing networks nodes. The manufacturing networks themselves, its attributes and its configurations appear as the quotient space of surrounding Kolmogoroff spaces (in terms of algebraic topology), which may arbitrarily “forget” or “remember” attached models allowing perfect procedures, for instance, to capture encapsulations, to fold and unfold properties, or to triggering on-off modes of self-organisation. Configurations may be modelled by indicators and attributes, and the views are expressed by “attached” tangent spaces to the nodes. In algebraic topology, the resulting set-up is referred to as a particular manifold with boundaries, where important attachments as well as all projections thereof are based upon homeomorphous mappings. Without going into further details, it can be comfortably postulated that the topological theory of manifolds has had impact on research into manufacturing networks’ research on the respective research communities already. By introducing topology and the theory of manifolds, many portions of manufacturing theories (e.g. generic elements, models and principles for social agents as well as software agents’ network interactions) may be reframed; other mappings permit designing novel steadily evolving network decision modes and such set-up facilitates exploiting the networks’ characteristics related to cooperative games and partnership for value optimisation. This example of topology stands for quite a number of similar approaches and structures, from various backgrounds and multiple perspectives, that have been observed to stimulate research around collaborative (manufacturing) networks.

Following the principles of engineering (and other applied sciences), any framework may be accepted as theory, if it addresses most problems (consummate with the principle of fecundity), also if it is currently solving problems at the highest rate (Matheson, 1996), no matter by which notation it is articulated or presented. In this respect, we go along with Mintzberg (2005), who advocates that theory can be seen “along a continuum, from lists (categories), to typologies (comprehensive lists), to impressions of relationships among factors, to causations between and patterns among these relationships, to fully explanatory models”. Management and manufacturing researchers often refer to their schemes as frameworks, taxonomies or typologies that allow identifying categories, even though this is a necessary but only first step for building theory. More generally speaking, enforced theory-building by rigorous problem solution strategies at the intersection of disciplines and research domains has already played a vital role for producing knowledge in the area of collaboration and organisation. The original call for papers for this special issue, strongly encouraged scholars to develop less narrow, more integral views on challenges related to complex organisational phenomena, to (re)configuration modes and to manufacturing network design. Moreover, this special issue intended to ascertains that more blind spots are exposed, that powerful new lenses will not get lost undeveloped and interesting theories can reach the potential for having a full impact on both relevant disciplines for CNs and practical work (covering practitioners and projects). By better theoretical integration of practical achievements in these disciplines, key problems in manufacturing networks eventually become tractable within established research fields. As a substantial consequence, developments for networked organisations will cumulate insight and adequate frameworks more swiftly and more coherently through engaging with more widely accepted impactful models embracing strong bodies of (ready for use) knowledge. 4. Actual contributions to interdisciplinary theory for collaborative networks in this special issue But what are practical steps to gain more theoretical insight from approaches and to integrate findings from other scientific disciplines into the domain of collaborative (manufacturing) networks? This special issue has chosen to encourage interdisciplinary contributions not only via a call for papers for an initial seminar but after peer reviews worked on hot spots or research with all participating authors (this is new for this kind of projects). What was aimed for, was a theoretical elaboration of the vast field of activities that we call interdisciplinary studies concerning collaborative (manufacturing) networks. Particularly, for manufacturing networks which may be considered as a human-governed and systematic combination by means of technological and conceptual procedures in order to transform inputs into outputs in the sense of marketable products, phenomena may be described in technological, socio-economical, social, ecological perspectives, etc. This enumeration already sketches some fundamental options to narrow down the problem domain to conventional disciplines, which is exactly not the intent of this special issue; rather, the contributions should be positioned strongly between disciplines. Based on that notion, this special issue aims to provide authors with a platform to make contributions to the debate based on current theoretical and empirical research. Papers addressing the following questions were considered relevant:

Some thoughts on interdisciplinarity

969

JMTM 23,8

Q1. What are elements of a general theory for collaborative (manufacturing) networks? Q2. What are the specific characteristics of networks (issues of theory-building)? Q3. How will interdisciplinary insight contribute to integrative theories for CNs?

970

Q4. How are networks to be optimised and controlled, especially in absence of a central decision-making unit? Q5. How are decisions taken (logic and transparency) and which methods should be deployed for (re-)linking units (connectivity)? Seven papers and a case study have been selected from work that had been handed in responding to calls for papers and after careful re-alignment in seminars and discussions with other authors as well as the guest editors in addition to the regular review process. The seven interdisciplinary contributions draw from mathematical models, complex adaptive systems, decision theory as well as “unexpected” fields of research, such as entropy, perturbations and information theory. All papers in this special issue are prominent examples for novel and innovative approaches arising from treading on interdisciplinary grounds and amalgamation of diverse theoretical constructs. 4.1 Extending supply chain management approaches beyond the traditional reach The first set of three papers is trying to reach beyond the traditional approaches in supply chain management. In their contribution, Ivanov and Sokolov demonstrate the necessity for even strategic networks, like supply chains, that further advances can only be made by integrating very different disciplines into one concept, which they call multi-structural cyber-physical networks. Very early in their paper, this becomes already obvious when they list the many aspects needed to account for in supply chains. Because of these many aspects, the dynamics for adaptations of planning and scheduling caused by continuous interactions and the (inter)dependencies for decision making, the authors draw on a wide variety of disciplines including operations research, control theory, system dynamics and artificial intelligence; but is also relies on the use of information and communications technology (ICT) to make it successful. The continued reliability of the information systems appears in the contribution by Durowoju et al. when they examine how to counter disruptions. For their thoughts they draw on entropy, a topic related to chaos theory, to study perturbations as they affect the performance of the supply chain under conditions of the supply chain structure, ordering options and integration level. They conclude that each supply chain structure might be affected in different ways. The structural integrity of supply chains appears also in the third contribution by Gerschberger et al. as they set to determine the complexity of the supply chain structure. They do so in order to determine which parts of that structure perform more weakly under conditions of uncertainty. In their view, the proposal is building on the theories of complexity and stretching the traditional cybernetic systems’ view to include the structural dynamics that also appear in the first contribution. Interestingly, they conclude that singular and consistent conceptualisations of complexity neither exist in general nor prevail with regard to a network perspective; that indicates that further works needs to be done. All three’s theoretical contributions are strongly linked to capturing parameters and to formalising concepts derived from complexity theory and chaos theory.

4.2 Introducing new approaches for collaborative (manufacturing) networks The second set of three papers offers new approaches from quite different perspectives. The fourth paper in this special issue by da Piedade Francisco et al. tackles network alignment by a management framework uniting strategic fit, predictive control and topological grounds. That results in a framework for a Collaborative Network Performance Management System. Remarkably, the performance measurement is based on instantiation, a principle rarely applied, but the contribution shows how this might be helpful for management of networks. The next contribution by Ma et al. sees collaboration in inter-firm networks as organisational change versus inertia tension field by implementing an emerging habitual domain theory that synthesises patterns for decision making and human behaviour. As point of departure they take that networks respond sluggishly to current changes and futures and are slow to respond to changes in the external environment. Most of us will perceive CNs being more agile, but does their stance raises the question that networks are subject to the same phenomena as individual organisations? And fundamentally, what do we exactly gain from networks? Part of that enquiry drives the sixth contribution by Eschenba¨cher and Zarvic´. They show that different stages of the life-cycle of CNs are best described by different organisational theories based on the relevant traits of these networks. Hence, these three contributions only demonstrate that reliance on a single theory might poorly advance understanding of collaborative (manufacturing) networks. 4.3 Looking back and forward at theoretical advances That point is picked up by us in the seventh contribution, an extensive outline on the role of interdisciplinarity research, intending to provide more statistical evidence about the research work around collaborative (manufacturing) networks. A structured literature review helped to gain insight in the rate of occurrence and the fields of knowledge surrounding collaborative (manufacturing) networks. A set of 202 papers has been retrieved, and by using statistics, clustering and categorisation we provide solely a clearer picture about focal points and main thrusts of research in this specific domain. However, our intention was not to praise or criticise other scholars’ work as outcome of mono-disciplinary or interdisciplinary research strategies, appropriate or inapt theories or as being informed indulgently by external disciplines to the domain or not. Is it not that every single published piece of research has its unique merits and limitations. In that spirit, we have cited and interpreted research by others, but we do so exclusively to illustrate how interdisciplinary scientific progress is actually made in this strand of manufacturing science. We do believe that this contribution as a review will contribute towards consolidation and credibility for researchers as well as managers promoting Distributed Manufacturing and collaborative (manufacturing) networks; the research agenda in that contribution testifies to that. 4.4 Case study A comprehensive case study concludes this special issue. It is a pleasure that Cheikhrouhou et al. have written up their experiences with the Swiss Microtech Enterprise Network, a well-known example of Collaborative Manufacturing Networks. Its set-up and evaluation, inspired by evolutionary biology and game theory, intensifies the thinking about life-cycle by drawing from complexity science and ICT as enabler. It is also a case in point for our earlier remarks that solution-oriented strategies might

Some thoughts on interdisciplinarity

971

JMTM 23,8

eventually also results in forming of theory, albeit through quite different strategies than theory-oriented research approaches. It is this wide variety of approaches and insight that has driven not only the call for this special issue but also the actual contributions. Hence, it underlines the point that progress can only be made at the intersection of disciplines and that these steps should be considered insprirational.

972 Acknowledgements for this special issue As much as we discuss collaborative (manufacturing) networks, this special issue is also a result of collaboration and most of it virtually. Not only by the authors but also by the reviewers, who have very patiently and carefully examined the propositions by authors, sometimes on numerous occasions. Excellent reviews that helped all of us move forward were written by: Hamideh Afsarmanesh, Henk Akkermans, David Bennett, Luis Camarinha-Matos, W.B. Lee, Jan Olhager, Egon Mu¨ller (assisted by Sebastian Horbach), Kulwant Pawar, Klaus-Dieter Thoben and Roger Warburton. The reviewers were drawn from the Scientific Committee that included more members (Jeff Butler, Afonso Fleury, Roger J. Jiao, Bernard Katzy, Bart MacCarthy, Laure Morel, Chihiro Watanabe). Furthermore, we received support from the Business School of the University of the West of Scotland for organising the seminar. Last but not least, we should mention David Bennett, not only in his role as Editor of this journal but also his support all through (reviews and decision-making), in fact, he was more or less the third guest editor. We sincerely hope that you will be inspired to contribute to further forming of theory and the intersection of domains of knowledge and will join us during further steps on this interdisciplinary journey. Note 1. The term “engineering” has a somewhat ambivalent meaning. Private correspondence with Kulwant Pawar (University of Nottingham), dated 17 October 2011, highlighted this ambiguous use of the word engineering and what it covers (for that reason “product design” was used instead of “engineering” in the original publication (Riedel and Pawar, 1991)). One might also refer to it as “new product development”. References Aboelela, S.W., Larson, E., Bakken, S., Carasquillo, O., Formicola, A., Giled, S.A., Haas, J. and Gebbie, K.M. (2006), “Defining interdisciplinary research: conclusions from a critical review of the literature”, Health Services Research, Vol. 42 No. 1, pp. 329-46 (Part 1). Alvesson, M. and Sko¨ldberg, K. (2000), Reflexive Methodology: New Vistas for Qualitative Research, Sage, Thousand Oaks, CA. Bertrand, J.W.M. and Fransoo, J.C. (2002), “Operations management research methodologies using quantitative modeling”, International Journal of Operations & Production Management, Vol. 22 No. 2, pp. 241-64. Boyer, K.K. and Swink, M.L. (2008), “Empirical elephants – why multiple methods are essential to quality research in operations and supply chain management”, Journal of Operations Management, Vol. 26 No. 3, pp. 337-48. Buffa, E.S. (1980), “Research in operations management”, Journal of Operations Management, Vol. 1 No. 1, pp. 1-7.

Camarinha-Matos, L.M. and Afsarmanesh, H. (2005), “Collaborative networks: a new scientific discipline”, Journal of Intelligent Manufacturing, Vol. 16 Nos 4-5, pp. 439-52. Campbell, L.M. (2005), “Overcoming obstacles to interdisciplinary research”, Conservation Biology, Vol. 19 No. 2, pp. 574-7. Carnap, R. (1966), Philosophical Foundations of Physics: An Introduction to the Philosophy of Science, Basic Books, New York, NY. Christopher, M. (2010), Logistics and Supply Chain Management, Prentice-Hall, Harlow. Coughlan, P. and Cogghlan, D. (2002), “Action research for operations management”, International Journal of Operations & Production Management, Vol. 22 No. 2, pp. 220-40. Davies, A.J. and Kochhar, A.K. (2002), “Manufacturing best practice and performance studies: a critique”, International Journal of Operations & Production Management, Vol. 22 No. 3, pp. 289-305. Dekkers, R. (2013), Applied Systems Theory, Springer, London. Dekkers, R. and Bennett, D. (2009), “Industrial networks of the future: review of research and practice”, in Dekkers, R. (Ed.), Dispersed Manufacturing Networks: Challenges for Research and Practice, Springer, Heidelberg, pp. 13-34. Filippini, R. (1997), “Operations management research: some reflections on evolution, models and empirical studies in OM”, International Journal of Operations & Production Management, Vol. 17 No. 7, pp. 655-70. Flynn, B.B., Sakakibara, S., Schroeder, R.G., Bates, K.A. and Flynn, E.J. (1990), “Empirical research methods in operations management”, Journal of Operations Management, Vol. 9 No. 2, pp. 250-84. Forrester, J.W. (1994), “System dynamics, systems thinking, and soft OR”, System Dynamics Review, Vol. 10 Nos 2-3, pp. 245-56. Forza, C. (2002), “Survey research in operations management: a process-based perspective”, International Journal of Operations & Production Management, Vol. 22 No. 2, pp. 152-94. Foucault, M. (1969), L’Arche´ologie du Savoir, Gallimard, Paris. Glaser, B.J. and Strauss, A.L. (1967), The Discovery of Grounded Theory, Aldine, Chicago, IL. Hartley, J. and Benington, J. (2000), “Co-research: a new methodology for new times”, European Journal of Work and Organizational Psychology, Vol. 9 No. 4, pp. 463-76. Hutcheon, L. (1997), “Disciplinary formation, faculty pleasures, and student risks”, ADE Bulletin, Vol. 117, Fall, pp. 19-22. Jacob, D., Bergland, S. and Cox, J. (2009), Velocity: Combining Lean, Six Sigma and the Theory of Constraints to Achieve Breakthrough Performance, The Free Press, New York, NY. Kuehnle, H. (2007), “Post mass production paradigm (PMPP) trajectories”, Journal of Manufacturing Technology Management, Vol. 18 No. 8, pp. 1022-37. Kuehnle, H. (2012), “Towards production network (PN) theory: contributions from systems of models, concurrent enterprising and distributed manufacturing”, International Journal of E-Business Development, Vol. 2 No. 2, pp. 53-61. Kuhn, T.S. (1962), The Structure of Scientific Revolutions, University of Chicago Press, Chicago, IL. Luhmann, N. (1990), Die Wissenschaft der Gesellschaft, Westdeutscher Verlag, Frankfurt am Main. McCutcheon, D.M. and Meredith, J.R. (1993), “Conducting case study research in operations management”, Journal of Operations Management, Vol. 11 No. 3, pp. 239-56.

Some thoughts on interdisciplinarity

973

JMTM 23,8

974

Matheson, C. (1996), “Historicist theories of rationality”, The Stanford Encyclopedia of Philosophy, 2012, from http://plato.stanford.edu/entries/rationality-historicist/ (accessed 24 January 2008). Meredith, J. (1998), “Building operations management theory through case and field research”, Journal of Operations Management, Vol. 16 No. 4, pp. 441-54. Mintzberg, H. (2005), Developing Theory about the Development of Theory, Oxford University Press, Oxford. Mittelstraß, J. (1987), “Die Stunde der Interdisziplinarita¨t?”, in Kocka, J. (Ed.), Interdisziplinarita¨t: Praxis, Herausforderung, Ideologie, Suhrkamp, Frankfurt am Main, pp. 152-8. Pickett, S.T.A., Burch, J., William, R. and Grove, J.M. (1999), “Interdisciplinary research: maintaining the constructive impulse in a culture of criticism”, Ecosystems, Vol. 2 No. 4, pp. 302-7. Popper, K.R. (1959), Logic of Scientific Discovery, Hutchinson, London. Popper, K.R. (1982), Logik der Forschung, J.C.B. Mohr, Tu¨bingen. Rhoten, D. and Parker, A. (2004), “Risks and rewards of an interdisciplinary research path”, Science, Vol. 306 No. 5704, p. 2046. Riedel, J. and Pawar, K.S. (1991), “The strategic choice of simultaneous versus sequential engineering for the introduction of new products”, International Journal of Technology Management, Vol. 6 Nos 3/4, pp. 321-34. Rungtusanatham, M.J., Choi, T.Y., Hollingworth, D.G., Wu, Z. and Forza, C. (2003), “Survey research in operations management: historical analyses”, Journal of Operations Management, Vol. 21 No. 4, pp. 475-88. Simon, H. (2012), Hidden Champions – Aufbruch nach Globalia: Die Erfolgsstrategien unbekannter Weltmarktfu¨hrer, Campus Verlag, Frankfurt am Main. Sousa, R. and Voss, C.A. (2008), “Contingency research in operations management practices”, Journal of Operations Management, Vol. 26 No. 6, pp. 697-713. Stuart, I., McCutcheon, D., Handfield, R., McLachlin, R. and Samson, D. (2002), “Effective case research in operations management: a process perspective”, Journal of Operations Management, Vol. 20 No. 5, pp. 419-33. Suddaby, R. (2006), “From the editors: what grounded theory is not”, Academy of Management Journal, Vol. 49 No. 4, pp. 633-42. Sutton, R.I. and Staw, B.M. (1995), “What theory is not”, Administrative Science Quarterly, Vol. 40 No. 3, pp. 371-84. Swamidass, P.M. (1991), “Empirical science: new frontier in operations management research”, Academy of Management Review, Vol. 16 No. 4, pp. 793-814. Thagard, P.R. (1988), Computational Philosophy of Science, MIT Press, Cambridge, MA. Timpf, S. (1999), “Abstraction, levels of details, and hierarchies in map series”, in Freksa, C. and Mark, D.M. (Eds), Spatial Information Theory: Cognitive and Computational Foundations of Geographic Information Science, Vol. 1661, Springer, London, pp. 125-40. Tranfield, D., Denyer, D., Marcos, J. and Burr, M. (2004), “Co-producing management knowledge”, Management Decision, Vol. 42 Nos 3/4, pp. 375-86. Venohr, B. and Meyer, K.E. (2007), “The German miracle keeps running: how Germany’s hidden champions stay ahead in the global economy”, Working Paper No. 30, Institute of Management Berlin, Berlin School of Economics, Berlin. von Bertalanffy, L. (1950), “The theory of open systems in physics and biology”, Science, Vol. 111 No. 2872, pp. 23-9.

von Bertalanffy, L. (1973), General System Theory, George Braziller, New York, NY. Voss, C., Tsikriktsis, N. and Frohlich, M. (2002), “Case research in operations management”, International Journal of Operations & Production Management, Vol. 22 No. 2, pp. 195-219. Wacker, J.G. (1998), “A definition of theory: research guidelines for different theory-building research methods in operations management”, Journal of Operations Management, Vol. 16 No. 4, pp. 361-85. Weick, K.E. (1995), “What theory is not, theorizing is”, Administrative Science Quarterly, Vol. 40 No. 3, pp. 385-90. Westbrook, R. (1995), “Action research: a new paradigm for research in production and operations management”, International Journal of Operations & Production Management, Vol. 15 No. 12, pp. 6-20. Wiendahl, H.-P. and Scholtissek, P. (1994), “Management and control of complexity in manufacturing”, Annals of the CIRP, Vol. 43 No. 2, pp. 533-40. About the authors Hermann Ku¨hnle is Full University Professor for Factory Operations and Production Systems, at the Otto-von-Guericke-University of Magdeburg, Germany, and has been Executive Director of the Institute for Ergonomics, Manufacturing Systems and Automation since 1994. From 1994 to 2001 he also was Foundation – and Executive Director of the Fraunhofer Institute for Factory Operation and Automation IFF, Magdeburg. Since 1995, he has been the spokesperson for the research field “Advanced Production Systems in Saxony-Anhalt” and board member of several companies and venture capital groups. From 1980 to 1994 he worked for the National Fraunhofer Institute for Production Engineering and Automation (IPA), Stuttgart, on Material Flow Planning, Enterprise Planning and Organisation, Computer Integrated Manufacturing, and since 1991 as Research Director and Head of the division “Enterprise Planning and Control”. During this period he initiated, built up and managed the CIM-Technology Transfer Centre for the University of Stuttgart. Since 1987, he has led a number of global, European, national and regional research programmes, as well as research consortia with leading companies, research partners and national institutes. Hermann Ku¨hnle is the corresponding author and can be contacted at: [email protected] Rob Dekkers is Reader in Industrial Management at the University of the West of Scotland (since 2006), after having been a Senior Lecturer at Delft University of Technology from 1992 onwards. Before that, he worked in industry as internal consultant, production manager and senior project manager. He holds a Master’s degree in Mechanical Engineering and a Doctoral degree (both from Delft University of Technology). He has (co-)authored about 100 publications on innovation and technology management, transitions in companies, manufacturing strategy, outsourcing and industrial networks. He is board member of the International Foundation for Production Research and Director of the International Association for Management of Technology, and serves on review panels and committees, e.g. the (International) Review Panels for the EPSRC. Main areas of research cover innovation and technology management, changes and transitions in companies, manufacturing strategy, outsourcing models, and industrial networks, underpinned by systems theories, science of complexity and evolutionary (biological) models.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Some thoughts on interdisciplinarity

975

The current issue and full text archive of this journal is available at www.emeraldinsight.com/1741-038X.htm

JMTM 23,8

976

The inter-disciplinary modelling of supply chains in the context of collaborative multi-structural cyber-physical networks Dmitry Ivanov Berlin School of Economics and Law, Berlin, Germany, and

Boris Sokolov Intelligent Information Technology Lab, St Petersburg Institute of Informatics and Automation of the RAS (SPIIRAS), St Petersburg, Russia Abstract Purpose – On modern markets, supply chains (SC) shape the competition landscape. At the same time, considerable research advancements have been recently achieved in the area of collaborative networks. Trends in information technology progress for networked systems include development of cyber-physical networks, cloud service environments, etc. The purpose of this paper is to identify an inter-disciplinary perspective and modelling tools for new generation SCs which will be collaborative cyber-physical networks. Design/methodology/approach – This study addresses the above-mentioned research goal by first, developing a methodical vision of an inter-disciplinary modelling framework for SCM based on the existing studies on SC operations, control and systems theories; and second, by integrating elements of different structures with structures dynamics within an adaptive framework based upon the authors’ own research. Findings – The inter-disciplinary modelling framework for multi-structural SCs has been developed. A new inter-disciplinary level of model-based decision-making support in those SCs is claimed based on the integration of previously isolated problems and modelling tools developed in such disciplines like operations research, control theory, system dynamics, and artificial intelligence. Originality/value – The novelty of this paper is the consideration of SC modelling in the context of collaborative cyber-physical systems. This topic is particularly relevant for researchers and practitioners who are interested in future generation SCs. Particular focus is directed towards the multi-structural SC modelling, structure dynamics, and inter-disciplinary problems and models in future SCs. Challenges of integrated optimization in the organizational and informational context are discussed. Keywords Supply chain management, Information technology, Control theory, System dynamics, Operations research, Cyber-physical system, Collaborative network, Structure dynamics, Cloud service environment, Integration Paper type Research paper

Journal of Manufacturing Technology Management Vol. 23 No. 8, 2012 pp. 976-997 q Emerald Group Publishing Limited 1741-038X DOI 10.1108/17410381211276835

The research described in this paper is partially supported by grants from the Russian Foundation for Basic Research (grants 12-07-00302-a, 09-07-11004, 11-08-01016-a, 10-07-00311, 11-08-00767-a), Department of Nanotechnologies and Information Technologies of the RAS (project 2.11), and program ESTLATRUS (project 2.1/ELRI-184/2011/14). The authors also thank Dr Semen Potryasaev and Dr Alexander Pavlov for assisting them in performing the experimental tests. The authors would also like to thank the Editor and anonymous referees for their comments and suggestions which contributed to the progress of this paper invaluably.

Abbreviations CN

– Collaborative network.

SC

– Supply chain.

IT

– Information technology.

OR

– Operations research.

MP – Mathematical programming. CT

– Control theory.

OPC – Optimal program control. DSS – Decision support system. 1. Introduction On modern markets, supply chains (SC) shape the competition landscape (Christopher, 2012). The SC forms are manifold and include processing industry (e.g. car manufacturing), process industries (e.g. petrochemistry), gas and energy supply. SCs integrate and coordinate suppliers, producers, and distributors from the customer point of view. From the very begin of the research on SCs, the role of information technology (IT) in the SC management (SCM) has been increasingly investigated (Dedrick et al., 2008; Lee et al., 2011). IT has been seen to make SCs more agile, flexible, and adaptable. In particular, agility in organizational and functional SC structures has been considered (Christopher and Towill, 2001). At the same time, considerable research advancements have been achieved in the area of collaborative networks (CN) (Camarinha-Matos, 2009). Research on CN mixes contributions from computer science, engineering, economics, management or socio-human communities (Camarinha-Matos and Macedo, 2010). According to Camarinha-Matos and Afsarmanesh (2005a, b, 2008): [. . .] a collaborative network is a network consisting of a variety of entities (e.g. organizations and people) that are largely autonomous, geographically distributed, and heterogeneous in terms of their operating environment, culture, social capital and goals, but that collaborate to better achieve common or compatible goals, and whose interactions are supported by computer networks.

Research on CN focuses on the structures and their dynamics evolving through autonomous behaviour of entities that collaborate to better achieve common goals. Current development of IT for networked systems and forecasts on future trends in IT progress include development of cyber-physical systems, cloud service environments, pervasive computing, and data and knowledge mining, to name a few ( Jain et al., 2009; Bardhan et al., 2010; Zhuge, 2011). Since SCM depends on the IT development, these new IT will change in next years the landscape of decision support systems (DSS) for SCM. In particular, SC information structure, and moreover, its dynamics, will have more and more impact on organizational, functional, topological, product, and financial SC structures. If the importance and the impacts of the energy structure on the above-named structures are considered additionally, the necessity of inter-disciplinary research on SCs as multi-structural dynamic cyber-physical networks becomes quite obvious.

Modelling of supply chains

977

JMTM 23,8

978

If so, it is now a timely and crucial topic to identify modeling tools for new generation of SCM where SCs are considered as collaborative cyber-physical networks (Figure 1). Cyber-physical systems incorporate elements from both information and material (physical) subsystems and processes. These subsystems and processes are integrated and decisions in them are cohesive. Elements of physical processes are supported by information services. Cyber-physical systems are characterized by decentralization and autonomous behavior of their elements. In addition, such systems evolve through adaptation and reconfiguration of their structures, i.e. through structure dynamics. The top part of Figure 1 shows the conventional SC structure. Three independent information resources (IR) are used by two users for decision support. In the bottom part, the cyber-physical SC is presented. IRs communicate with each other and either take self-control decisions or provide a service for a user who will take the final decision. These services incorporate data from a number of heterogeneous IRs. As all the SC structures become more and more interconnected with each other and IT structure begins to play one of the most important roles in SC structure dynamics, future generation SCM will be characterized by an increasing number of decisions taken in the IT structure without human interference. In light of this, inter-disciplinary models will be required in order to correctly represent integrated elements in different structures (e.g. supply batches, enterprise interests, cloud services, and product structure dynamics). The goal of this paper is: . to identify modeling methods for different SC structures from recent literature; . to justify the necessity of multi-structural dynamic SC modeling and inter-disciplinary models; and . to delineate and systemize methodical issues in future generation SCs.

Figure 1. Next generation supply chain as a cyber-physical system

Information structure

Material flow structure

Conventional Supply Chain

Cyber-Physical Supply Chain

The developments of this study do not represent a finished discussion, but rather launch a new discussion on DSS for future SCs in the context of CNs, collective adaptive systems, and cyber-physical networks. This study addresses the above-mentioned research issues with the help of: . developing a methodical vision of an inter-disciplinary modeling framework for SCM based on the existing studies on SC operations, control and systems theories; and . integrating elements of different structures with structures dynamics within an adaptive framework based upon our own research. The rest of this paper is organized as follows. In Section 2, state-of-the-art is analyzed and the needs for future research on inter-disciplinary modeling are derived. Section 4 presents the vision and some model examples in an integrated inter-disciplinary modeling framework. Section 3 is devoted to SC structure dynamics control (SDC) and exemplified structural interrelation in cyber-physical SCs with the help of cloud information services. In Section 5, challenges of integrated optimization in the organizational and informational context are discussed. The paper is concluded by summarizing the most important features of this study and giving an outlook about future research needs in Section 6. 2. State-of-the-art In the study by Ivanov and Sokolov (2012a), the multi-structural framework for SC structure dynamics was presented. It identified the main SC structures as follows: . product structure (bill-of-materials); . functional (structure of management functions and business-processes); . organizational (structure of facilities, enterprises, managers and workers); . technical-technological (structure of technological operations for product production and structure of machines, and devices); . topological (roads and transportation ways) structure; . informational (information flows according to a coordination strategy); and . financial (structure of costs and profit centres). In this paper, this framework will be used in order to systematically represent recent literature on modeling methods in different SC structures as well as for structural interrelations. Organizational and functional structures: material flows Stable SC processes in a complex environment support enterprise competitiveness. On the contrary, the “overheated” SCs lack resilience and stability and become more and more vulnerable. Therefore, the achievement of the planned SC performance can be inhabited by changes and perturbation impacts in a real execution environment (Kleindorfer and Saad, 2005). This forces the research on SC replanning and rescheduling (Krajewski et al., 2005) to make SCs reliable and flexible enough to be able to adapt their behaviour in the case of perturbations impacts and to remain stable and resilient by recovering disruptions.

Modelling of supply chains

979

JMTM 23,8

980

In research on SCM, the research community faces the challenges of governing SC dynamics (Graves and Willems, 2005; Kouvelis et al., 2006; Safaei et al., 2010). After a long period of research on increasing SC agility, speed, and performance, the research focus has been shifting to a paradigm that the performance of SCs is to interrelate to dynamics, adaptability, stability, and crisis-resistance (Disney and Towill, 2002; Braun et al., 2003; Disney et al., 2006; Gunasekaran et al., 2008; Sarimveis et al., 2008; Ivanov et al., 2012). Quantitative research on SC planning has a long and successful history and is typically based on operations research (OR) techniques where practitioners and researchers are typically seeking to optimize efficiency and/or responsiveness/flexibility of SCs (Chen and Paulraj, 2004; Klibi et al., 2010). SCM has been a very visible and influential topic in the field of quantitative modeling. Tayur et al. (1999), de Kok and Graves (2004) and Simchi-Levi et al. (2004) provide a systematic summary of OR-based quantitative models of the SCM. Along with OR methods, system dynamics and control theoretic approaches have been applied to taking into account execution dynamics while SC planning. The study by Akkermans and Dellaert (2005) presented a coherent framework which succeeds in mapping SC issues with formal methods. Barlas and Gunduz (2011) presented demand forecasting and sharing strategies to reduce fluctuations and the bullwhip effect in SCs based on system dynamics approach. Control theory (CT) as a base for studying multi-stage, multi-period dynamic systems is an interesting research avenue to extend existing results while taking into account the intrinsic peculiarities of modern SCs. CT contains a rigor quantitative basis for planning optimal control policies including differential games and stochastic systems, stability of controlled processes and non-linear systems, controllability and observability, and adaptation (Disney et al., 2006; Sarimveis et al., 2008; Schwartz and Rivera, 2010). A popular technique of SC control is the model predictive control (MPC). MPC is a control strategy based on the explicit use of a process model to predict the process output (performance) over a long period of time (Camacho and Bordons, 2004). The model attempts to predict the control variables for a set of time periods. Predicted control variables depend on disturbance forecasts (i.e. demand, prices and interest rates) and also on a set of given parameters that are known in the control literature as control inputs. Applications of MPC to multi-echelon production-inventory problems and SCs have been examined previously in the literature (Perea et al., 2000). Braun et al. (2003) developed a decentralized MPC implementation for a six-node, two-product, three-echelon demand network problem developed by Intel Corporation that consists of interconnected assembly/test, ware-house, and retailer entities. In order to achieve resemblance of the models and SCs, complex adaptive systems (CAS) and multi-agent systems (MAS) have been extensively applied to SCM domain so far (Swaminathan et al., 1998; Choi et al., 2001; Surana et al., 2005). The attraction of the agent-based modeling makes it possible to reflect the decentralized decision making in SCM taking into account individual human behavior subjectivism, individual risk perceptions, etc. In addition, as in SCs it is practically impossible to develop a model structure with the defined input-output interrelations, one of the possible approaches is to apply behavioral frameworks (Polderman and Willems, 1998). Gjerdrum et al. (2001) investigated the possibilities of how expert systems techniques for distributed decision

making in terms of agents can be combined with contemporary numerical optimization techniques for the purposes of SC optimization. Kuehnle (2008) considered the application of MAS for integrated modeling production networks. In view of the above-mentioned possible limitations of OR and CT, new techniques of planning and control intellectualization such as evolutionary (meta)-heuristics, fuzzy-neural systems, knowledge-based self-organizing networks should enrich classical OR and CT in order to be widely applicable to SCM. Intellectualization of control can be seen as CT and can become the area where the knowledge of SC managers and control specialists can be effectively integrated by taking advantages of intelligent information systems (e.g. radio frequency identification (RFID) and navigation systems). E.g. CT can provide a wider possibility for investigating SC dynamics in different environments (and not only in discrete or stochastic). Besides, application of RFID can extend the existing approaches to SC adaptation (Lee and ¨ zer, 2008). O Interrelating information and material structures The impact of IT on the material processes in SCs becomes more and more crucial. Recent research indicated that an aligning of business processes and IT may potentially provide new quality of decision-making support and an increased SC performance (Dedrick et al., 2008; Jain et al., 2009). That is why it becomes a timely and crucial topic to consider SCs as collaborative cyber-physical systems. Such SCs are common not only in manufacturing but also in different cyber-physical systems, e.g. in networks of emergency response units, city traffic control, and security control systems. At the same time, scientist from OR and SCM underline that the increase in complexity and multi-dimensionality of operations management problems, including SCM, necessitate that OR methodology integrates with systems science, CT, artificial intelligence, and informatics (Taha, 2002; Barabasi, 2005; Kuehnle, 2008). In this setting, a possible approach to problem dimensionality can be a “relief” of the main model by distributing its certain elements to another model. On the other hand, embedding elements of other techniques (e.g. those from CT) can enrich the world of discrete mathematics by new categories and integrate the operability objectives such as robustness and flexibility into SC design. In many branches, hierarchical SCs with pre-determined supplier structure and product programs evolve into customer-oriented, temporary networking of core competences. Dekkers and van Luttervelt (2006), Kuehnle (2010) and Ivanov et al. (2010) pointed out that such agile networking is subject to structure dynamics that should be taken into account. These concepts, known as virtual enterprises, factory-on-demand, smart SCs, responsive SCs, dispersed manufacturing networks or distributed manufacturing become a popular form of competitive organizations (Noori and Lee, 2002, 2009; Camarinha-Matos and Afsarmanesh, 2005a, b; Camarinha-Matos et al., 2005; Dekkers and van Luttervelt, 2006; Mu¨ller et al. 2008; Kuehnle, 2008; Gunasekaran et al., 2008; Dekkers, 2009a, b; Kuehnle, 2010). SC structure dynamics is one of the main challenges in modern SCM that has been addressed in recent literature. Earlier literature presents several optimization-based approaches to SC configuration and planning with structure dynamics. Chauhan et al. (2006) address the problem of short-term SC design using the idle capacities of qualified

Modelling of supply chains

981

JMTM 23,8

982

partners in order to seize a new market opportunity. Sarkis et al. (2007) presented a strategic model for agile virtual enterprise partner selection. Dekkers and van Luttervelt (2006) developed methods for adapting network processes and structures to changing circumstances, expressed by the criteria of innovation, speed and flexibility, and changeability. Helo et al. (2009) presented a system dynamics-based concept and a tool for designing and modeling agile supply-demand networks. The study by Ivanov and Sokolov (2010) developed the quantitative framework for SC SDC based on multi-structural macro-states of an SC which are composed of the different SC structures and their interrelations. At different stages of the SC evolution, the elements, parameters, and structural interrelations change. In these settings, an SC can be considered a multi-structural process. Other structural interrelations Previous research has episodically tackled the interconnections of the SC structures. First, incorporation of product structure dynamics into manufacturing or logistics decisions (Zhang et al., 2009) as well as in organizational structuring while locating facilities (Nepal et al., 2012) should be stated. Second, decisions in information and organizational structures have been considered in an integrated manner (Vickery et al., 2010). Third, financial flows have been analyzed subject to an integration with material flows (Guille´n et al., 2006) and product structures (Laı´nez et al., 2009). Finally, SC topological structure and its robustness have been investigated (Nair and Vidal, 2011; Vahdani et al., 2011). The robustness considerations have also been incorporated in the product structure analysis (Yadav et al., 2011). Observations One of the main SC features is the multiple structure design and changeability of structural parameters because of objective and subjective factors at different stages of the SC life cycle. The literature analysis shows that the dynamic characteristics of SCs are distributed upon different structures, e.g. organizational (i.e. agile supply structure (Sarkis et al., 2007)), functional (i.e. flexible competencies), product based (i.e. product flexibility (Graves and Willems, 2005)), informational (i.e. fluctuating information availability (Bensoussan et al., 2007)), financial (i.e. cost and profit sharing (Cachon and Lariviere, 2005)). This multi-dimensional dynamic space along with the coordinated and distributed decision making lead us to the understanding of modern SCs as multi-structural active systems with structure dynamics. The literature review outlined above allows the conclusion that two basic issues exist while considering SC dynamics. First, in each of the SC structures, different parameters change during the plan execution. Second, decisions and parameters in all the structures are interconnected and influence each other both during the planning and while replanning. To the best of our knowledge, these two interrelated domains have never been explored simultaneously in the context of SCs. The above-mentioned necessitates, the description of SCs as dynamic systems with structural changes that explicitly incorporates structure dynamics aspects which inevitably exist in all the CN applications. In addition, the structure dynamics explicitly relates different structures which are tightly interlinked with each other in practice. Some examples of the structural interrelations follow. Business processes are

designed in accordance with SC goals and are executed by organizational units. These units fulfil management operations and use certain technical facilities and information systems for planning and coordination. Business processes are supported by information systems. Organizational units have a geographical (topological) distribution that also may affect the planning decisions. Collaboration and trust (the so-called “soft facts”) in the organizational structure do affect other structures, especially the functional and informational structures. Managerial, business processes (distribution, production, replenishment, etc.), technical and technological activities incur SC costs, which also correspond to different SC structures. So the representation of SCs as complex multi-structural dynamic systems can be favourable both for the identifying structures and corresponding models and for identification of different structural relations from the static and dynamic points of view. In addition, the above-mentioned studies along with some other research work (Beamon, 1998; Simchi-Levi et al., 2004; Ivanov, 2009) indicated that SCM problems are tightly interlinked with each other with regard to different levels, different structures, and in dynamics and have multi-dimensional characteristics that require the application of different integrated frameworks of decision-making support. If so, inter-disciplinary modeling frameworks for future generation SCs are needed. 3. Supply chain structure dynamics control SDC approach is inter-disciplinary and reaches beyond the classical borders of CT and mathematical optimization. It is based on a combined application of optimal program control (OPC) theory and mathematical programming (MP), and extends their classical borders by their mutual integration and by decentralization of system description with the help of active modeling objects (AMO). The main idea of the SDC-based models is the dynamic interpretation of planning in accordance with the natural logic of time with the help of OPC. The solution procedure is transferred to MP. In this setting, the solution procedure do not depend on the continuous optimization and can be of a discrete nature, e.g. an integer linear program. The SDC is based on the dynamic representation where the decisions on SC planning are taken for certain intervals of structural constancy and regarding problems of significantly smaller dimensionality. For each interval, a static optimization problem of a smaller dimensionality can be solved with the help of MP. The transitions between the intervals are modeled in the dynamic OPC model. Besides, the a priori knowledge of the SC structure, and moreover, structure dynamics, is no longer necessary – the structures and corresponding functions are optimized simultaneously as the control becomes a function of both states and structures. The splitting of the planning period into the intervals occurs according to the natural logic of time and events. As the SDC is based on CT, it is a convenient approach to describe services due to the abstract nature of state variables which can be interpreted as abstract service volumes. The constructive implementation of the SDC principles is based on the usage of Pontryagin’s maximum principle (Pontrayagin et al., 1964). According to the maximum principle, the problem consists in finding an OPC that transits a dynamic system from an initial to a desired final state subject to constraints on control and states and minimizing some objective functionals. The OPC calculation is based on the Hamiltonian function.

Modelling of supply chains

983

JMTM 23,8

984

In integrating the main and the conjunctive equation systems, the values of variables in both of the systems can be obtained at each point of time. The maximum principle guarantees that the optimal solutions (i.e. the solution with maximal values) of the instantaneous problems (i.e. at each point of time) give the optimal solution to the overall problem. This principle is a convenient approach to naturally decompose a problem into some sub-problems. For these sub-problems, optimal solutions can be found, e.g. with the help of MP. Then these solutions are linked into an OPC. The study by Ivanov and Sokolov (2012b) has proposed an original model to represent SC schedules as OPC. The sub-problems of a small size appear through the dynamic job releases, i.e. according to the natural time events. Thus, such decomposition is based on the natural (objective) time logic, and not on any artificial (subjective) rules. Therefore, no admissible solution can be lost during the optimization procedure which allows us to consider the proposed procedure as an optimal approach that is based on the proved axiomatic of OPC theory. Finally, SCs are characterized by decentralization. SC elements are active (they can compete and have conflicting aims, interests and strategies). The preliminary investigations confirm that the most convenient concept for the formalization of SC SDC processes is the concept of an AMO. In a general case, it is an artificial object, acting in space and time and interacting (by means of information, financial, energy, and material flows) with other AMO in the system and the environment. The AMO activities are based on their own operating policies, interests and goals, replenishment and consumption of resources (i.e. materials, energy, etc.), and collaboration with other enterprises in the SC and customers (Figure 2). All of the four AMO functions (components) significantly differ in their nature, but the joint execution of these functions, the collaboration being the main one, provides the AMO with new characteristics which distinguish the AMO as a specific object of investigation and control and basically differentiates the SC SDC from classical automatic control or physics tasks. The proposed structure of an AMO can be interpreted widely. For example, the multi-agent or CAS ideology can be considered as a basis for the AMO interaction modeling. Environment dynamics disturbances Collaboration rules

COLLABORATION

Collaboration state disturbances

Competence profiles Operating policies

Figure 2. Concept of active modeling objects

COMPETENCIES

Competence state disturbances Execution state

OPERATION disturbances

Replenishment and processing

RESOURCES

Resource state

Example In recent years, studies on SC dynamics were broadened by developments in IT such as RFID, supply chain event management and mobile business which provide a constructive basis to incorporate the stages of SC planning and execution control ¨ zer, 2008). If so, the basic idea of future generation SCs, i.e. the cyber-physical (Lee and O SCs, can be implemented. Let us consider an example of an integrated scheduling in functional and informational SC structures under consideration of a cloud information service environment. In these settings, two questions may be raised:

Modelling of supply chains

985

Q1. What is the optimal volume of information services needed to ensure operation of physical systems? Q2. How these services shall be scheduled at the planning stage and re-scheduled (adapted) in dynamics at the execution control stage? The first problem analyses investments into information infrastructures, and the second problem matches actual physical processes with information services. Conventionally, the above-described two problems are solved step-by-step. With the help of SDC, a special dynamic representation of multi-structural networks is proposed where such problems can be solved simultaneously. In addition, due to the increasing role of information services in different forms, e.g. cloud computing, the service-based approaches to integrated planning and scheduling of both material and information flows in collaborative SCs are needed. Such integration is also important to prevent failures of IT-enabled SCs. Although recent research has extensively dealt with SC scheduling and IT scheduling (see, e.g. works on scheduling telecommunications) in isolation, the integrated scheduling of both material and information flows still represents a research gap. A simple example of the interrelations among business processes, services, and IRs is shown in Figure 3. The problem is to find a joint schedule, i.e. taking into account the IS modernization, four schedules should be generated in a coordinated manner, i.e.:

Information Resource

Services

Business process

Concept model – BP – Services – Information Resources

Stocking

Stock receipt

Receipt of data

ERP module “Procurement

Supply documents

Terminals

Stock placement

Delivery

Receipt of pick- up data

Pick-up placement

ERP module “Sales”

RFID

Delivery documents

Printer

Figure 3. Interrelations among business processes, services, and IRs

JMTM 23,8

986

(1) (2) (3) (4)

a a a a

schedule schedule schedule schedule

for for for for

the material supply processes in the SC (model M1); services (model M2); the IRs (model M3); and the IR modernization (model M4).

Goals are measured by the job’s delivery times to customers and the volume of the delivered jobs. Jobs are to be scheduled subject to maximal customer service level, minimal backlogs, minimal idle time of services, and minimal costs of IT (including, fixed, operation, and idle cost). Customer service level is measured by a function of the times when the jobs are delivered to the customers. The SC is modeled as a networked controlled system described through a dynamic interpretation of the operations’ execution. Control models (M1-M2) are first used to assign and sequence services to business operations. Then M2-M3 are employed to assign and schedule services to IR. Finally, M3-M4 are launched to schedule IT modernization (reconfiguration). The basic interaction of these models is that after the solving the conjunctive system for M1, the found control variables are used in the constraints of the conjunctive system for M2. Analogously, M2, M3, and M4 are interconnected. In solving the main systems, the interaction of the models is organized in the reverse way, from M4 to M1. Note that in the calculation procedure, the models M1-M4 will be solved simultaneously, i.e. the materials, services, IRs, and modernization scheduling will be integrated. The peculiarity of the proposed approach is the dynamic interpretation of scheduling based on a natural dynamic decomposition of the problem and its solution with the help of a modified form of continuous maximum principle blended with combinatorial optimization. In following a common approach consisting of decomposing the solution space and using exact methods over its restricted sub-spaces, the use of the OPC theory for the dynamic scheduling problem decomposition is proposed. The usage of the OPC theory for this decomposition rather than heuristics procedures is proposed. For the solution to the above-mentioned scheduling problem, maximum principle in continuous form blended with combinatorial optimization will be applied. According to the maximum principle, the optimal solution of the instantaneous problems can be shown to give the optimal solution to the overall problem (Pontrayagin et al., 1964). In the approach proposed, the calculation procedure is transferred to MP methods and is therefore independent of the OPC. The solution at each point of time is calculated with MP. OPC is used for modeling the execution of operations and interlinking the MP solutions during the planning horizon. Hence, the solution procedure becomes independent of the continuous optimization algorithms and can be of a discrete nature, e.g. an integer assignment problem. Methods of discrete optimization can be used to obtain optimal schedules for those sub-problems of small dimensions. 4. Inter-disciplinary modeling framework In this section, a vision of an inter-disciplinary modeling framework for SCM will be developed. The vision of the framework is based on the challenges of SC problems and advantages/limitation of different modeling techniques to approach these problems as described in Section 2. With regard to SCM, different quantitative frameworks have been constituted so far. Beamon (1998) distinguished between deterministic analytical models, stochastic analytical models, economic models and simulation models.

Gunasekaran and Ngai (2009) classified the models for SC models based on the nature of the decision-making areas and sub-classified to focus on solving problems. It can be observed that different methods for SC optimization have different application areas, advantages and disadvantages. These possible limitations reduce the application areas and lead to may unrealistic assumptions in the models which rarely reflect the multi-stage, multi-period and multi-product nature of real decision-making problems. For example, lead times are not fixed and are not known with high accuracy, inventory levels should be bounded below by zero and above due to warehouse capacities, and the production rates which are limited by the machinery capacities. Another limitation is that single stage or at best dyadic systems are usually studied, assuming production of a single product or aggregated production. In real life systems, various products are produced and moved with different production rates, order sizes and different lead times, which, however, share common machinery and storage facilities. Horizontal integration is often represented by considering the SC stages in the raw, while interconnections between different level and same level stages are ignored (Sarimveis et al., 2008). Finally, the dynamics of transport tariffs, raw material costs, labor costs and inventory costs are rarely taken explicitly into account. At the same time, the possible limitations of certain methods are frequently exactly compensated in the advantages of other methods. Although those methods appear to differ in targets, presumptions, application areas, enabling technologies, and research methodologies, each compliments the others and endeavors to improve decision making. In the light of this, the inter-disciplinary contributions can be stated as a future research avenue to reflect multi-dimensional dynamic characteristics of SC processes. Table I considers an example of three methods, namely MP, OPC and MAS. Gjerdrum et al. (2001) and Ivanov et al. (2007) investigated the possibilities of how MAS techniques for distributed decision-making in terms of agents can be combined with contemporary numerical optimization techniques for the purposes of SC optimization. Kuehnle (2008) considered the application of MAS for integrated modeling production networks. To overcome the limitations of CT, OC is frequently Approach Application areas

Advantages

Disadvantages

MP

Models of SC design and planning with steady-state conditions and limited dimensionality

Optimal solutions Clarity Available software

OPC

Dynamic models of SC design, planning, and scheduling SC execution control and adaptation

Multi-stage, multi-period, multi-commodity SC dynamics Time-independent problem dimension Non-stationary

MAS

Decentralized planning and coordination of SCs

Decentralized decisionmaking Human-driven decisions

Dimensionality and complexity of real problems Dynamics and uncertainty of system and model evolution Poor flexibility Elaborated for automatic control systems Centralized control Non-linear mathematics is not always suitable for SCs Difficulties in handling discrete event systems The quality of a found solution with regard to potential optimum usually remains unknown

Modelling of supply chains

987

Table I. Analysis of methods for SC dynamics

JMTM 23,8

988

combined with OR techniques. A possible approach can be a “relief” of the dimensionality of OR models by distributing some of their elements to dynamic control models. In recent literature, a trend for a combination of different approaches to modeling and simulation can be stated. For example, Kuehnle (2008) considered agents as a part of the complex interrelated models for CNs planning. The existing frameworks provide the evidence that isolated application of only one solution method leads to a narrowing in problem formulation, overdue constraints and sometimes unrealistic or impracticable goals. The main motivation for this research approach is to combine the possibilities of different decision-making techniques, e.g. OR and CT, to achieve new quality in the DSS for SCM. For example, in applying the proved fundamentals of CT to the SCM domain, the conventional OR-based modeling techniques for SCM can be enriched by new viewpoints on dynamics, stability, adaptability, consistency, non-linearity, and high dimensionality of the complex system. The mathematics of the CT can help in revealing new conformities to natural laws that remain unrevealed within the OR field. Hence, the conventional SCM problems may be considered from a different viewpoint and new problems that are near to the real-world settings may be revealed and formulated. The developed framework takes into account the principles of SC elements’ activity, multiple modeling, integration, and decentralization. To reflect the active nature of decision-making elements in SCs, the agents are expressed as conceptual modeling entities or AMO. Integration is considered from four perspectives: the integration of various modeling approaches and frameworks, the integration of planning and execution models, the integration of decision-making levels, and the implementation of integration throughout: “conceptual model ! mathematical model ! software”. Decentralization in the decentralized integrated modeling approach methodology considers the main principle of management and decision making in CN. This means that all the models contain elements of decentralized decision making and CN elements’ activity. Decisions about CN management are not established and optimized “from above” but are a product of iterative coordinating activities of the enterprises (agents) and a coordinator (e.g. a 4PL provider). In the research framework, different modeling techniques are not set off with each other but considered in their mutual interrelations. As problem situations usually consist of different partial problems with different changeability of the data nature, structures and requirements for output representation, these partial problems can be solved by means of different modeling techniques. The selection of a solution method depends on the data fullness, problem scale, one or multiple objectives, requirements on output representation and the interconnection of a problem with other problems. Different approaches from the OR, CT, and agent-based modeling have a certain application area and a certain solution procedure. Therefore, the basic principles of system modeling – decomposition and integration – can be applied. OR techniques are used for optimization of clear-definable dynamic planning and scheduling problems. In practice, these problems are tightly interlinked within the so-called problem situations. To integrate the partial problems into problem situations, methods of systems science are applied. A further extension is the attraction of CT to reflect the real-time dynamics and dynamic optimization of SC processes in the real-time mode. Besides, CT makes it possible to incorporate the analysis of stability, adaptation, and real-time dynamic feedbacks. Finally, the attraction of the behavioral

frameworks and agent-based modeling makes it possible to reflect the decentralized decision-making in SCM taking into account individual human behavior subjectivism, individual risk perceptions, etc. in order to intellectualize planning and control. The coordinated application of different models improves modeling quality, as disadvantages of one model class are compensated for by advantages of other classes, which enable the examination of the determined classes of modeling tasks. In integrating elements of different modeling techniques, two basic ways exist. First, a basic solution method for all the parts of the integrated problem can be chosen and reinforced by attracting elements from other methods at certain places (Ivanov et al., 2011). This is the preferable way. If this is impossible, different methods for different parts of the problems shall be used and their input-output results shall be integrated. Examples of integrated models and generic problem situations Let us consider an example of the integrated SC planning (SCP). In SCP models, different goals, variables and constraints can be expressed either in static or dynamic form. In applying only one solution method, e.g. MP or optimal control, significant difficulties in representing both static and dynamic aspects are frequently encountered. Therefore, it can be sensible to distribute static and dynamics goals, variables and constraints among different models where the corresponding static or dynamic elements can be expressed in the best way. Although SC structures are subject to dynamics, they do not change permanently. There are certain time intervals when these structures remain unchanged. Hence, for these intervals, conventional MP models can be applied while structure dynamic can be presented in a CT dynamic model (Ivanov et al., 2011). An SC is considered that consists of raw material suppliers, production plants, distribution centers, and customers. The structure and the parameters of SC undergo changes at discrete time points. These points divide the planning interval into subintervals. The SC structure does not vary at each subinterval. The changes of SC structure (e.g. because of changes in transportation costs, contracts with suppliers, unplanned transportation or production resource unavailability) are referred as SC structure dynamics. With SC structure dynamics, the scenarios of SC execution are set up. It is assumed that each network element within the subintervals (structure constancy intervals) is characterized by the following given characteristics: the maximal inventory at a node; the capacity of the node; the capacity of the inter-node channel. The problem consists of finding a transportation plan (time and intensity of deliveries) and a production plan (processing times and lot-sizes) with the goal of minimizing SC costs (inventory costs, production costs, and transportation costs) along with the maximizing SC service level (the volume of fulfilled customer orders). Conventionally, this problem is solved as two isolated models. First, an optimal aggregate plan is found with the help of transportation model. Subsequently, the outputs of this model are used as inputs into lot-size and routing planning (Chen, 2010). The integration of these two problems within only one method, i.e. the MP, leads to a significant increase in the number of integer variables and constraints and would force us to make some unrealistic assumptions regarding dynamic changes in control and structural parameters. In addition, structural changes and inventory dynamics in different periods do not allow direct application of one-period transportation models.

Modelling of supply chains

989

JMTM 23,8

990

In the developed SDC-based model, structure dynamics (i.e. the execution scenarios according to different structural states), inventory dynamics, and transitions between the intervals are modeled in the dynamic OPC model. The assumption on the intervals of structural constancy allows transition from the dynamic to static models. For each interval of structural constancy, a static discrete optimization problem of a smaller dimensionality can be solved with the help of an MP method. In particular, an integrated inventory-transportation model is formulated and solved (Ivanov et al., 2011). Its particular feature is inventory consideration as a link between the time intervals. According to the above-mentioned methodical approach, we propose to distribute static and dynamic elements between two models in order to avoid the overloaded dimensionality of a general model and to represent the corresponding variables, constraints and goals in the most appropriate model class. The first model (MP) describes the aggregate production and transportation planning within each of the l-subintervals. The second model (optimal control): . interrelates these subintervals in dynamics according to the natural time logic of changes in the SC; and . describes dynamics of SC processes (e.g. frequency and quantities of transportation) within the subintervals of SDC. 5. Challenges of integrated optimization in the organizational and informational context Large-scale centralized models for planning the whole time horizon are very sensitive to changes in data availability. SCM is based on information sharing and coordination, and many SC optimization models assume full information availability. However, due to dynamic changes and coordination problems in the SC it is frequently impossible. If such a disturbance takes place two issues occur: is the SC able to continue its operation? Can mathematical models work with incomplete or delayed information? In the light of the above-mentioned practical challenges, the application of the SCD approach to SC optimization can be very favorable. In the case of partial information unavailability, it affects only the current interval and does not break the work of the whole model. The practical projects have also shown the following challenges of integrated SC optimization. First, data come from different departments. Formulation of multi-criteria integrated mathematical problems is not always compatible with local performance criteria in different departments. It is also not always obvious who can implement the results of integrated optimization because of limited management competencies. Finally, data availability in information systems shall be checked before starting integrated optimization. Nevertheless, some possible applications of SDC-based integrated SC optimization can be identified as follows: . Using “ideal” solutions as a comparison base. . Re-structuring of management competencies. . Justifying interrelations among different decision domains in order to proof coordination and integration efficiency. . Alignment of performance management in the internal and external SC.

Through the development of the planning models in terms of control and their interconnection with planning decisions, the transition from large-scale centralized models to decentralized flexible models with a distribution of local and system-wide information can be realized. In order to approach those issues, a combination of quantitative methods, SCM knowledge, and intelligent IT can be very promising in industrial applications.

Modelling of supply chains

991 6. Conclusions and future research outlook SCM integrates, coordinates and harmonizes partial logistics processes and their links to the corresponding production processes from the perspective of improving the total value-adding chain performance. In many branches, hierarchical SCs with pre-determined supplier structure and product programs evolve into customer-oriented, temporary networking of core competences. With the development of SCM, new specific problems and integrated problems from production and logistics management such as distribution-production or production-inventory problems have appeared. This is quite naturally since the integration and coordination as the reflection of complex interrelations in SCs are the central ideas of SCM. Therefore, integration and coordination should also become a kernel idea in decision support models. In approaching integrated SCM problems, the application of only one modeling technique is frequently insufficient. In this setting, a new modeling level – the model integration and coordination level – shall be developed in future. In addition, SCs are multi-structural systems. The proposed original framework explicitly incorporates structure dynamics aspects which inevitably exist in all the CN applications. In addition, the SC structure dynamics explicitly relates different structures which are tightly interlinked with each other in practice. SC information structure, and moreover, its dynamics, will have more and more impact on organizational, functional, topological, product, and financial SC structures. This necessitates the inter-disciplinary research on SCs as multi-structural dynamic cyber-physical networks. This paper has undertaken an attempt to identify an inter-disciplinary perspective and modeling tools for new generation SCs which will be collaborative cyber-physical networks. The developed quantitative framework makes it possible to consider conventionally isolated SC design, planning, and execution problems taking into account SC structure dynamics. An additional advantage of attracting systems and control sciences to the SCM domain is that generic SC properties like stability, adaptability, redundancy, and feedback control can be investigated in their fullness and consistency with operations planning and control within a conceptually and mathematically integrated framework. In this paper, the vision of integrated inter-disciplinary modeling has been presented and elaborated in one specific context, namely the applicability of CT approaches to SCM. The combined application of CT with OR, system dynamics and artificial intelligence enriches the possibilities of modeling and optimization DSS for future generation SCs. In future, when validating different models and applying them to different SCM problems, we will be able to develop generic methodical patterns and recommendations on using multiple-model theoretic frameworks, theorems and tools in approaching different classes of SCM problems. As a result, a taxonomy for application of OR, control and system theoretic frameworks, theorems and tools for the SCs as cyber-physical

JMTM 23,8

992

systems can be developed. In doing so, the basic limitation of this research in its current state, i.e. the lack of technological infrastructure for validating the multiple-model by a wider research community shall be eliminated. Another crucial topic of the future research is the impact of uncertainty and disruptions. Even in this area, the potential of SDC can be applied to SCM to a great extent. For example, it becomes possible to analyse the flexibility of SC configurations or to investigate the SC schedule robustness. In addition, information flow disruptions can be addressed. References Akkermans, H. and Dellaert, N. (2005), “The rediscovery of industrial dynamics: the contribution of system dynamics to supply chain management in a dynamic and fragmented world”, System Dynamics Review, Vol. 21 No. 3, pp. 173-86. Barabasi, A.L. (2005), “Network theory – the emergence of the creative enterprise”, Science, Vol. 308, pp. 639-41. Bardhan, I.R., Demirkan, H., Kannan, P.K., Kauffman, R. and Sougstad, R. (2010), “An interdisciplinary perspective on IT services management and service science”, Journal of Management Information Systems, Vol. 26 No. 4, pp. 13-64. Barlas, Y. and Gunduz, B. (2011), “Demand forecasting and sharing strategies to reduce fluctuations and the bullwhip effect in supply chains”, Journal of Operational Research Society, Vol. 62 No. 3, pp. 458-73. Beamon, B.M. (1998), “Supply chain design and analysis: models and methods”, International Journal of Production Economics, Vol. 55 No. 3, pp. 281-94. Bensoussan, A., C¸akanyildirim, M. and Sethi, S. (2007), “Optimal ordering policies for inventory problems with dynamic information delays”, Production and Operations Management, Vol. 16 No. 2, pp. 241-56. Braun, M.W., Rivera, D.E., Flores, M.E., Carlyle, W.M. and Kempf, K.G. (2003), “A model predictive control framework for robust management of multi-product, multi-echelon demand networks”, Annual Reviews in Control, Vol. 27, pp. 229-45. Cachon, G.P. and Lariviere, M.A. (2005), “Supply chain coordination with revenue-sharing contracts: strengths and limitations”, Management Science, Vol. 51 No. 1, pp. 30-44. Camacho, E.F. and Bordons, C. (2004), Model Predictive Control, Springer, London. Camarinha-Matos, L.M. (2009), “Collaborative networked organizations: status and trends in manufacturing”, Annual Reviews in Control, Vol. 33 No. 2, pp. 199-208. Camarinha-Matos, L.M. and Afsarmanesh, H. (2005a), “Collaborative networks: a new scientific discipline”, Journal of Intelligent Manufacturing, Vol. 16, pp. 439-52. Camarinha-Matos, L.M. and Afsarmanesh, H. (2005b), “The emerging discipline of collaborative networks”, Journal of Intelligent Manufacturing, Vol. 16 Nos 4-5, pp. 439-52. Camarinha-Matos, L.M. and Afsarmanesh, H. (2008), Collaborative Networks: Reference Modeling, Springer, New York, NY. Camarinha-Matos, L.M. and Macedo, P. (2010), “A conceptual model of value systems in collaborative networks”, Journal of Intelligent Manufacturing, Vol. 21 No. 3, pp. 287-99. Camarinha-Matos, L.M., Afsarmanesh, H. and Ortiz, A. (Eds) (2005), Collaborative Networks and Their Breeding Environments, Springer, New York, NY.

Chauhan, S.S., Proth, J.M., Sarmiento, A.M. and Nagi, R. (2006), “Opportunistic supply chain formation from qualified partners for a new market demand”, Journal of the Operational Research Society, Vol. 57, pp. 1089-99. Chen, I.J. and Paulraj, A. (2004), “Understanding supply chain management: critical research and a theoretical framework”, International Journal of Production Research, Vol. 42 No. 1, pp. 131-63. Chen, Z.-L. (2010), “Integrated production and outbound distribution scheduling: review and extensions”, Operations Research, Vol. 58 No. 1, pp. 130-48. Choi, T.Y., Dooley, K.J. and Rungtusanatham, M. (2001), “Supply networks and complex adaptive systems: control versus emergence”, Journal of Operations Management, Vol. 19 No. 3, pp. 351-66. Christopher, M. (2012), Logistics and Supply Chain Management: Creating Value-adding Networks, Financial Times, Prentice-Hall, Dorchester. Christopher, M. and Towill, D.R. (2001), “An integrated model for the design of agile supply chains”, International Journal of Physical Distribution and Operations Management, Vol. 31, pp. 235-44. Dedrick, J., Xu, S. and Zhu, K. (2008), “How does information technology shape supply-chain structure? Evidence on the number of suppliers”, Journal of Management Information Systems, Vol. 25 No. 2, pp. 41-72. Dekkers, R. (Ed.) (2009a), Dispersed Manufacturing Networks: Challenges for Research and Practice, Springer, London. Dekkers, R. (2009b), “Distributed manufacturing as co-evolutionary system”, International Journal of Production Research, Vol. 47 No. 8, pp. 2031-54. Dekkers, R. and van Luttervelt, C.A. (2006), “Industrial networks: capturing changeability?”, International Journal of Networking and Virtual Organisations, Vol. 3 No. 1, pp. 1-24. de Kok, A.G. and Graves, S.C. (2004), Supply Chain Management: Design, Coordination and Operation, Elsevier, Amsterdam. Disney, S.M. and Towill, D.R. (2002), “A discrete linear control theory model to determine the dynamic stability of vendor managed inventory supply chains”, International Journal of Production Research, Vol. 40 No. 1, pp. 179-204. Disney, S.M., Towill, D.R. and Warburton, R.D.H. (2006), “On the equivalence of control theoretic, differential, and difference equation approaches to modeling supply chains”, International Journal of Production Economics, Vol. 101, pp. 194-208. Gjerdrum, J., Shah, N. and Papageorgiou, L.G. (2001), “A combined optimization and agent-based approach to supply chain modeling and performance assessment”, Production Planning & Control, Vol. 12 No. 1, pp. 81-8. Graves, S.C. and Willems, S.P. (2005), “Optimizing the supply chain configuration for new products”, Management Science, Vol. 51 No. 8, pp. 1165-80. Guille´n, G., Badell, M., Espun˜a, A. and Puigjaner, L. (2006), “Simultaneous optimization of process operations and financial decisions to enhance the integrated planning/scheduling of chemical supply chains”, Computers & Chemical Engineering, Vol. 30 No. 3, pp. 421-36. Gunasekaran, A. and Ngai, N.W.T. (2009), “Modeling and analysis of build-to-order supply chains”, European Journal of Operational Research, Vol. 195 No. 2, pp. 319-34. Gunasekaran, A., Lai, K.-H. and Edwin Cheng, T.C. (2008), “Responsive supply chain: a competitive strategy in a networked economy”, Omega, Vol. 36 No. 4, pp. 549-64.

Modelling of supply chains

993

JMTM 23,8

Helo, P., Kitaygorodskaya, N., Salminen, S. and Jiao, R.J. (2009), “Designing and modeling agile supply-demand networks”, in Dekkers, R. (Ed.), Dispersed Manufacturing Networks: Challenges for Research and Practice, Springer, London. Ivanov, D. (2009), “DIMA – a research methodology for comprehensive inter-disciplinary modeling of production and logistics networks”, International Journal of Production Research, Vol. 47 No. 5, pp. 1133-55.

994

Ivanov, D. and Sokolov, B. (2010), Adaptive Supply Chain Management, Springer, London. Ivanov, D. and Sokolov, B. (2012a), “Dynamic supply chain scheduling”, Journal of Scheduling, Vol. 15 No. 2, pp. 201-16. Ivanov, D. and Sokolov, B. (2012b), “Structure dynamics control approach to supply chain planning and adaptation”, International Journal of Production Research. Ivanov, D., Arkhipov, A. and Sokolov, B. (2007), “Intelligent planning and control of manufacturing supply chains in virtual enterprises”, International Journal of Manufacturing Technology and Management, Vol. 11 No. 2, pp. 209-27. Ivanov, D., Sokolov, B. and Dolgui, A. (2012), “Applicability of optimal control theory to adaptive supply chain planning and scheduling”, Annual Reviews in Control, Vol. 36 No. 1, pp. 73-84. Ivanov, D., Sokolov, B. and Kaeschel, J. (2010), “A multi-structural framework for adaptive supply chain planning and operations with structure dynamics considerations”, European Journal of Operational Research, Vol. 200 No. 2, pp. 409-20. Ivanov, D., Sokolov, B. and Ka¨schel, J. (2011), “Integrated supply chain planning based on a combined application of operations research and optimal control”, Central European Journal of Operations Research, Vol. 19 No. 3, pp. 219-317. Jain, V., Wadhwa, S. and Deshmukh, S.G. (2009), “Revisiting information systems to support a dynamic supply chain: issues and perspectives”, Production Planning & Control, Vol. 20 No. 1, pp. 17-29. Kleindorfer, P.R. and Saad, G.H. (2005), “Managing disruption risks in supply chains”, Production and Operations Management, Vol. 14 No. 1, pp. 53-68. Klibi, W., Martel, A. and Guitouni, A. (2010), “The design of robust value-creating supply chain networks: a critical review”, European Journal of Operational Research, Vol. 203 No. 2, pp. 283-93. Kouvelis, P., Chambers, C. and Wang, H. (2006), “Supply chain management research and production and operations management: review, trends, and opportunities”, Production and Operations Management, Vol. 15 No. 3, pp. 449-69. Krajewski, L., Wei, J.C. and Tang, L.-L. (2005), “Responding to schedule changes in build-to-order supply chains”, Journal of Operations Management, Vol. 23, pp. 452-69. Kuehnle, H. (2008), “A system of models contribution to production network (PN) theory”, Journal of Intelligent Manufacturing, Vol. 18 No. 5, pp. 543-51. Kuehnle, H. (Ed.) (2010), Distributed Manufacturing: Paradigm, Concepts, Solutions and Examples, Springer, London. Laı´nez, J.M., Puigjaner, L. and Reklaitis, G.V. (2009), “Financial and financial engineering considerations in supply chain and product development pipeline management”, Computers & Chemical Engineering, Vol. 33 No. 12, pp. 1999-2011. Lee, H.L. and O¨zer, O¨. (2008), “Unlocking the value of RFID”, Production and Operations Management, Vol. 16 No. 1, pp. 40-64.

Lee, J., Palekar, U.S. and Qualls, W. (2011), “Supply chain efficiency and security: coordination for collaborative investment in technology”, European Journal of Operational Research, Vol. 210, pp. 568-78. Mu¨ller, E., Horbach, S. and Ackermann, J. (2008), “Integrative planning and design of logistics structures and production plants in competence-cell-based networks”, International Journal of Service Operations and Informatics, Vol. 3 No. 1, pp. 40-52. Nair, A. and Vidal, J.M. (2011), “Supply network topology and robustness against disruptions – an investigation using multi-agent model”, International Journal of Production Research, Vol. 49 No. 5, pp. 1391-404. Nepal, B., Monplaisir, L. and Famuyiwa, O. (2012), “Matching product architecture with supply chain design”, European Journal of Operational Research, Vol. 216 No. 2, pp. 312-25. Noori, H. and Lee, W.B. (2002), “Factory-on-demand and smart supply chains: the next challenge”, International Journal of Manufacturing Technology and Management, Vol. 4 No. 5, pp. 372-83. Noori, H. and Lee, W.B. (2009), “Dispersed network manufacturing: an emerging form of collaborative networks”, in Dekkers, R. (Ed.), Dispersed Manufacturing Networks: Challenges for Research and Practice, Springer, London, pp. 39-58. Perea, E., Grossmann, I., Ydstie, E. and Tahmassebi, T. (2000), “Dynamic modeling and classical control theory for supply chain management”, Computers & Chemical Engineering, Vol. 24, pp. 1143-9. Polderman, J.W. and Willems, J.C. (1998), Introduction to Mathematical Systems Theory: A Behavioral Approach, Springer, New York, NY. Pontrayagin, L.S., Boltyanskii, V.G., Gamkrelidze, R.V. and Mishchenko, E.F. (1964), The Mathematical Theory of Optimal Processes, Pergamon Press, Oxford. Safaei, A.S., Moattar Husseini, S.M., Farahani, R.Z., Jolai, F. and Ghodsypour, S.H. (2010), “Integrated multi-site production-distribution planning in supply chain by hybrid modeling”, International Journal of Production Research, Vol. 48 No. 14, pp. 4043-69. Sarimveis, H., Patrinos, P., Tarantilis, C.D. and Kiranoudis, C.T. (2008), “Dynamic modeling and control of supply chain systems: a review”, Computers & Operations Research, Vol. 35, pp. 3050-561. Sarkis, J., Talluri, S. and Gunasekaran, A. (2007), “A strategic model for agile virtual enterprise partner selection”, International Journal of Operations & Production Management, Vol. 27 No. 11, pp. 1213-34. Schwartz, J.D. and Rivera, D.E. (2010), “A process control approach to tactical inventory management in production-inventory systems”, International Journal of Production Economics, Vol. 125 No. 1, pp. 111-24. Simchi-Levi, D., Wu, S.D. and Zuo-Yun, S. (2004), Handbook of Quantitative Supply Chain Analysis, Springer, New York, NY. Surana, A., Kumara, S., Greaves, M. and Raghavan, U.N. (2005), “Supply-chain networks: a complex adaptive systems perspective”, International Journal of Production Research, Vol. 43 No. 20, pp. 4235-65. Swaminathan, J.M., Smith, S.F. and Sadeh, N.M. (1998), “Modeling supply chain dynamics: a multiagent approach”, Decision Science, Vol. 29 No. 3, pp. 607-32. Taha, H.A. (2002), Operations Research: An Introduction, 7th ed., Prentice-Hall, New York, NY.

Modelling of supply chains

995

JMTM 23,8

996

Tayur, S., Ganeshan, R. and Magazine, M. (1999), Quantitative Models for Supply Chain Management, Kluwer Academic Publishers, Boston, MA. Vahdani, B., Zandieh, M. and Roshanaei, V. (2011), “A hybrid multi-stage predictive model for supply chain network collapse recovery analysis: a practical framework for effective supply chain network continuity management”, International Journal of Production Research, Vol. 49 No. 7, pp. 2035-60. Vickery, S.K., Droge, C., Setia, P. and Sambamurthy, V. (2010), “Supply chain information technologies and organisational initiatives: complementary versus independent effects on agility and firm performance”, International Journal of Production Research, Vol. 48 No. 23, pp. 7025-42. Yadav, S.R., Mishra, N., Kumar, V. and Tiwari, M.K. (2011), “A framework for designing robust supply chains considering product development issues”, International Journal of Production Research, Vol. 49 No. 20, pp. 6065-88. Zhang, L., You, X., Jiao, J. and Helo, P. (2009), “Supply chain configuration with co-ordinated product, process and logistics decisions: an approach based on Petri nets”, International Journal of Production Research, Vol. 47 No. 23, pp. 6681-706. Zhuge, H. (2011), “Semantic linking through spaces for cyber-physical-socio intelligence: a methodology”, Artificial Intelligence, Vol. 175 Nos 5-6, pp. 988-1019. Further reading Drenick, R.F. and Shahbender, R.A. (1957), “Adaptive servomechanism”, AIEE Transactions, Vol. 76, pp. 286-92. Ivanov, D. (2010), “Adaptive aligning of planning decisions on supply chain strategy, design, tactics, and operations”, International Journal of Production Research, Vol. 48 No. 13, pp. 3999-4017. Kalinin, V.N. and Sokolov, B.V. (1985), “Optimal planning of the process of interaction of moving operating objects”, International Journal of Differential Equations, Vol. 21 No. 5, pp. 502-6. Kim, D.H. (2007), “Towards an architecture modeling language for networked organizations”, in Camarinha-Matos, L., Afsarmanesh, H., Novais, P. and Analide, C. (Eds), Establishing the Foundation of Collaborative Networks, Springer, Boston, MA, pp. 309-16. Lee, H.L. (2004), “The triple-A supply chain”, Harvard Business Review, Vol. 82, October, pp. 102-12. Puigjaner, L. and Lainez, J.M. (2008), “Capturing dynamics in integrated supply chain management”, Computers & Chemical Engineering, Vol. 32, pp. 2582-605. Zhang, W.-B. (2002), “Theory of complex systems and economic dynamics”, Nonlinear Dynamics, Vol. 6 No. 2, pp. 83-101. About the authors Dmitry Ivanov is Full Professor for International Supply Chain Management at Berlin School of Economics and Law, Germany. He studied Industrial Engineering and Business Administration in St Petersburg and graduated in 2000 with distinction. From 2000-2005 he was mainly engaged in industry and consulting, especially for ERP systems. He gained his PhD (Drrer.pol.), Doctor of Economics Science, and Habilitation (Doctor of Science) degrees in 2006, 2008, and 2011, respectively. In 2010-2011 he was Professor and Acting Chair of Operations Management at University of Hamburg. He is the (co)-author of more than 170 scientific works, including the monograph Adaptive Supply Chain Management. His research interests lie in the area of adaptive supply chains, applied optimal control theory, MP, socio-informational systems, neo-cybernetics, and collaborative networks. He is a Member of the IFAC TC 5.2 and Chair of German-Russian logistics scientific society DR-LOG. His works have been published in various academic journals, including International Journal of Production Research, Annual Reviews in Control,

European Journal of Operational Research, Journal of Scheduling, etc. Dmitry Ivanov is the corresponding author and can be contacted at: [email protected] Boris Sokolov is Professor and a Deputy Director for Research at the Saint-Petersburg Institute of Informatics and Automation (SPIIRAS) of the Russian Academy of Science. He is the author of a new scientific lead: optimal control theory for structure dynamics of complex systems. His research interests include basic and applied research in mathematical modelling, optimal control theory, mathematical models and methods of support and multi-criteria decision making in complex organization-technical systems under uncertainties. He is the author and co-author of three books on systems and control theory and of more than 180 scientific papers. His works have been published in various academic journals, including European Journal of Operational Research, Journal of Scheduling, etc.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

Modelling of supply chains

997

The current issue and full text archive of this journal is available at www.emeraldinsight.com/1741-038X.htm

JMTM 23,8

Entropy assessment of supply chain disruption

998

Norwich Business School, University of East Anglia, Norwich, UK, and

Olatunde Amoo Durowoju and Hing Kai Chan Xiaojun Wang School of Economics Finance and Management, University of Bristol, Bristol, UK Abstract Purpose – Manufacturing organizations and networks are heavily dependent on the flow of information within and across organization boundaries. A disruption in information flow might interrupt the operations of the organization and make management even more difficult. The purpose of this paper is to incorporate information theory approach to investigate the perturbation introduced into a manufacturing organization as a result of disruption in the flow of critical information needed in manufacturing operations. Design/methodology/approach – This study proposes the use of entropy theory to assess the level of risk introduced by different sources of perturbation into the material flow stream and the use of discrete event simulation to investigate the impact of the resulting disruption on collaborating members. Findings – The result of the analysis carried out on the effect of system failure on supply chain performance revealed that the retailer experiences the most uncertainty in the supply chain while the holding cost constitutes the most unpredictable cost measure when a system failure breach occurs. For the manufacturer and wholesaler, the holding cost is responsible for most of the uncertainty in predicting the impact of the threat on inventory management cost, while the backlog cost holds the highest complexity level for the retailer. Practical implications – Once this methodology is well developed for use in industrial networks, it can serve as a risk assessment, risk monitoring and risk prediction tool. The paper also calls for a proactive approach to disruption risk management. Originality/value – This paper proposes a novel approach to assess the impact of information disruption, using entropy theory coupled with simulation methodology. Keywords Manufacturing industries, Supply chain management, Operations and production management, Costs, Information technology, Risk impact assessment, Perturbation, Simulation, Supply chain disruption, Operations cost, Operational complexity, Entropy assessment Paper type Research paper

Journal of Manufacturing Technology Management Vol. 23 No. 8, 2012 pp. 998-1014 q Emerald Group Publishing Limited 1741-038X DOI 10.1108/17410381211276844

1. Introduction Traditionally, supply chain management (SCM) focused around the manufacturer and their immediate suppliers. But in today’s world, SCM focuses on the optimization of all movement of goods and/or services and the flow of information, starting with the suppliers’ supplier all the way through to the customers’ customer (Plenert, 2002). In other words, firms have seen the need to manage the physical flows and the information flows up and down the supply stream in a coordinated manner. Management of these flows are increasingly becoming more challenging as there is currently a marked increase in the geographical dispersion of manufacturing sites, suppliers, warehouses and customers in today’s manufacturing networks (Colotla et al., 2003; Tse et al., 2011). However, information

technology (IT) has been effective in improving the efficiency of inter-organizational operations by mitigating uncertainties inherent in collaborative networks via efficient transmission of information. IT has been used extensively by manufacturing networks to enhance the efficiency and effectiveness of manufacturing operations. Beyond that, it has been used as a vehicle for both internal and external integration. For most organizations, IT is core to their competitive advantage. It is safe to say that most organizations are now extending the way IT is being utilized to improve their competitiveness in this highly competitive global market. For example, the concept of cloud computing is a somewhat recent development in the way IT is being exploited where hardware or software resources, or a combination of both, are accessed anywhere in the world by an organization or an individual via the internet (Amir, 2009; Smith, 2009; Armbrust, 2010). These resources are shared amongst many users, abstracted, available on demand, scalable, and configurable (Marston et al., 2011). Some have even extended the concept to manufacturing where product design, manufacturing, testing, management, and all other stages of a product life cycle are encapsulated into cloud services and managed centrally (Xu, 2012) similar in principle to (but not the same as) distributed manufacturing described in Bennett and Dekkers (2009). Essentially the idea is to pay for what you use as opposed to renting where you pay for the specified period irrespective of use (Subashini and Kavitha, 2011), which can be quite expensive. These growing advantages of leveraging processes using IT and the fact that IT cost is becoming more and more inexpensive have led to the increasing level of collaboration found in many networks. There is an increase in the level of connectivity and interdependency (referred to as complexity) found in the network as businesses come together in the spirit of collaboration being geographically distant from one another. These complexities affect the dynamics of the network and thus require appropriate network management approach. It is perceptible that as complexity increases so does the level of uncertainty enveloping the business or supply chain increases. A compromise in the operations of one member could affect other interconnected members of the network, the extent to which depends on how reliant their operations are on the compromised company (Craighead et al., 2007). It is therefore apparent that strategies should change to accommodate the increased or increasing complexity present in the supply chain. Consequently, there should be a pro rata increase in the level of information required to monitor and control the operations of the network. Nowadays, it is difficult to separate information from manufacturing as the absence of information results in the poor performance of manufacturing operations. Managing manufacturing operations effectively requires efficient management and use of information. One cannot be managed with total disregard to the other, it will only spell disaster. The fusion of both is crucial to business survival and serves as the foundation for competitive advantage. In this study we try to examine an aspect of network management which is information management by looking at the disruption in the flow of information (defined here as perturbation) and how this affects manufacturing practices. It will be interesting to know which areas of operation are most vulnerable. There are various sources of perturbation (also referred to here as threat) and these sources have varying level of impact on supply chain performance. By examining the impact of each threat, we can lay a foundation upon which appropriate manufacturing decisions are made. It is important to understand what level of integration in the network is appropriate considering the perturbations these threats have and how it affects performance of the entire network (Durowoju and Chan, 2012). Equally important is the

Supply chain disruption

999

JMTM 23,8

1000

way the entire network reacts in the event of perturbation (information flow disruption). For example, which ordering policy would reduce the impact of the perturbation on the network and how does perturbation at each tier in the supply chain affect the entire network? Also how are the different network structures affected by these perturbations. Beyond this, management should concern themselves with the areas of operation that requires the most attention depending on their vulnerability. Combining the answers to these questions would help create understanding of the dynamics of the network and the interaction between information management and operations management. This would inform an appropriate strategy designed to cater for the needs of specific collaborative networks. This study stems from the need to develop concepts suitably adapted to network management which is different from those intended for individual organizations. Understanding some of the dynamics mentioned earlier provides valuable insight and is a positive step in the right direction. This study utilizes the concept of entropy theory to analyze the degree of perturbation each perturbation source (threat) introduces to the supply chain and its members and how this affects supply chain performance. Discrete event simulation is recommended to investigate the impact each perturbation source has on supply chain performance such as supply chain cost and customer service level. The cost being investigated by this study are operations costs which includes the inventory holding cost, backlog cost and ordering cost, and do not include cost associated with damage to company’s image or regulatory fines. This paper is organized as follows. Section 2 presents a review of relevant literature. In Section 3, the measurement of perturbation using entropy theory is described and the proposed framework is illustrated. Section 4 summarizes the study and presents implication for future research. 2. Literature review The trend in competitiveness has moved away from individual competition to more collaborative network competition. Supply chains now compete against each other and there is an increasing move from individual objective to a more network objective. Organizations are coming together forming collaborative networks in order to take advantage of the strengths and opportunities that each one has. Therefore, competition in the global market place has grown from inter firm competition to highly efficient collaborative supply chain network within and between industries (Lancioni et al., 2003). Organizations no longer compete on individual basis alone but now compete on a supply chain basis. This collaboration has been achieved by integrating business processes using IT and information communication technologies (ICT). Integration has been advocated for in literature and this has been linked with improved business performance (Flynn et al., 2010; Wang and Chan, 2010). As described by Waters (2006), integration has been improved both internally and externally by the use of IT. Internally, by using bar codes, magnetic stripes and radio frequency identification to track individual packages; monitoring vehicles through telematics; controlling warehouses through automatically guided vehicles; monitoring transactions and planning operations – and a host of other functions. Externally, by allowing vendor-managed inventory; collaborative planning, forecasting and replenishment; synchronized material movement through the whole supply chain, payments by electronic fund transfer, roadside detectors to monitor traffic conditions and route vehicles around congestion – and so on.

IT has aided collaborative practices in the areas of research and development, product innovation, SCM, global networking, governance, etc. and the benefits derivable from its use has driven its further development. IT especially has been extended to increase the amount of benefits it can offer in terms of ease and speed of communication. Communication between businesses has greatly improved over the years and huge efforts have been invested into communication with customers as well. Fuelling this vast improvement is the plethora of investigations into the benefits of communication and information sharing (Bourland et al., 1996; Chan and Chan, 2009; Cheng, 2011; Huang and Lin, 2010; Li et al., 2006; Li and Lin, 2006; Yang et al., 2009; Zhou and Benton Jr, 2007; Yu et al., 2001; Raschke, 2010), and as such most organizations now heavily depend on information systems (IS). At a network or supply chain level, competitiveness is no longer a case of sharing information but how efficiently it can be transmitted. The development in IT seriously helped this course as the introduction of IS such as electronic data interchange; internet, and world wide web aided the timely exchange of data (Gunasekaran and Ngai, 2004). This not only helped mitigate the effect of supply chain uncertainties but enabled intra-organization information sharing as well as inter-organization transactions (Kappelman et al., 1996; Gallear et al., 2008). 2.1 The role of information in supply chain operations Concerns over the efficiency and effectiveness of supply chain operations has been raised and assessed over the years by academics as well as practitioners. These inefficiencies have been perceived to be as a result of uncertainties in key aspect of operations, which affect the flow of information and material along the supply chain. For example, uncertainties in demand, for a long time, has lead to distortion in the accurate capture of demand information and the ability to respond to them in a timely and efficient fashion. Traditionally, due to the absence of the integration initiative at the time, most members in each tier of the chain relied on historical demand or order data to forecast future demand. The traditional forecasting method could not accommodate the uncertainties of demand, as a result supply chain members had to order and produce surplus to be able to accommodate these uncertainties. An interesting pattern emerged as the forecasted demand was amplified the further you go upstream the chain (Lee et al., 2004). The consequence of this action was that there tend to be more inventories in store than is needed, which can be very costly and even worse if the products are perishable. These uncertainties have led to poor performance of the supply chain and the reduction in the quality of product and/or services offered. The idea of information integration (in which IT is used to leverage operational activities) later emerged where supply chain performance is improved not only by sharing demand information, but other types of information such as production; inventory; capacity; and lead time information, as you go upstream (Mukhopadhyay and Kekre, 2002; Kulp et al., 2004; Devaraj et al., 2007; Yu et al., 2010; Lau et al., 2004). This enabled supply chain members to keep just the amount of inventory needed to efficiently cater for demand. As these technologies evolved, it later became apparent that the information being shared should be of good quality and be passed in a timely manner to help mitigate the effect of demand uncertainties (Bourland et al., 1996; Wiengarten et al., 2010). The flow of information is crucial to business survival. For materials to move down the supply chain there has to be a prior movement of information up the chain.

Supply chain disruption

1001

JMTM 23,8

A disruption in this flow of information would have an effect on the movement of materials down the chain and hence lead to customer dissatisfaction. There is of course pressing need for operations managers to carefully protect and manage the flow of information and plan appropriately for disruption in this flow. It is not surprising that a study by Munoz and Clements (2008) suggests information flow delays play a larger role as a contributor to lost sale revenues in the supply chain than material delay does.

1002 2.2 Information quality and the internet Li et al. (2006) surmised that the advantage of using inter-organizational IS does not only come from efficient transaction processing and improved monitoring and information processing capacity as previous studies indicate, but also from sharing and improved access to key business information. Studies in literature have shown that sharing information is important in supply chains, however, the derived benefits only come under the right conditions, that is, right information being shared at the right time in the right format by the right entities within the right environment (Huang and Lin, 2010; Chan and Chan, 2009). According to Wiengarten et al. (2010) information quality plays a pivotal role in the success of collaborative practices, and this is due to quality factors such as timeliness, accuracy, relevance and added value of the shared information. Security concerns as it relates to information sharing are privacy, protection of proprietary information, and preservation of the quality of information. Therefore, a disruption in the flow of critical information presents a risk to information quality. 2.3 Perturbation assessment, threat and cost A number of studies have approached disruption risk in different ways ranging from conceptual, empirical, simulation, survey, case study and review or a combination of these (Rao and Goldsby, 2009; Olson and Wu, 2010; Rainer et al., 1991). Rao and Goldsby (2009) conducted an extensive review of supply chain disruption literature and created a typology of disruption types and sources. However, they did not mention IT itself as a source of disruption. Most risk studies have been primarily based on threats other than those from IT and ICTs while a few studies have suggested IT security as a potential risk to the supply chain (Schmitt and Singh, 2009; Kim et al., 2011; Rees et al., 2011). A proper classification of threats to information is therefore needed to gain a more comprehensive typology of threat sources. A few studies have examined the impact of disruption on supply chain operations. While some of these studies have examined the effect of physical disruption such as natural disasters (Samir, 2008), interestingly, a few others have examined the effect of IT security incidents (Deane et al., 2009; Kim et al., 2011; Loch et al., 1992; Pisello, 2004). Some approach have focused primarily on disruption effect (in terms of delay in information flow) on supply chain without any regard to cause (Munoz and Clements, 2008; Schmitt and Singh, 2009) while others have looked more specifically at how specific disruption types (threats) affect the supply chain (Altay and Ramirez, 2010; Craighead et al., 2007). While the former approach give a more general assessment of the impact of perturbation, the latter gives clearer understanding of the dynamics of threats and how they impact the chain. From Altay and Ramirez, we understand that disasters lead to disruption which affects all sectors of the chain but certain threats have more impact than others. This threat-type impact study seems to be lacking in information disruption literature. To know how these threats affect the performance of the network is crucial to appropriate disruption risk planning and management.

There are various types and causes of information disruption (perturbation) that can occur in the supply chain/network. There is however a need for systematic classification of information threats and how these threats impact the entire network. Although some classification of threat exist emanating from surveys done by practitioners (Baker et al., 2010; Potter and Beard, 2010); institutions (Stoneburner et al., 2002; Richardson, 2009) and; academic sources (Samy et al., 2010; Warren, 2000; Whitman, 2003; Loch et al., 1992; Kim et al., 2011), these are partial lists and there is need for a complete list of threats to information and how these impact the network and how the network characteristics can be leveraged to contain these impact. It is therefore essential to have a systematic and in-depth investigation into the impact of information management on network operations and collaboration success. To set the ball rolling, this study proposes a methodology to investigate the impact of established threats to information flow, using entropy theory, on the performance of supply chain and how network characteristics moderate these impacts. The network characteristics referred to in this study are the structure of the supply chain, level of integration and ordering policy. These will be explained in more details in Section 3. 3. Measuring IT perturbations IT perturbations for the most part represent the uncertainties inherent in an IT system. In order to plan for a formidable IT strategy for specific supply networks (SCN), the impact of IT perturbations on supply chain operations must be well understood. Knowing the impact of information uncertainty on an organization and how this affects other members would appear to be a requisite foundation to any sustainable information management policy. This will help the supply chain prioritize “what?” and “where?” along the supply chain require intensified monitoring and what appropriate mitigation solution should be adopted in the event of disruption. This way, the supply chain can effectively and efficiently plan its operations, optimally prepared for any eventualities. There have been many studies looking at IT risks to an organization and a very few have studied IT risk to the supply chain. However, there is yet to be an academic study on the impact of IT disruptions on the supply chain at an operational level by considering deliverables such as inventory cost, production cost, customer service level, etc. It is not yet evidenced whether this will have a ripple effect (similar in principle to bullwhip effect where there is significant increase in impact) or a trickle down effect (i.e. where there is not a significant increase in impact) on members as one goes upstream the supply chain. 3.1 Entropy assessment of IT risk While previous studies have estimated information security risks as a function of threat occurrence and the associated financial loss (Rees et al., 2011; Deane et al., 2009), only a few have employed the entropy theory. Although the concept of using entropy as an approach to determine uncertainty has been used in literature by Frizelle and Woodcock (1995), Frizelle and Efstathiou (2002), Ronen and Karp (1994), Sivadasan et al. (2002) and Martı´nez-Olvera (2008), the closest previous studies have come to applying entropy as an assessment tool in security studies has been in data privacy studies such as disclosure risk assessment (Airoldi et al., 2011), measuring anonymity (Bezzi, 2007; Deng et al., 2007). The argument is that since complexity of a system (characterized by the uncertainty of a system) can be measured using entropy approach

Supply chain disruption

1003

JMTM 23,8

1004

(Frizelle and Woodcock, 1995; Martı´nez-Olvera, 2008), and, the little is known about a random variable the more the entropy of that variable, hence the level of entropy of a perturbation source can be determined once the probability of occurrence is known (Airoldi et al., 2011). The approach is to evaluate disruption threats using established threat occurrence to work out the level of entropy each threat introduces into the system. This will help identify those threats that are hot spots to guide management decision in selecting appropriate countermeasures and mitigation solutions. Frizelle and Efstathiou (2002) explained that high entropy can impede flow by introducing obstacles that makes supply chain operations less predictable. By inference, disruption in information flow introduces obstacles to the flow of operation and the predictability of these disruptions can help evaluate the level of chaos they introduce into the system. Airoldi et al. (2011) demonstrated that using entropy approach in estimating risk is very effective but this has not been applied to impact assessment. The mathematical definition of entropy as prescribed by Shannon (1948) is a quantitative measure of uncertainty (Martı´nez-Olvera, 2008; Sivadasan et al., 2002): H ðSÞ ¼ 2

n X

pi log2 pi

ð1Þ

i¼1

H(S) is the entropy level of the system, defined here as the expected amount of information needed to describe the state of the system S, and pi is the probability of breach impact i (i ¼ 1, . . . ,n) occurring, where pi $ 0 and n X

pi ¼ 1

i¼1

Sivadasan et al. (2002) established entropy as a measure of complexity which derives from variation in information and material flow between a supplier and a customer. They obtained data from two organizations and measured the variation between sales forecast and sales order; sales order and actual dispatch; purchasing forecast and purchasing orders; purchasing orders and actual deliveries. These variations were sources of uncertainty which result in operational complexities that can be passed on from one business to the other (Sivadasan et al., 2002). They proposed two models for determining operational complexity under two conditions. First is complexity associated with knowing whether the system is “in control” or “not in control” denoted by the “in or not in control operational complexity index,” OCI (S INC) as shown in equation (2), where P is the probability of being in control. OCI (S INC) is a measure of the amount of information needed to describe the “in-control” or “not in control” state of the system. The closer the probability of incidence is to 0.5, the closer the OCI (S INC) value is to one (Durowoju et al., 2011): OCI ðS INC Þ ¼ 2P log2 P 2 ð1 2 PÞlog2 ð1 2 PÞ

ð2Þ

Second is the complexity associated with not-in-control states, given that the system is not in control, i.e. a breach has occurred. This is denoted by the “not in control operational complexity index,” OCI (S NC) shown in equation (3), where pi/j is the conditional probability computed over the “not in control” state with states i (i ¼ 1, . . . , n) at nodes j ( j ¼ 1, . . . , M). This index is a measure of the amount of information needed to monitor the extent to which the system is not in control:

OCI ðS NC Þ ¼ 2ð1 2 PÞ

M X n X Pi Pi log2 j j j¼1 i¼1

ð3Þ

According to Sivadasan et al. (2002), the sum of equations (2) and (3) is the total operational complexity denoted by the operational complexity index, OCI (Stotal). It follows that the higher the operational complexity index, the higher the entropy introduced by the breach into the system and hence the more the associated information needed to manage the system and vice versa. This has been adapted to reflect the level of perturbation created by disruptive threat in a supply chain. However, in our study, variation between the performance measures with and without breach was used. 3.2 Threat, entropy and supply chain performance To illustrate this concept, we use the simulation approach of a previous study (Durowoju and Chan, 2012) on the impact of a single threat (system failure) on a basic supply chain. Result from the study is not reliable due to some limitations in the simulation approach. These limitations were improved upon in this study. Using arena simulation software, we modified the basic supply chain scenario comprising of three agents, the retailer; wholesaler and manufacturer. Each agent use an order-up-to stock policy (s, S) and when the inventory position (IP) falls below the re-order point, s, an order (S-IP) is placed to the adjacent agent upstream the chain. We increased the run to 600 days and a warm-up period of 120 days was estimated using a time-series inspection method, giving an effective simulation period of 480 days. The number of replication was also extended to 50 and this presented us with reliable simulation result (at 95 percent confidence level) on the impact of system failure on supply chain members. From the experimental result, we first establish the impact on supply chain performance indices such as backlog cost, holding cost and ordering cost for each supply chain member/agent by comparing between the operational costs of each agent when breach occurs (i.e. system out of control) to when there is no breach (i.e. system under control). Comparison is between same agents, i.e. manufacturer in the breach scenario (BS) and manufacturer in the non-breach scenario (n-BS). This is done by computing the difference between the outputs of each replication in both scenarios. Once this is completed, the difference is categorised into bins (which is also referred to as states) and the frequency is computed. Table I is an example of this computation for the manufacturer only. Negative indicates there is reduction in cost when breach occurs and positive indicates there is increase in cost when breach occurs. Since risk assessment is based on the negative impact of threat, we therefore categorise values in the range of 2 25 to 0 to indicate a state of control (in-control) where breach has no negative impact on operation while those of 1 and above indicate an out-of-control state. Second, we calculate the probability of each state occurring and finally compute the entropy score for each state using equations (2) and (3) and the total entropy score OCI (Stotal) is then obtained. The OCI (S NC), OCI (S INC) and OCI (Stotal) values for the example in Table I are 1.62, 0.8 and 2.42, respectively. Table II shows the various entropy values for the entire supply chain agents. From the result in Table II, we see that the retailer has the most total entropy value of 5.16. This shows that the retailer experiences the most uncertainty, hence requires more attention paid. This complexity is largely due to uncertainties in backlog and holding cost and the impact on ordering cost is more predictable than the other two costs.

Supply chain disruption

1005

JMTM 23,8

1006

Table I. The manufacturer example

Replication

Difference in cost (BS minus n-BS)

Replication

Difference in cost (BS minus n-BS)

25.73 2.92 11.84 1.36 5.68 12.01 2.53 21.84 25.90 20.55 20.83 3.02 16.88 30.19 5.14 6.69 6.97 11.72 4.04 9.48 2.59 20.35 22.81 9.92 2.40

26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

12.48 2 0.90 11.78 9.75 4.12 4.59 14.87 7.37 2 4.88 12.84 5.47 16.80 5.74 43.52 5.87 2 5.29 9.20 2 10.91 6.60 2 0.36 3.65 6.41 6.69 2 5.34 2.60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Supply chain agent Retailer Wholesaler Manufacturer Table II. Operational complexity introduced by system failure

Total score

INC

OCI (S ) OCI (SNC) OCI (Stotal) OCI (SINC) OCI (SNC) OCI (Stotal) OCI (SINC) OCI (SNC) OCI (Stotal) OCI (SINC) OCI (SNC) OCI (Stotal)

Bin

Frequency

220 215 210 25 0 5 10 15 20 25 30 35 40 45 50

0 0 1 4 7 11 15 7 2 1 0 1 0 1 0

Backlog cost

Holding cost

Ordering cost

Total score

0.72 1.27 1.99 0.94 0.00 0.94 0.86 0.43 1.28 2.52 1.7 4.21

0.80 1.17 1.97 0.86 1.71 2.57 0.80 1.62 2.42 2.46 4.5 6.96

1.00 0.20 1.20 1.00 0.20 1.20 0.99 0.19 1.18 2.99 0.59 3.58

2.52 2.64 5.16 2.8 1.91 4.71 2.65 2.24 4.88 7.97 6.79 14.75

For the retailer, the backlog cost represents the highest uncertainty with OCI (Stotal) of 1.99. This can be interpreted as the retailer will be less certain of what the impact of system failure will be on the backlog cost it incurs when the breach occurs. Second to the retailer is the manufacturer with OCI (Stotal) of 4.88. This level of uncertainty is largely due to the holding cost in this case. The holding cost represents an OCI (Stotal) of 2.42 for the manufacturer and that means it is more difficult to ascertain the level of impact such

threat would have on the manufacturer’s inventory holding operation. Hence the manufacturer would require more information or effort to attend to this. The same is the case for the wholesaler as the holding cost is the source of the highest complexity when compared to its backlog and ordering cost. We can also compare scores for each supply chain performance measures. In this case, the holding cost (OCI (Stotal) of 6.96) constitutes the highest source of uncertainty when looking at the supply chain as a whole. This reveals that holding cost is the biggest hot spot and more attention should be paid to ensure that the impact of a threat on it should be minimized. A closer look at the result reveals that the only type of complexity experienced by the wholesaler in terms of backlog cost is only due to the uncertainty in OCI (S INC). The OCI (S NC) value is zero in this case which indicates that when the wholesaler experiences system failure, it is most likely that the extent of impact will be predictable. This is less predictable when it concerns ordering cost and most unpredictable when it concerns holding cost of the wholesaler. For the manufacturer, the impact on ordering cost is more predictable than backlog cost and holding cost, with holding cost being the most unpredictable when a breach occurs. While this study is not an optimality study, it provides a useful methodology to disruption risk assessment from a process-based view where the impact on supply chain members and supply chain performance indices can be assessed. Coupling the knowledge of the relative importance of each index with the knowledge of how each performance index is affected would be invaluable to establishing the appropriate mitigation strategy. The overall total entropy can be computed for each threat incidence and this can be used to identify and classify the threats that are more problematic and as such would need to be prioritised. In this example, a system failure represent an uncertainty level of 14.75. It is also important to understand the role supply chain characteristics play in either mitigating the impact of these threats or complicating them. For this reason, we propose the research framework shown in Figure 1. This approach also helps to understand the priority areas in supply chain design. The operational measures in the framework represent the performance measure at the agent level while the strategic measures looks at the supply chain as a whole by averaging the measures of all supply chain members. Perturbation versus order placement. Disruption of IS causes unavailability of service. The duration of which depends on the severity of the perturbation. In a scenario where an organization captures and processes market demand information with the aid of IS such as electronic point of sale, disruption of this service would mean that demand information would be inaccessible for a period of time. Consequently, the organization cannot place accurate order with its supplier and the ordering policy in place can either help the situation or make it worse. The framework is proposed to evaluate various ordering policies to understand which or a combination of which policies will be best to absorb the impact. Perturbation versus integration. The impact of disruption at varying level of integration can be investigated by analyzing scenarios where demand and inventory information is being shared by adjacent tiers. Integration can be modelled as sharing real time information (Durowoju and Chan, 2012) and the scenarios are presented as follows. The first scenario, RD-M, represents a situation where demand and inventory information is being shared between the retailer and distributor.

Supply chain disruption

1007

JMTM 23,8

Perturbation by Threat

Ordering Options

1008 Level of Integration RD-M

R-DM

RDM

Supply Chain Structure I

II

III

IV

V

Serial SC

Manufacturer’s SC

Distributor’s SC

Retailer’s SC

Network

Performance Measures

Figure 1. Research framework

Operational

Strategic

Backlog cost Ordering cost Holding cost Fill rate

Supply chain cost SC order fill rate

In R-DM, integration is between the distributor and the manufacturer only. The distributor receives order from the retailer and the manufacturer is able to see this order and the inventory at the distributor. In the last scenario, RDM represents integration between the retailer, the distributor, and the manufacturer where the manufacturer and distributor is able to capture real time demand data and inventory information of downstream members. Perturbation versus supply chain structure. A typical supply chain has four tiers: retailers; distributors; manufacturers and suppliers. In this experiment, we try to see the effect of perturbation on a serial supply chain; a retailer supply chain (SCR) (convergent); a manufacturer supply chain (SCM) (divergent); a distributor supply chain (SCD) and a SCN. A SCR is such that has one retailer and multiple agents up stream of the supply chain, while a SCM has one manufacturer instead with multiple agents down stream the supply chain. The SCD has one distributor in the chain alongside multiple agents up stream and down stream, while a SCN has multiple agents in all of the tiers of supply. A serial chain is one with a single agent in each tier of supply. To investigate the nature of the effect of these perturbations on partners, whether ripple effect or trickle down effect, we propose to examine the performance of each member of the supply chain relative to one another.

4. Conclusion and implication for future research Incidence of threat changes from year to year some worse than others. It is believed that many incidences of threats go unnoticed perhaps due to poor monitoring system or the absence of it. Organizations therefore need to adopt a proactive approach rather than a reactive one to manage risk. An organization needs to be able to detect any incidence of perturbation within its premises and must be able to log it. The ability to profile these perturbations and recognize “hot spots” is paramount in order to establish a formidable risk policy. The incidence and type of threat changes year in year out and as such, businesses should be able to contain any changes using an appropriate risk management system that can analyze these changes. Incidents that were no longer an issue a year ago, being mitigated by the measures already put in place, might become a serious issue the next year rising from the increased level of sophistication with which these threats manifest. In the same vein, threats considered benign the previous year might become malignant the next year following an increase in the number of occurrence. In assessing the emergent impact of the threats in IS, entropy theory is suggested. The theory quantifies the level of chaos each incidence of threat poses as an indication of threat level. Besides profiling the occurrence of perturbations within an organization along with the threats responsible for them and learning from one’s failure, it is equally important to learn from the failure of others. This is why survey information is important so that decisions can be made not just based on internal data but on the likelihood that external data represent. Many organizations have been felled by the incidence of threats they have never experienced before and so it would be intuitive to pay attention to the threats plaguing others even if it has not happened to an organization yet. Not doing this constitutes externality blindness and an organization or network will be ill prepared to manage the disruption should the threat occur (Mitroff and Alpaslan, 2003; Altay and Ramirez, 2010). This study suggests an assessment of the threat level of each security breach type as a function of entropy score. This is also termed operational complexity of the breach and it investigates the impact of these perturbations as they affect the performance of the supply chain under different supply chain characteristics or conditions. These conditions are supply chain structure, ordering options and integration level. The study tries to see how supply chain structures are affected by the perturbations that the threats introduce and to demonstrate that each supply chain structure might be affected in different ways. The implication of this is that one size does not fit all, and supply chains would need to develop bespoke security risk management system suited to individual supply chain structures. Also the impact of just one threat, system failure, has been demonstrated thus far and an extension of this approach to other sources of perturbation is one of our future endeavours. However, the example used in this study to illustrate this concept has revealed the potential of entropy theory as a useful risk impact assessment tool, though entropy has been established in literature as a valid measurement of complexity in supplier-customer systems. It has also been validated as a very useful approach in other risk studies (Airoldi et al., 2011; Bezzi, 2007). Future research would be to validate this methodology by completing the experiments to see real evidence of this in practice. Further investigation can be done to see the impact of threats to information under other supply chain characteristics such as lean or agile, type of information being shared, etc. For instance, the effect of production information such as capacity or lead

Supply chain disruption

1009

JMTM 23,8

time information can be investigated. Future investigation could also be done on how entropy theory can be applied as an optimality study where the entropy level of each threat-mitigating countermeasure can be assessed to produce an optimal combination of countermeasure that effectively provide adequate security at a minimal cost.

1010

References Airoldi, E.M., Bai, X. and Malin, B.A. (2011), “An entropy approach to disclosure risk assessment: lessons from real applications and simulated domains”, Decision Support Systems, Vol. 51 No. 1, pp. 10-20. Altay, N. and Ramirez, A. (2010), “Impact of disasters on firms in different sectors: implications for supply chains”, Journal of Supply Chain Management, Vol. 46 No. 4, pp. 59-80. Amir, M.S. (2009), “It’s written in the cloud: the hype and promise of cloud computing”, Journal of Enterprise Information Management, Vol. 23 No. 2, pp. 131-4. Armbrust, M. (2010), “A view of cloud computing”, Communications of the ACM, Vol. 53 No. 4, pp. 50-8. Baker, W., Goudie, M., Hutton, A., Hylender, C.D., Niemantsverdriet, J., Novak, C., Ostertag, D., Porter, C., Rosen, M., Sartin, B. and Tippett, P. (2010), 2010 Data Breach Investigations Report, Verizon RISK Team in cooperation with the United States Secret Service, Washington, DC. Bennett, D.J. and Dekkers, R. (2009), “Industrial networks of the future: review of research and practice”, in Dekkers, R. (Ed.), Dispersed Manufacturing Networks, Springer, Heidelberg. Bezzi, M. (2007), “An entropy based method for measuring anonymity”, Proceedings of the Third International Conference on Security and Privacy in Communications Networks and the Workshops, IEEE, 17-21 September, pp. 28-32. Bourland, K.E., Powell, S.G. and Pyke, D.F. (1996), “Exploiting timely demand information to reduce inventories”, European Journal of Operational Research, Vol. 92 No. 2, pp. 239-53. Chan, H.K. and Chan, F.T.S. (2009), “Effect of information sharing in supply chains with flexibility”, International Journal of Production Research, Vol. 47 No. 1, pp. 213-32. Cheng, J.-H. (2011), “Inter-organizational relationships and information sharing in supply chains”, International Journal of Information Management, Vol. 31 No. 4, pp. 374-84. Colotla, I., Shi, Y. and Gregory, M.J. (2003), “Operation and performance of international manufacturing networks”, International Journal of Operations & Production Management, Vol. 23 No. 10, pp. 1181-206. Craighead, C.W., Blackhurst, J., Rungtusanatham, M.J. and Handfield, R.B. (2007), “The severity of supply chain disruptions: design characteristics and mitigation capabilities”, Decision Sciences, Vol. 38 No. 1, pp. 131-56. Deane, J., Ragsdale, C., Rakes, T. and Rees, L. (2009), “Managing supply chain risk and disruption from it security incidents”, Operations Management Research, Vol. 2 Nos 1/4, pp. 4-12. Deng, Y., Pang, J. and Wu, P. (2007), “Measuring anonymity with relative entropy”, in Dimitrakos, T., Martinelli, F., Ryan, P. and Schneider, S. (Eds), Formal Aspects in Security and Trust, Springer, Berlin. Devaraj, S., Krajewski, L. and Wei, J.C. (2007), “Impact of ebusiness technologies on operational performance: the role of production information integration in the supply chain”, Journal of Operations Management, Vol. 25 No. 6, pp. 1199-216.

Durowoju, O. and Chan, H.K. (2012), “The role of integration in information security breach incidents”, Proceedings of the Seventeenth International Working Seminar on Production Economics, Innsbruck, Austria, 20-24 February, pp. 129-40. Durowoju, O., Chan, H.K. and Wang, X. (2011), “The impact of security and scalability of cloud service on supply chain performance”, Journal of Electronic Commerce Research, Vol. 12 No. 4, pp. 243-56. Flynn, B.B., Huo, B. and Zhao, X. (2010), “The impact of supply chain integration on performance: a contingency and configuration approach”, Journal of Operations Management, Vol. 28, pp. 58-71. Frizelle, G. and Efstathiou, J. (2002), Seminar Notes on Measuring Complex Systems, London School of Economics, London. Frizelle, G. and Woodcock, E. (1995), “Measuring complexity as an aid to developing operational strategy”, International Journal of Operations & Production Management, Vol. 15, pp. 26-39. Gallear, D., Ghobadian, A. and O’Regan, N. (2008), “Digital/web-based technology in purchasing and supply management: a UK study”, Journal of Manufacturing Technology Management, Vol. 19 No. 3, pp. 346-60. Gunasekaran, A. and Ngai, E.W.T. (2004), “Information systems in supply chain integration and management”, European Journal of Operational Research, Vol. 159, pp. 269-95. Huang, C.-C. and Lin, S.-H. (2010), “Sharing knowledge in a supply chain using the semantic web”, Expert Systems with Applications, Vol. 37, pp. 3145-61. Kappelman, L.A., Richards, T.C. and Tsai, R.J. (1996), “A manager’s guide to electronic data interchange: doing business on the information superhighway”, Logistics Information Management, Vol. 9 No. 1, pp. 12-17. Kim, W., Jeong, O.-R., Kim, C. and So, J. (2011), “The dark side of the internet: attacks, costs and responses”, Information Systems, Vol. 36, pp. 675-705. Kulp, S.C., Lee, H.L. and Ofek, E. (2004), “Manufacturer benefits from information integration with retail customers”, Management Science, Vol. 50, pp. 431-44. Lancioni, R.A., Smith, M.F. and Schau, H.J. (2003), “Strategic internet application trends in supply chain management”, Industrial Marketing Management, Vol. 32, pp. 211-17. Lau, J.S.K., Huang, G.Q. and Mak, K.L. (2004), “Impact of information sharing on inventory replenishment in divergent supply chains”, International Journal of Production Research, Vol. 42, pp. 919-41. Lee, H.L., Padmanabhan, V. and Whang, S. (2004), “Information distortion in a supply chain: the bullwhip effect”, Management Science, Vol. 50, pp. 1875-86. Li, J., Sikora, R., Shaw, M.J. and Woo Tan, G. (2006), “A strategic analysis of inter organizational information sharing”, Decision Support Systems, Vol. 42, pp. 251-66. Li, S. and Lin, B. (2006), “Accessing information sharing and information quality in supply chain management”, Decision Support Systems, Vol. 42, pp. 1641-56. Loch, K.D., Carr, H.H. and Warkentin, M.E. (1992), “Threats to information systems: today’s reality, yesterday’s understanding”, MIS Quarterly: Management Information Systems, Vol. 16, pp. 173-86. Marston, S., Li, Z., Bandyopadhyay, S., Zhang, J. and Ghalsasi, A. (2011), “Cloud computing: the business perspective”, Decision Support Systems, Vol. 51, pp. 176-89. Martı´nez-Olvera, C. (2008), “Entropy as an assessment tool of supply chain information sharing”, European Journal of Operational Research, Vol. 185, pp. 405-17.

Supply chain disruption

1011

JMTM 23,8

1012

Mitroff, I.I. and Alpaslan, M.C. (2003), “Preparing for evil”, Harvard Business Review, Vol. 81, pp. 109-15. Mukhopadhyay, T. and Kekre, S. (2002), “Strategic and operational benefits of electronic integration in b2b procurement processes”, Management Science, Vol. 48, pp. 1301-13. Munoz, A. and Clements, M.D. (2008), “Disruptions in information flow: a revenue costing supply chain dilemma”, Journal of Theoretical and Applied Electronic Commerce Research, Vol. 3, pp. 30-40. Olson, D.L. and Wu, D.D. (2010), “A review of enterprise risk management in supply chain”, Kybernetes, Vol. 39, pp. 694-706. Pisello, T. (2004), “Is there a business case for it security?”, Security Management (accessed 22 April 2011). Plenert, G.J. (2002), International Operations Management, Copenhagen Business School Press, Copenhagen. Potter, C. and Beard, A. (2010), “Information security breach survey 2010”, available at: www. infosec.co.uk/files/isbs_2010_technical_report_single_pages.pdf Rainer, R.K., Snyder, C.A. and Carr, H.H. (1991), “Risk analysis for information technology”, Journal of Management Information Systems, Vol. 8, pp. 129-47. Rao, S. and Goldsby, T.J. (2009), “Supply chain risks: a review and typology”, International Journal of Logistics Management, Vol. 20, pp. 97-123. Raschke, R.L. (2010), “Process-based view of agility: the value contribution of it and the effects on process outcomes”, International Journal of Accounting Information Systems, Vol. 11 No. 4, pp. 297-313. Rees, L.P., Deane, J.K., Rakes, T.R. and Baker, W.H. (2011), “Decision support for cybersecurity risk planning”, Decision Support Systems, Vol. 51 No. 3, pp. 493-503. Richardson, R. (2009), “14th annual CSI computer crime and security survey”, in Peters, S. (Ed.), CSI Computer Crime and Security Survey, Computer Security Institute, New York, NY. Ronen, B. and Karp, R. (1994), “An information entropy approach to the small-lot concept”, IEEE Transactions on Engineering Management, Vol. 41, pp. 89-92. Samir, D. (2008), “Predicting and managing supply chain risks”, International Series in Operations Research and Management Science, Vol. 124, pp. 53-66. Samy, G.N., Ahmad, R. and Ismail, Z. (2010), “Security threats categories in healthcare information systems”, Health Informatics Journal, Vol. 16 No. 3, pp. 201-9. Schmitt, A.J. and Singh, M. (2009), “Quantifying supply chain disruption risk using monte carlo and discrete-event simulation”, in Rossetti, M.D., Hill, R.R., Johansson, B., Dunkin, A. and Ingalls, R.G. (Eds), Proceedings of the 2009 Winter Simulation Conference, IEEE, Austin, TX. Shannon, C.E. (1948), “A mathematical theory of communication”, The Bell System Technical Journal, Vol. 27, pp. 379-423, 623-56. Sivadasan, S., Efstathiou, J., Frizelle, G., Shirazi, R. and Calinescu, A. (2002), “An information-theoretic methodology for measuring the operational complexity of supplier-customer systems”, International Journal of Operations & Production Management, Vol. 22, pp. 80-102. Smith, R. (2009), “Computing in the cloud”, Research Technology Management, Vol. 52, pp. 1-7.

Stoneburner, G., Goguen, A. and Feringa, A. (2002), “Risk management guide for information technology systems”, Recommendations of the National Institute of Standards and Technology, National Institute of Standards and Technology, Gaithersburg, MD. Subashini, S. and Kavitha, V. (2011), “A survey on security issues in service delivery models of cloud computing”, Journal of Network and Computer Applications, Vol. 34, pp. 1-11. Tse, Y.K., Tan, K.H., Chung, S.H. and Lim, M.K. (2011), “Quality risk in global supply network”, Journal of Manufacturing Technology Management, Vol. 22 No. 8, pp. 1002-13. Wang, W.Y.C. and Chan, H.K. (2010), “Supply chain planning and configuration in the global arena – a syncretic perspective”, International Journal of Production Economics, Vol. 127, pp. 211-14. Warren, M. (2000), “Cyber attacks against supply chain management systems: a short note”, International Journal of Physical Distribution & Logistics Management, Vol. 30 Nos 7/8, pp. 710-16. Waters, D. (2006), Global Logistics, Kogan Page, London. Whitman, M.E. (2003), “Enemy at the gate: threats to information security”, Communications of the ACM, Vol. 46, pp. 91-5. Wiengarten, F., Humphreys, P., Cao, G., Fynes, B. and Mckittrick, A. (2010), “Collaborative supply chain practices and performance: exploring the key role of information quality”, Supply Chain Management: An International Journal, Vol. 15, pp. 463-73. Xu, X. (2012), “From cloud computing to cloud manufacturing”, Robotics & Computer-Integrated Manufacturing, Vol. 28, pp. 75-86. Yang, T., Wen, Y.-F. and Wang, F.-F. (2009), “Evaluation of robustness of supply chain information-sharing strategies using a hybrid taguchi and multiple criteria decision-making method”, International Journal of Production Economics, Vol. 134 No. 2, pp. 458-66. Yu, M.M., Ting, S.-C. and Chen, M.-C. (2010), “Evaluating the cross-efficiency of information sharing in supply chains”, Expert Systems with Applications, Vol. 37, pp. 2891-7. Yu, Z., Yan, H. and Cheng, T.C.E. (2001), “Benefits of information sharing with supply chain partnerships”, Industrial Management & Data Systems, Vol. 101 No. 3, pp. 114-21. Zhou, H. and Benton, W.C. Jr (2007), “Supply chain practice and information sharing”, Journal of Operations Management, Vol. 25 No. 6, pp. 1348-65. About the authors Olatunde Amoo Durowoju is currently a PhD student at the Norwich Business School, University of East Anglia. He has a Master’s degree in Food Production Management. His research interests are in operations strategy and supply chain management with keen interest in simulation modelling and combinatorial optimization. He is also interested in information management and IT security. His current research is on understanding the impact of information security breach on the performance of the supply chain and the mitigating effect of supply chain dynamics. He has work experience in the food industry, hospitality as well as the healthcare industry. He has published in Journal of Electronic Commerce Research and is currently working on publishing in more leading journals before the completion of his PhD. Dr Hing Kai Chan is a Senior Lecturer in the Norwich Business School, University of East Anglia, UK. His current research interests include industrial informatics, applications of soft computing on intelligent industrial systems and supply chains, simulation and modeling

Supply chain disruption

1013

JMTM 23,8

1014

of industrial systems. He has published in various IEEE Transactions, Communications of the ACM, Expert Systems with Applications, International Journal of Production Research, International Journal of Production Economics, among others. Xiaojun Wang received the MSc degree in e-Business Management from University of Warwick and the PhD degree in Management Studies from University of Liverpool. He is currently a Lecturer in Management at School of Economics, Finance and Management, University of Bristol. His research interests include operations and supply chain management, modelling and analysis of production management systems, sustainable supply chain, and multiple criteria decision making. His research has been published in Omega, International Journal of Production Economics, International Journal of Production Research and Production Planning and Control.

To purchase reprints of this article please e-mail: [email protected] Or visit our web site for further details: www.emeraldinsight.com/reprints

The current issue and full text archive of this journal is available at www.emeraldinsight.com/1741-038X.htm

A model to determine complexity in supply networks

Complexity in supply networks

Markus Gerschberger and Corinna Engelhardt-Nowitzki LOGISTIKUM, School of Economics, Upper Austria University of Applied Sciences, Steyr, Austria

1015

Sebastian Kummer Institute for Transport and Logistics Management, Vienna University of Economics and Business, Vienna, Austria, and

Franz Staberhofer LOGISTIKUM, School of Economics, Upper Austria University of Applied Sciences, Steyr, Austria Abstract Purpose – The purpose of this paper is to further advance an existing supplier evaluation model for the purpose of identifying those supplier relations which predominantly threaten or worsen a company’s performance. A defined basic set of parameters to determine complexity facilitates the identification of critical locations within a supply network (SN) under certain business conditions. Design/methodology/approach – This paper is based on a structured literature review in scientific periodicals in logistics/supply chain management between 2000 and 2009. Articles are analysed based on a structured framework and the identified complexity parameters are operationalised using quantitative and summable measures. The conceptual model is applied within a multiple case study in the Austrian agricultural industry. Findings – This paper illustrates how complexity in SNs can be operationalised in a company-specific configuration in order to achieve concrete managerial recommendations. Hence, the model allows evaluating SN-partners based on selected parameters to determine the contribution of a single partner to the overall complexity. Research limitations/implications – Due to the literature review executed and the case study approach chosen, the research may lack generalisability. Therefore, continued validation by means of implementing a greater amount of use cases in other companies and industries is advisable. Practical implications – Applying the model, a company is able to determine tier-1 to tier-n suppliers which are predominantly affecting its business from a complexity perspective. Originality/value – Unlike typical current complexity evaluation approaches, the proposed model respects rapid and continuous applicability, profound conceptualisation and practical feasibility. Keywords Austria, Agriculture, Suppliers, Channel relationships, Supply chain management, Complexity, Supplier evaluation, Supply networks Paper type Research paper

This project was supported by the Upper Austrian government within the research program AGTIL, especially in the project “ASC – Adaptive Supply Chain”. The authors would like to thank all partners for their contributions. They would also like to thank the anonymous referees and the editors for their valuable comments that considerably helped to improve this paper.

Journal of Manufacturing Technology Management Vol. 23 No. 8, 2012 pp. 1015-1037 q Emerald Group Publishing Limited 1741-038X DOI 10.1108/17410381211276853

JMTM 23,8

1016

1. Introduction Dynamic environmental conditions often lead to an increase in complexity of economic supply networks (SN). In particular, market uncertainty has increased the operational complexity of supplier-customer systems; however, no adequate measures for complexity quantification are available so far (Sivadasan et al., 2006). Although practitioners and scientists are aware of this, suitable approaches to successfully handle complexity in SNs are missing (Schuh et al., 2008a). Especially the identification of the most relevant network segments from a complexity perspective is a critical success factor for a company’s competitiveness, since it provides information to management about what kind of improvement measures might be most urgent and might deliver the most favourable impact in a certain economic situation. In the context of this paper, we subsume all inter-connected companies that exist upstream to a company under the term “supply network” according to Choi and Krause (2006). Due to length and focus constraints and in order to achieve a sound managerial relevance, we consider the fact that SN-related decisions are mainly taken within singular companies and here in separate departments for buy-side and customer-side operations. Hence, we assume a hierarchical network topology, being aware of the fact that other network types exist (e.g. heterarchical value networks, cp. Johannessen and Hauan, 1993; Hedlund, 1986). Also customer-side flows are purposely excluded, paying respect to the current practice that in most companies sourcing and supply chain management (SCM) have more distinct organisational bonds compared to the link between sales/marketing and SCM. Especially in rapidly changing markets, a corresponding method for decision support has to be easy to apply and must allow for fast execution to enable continuous and iterative application. However, a carefully defined approach is required to analyse complex SNs (Jain and Benyoucef, 2008). Thus, the first purpose of this paper is to operationalise parameters which determine SN-complexity. Second, possible methods how to implement these parameters within a respective decision model are proposed, also including advice on managerial implementation (obtained from an industrial case study). In doing so, the main intention of the present paper is to focus on model development and parameter operationalisation, restricting empirical evidence to singular observations that are discussed extensively in a doctoral thesis (Gerschberger, 2012). Thus, a decision model is developed to identify those partners in the SN that are of main importance for the focal company under varying conditions from a complexity-oriented perspective. The paper is organised as follows: first, a brief overview of the relevant complexity literature is given, second, the basic concept of the applied SN decision model is explained, finally, existing attempts to operationalise complexity parameters are discussed. On occasion, existing practical evidence that may impact the operationalisation of complexity parameters are discussed. 2. Conceptual background – complexity in supply networks Section 2 presents the fundamental concepts for the present paper regarding complexity in SNs. Since several authors (Bozarth et al., 2009; Wycisk et al., 2008; Meyer, 2007; Choi and Krause, 2006) have analysed different approaches to explain the term “complexity” in detail, this is not reproduced here. Summarising, the main findings from these sources are that the theoretical foundation regarding complexity is abstract, heterogeneous and often vague. We support the opinion that the abstract construct “complexity” should be

characterised and further operationalised indirectly using substantive complexity parameters (Kirchhof and Specht, 2003), herewith accepting the fact that increasing, and especially company-specific concretisation may be attended by a loss of generalisability. Based on the identification of a generic set of parameters in Section 4, a working definition of complexity is developed. We use the term “complexity parameter” instead of “complexity driver”, although the second is more commonly used, because not all components which determine complexity in SNs are actually driving complexity. Some parameters are determined by the structure of the network than driving complexity, e.g. the number of elements (the number of suppliers in the SN) or interrelations (the number of connections between companies in the SN). The review of existing attempts to determine complexity was conducted in the face of two central challenges: first, to find a way to determine complexity from a practical point of view using a manageable set of parameters and second, to develop a model that can be applied with an extended (not only a dyadic – evaluating direct suppliers) focus. The majority of contributions to complexity-related research originate from different disciplines, e.g. sociology, physics, biology or chemistry. Meanwhile many fields of study have touched upon complexity issues, initially starting to investigate simple systems and gradually maturing towards analysing more complex systems (Shahabi and Banaei-Kashani, 2007), often subsumed under the term “complex systems theory” (CST) (Bar-Yam, 2003). CST, as a quite young research discipline, can be seen as a meta-theory combining research efforts dealing with the investigation of complex systems (Shahabi and Banaei-Kashani, 2007). CST is closely connected to cybernetics and systems theory. Beyond this, findings originate from a series of monodisciplinary sciences, trying to extract commonalities in order to find general coherences and laws (Kappelhoff, 2000). Figure 1 exemplarily shows these interrelations. Figure 1 shows the connecting factors between system theory and cybernetics. While system theory investigates the structure, the behaviour and the relevant system determining factors, a main interest in the field of cybernetics regards the principles of information exchange and control within a system. Indeed both aspects are highly relevant for the discussion of complexity issues. Comparable to CST, the conceptual roots of “complexity” are neither singular nor monolithic, but rather come from a multifaceted range of disciplines (Manson, 2001; Tarride and Zuniga, 2010). Taking a closer look at the framework in Figure 1, the intention to condense one general and unique complexity conceptualisation has not been achieved, or, more likely, is only feasible at a high level of abstraction. Hence, it becomes obvious, why complexity is defined in such heterogeneous notions within singular disciplines. For instance, must a complexity interpretation necessarily have different implications from the perspective of: . An information theoretical conceptualisation (Sivadasan et al., 2006; Frizelle and Woodcock, 1995; Shannon and Weaver, 1963), often using entropy-related approaches. . System theory (Bliss, 2000; Reiß, 1993), though SCM often applies a high level of abstraction to reduce real world complexity (Caddy and Helou, 2007). . Evolutionary approaches describing interactions between organisms and the environment (Li et al., 2010; Meszena et al., 2001; Kauffman, 1993), even discussing a “evolutionary complexity of complex adaptive supply networks”, emphasising the network structure and collaboration mechanisms.

Complexity in supply networks

1017

JMTM 23,8

Structure Structural Change …

1018 Philosophy

Economics

Engineering

Jurisprudence

Sociology

Political science

Biology

Chemistry

Cybernetics

Evolutionary principles Patterns Simulation

Philosophical systems

Economic system

Electrical systems

Legal system

Social systems

Political systems

Biological systems

Chemical systems

Complex Systems Theory Physical systems

Figure 1. A reference framework of the complex systems theory

Physics

Monodisciplinary sciences

Systems theory

Control/handling

Behaviour

Information

Data

Communication

Variables

Source: Gomez (1981) .

The field of complex adaptive systems (Choi and Krause, 2006; Choi et al., 2001; Holland, 1992).

Although CST has discussed relevant phenomena related to complexity, such as emergence and non-linear circularities, SCM approaches have not successfully interpreted respective mechanisms. Most existing approaches focus on internal complexity with a static character, which is not at all applicable to a SN that consists of multiple information and material flows (Dekkers et al., 2005). None of the approaches have been adequately transferred to the domain of SNs. This is especially valid for concepts from systems theory (Dekkers et al., 2005). In principle, CST coherences like encapsulation provide valuable stimulation, but have not yet been operationalised for the use in a SN design context. According to Schuh et al. (2008a) the challenge for investigating complexity in SNs is to identify adequate methods for system representation, to investigate interdependencies between core elements, and the specification of adequate complexity parameters, that also respect complexity emerging from the external environment. An interesting issue is the question how an observer (here, a decision taker or manager) could determine the contribution or impact of a single element (here a single supplier) to the overall complexity in the SN. Since complexity causes and impacts on the network may change over time – and even excursively in turbulent markets – the capability to quickly identify such network locations (supplier or groups of suppliers) under given business conditions is essential for a company. Since the multifaceted topology of a SN notably influences network complexity, some attempts have been made to describe the generic structures of SNs. For instance, Lambert et al. (1998) have proposed a widely agreed model, tying a tree-shaped

converging SN structure to the inbound side and another diverging structure to the outbound side of a focal company. Although comprehensible, this model does not respect the potential existence of further linkages. A supplier might, for example, deliver the same goods to competing customers or a customer might not purchase a product exclusively from the focal company. The theoretically ideal network configuration would exhaustively respect the maximum of existing elements and linkages, however at the expense of practical feasibility. For instance, when considering (realistic) effects such as incomplete information, unpredictability or opportunism in the SN, there will be presumably low data and information accessibility when extending the focus beyond a dyadic buyer-seller relationship. Thus, the potential exhaustive model might be valid in principle but is not feasible for computation. In addition, the differing views of each SN company regarding, for example, the mutual importance of SN participants increase complexity. Altogether, complexity from a SN perspective remains multifaceted and has to be discussed based on a multiple set of determinants. Thus, the first step within the concrete investigation of a certain system is to set-up the SN topology that is applicable within this specific setting. In this context, a SN segment in principle consists of all suppliers upstream to a company that deliver parts to one specific product group. Since not all suppliers are of the same relevance when taking managerial SN-decisions and a complete enumeration is neither feasible nor reasonable, the model application approach includes a purposeful elimination of less relevant suppliers. Depending on the concrete practical decision details, for instance, a product group analysis based on financial aspects (turnover, profit margin) or quantity aspects (sold items per period) could be a first step. Further recommendations regarding the determination of parameters to distinguish relevant from less relevant suppliers can be found in Traxler et al. (2011). The integration of the best possibly operationalised complexity parameters into this network structure has to be the subsequent step. Within previous research, we have identified a relevant supplier network management model developed by Mu¨ssigmann (2007) that fits two requirements: network structure representation and integration of multiple parameter sets. In combination with the vector-based approach to operationalise complexity in production systems introduced by Windt et al. (2008), this is the basic concept of the model described in the next section. 3. Basic idea of the model to determine complexity The application of vectors for the purpose of determining complexity has been proposed by Windt et al. (2008), though in the context of production systems, not SNs. A bundle of complexity parameters is proposed to represent the complexity of the production system. The basic idea is to compare different systems or different states of one system based on the difference of two vectors, applying an ordinal scale. Another useful model is proposed by Mu¨ssigmann (2007), who applies graph theory to represent a supplier network or a selected segment of it (excluding less relevant suppliers). Though other network types exist, Mu¨ssigmann assumes a converging material flow as, for instance, applicable in the automobile industry. Each supplier relation(ship) (consisting of one supplier and the relation to his customer) in the examined network segment shown in Figure 2 is represented throughQcomplexity Q parameters ðk1; j;l ; . . . ; k5; j;l Þ which are bundled in an evaluation vector ðk1;l ; . . . ; k6;l Þ (step 1 in Figure 2). The vector consists of five complexity parameters which have been

Complexity in supply networks

1019

JMTM 23,8

1020

Figure 2. Conceptual model design

Step 1:Creation of evaluation vector for each supplier relation

k1, 4, l k2, 4, l k3, 4, l 4 k 4, 4, l k5, 4, l

k1, 1, l k2, 1, l k3, 1, l 1k 4, 1, l k5, 1, l

k1, 5, l k2, 5, l 5 k 3, 5, l k4, 5, l k5, 5, l

k1, 2, l k2, 2, l 2 k 3, 2, l k4, 2, l k5, 2, l

k1, 6, l 6 k2, 6, l k3, 6, l k4, 6, l k5, 6, l

k1, 3, l 3k 2, 3, l

k3, 3, l k4, 3, l k5, 3, l

Step 2: Creation of evaluation vector of the entire supply network (segment)

–

0

… l… k0, l… ki, j, l… k1, j, l… k2, j, l… k3, j, l… k4, j, l… k5, j, l… w… w1, l… w2, l… w3, l… w4, l… w5, l… E0, l…

k0, l =

k1, 0, l k2, 0, l k3, 0, l k4, 0, l k5, 0, l

Step 3: Interpretation of weighting vector

– = w

w1, l w2, l w3, l w4, l w5, l

Step 4: Identification of relevant supply network segments

–

–Tk E0, l = w 0, l

supplier relation (consisting of one supplier and the relation between two companies) supply network (SN) evaluation vector for the artificial supplier, representing a product group complexity parameter i for supplier relation j in SN l complexity parameter for the number of elements of supplier relation j in SN l complexity parameter for the number of interrelations of supplier relation j in SN l complexity parameter for the uncertainty caused by supplier relation j in SN l complexity parameter for the influence of supplier relation j on the ease of handling variety in SN l complexity parameter for geographical components of supplier relation j in SN l weighting vector weight for complexity parameter number of elements in SN l weight for complexity parameter number of interrelations in SN l weight for complexity parameter uncertainty in SN l weight for complexity parameter variety in SN l weight for complexity parameter geographical components in SN l position index of SN l

identified in the literature review described later on. The network segment in Figure 2 consistsQof six supplier relations that deliver components for one specific product group ðk# ð0; l ÞÞ. It is advantageous to use parameters that might be aggregated mathematically if a company intends to process several network scenarios using, for example, spreadsheet-simulation. Alternatively, an ordinal comparison is possible, but might deliver less comprehensive findings. If, reasonable parameter quantification is impossible (e.g. due to data acquisition issues), it is not advisable to quantify for the pure purpose of mathematical computation. The vectors for each supplier relation Qare consolidated in one overall evaluation vector for the network segment examined ðk# ð0; l ÞÞ (step 2 in Figure 2). With respect to the type of quantification used different aggregation methods are needed. The first two parameters ðk1;j;l ; k2;j;l Þ count system elements and therefore a cumulative consolidation suggests itself. For the remaining three parameters ðk3;j;l ; k4;j;l ; k5;j;l Þ besides an additive aggregation a multiplicative consolidation is possible or a min/max-function can determine the parameter value ðk3;0;l ; k4;0;l ; k5;0;l Þ in the overall evaluation vector. For example, the degree of uncertainty that enters the SN because of a single supplier relation can be measured based on the supplier’s delivery performance. In that case the minimum value ðk3;1;l ; k3;2;l ; . . . ; k3;6;l Þ of the six supplier relations or the minimum product of a path (e.g. supplier relation four and supplier relation two) can determine the parameter uncertainty ðk3;0;l Þ.

Due to the fact that the described model is a multi-criteria model, the importance of parameters can differ from one company to another. This is respected by means of a Q weighting vector ðwÞ that allows for situational parameter prioritisation (step 3 in Figure 2). Finally, a position index ðE 0;l Þ for each network segment can be created prioritising the most important segments (step 4 in Figure 2) through multiplication of the Q evaluation with the transposed weighting vector ðwT Þ. From a company perspective, the SN segment with the highest position index can be regarded to be of high priority for management activities (Gerschberger et al., 2010). We assume that for most cases the model implementation will proceed iteratively: starting with the focal company, first relevant direct supplier relations should be inspected. The next step is to identify a substantial set of parameters to sufficiently describe complexity in SNs. Depending on these parameters and the expected data structure in companies that would use the developed model, parameter operationalisation is an important issue. 4. Identification of the parameters to determine complexity in supply networks The challenge of this research step was to find a way to traceably identify basic parameters to determine complexity in SNs. The idea behind this review was to define parameters which are mentioned most in literature and therefore are supposed to be most relevant. This literature review was done selectively with the distinct intention of including conceptual and empirical sources. To keep biases low, a triangulation setting (Jonsen and Jehn, 2009) was used: the data sources were diversified into three streams within a thoroughly structured review approach: recognised academic journals, monographs and two substantial current research projects which were analysed independently from each other. The findings of stream one (systematically selected journal publications) are the basis and the remaining streams (monographs and research projects) were used to exploratively cross-check and further validate the results of the first stream. Finally, all three streams showed a high congruence with each other. We strongly believe that concrete practical applications will require further and context-specific modifications of this parameter set. Case studies can be a promising procedure to add empirical evidence to the current result. Hence, we decided to dedicate the first step of our research (to be published in the present article) to mainly conceptual considerations, thus developing a profound basis for subsequent empirical investigation that respects model-specific assumptions rather than only generic complexity considerations. For the review of academic journals (stream one) the 12 top-ranked in logistics and SCM (Menachof et al., 2009) – displayed in Table I – have been reviewed. The research pairs “complexity and supply chain”, “complexity and network” and “complexity and system” were used as formalised key strings resulting in 157 papers. After an elaborate content analysis, 19 papers (marked with * in the references) were identified that explicitly treat the topic of complexity in SNs and focus on complexity parameters. Table II lists five resulting parameters with a sufficient number of mentions. Three additional parameters, “heterogeneity of elements”, “heterogeneity of relations” and “dynamics” were added to the table despite a low number of mentions following respective evidence from streams one and two. The complexity parameters identified

Complexity in supply networks

1021

JMTM 23,8

Ranking Journal 1 2

1022

Table I. Journal selection according to their research usefulness index

3 4 5 6 7 8 9 10 11 12

Journal of Business Logistics – JBL International Journal of Physical Distribution and Logistics Management – IJPDLM International Journal of Logistics Management – IJLM Journal of Operations Management – JOM Supply Chain Management Review – SCMR Transportation Journal – TJ Harvard Business Review – HBR Management Science – MS Transportation Research Part E, Logistics and Transportation Review – TR-E Supply Chain Management: An International Journal – SCM-IJ International Journal of Operations and Production Management – IJOPM International Journal of Logistics: Research and Applications – IJL-RA

Research usefulness index 41.921 33.206 29.071 19.052 17.875 17.525 15.108 14.249 13.899 12.309 11.864 11.546

within the journal review confirm those identified by Bozarth et al. (2009), Masson et al. (2007) and Milgate (2001). For all of them, empirical evidence is available (cp, especially Jonsson et al., 2007; Masson et al., 2007; Perona and Miragliotta, 2004). Table III displays the aggregated review of review streams one and two. To ensure the quality of the monographs (stream two) the focus was mainly on doctoral theses and references cited there. Finally (stream three), the parameter set was matched with those identified in the research project “Coll-Plexity” (Schuh et al., 2008a, b) and with findings from the research association the “Manufacturing Complexity Network” (Frizelle and Huw, 2002; Sivadasan et al., 2002a, b). This again confirmed the integrity of the identified parameters. A further interesting aspect was that authors within the “Manufacturing Complexity Network” focused on the determination of complexity using entropy measures (Sivadasan et al., 2006), and mentioned several interconnected aspects to describe complexity (number of elements, degree of variety, degree of system predictability and uncertainty, degree of interaction between elements and degree of order within the structure). According to these findings complexity is determined by five generic core parameters: (1) The number of elements and interrelations that define the system. (2) The degree of uncertainty that enters the system. (3) The influence of the supplier on the customer-requested product variety. (4) Geographical components that act on the system. To facilitate concrete managerial concerns a more context-specific operationalisation of these parameters is needed to replace the generic identifiers ðk1;j;l ; . . . ; k5;j;l Þ in Figure 2 with substantive real-world measures.

Bozarth et al. (2009) Choi and Hong (2002) Choi and Krause (2006) Christensen et al. (2007) Closs et al. (2008) Guide et al. (2003) Hofer and Knemeyer (2009) Jonsson et al. (2007) Kinra and Kotzab (2008) Masson et al. (2007) Meepetchdee and Shah (2007) Milgate (2001) Oke (2003) Prater et al. (2001) Sanchez and Perez (2005) Sivadasan et al. (2002b) Stonebraker and Liao (2006) Vickery et al. (2004)) Wycisk et al. (2008) Total

Author

13

1

6

0

Structural complexity-determining parameters Number of Heterogeneity Number of Heterogeneity of elements of elements interrelations interrelations

9

3

8

9

Complexity-influencing parameters Customer-driven Geographical Uncertainty Dynamic product variety components

Complexity in supply networks

1023

Table II. Identified parameters of complexity

Table III. Review of the results against parameters identified within monographs

Blecker et al. (2005) Bohne (1998) Csa´ji and Monostori (2008) Engelhardt-Nowitzki and Zsifkovits (2006) Gino (2002) Gomez and Probst (1999) Hasenpusch et al. (2004) Heina (1999) Ivanov (2006) Kaluza et al. (2006) Kirchhof andSpecht(2003) Klaus (2005) Lawrenz (2001) Liening (1999) Luhmann (1976) Mainzer (2008) Malik (2000) Moder (2008) Piller (2006) Scherf (2003) Schuh (2005) Sivadasan et al. (2002a) Ulrich and Probst (1988) Vester (2003) Mentions in monographs Mentions in journal publications (Table II) Total number of mentions 9 1 10

32

Heterogeneity of elements

19 13

Number of elements

20

14 6

Number of interrelations

6

6 0

17

8 9

17

14 3

Heterogeneity of interrelations Uncertainty Dynamic

21

13 8

Customer-driven product variety

1024

Author

14

5 9

Geographical components

JMTM 23,8

5. Description and operationalisation of complexity parameters In this section, the so far identified parameters are explained in more detail including possible operationalisation means. Since we strongly believe that a managerial discipline has to align its research efforts to concrete and practical usable findings, the selection of potential operationalisation possibilities did not exhaustively analyse all thinkable ways of operationalisation, but was rather done with respect to the specific situation in SNs (in particular data availability, necessary time for data acquisition and easy model implementation ability). Thus, other operationalisation attempts are supposable, though we recommend not disclaiming practical applicability and hence managerial relevance in favour of too exhaustive theoretical considerations. In case of doubt, a sensitivity analysis regarding the impact of a singular parameter on the overall model results may serve to prove its significance for the concrete setting. 5.1 Number of elements and interrelations The differentiation between complexity of elements and interrelations originates from systems theory (Reiß, 1993; Figure 1). An element is represented by a node in the system, in a supply context by one company. The number of interrelations stands for the number of edges (connections between companies) in the SN, hence indicating the interconnectivity within the network and, also the geographical dispersion (Wycisk et al., 2008). A larger number of suppliers usually results in a larger number of potentially competing business objectives (Choi and Hong, 2002), higher transaction costs and information asymmetries (Williamson, 1981), causing obsolete stock. The higher the number of elements and interrelations, the higher is the number of relevant logistical requirements and the likeliness of errors (Mu¨ssigmann, 2007). One possibility to operationalise the “number of elements” is to set k1;j;l to 1 in each single evaluation vector ðk1;1;l ; k1;2;l ; k1;3;l – only direct suppliers are evaluated in a first step) as one supplier (one element) is the basis of this vector. If, for example, supplier relation three is in this case of primary importance the next step is to convince (in collaboration with this supplier) the tier-2 suppliers (five and six) to participate in the supplier evaluation. In this case, the aggregated value for the parameter “number of elements” for supplier relation three increases from one to three because of the two additional suppliers on the second level. The same procedure can be applied for the second parameter “number of interrelations” representing the connections between the companies within this SN segment. We are aware of the fact that this operationalisation is a simplification with regard to practical applicability. If a company has the respective resources and data or if in a specific context the pure number of suppliers and relationships do not yield valid results, a more specified manner of operationalisation could be applied. However, for efficiency reasons, sensitivity analyses, for example, should be done in order to make sure that the enhanced effort of abstract completeness is attended by enhanced result validity. For instance, suppliers with a high degree of technical uniqueness and risk could deliver a disproportionally higher complexity compared to commodity vendors. Accordingly, a rating-based or weighting could be applied to represent this coherence beyond pure supplier count. Further operationalisation attempts have been extensively discussed in the literature, e.g. the degree of inter-connectedness (Meepetchdee and Shah, 2007), the minimum spanning tree or the average/minimum path lengths (Jungnickel, 1994).

Complexity in supply networks

1025

JMTM 23,8

1026

5.2 Uncertainty Uncertainty exists when the probability of an event occurring is not zero or one. In SNs numerous decisions have to be taken without knowing all relevant variables. Thus, uncertainty has a significant overall effect on delivery time and reliability (Milgate, 2001). In principle, three potential uncertainty sources exist: suppliers, production and customers (Davis, 1993). Considering possible dimensions in which these sources can cause uncertainty, numerous classifications (Ivanov, 2006; van der Vorst and Beulens, 2002) and, as a consequence, attempts to determine and measure uncertainty were developed. With respect to the supply side focus of the present model, operationalisation attempts range from simple descriptive procedures to quantitative estimation of uncertainty, as well as to formalised decision procedures. The analysis may be qualitative or quantitative, depending on the level of resolution required and the amount of information available and including a variety of available statistical means. In this context, uncertainty is often summarised in terms of bias, variance, and measures based on statistical distribution (Smith, 2002). Beyond this general classification, uncertainty measures with a SN focus were identified. For instance, Milgate (2001) uses three variables to determine uncertainty. Upstream uncertainty can be manifested through late supplier deliveries and poor quality of incoming materials (Davis, 1993). Downstream uncertainty takes the form of unforeseen demand variability that causes problems in the SN. Vickery et al. (1999), based on Miller and Dro¨ge (1986) use five items (volatility in marketing practices, product obsolescence rate, unpredictability of competitors, unpredictability of demand and tastes and change in production or service modes) to determine uncertainty on a Likert-scale. Additionally, John and Weitz (1988) defined five uncertainty measures (market share volatility, ease or difficulty of monitoring trends, industry volume volatility, sales forecast accuracy, and predictability) which have been used or modified by several other authors (Christensen et al., 2007; Celly and Frazier, 1996). Forecasting methods try to estimate the future demand by extrapolation of historical data (Weiss, 2007). The aim is to deduce based on historical data; past, current and future forecasts and the gap to real demand (forecast error) to continuously develop more accurate demand information for upcoming periods. In context of this research, the parameter “uncertainty” has to express the extent to which a single supplier relation has the potential to influence the degree of uncertainty in the SN. Delivery reliability is a well-known measure of this, which has been surveyed in many companies. With respect to Arnold et al. (2008) and close to Milgate (2001), an adapted form of delivery reliability for each supplier relation ðLT j;l Þ is proposed to operationalise the parameter “uncertainty” and is calculated in the following way: k3; j;l ¼ 1 2 LT j;l LT j;l is the number of orders supplied on time in relation to all deliveries made by supplier relation j in network segment l ( ¼ for one product group). Keeping the classification (0 ¼ no contribution to the degree of complexity, 1 ¼ max. contribution) the delivery reliability is deducted from 1, because the higher the delivery reliability the lower the potential impact of this single supplier relation has on the degree of uncertainty. When realising model step 2 – creation of an evaluation vector of the entire network segment – one aggregation option for the parameter “uncertainty” ðk3;0;l Þ is, like for quality measures, that the minimum or maximum figure determines the value for the

entire network segment. In this example the maximum value is important and can be identified as follows:

Complexity in supply networks

k3;0;l ¼ max {k3; j;l } j

Independent of the measurement approach used, uncertainty impacts may be amplified through the contributions of multiple network elements and interrelations. This is represented via the aforementioned parameter “number of elements and interrelations” and hence not redundantly considered in the category “uncertainty”. When proceeding towards practical implementation, a company has to carefully validate the context-specifically appropriate operationalisation attempt. This will depend, for example, on the number of direct suppliers, data availability and personnel resources. 5.3 Customer-driven product variety From a company’s perspective, complexity arises because of interdependent variety influences within the SN. For instance, the decision to globally sell a rich variety of products may require mastering various sales channels, technologies, components and suppliers. Subsequently, non-repetitive manufacturing in small lots could cause process automation constraints. Thus, a substantial amount of work has been devoted to discussing competitive advantages, but also increased complexity caused through high variety (Perona and Miragliotta, 2004). Furthermore, it is important for companies to evaluate how the complexity of the manufacturing system or the product arrangement system changes when new product variants are introduced to or removed from the production program (Blecker et al., 2006). Also procedures to determine the appropriate degree of variety (Er, 2004; Randall and Ulrich, 2001; Lindsley et al., 1991; Bental and Spiegel, 1984) and attempts to ease variety handling in internal production processes (Blecker and Abdelkafi, 2006; Strassner, 2006) are manifold in the literature. However, most concepts show a limited applicability because in general, the appropriate degree of variety as well as the desired product design is determined by customer-driven requirements and strategic decisions of the focal company. As a result, the suppliers influence is very limited. Nevertheless, a single supplier relation can have a major impact on the ease of handling this defined degree of variety. Therefore, the aim of the present research is to quantify this potential impact of single supplier relations (SN segments) and selected operationalisation attempts are emphasised. In particular, the pivotal supplier index (PSI) treats the question of whether a specific supplier is required to be able to fulfil a certain customer demand (Lang, 2007). This supports the common trend of considering specificity issues in current logistics and SCM literature (Krause, 2008; Matthes, 2007). The easiest way to operationalise this is to answer this question with “yes” or “no”. “Yes” implies that this supplier is of major importance and therefore can have a large impact on the ease of handling variety. Keeping the range of the first two parameters, “yes” can be valued with one and “no” with zero. Further, the “product line breadth” approach is used in a series of studies to investigate variety impacts on business objectives (Lindsley et al., 1991; Yeh and Chu, 1991; Kekre and Srinivasan, 1990; Bental and Spiegel, 1984). The approach is to count the number of products in line – where each product combines distinct characteristics (Da Silveira, 1998). Therefore, the product line breadth is another way of labelling the number of variants within a product group or branch. Further developed for the purpose of this research, the number of variants within the product group (the network segment) where supplier

1027

JMTM 23,8

1028

relation one to six provides components in relation to the PLB can be an applicable measure to evaluate the influence of a single supplier relation on the ease of handling variety. For the prototypical model implementation, a combination of the PSI and the adapted product line breadth is used to get an insight into a single supplier’s importance from at least two perspectives (how specific are parts delivered by the supplier relation and in how many product variants in the network segment are these parts needed). Let’s assume that supplier relation j delivers specific parts (PSI ¼ 1) and PLB is the relation between the number of variants in which parts of j are implemented ðdj;l Þ, and the overall number of variants in the product group l (dl), the influence of j on handling the variety ðk4; j; l Þ can be calculated as follows: k4;j;l ¼ PSI £ PLB ¼ 1 £

d j;l dl

This general procedure is applicable for all single supplier evaluation vectors. A further development is to find a more meaningful and differentiated way to operationalise PSI. An option can be to let company experts estimate the time and money needed to shift supply from the current supplier relation to the “next best”. The greater the time and money needed, the higher the specificity of parts and the higher the potential impact of this supplier relation on the ease of handling variety in the network segment. Unsolved up to now for the parameter “variety” is how the single values in the evaluation vectors of supplier relations ðk4;1;l ; . . . ; k4;6;l Þ can be aggregated to one value for the SN segment ðk4;0;l Þ. Once again similar to the parameter “uncertainty” the maximum value can count. In this context, it is advisable to use a combined value for supplier relations of one path. The reason is that a high importance of supplier4 for supplier1 does not necessarily result in a high importance of supplier1 for the focal company. Therefore, when not only the direct supplier relations are the subject of analysis, the highest path product can be a way of how to define the parameter for this network segment. Realising
 this, the number of supplier relations j without a pre-supplier K el ; K el ¼ {4; 5; 6; 2} in Figure 2) has to be defined as well as the number of supplier relations j on the path from j to 0 ðK *j; l Þ. Having done this k4;0;l can be calculated as follows: 8 9