Dynamic Modelling of Strategic Supply Chains

Dynamic Modelling of Strategic Supply Chains Antonia Albani, Alexander Keiblinger, Klaus Turowski, Christian Winnewisser Chair of Business Information Systems (Wirtschaftsinformatik II) University of Augsburg Universtitätsstraße 16, 86159 Augsburg, Germany {antonia.albani, alexander.keiblinger, klaus.turowski, christian.winnewisser}@wiwi.uni-augsburg.de Abstract. The high level of complexity connected with the modelling of supplier networks often diminishes the practical impact of supply chain management approaches. Furthermore, many of those approaches primarily focus on operative tasks. This paper proposes a concept for the dynamic modelling of supply chains. In addition, the business domain of strategic supply chain management is introduced, extending the traditional frame of reference in strategic sourcing from a supplier-centric to a supply-chain-scope. Concluding, a business component model for the domain of strategic supply chain development is laid out to illustrate the feasibility of the proposed concept. 1 Introduction The supply chain is one of the most analysed objects of reference in research centred around inter organisational systems (IOS), virtual enterprises and value networks, especially with respect to the perceived business value. However, failed initiatives, primarily in the field of supply chain management, have spurred concern about the practicability of present approaches and theories and have shown the need for further refinement and adaptation. According to [11], one of the most important reasons for failure can be found in the high degree of complexity that is connected with the identification of supply chain entities and the modelling of the supply chain structure, as well as the high coordination effort. In addition, many research efforts in supply chain management have been based on the more operative interpretation of [8, S. 22-38, 9, S. 16-22], primarily focusing on the optimisation of forecast and planning accuracy, and the optimisation of material flows over the whole supply chain, whereas more strategic tasks such as long term supplier development have not been widely discussed in a network perspective yet. Therefore, this paper introduces the concept of strategic supply chain development, which extends the traditional frame of reference in strategic sourcing from a suppliercentric to a supply-chain-scope. To address the complexity issues in supply chain modelling, the proposed concept includes an outline for the dynamic modelling of strategic supply-chains. The respective business domain will be described in section 2. In section 3, the concept of dynamic supply chain modelling will be detailed further and challenges of the technological development will be discussed. Based on these findings, a business component model for the domain of strategic supply chain development will be outlined in section 4. 2 From strategic sourcing to strategic supply chain development The relevance of the purchasing function in the enterprise has increased steadily over the past two decades. Till the 70ies, purchasing was widely considered an operative task with no apparent influence on long term planning and strategy development [13]. This narrow view was broadened by research that documented the positive influence that a targeted supplier collaboration and qualification could bring to a company's strategic options [2]. In the 80ies, trends such as the globalisation, the focus on core competencies in the value chain with connected in-sourcing and out-sourcing decisions, as well as new concepts in manufacturing spurred the recognition of the eminent importance of the development and management of supplier relationships for gaining competitive advantages. As a result, purchasing gradually gained strategic relevance on top of its operative tasks [10]. Based on these developments, purchasing has become a core function in the 90ies. Current empiric research shows a significant correlation between the establishment of a strategic purchasing function and the financial success of an enterprise, independent from the industry surveyed [3, S. 513]. One of the most important factors in this connection is the buyer-supplier-relationship. At many of the surveyed companies, a close cooperation between buyer and supplier in areas such as long-term planning, product development and coordination of production processes led to process improvements and resulting cost reductions that were shared between buyer and suppliers [3, S. 516]. In practice, supplier development is widely limited to suppliers in tier-1. With respect to the above demonstrated, superior importance of supplier development we postulate the extension of the traditional frame of reference in strategic sourcing from a supplier-centric to a supply-chain-scope, i.e., the further development of the strategic supplier development to a strategic supply chain development. This re-focuses the object of reference in the field of strategic sourcing by analysing supplier networks instead of single suppliers. Embedded in this paradigm shift is the concept of the value network that has been comprehensively described, e.g., [12, 19]. The main reason for the lack of practical implementation of strategic supply chain development can be found in the high degree of complexity that is connected with the identification of supply chain entities and the modelling of the supply chain structure, as well as the high coordination effort, as described by [11]. 2.1 Relation to supply chain management Many research efforts to supply chain management refer to [8, S. 22-38, 9, S. 1622], who give a rather operative interpretation. According to their work, main tasks include the optimisation of forecast and planning accuracy, and the optimisation of material flows over the whole supply chain. The main aim is the minimisation of in- ventory and lead time, which is also partially mirrored by many current initiatives based on the Supply Chain Council's SCOR-model. Therefore, primarily the tasks of the operative purchasing function are put in a network context. The concept of strategic supply chain development as introduced in this paper is based on the tasks of the strategic purchasing function and extends the frame of reference in strategic sourcing from a supplier-centric to a supply-chain-scope. It should not go unnoticed that [4] refer to current research that give an extended interpretation of supply chain management and partly consider supplier relationships as well. According to this broad definition, strategic sourcing could be regarded as part of supply chain management as well. However, the tasks of strategic purchasing are generally not discussed in a network context. Independent from this discussion, the problems described by [11] and the resulting hurdles for a practical implementation of the concept of strategic supply chain management remain imminent. 2.2 Description of the functional tasks of strategic supply chain development The functional tasks of strategic supply chain development are defined next. Those tasks will be derived from the main tasks of strategic sourcing. The most evident changes are expected for functions with cross-company focus. The functional tasks of strategic supply chain development have been illustrated in a function decomposition diagram (Figure 1). Processes and tasks that will be automated have been shaded. Following, only selected tasks will be described, focusing on changes to current tasks of strategic purchasing. Task „Model strategic supply chain“: The process "supplier selection" from strategic purchasing undergoes the most evident changes in the shift to a supply chain centric perspective. The expansion of the traditional frame of reference in strategic sourcing requires more information than merely data on existing and potential suppliers in tier-1. Instead, the supply chains connected with those suppliers have to be identified and evaluated, e.g., by comparing alternative supply chains in the production network. As a consequence, the task "supplier selection" is only part of the process that leads to the modelling of strategic supply chains. According to the assumptions described above, the rating of supply chains requires the evaluation of networks instead of single suppliers. There has been preparatory work on evaluation methods for business networks (e.g., [17, 15]) on which we have based initial methods for the described application. However, there is need for further research, especially in the area of aggregation of incomplete information. For the time being, the problem has been tackled by identifying strategic, "mission critical" suppliers through a multi-dimensional mix of evaluation criteria (e.g., in the area of volume, quality, service levels, processes) and by aggregating the evaluation results for these suppliers as representatives for the whole supply chain. In the first round of implementation, the selection of suppliers will not be automated by the application. Strategic supply chain development deals with long-term supplier relationships. An automation of respective fundamental contract negations seems neither feasible nor desirable in the short term. In fact, the results from automated supply chain identification and evaluation should be used as decision support for supplier selection. Fig. 1. Functional decomposition diagram for the supply chain development Task „Qualify strategic supply chain“: In addition to the selection of suitable supply chains and composition of alternative supply chains, the performance improvement of strategic important supply chains is one of the major goals of strategic supply chain development. Main prerequisite is the constant evaluation of the actual performance of selected supply chains by defined benchmarks. The application supports respective evaluation methods and enables the user to identify imminent problems in the supply chain and to initiate appropriate measures for qualification of supply chain partners. This is important because of the long-term character of strategic supply chain relationships. As a result of the long-term perspective, qualification measures – e.g., along the dimensions product, processes and management abilities – require deployment of resources on the buyer side as well. Because of this effort, problems in the supply chain should be identified proactively and qualification measures should be tracked. Task "Plan strategic demand": Strategic planning, i.e., analysing, scheduling and grouping of long-term demand, primarily affects intra-company processes that will not change significantly by switching to a supply chain perspective in strategic purchasing and therefore will not be included in the first stage of the application. 3 Concept and technological issues in self-modelling supply chains The concept of the supply chain as a dynamic, self-modelling network constitutes the basis for the identification of strategic supply chains, as described in section 2, for the domain of strategic supply chain development. Based on a sample supply chain, as shown in Fig. 2, the concept and the technological issues for developing selfmodelling supply chains will be explained and illustrated in this section. Fig. 2 on the left shows the complete network of existing and alternative supply chains for a specific demand, with the currently active supplier (supplier 1-2) being highlighted. In order to be able to visualise and evaluate the network at a particular date in a structured way, a specific strategic demand is communicated from the producer to existing and selected alternative suppliers in tier-1. Subsequently, the suppliers in tier-1 report the demand to their own respective suppliers. E.g., for supplier 1-2, these are the suppliers 2-2, 2-3 and 2-4 in tier-2. In the following, these suppliers report the demand to their related suppliers in tier-3, which split-lot transfer the requested information. The requestors aggregate the received information with the own information and send it back to the supplier 1-2 in tier-1. Having aggregated the information of all suppliers, the supplier 1-2 adds its own information before split-lot transferring it to the producer. Fig. 2. Left: Supplier network Right: Alternative supply chain With the suppliers’ data locally available, the producer can visualise the selected sub-chain as a network, in which each participant of the sub-chain constitutes a net- work hub. Based on that data, the producer is able to evaluate the performance of the selected sub-chain by self defined benchmarks. In order to optimise sub-chains, alternative supply chains can be visualised and modelled by applying the same concept as described above. Fig. 2 on the right highlights an alternative virtual supply chain. In the event of this alternative supply chain being the best performing sub-chain in the network, the existing supply chain can be modified, substituting supplier 1-2 in tier-1 with the new supplier 1-1, while keeping supplier 2-2 in tier-2 and supplier 3-1 in tier3. For the development of self-modelling supply chains different technological issues, e.g., distributed systems, asynchronous communication, consistency of data and synchronisation need to be discussed, since the concept described above poses multiple technological challenges which have implications on the system design. These challenges will be described first. Later rationales for the chosen technologies in the domain model of strategic supply chain development will be given. Distributed Systems: As defined by [16], the network of independent systems that constitute strategic supply chains appears to the user of a single node inside that network as a single system and therefore represents a distributed system. It is an open peer group of loosely coupled systems. There are no server or directory infrastructures and apart from the fact, that every node implements such functions for its own view of the network, no hierarchy exists. Therefore there is no single point of failure in the system. Asynchronous Communication: To enable such loosely coupled networks, messaging can be used as transport constituting communication channels. Messaging systems encapsulate sending and receiving of messages and allow multiple transport mechanisms, e.g. SOAP/XML, JMS or even SMTP. By using the concept of conversations [7], based on messaging, pre-programmed patterns (conversational policies inside conversational contexts) can be implemented for flexible transactions of information between nodes. Conversation policies are used in the agent community for coupling internal states of agents but are used in this context to couple business process. Conversation policies are machine readable patterns of message exchange in a conversation between systems and are composed of a message schema, sequence and timing information. To support the direct conversation with suppliers in the different tiers of the network or with filtered groups of them, unicast and multicast methods of addressing peers or groups of peers are needed. For the discovery of new potential suppliers, the mechanism of broadcast messages is needed as well. The client-server paradigm, using synchronous invoke/return schemes would create unnecessary dependencies between systems and could lead to deadlock situations. A higher level of robustness can be achieved by applying peer-to-peer approaches and using conversation. Consistency of data: With loosely coupled information systems the problem of data consistency arises. Mechanisms need to be implemented in order to guarantee that at some specific point in time the data used to generate virtual supply chains is accurate. Synchronisation: The update of information and the synchronisation between systems are typical problems in peer-to-peer scenarios and have to be considered for the specific application domain. In the following it is shown how those technical issues have been addressed in the business application of strategic supply chain development. There are different configurations of systems in the network of strategic supply chain, ranging from a very simple configuration where business data is entered manually to an integrated solution, e.g. coupled with an enterprise integration system, the software composing the functionality of nodes has to be scalable to the demands of the company participating in the network. As a possible solution to build distributed networks, business component technologies [18] are suggested. The underlying idea of business components combines components from different vendors to an application which is individual to each customer. This principle of modular black-box design has been used in this system design and will be exemplified in section 4. For the issues of asynchronous communication, consistency of data and synchronisation, communication channels that support the message exchange between business components, and therefore between communication components of each participating node of the strategic supply chain network, are proposed. They act as a coordination instance and decrease the coupling between components. Potential technologies are software busses, event channels or tuple spaces [6, 16]. Tuple spaces support the data driven communication according to the pull-principle where an interested party has to request data, filtered by specific restrictions configured by the requesting party. Tuple spaces act as message buffer, allowing asynchronous communication by storing messages until they are explicitly deleted or fetched. Therefore they decouple sender and receiver of messages in a temporal manner. Tuple spaces are special associative memories where data is not chosen by keys but by patterns of the message content. If a network node has identified a suitable supply chain it may request information about changes in the respective supply chain. To support this, the Publisher Subscriber Pattern [5] has been combined with tuple spaces on a conception level. A subscriber gets an event if data changes within tuples of interest. This allows network participants to shape virtual supply chains out of the crowds of service and product suppliers participating in the network. Such extensions to tuple spaces complement them and create an integral solution for a data driven and reactive network for the exchange of strategic supply chain data. Having discussed tuple-spaces and business components as possible solutions for the development of self-modelling, distributed supply chains, in the next section an optimised business component model for the domain of strategic supply chain development is presented. The deployment of different business components on different network hubs is explained and the communication between those distributed business components is illustrated. 4 Business component model for the business domain of strategic supply chain development To illustrate the feasibility of the concepts described above, a business component model for the domain of strategic supply chain development has been derived, based on the Business Component Modelling (BCM) process [1]. The business component model is shown in Fig. 3, in accordance with the notation of Unified Modelling Language [14], with all component services listed and with the dependencies between the individual business components represented by arrows. Fig. 3. Business component model The model consists of seven components which provide services performing certain business tasks of the specified domain. The component supply chain development offers services for the specification of the demand and for the definition of application users. The supply chain administration is responsible for aggregating and managing the data received from the suppliers, whereas the visualisation and selection component visualises the virtual supply chains and allows selection of sub-chains in order to evaluate them. Services for the definition of evaluation criteria and methods and for rating supply chains are provided by the evaluation component. The performance data, resulting from evaluating supply chains, is administrated by the performance data administration component. Since the components are located on different network hubs, two additional components are needed for inter-component communication, namely the communication manager and the communication component. In order to illustrate the deployment of those components and the communication between them a deployment and a sequence diagram are introduced next. Fig. 4 shows the deployment of business component instances on different network hubs. For the domain of supply chain management different systems use the specified components, namely the producer system and all suppliers systems. The producer holds all components belonging to the strategic supply chain development system including the supply chain development and the evaluation components. Whereas the suppliers utilise only those components necessary for recording and sending the own data to the producer and for acquiring the data received from their own suppliers. Fig. 4. Instantiation and deployment diagram The dependencies and services calls are presented by arrows. For a more coherent display not all component services and dependencies are shown. The dependencies and the dataflow are given by means of an example and are illustrated in Fig. 5. Example: Using the services of the supply chain development component a producer is able to specify the demand and to communicate the demand to all suppliers in tier-n. Triggered by that request, the supply chain development component accesses the service request collaboration of the collaboration manager component which uses the send message service of the communication component in order to send the demand to the suppliers in tier-1. The service requests of the different components are visualised in Fig. 5 by arrows. The communication components of the suppliers in tier-1 receive the message sent by the producer and forward the request to their collaboration managers accessing the process collaboration request service. Each collaboration manager uses the communicate demand to existing and potential suppliers service of the supply chain development component to forward the demand to existing and potential suppliers. The collaboration manager and the communication components are responsible for communication between the systems. Supplier 1, having received the information data from supplier 3, stores the data using the service write complex monitoring objects in its own system. This information together with information about all the suppliers is sent back to the producer. At any time the producer receives the aggregated information from its suppliers in tier-1 and their relative suppliers. This information is given to the supply chain administration component in form of a complex monitoring object. Each user can then request to visualise the complex monitoring object. The complete supply chain is presented con- taining all information about the single supplier nodes. The producer is therefore able to evaluate and to develop the complete supply chain according to its requirements. Fig. 5. Sequence diagram 5 Conclusion In today's economy, formerly closely-linked value chains are more and more transforming into flexible networks, thus significantly changing the buyer/producer - supplier relationship. As laid out in this article, those points hold especially true for the field of strategic purchasing, where an extension from a chain to a network perspective is desirable - however, complexity issues connected with the modelling of suppliernetworks have to be overcome. Having identified the need for managing the complexity of supplier network modelling and for extending the frame of reference of strategic purchasing, the concepts of dynamic supply chain modelling and strategic supply chain development have been introduced. Conceptual and technological issues have been discussed and a business component model for strategic supply chain development, addressing the identified issues and proposing a concept and methodology to overcome them, has been identified. As pointed out in the article, further research has to be conducted in the area of network evaluation as well as for the assessment of the practicability of the proposed approach during the further prototypical implementation of the derived business component model for strategic supply chain development. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] A. Albani, A. Keiblinger, K. Turowski and C. Winnewisser, Domain based identification and modelling of business component applications, In Proceedings of the 7th East-European Conference on Advances in Databases and Information Systems, Dresden, Germany, September 2003. D. Ammer, Materials Management, Homewood, 1968. A. S. Carr and J. N. Pearson, Strategically managed buyer - supplier relationships and performance outcomes, Journal of Operations Management, 17 (1999), pp. 497 519. A. S. Carr, Smelzer, L.R., The relationship of strategic purchasing to supply chain management, European Journal of Purchasing & Supply Management, 5 (1999), pp. 43 - 51. E. Gamma, R. Helm, R. Johnson and J. Vlissides, Design patterns: elements of reusable object-oriented software, Addison-Wesley, Reading, Mass., 1995. D. Gelernter, Multiple tuple spaces in Linda, Proc. of PARLE '89, Springer Verlag (1989), pp. 20-27. J. E. Hanson, P. Nandi and S. Kumaran, Conversation support for Business Process Integration, Proceedings 6th IEEE International Enterprise Distributed Object Computing Conference (EDOC-2002), IEEE Press2002, pp. 65-74. J. B. Houlihan, International Supply Chain Management, International Journal of Physical Distribution and Logistics Management, 15 (1985), pp. 22-38. T. Jones and D. Riley, Using Inventory for Competitive Advantage through Supply Chain Management, International Journal of Physical Distribution and Logistics Management, 5 (1985), pp. 16-22. L. Kaufmann, Purchasing and Supply Management - A Conceptual Framework, Handbuch Industrielles Beschaffungsmanagement, Hahn, D, Kaufmann, L. (Hrsg.), Wiesbaden, 2002, pp. 3 - 33. D. M. Lambert and M. C. Cooper, Issues in Supply Chain Management, Industrial Marketing Management, 29 (2000), pp. 65-83. T. W. Malone and R. J. Lautbacher, The Dawn of the E-Lance Economy, Harvard Business Review (1998), pp. 145 - 152. R. McIvor, P. Humphreys and E. McAleer, The Evolution of the Purchasing Function, Journal of Strategic Change, Vol. 6 (1997), pp. 165 - 179. OMG, OMG Unified Modelling Language Spezification Version 1.4, 2001. M. Sawhney and J. Zabin, The Seven Steps to Nirvana: Strategic Insights into eBusiness Transformation, New York, 2001. A. S. Tanenbaum, Distributed Operating Systems, Prentice Hall, 1995. D. Tapscott, D. Ticoll and A. Lowy, Digital Capital: Harnessing the Power of Business Webs, Boston, 2000. K. Turowski, ed., Standardized Specification of Business Components, Gesellschaft für Informatik, Working Group 5.10.3 - Component Oriented Business Application Systems, Augsburg, 2002. H.-J. Warnecke, Vom Fraktal zum Produktionsnetzwerk. Unternehmenskooperationen erfolgreich gestalten, Berlin, 1999.