Modelling and Simulation of the supply chain: a problem of preventive transshipment Stefano Saetta, Lorenzo Tiacci Dipartimento di Ingegneria Industriale Università degli Studi di Perugia - Italy Keywords: Preventive Transhipment, Negotiation Mechanisms, auction algorithms, SCM Simulation, Lean Simulation Abstract The aim of this paper is to illustrate the Modelling and Simulation of a working architecture for a specific and highly meaningful contest: the supply-chain management in the fashion field. In particular, the use of transshipments between different Points of Sales in order to realign inventory balances is studied. A model for the case considered is proposed with the description of the model and their characteristics. A Negotiation Mechanism between agents based on a single-issue auction has been introduced. Results of a simulation campaign are discussed and further developments and extensions of the model proposed are considered. Simulation is implemented in JAVA ® , by using its multithreading tools. Simulation results can be used in Lean Simulation for a fast modelling of similar problems. In particular critical parameters, for the performance of the systems, are found. INTRODUCTION Modelling and simulation of the supply chain is nowadays necessary in order to study the performance of integrated and distributed logistic systems. In Supply Chain modelling in recent years, new software architectures, based on the multi-agent paradigm, have been introduced in production and distribution problem solving (eg. Barbucenau and Fox, 1995, Krothapalli and Deshmukh, 1999, Hu et al., 2001). Problems involving a high number of complex interacting entities are the most suitable to be structured according to such paradigm. Basically this approach is indeed different from solutions where there is a fully centralised control of the single entities of the system. In the agent-paradigm, in fact, single entities can autonomously concur to the optimal performance of the system, by mean of a higher or lower degree of cooperation with the other entities. These studies have brought to a more profound knowledge of both theoretical and practical aspects of single topics. Nevertheless, proposals of a multi-agent architecture applied to a certain class of problems (which involve the choice of a specific language, structure, negotiation protocol etc.) for testing possible advantages of the multi- agent paradigm are still rare. The aim of this paper is to Model and Simulate a working multi-agent architecture for a specific and highly meaningful contest: the distribution system in the fashion industry. In particular the management of transhipments between different Points of Sales (PS) is treated. In the case considered, transhipments are preferred to costly and urgent shipments from the Central Warehouse (CW). A Negotiation Mechanism between agents, based on a single-issue auction, has been introduced to maximise the transhipments. In the following, the fashion industry case studied is considered and possible advantages of a multi-agent architecture for the coordination of supply policies are discussed. Afterwards, characteristics of agents are described and negotiation mechanism is treated. The implementation of the simulation model, based on the Java process oriented simulation software, is synthetically shown. Finally, results of a simulation campaign are discussed and conclusions are drawn. Results can be used in meta modelling in order to obtain effective procedures for similar problems. In particular the inter relations between critical parameters could be useful in developing Lean Simulation tools (Bruzzone and Saetta, 2002). Those tools can allow a rapid answer in the application of Modelling and Simulation in real cases, where usually time responsiveness must be maximised. In problems concerning Small and Medium Enterprises or in Large programs, Lean Simulation could spread out the using of Modelling and Simulation techniques. . SUPPLY CHAIN CONSIDERED The supply chain structure of the case considered is shown in figure 1. A certain number of producers supply products to a CW, which delivers them to all the PSs spread throughout the region. ISBN: 1-56555-268-7 835 SCSC '03