Methods and Models for the Simulation of a Virtual Network

This work describes the development of a protocol of analysis of a virtual network of manufacturing enterprises, with particular interest in the framework to support the distributed simulation. This wants to propose innovative solutions for the evaluation of the performances of manufacturing systems operating in uncertainty conditions and to simulate the interactions between the enterprises. Even if many simulation tools are commercially available (many especially suited for the manufacturing sector), a strong limitation in using such tools to evaluate performances is given by the complexity of their use in an integrated and coordinated environment. The development of cooperation and integration methodologies among different simulation tools (i.e. distributed simulation) can widen the application context of simulation. In particular the developed methodology allows to define a network of connections between the enterprises of a "supply chain" evidencing: (i) exchanges of information about of the state of the various enterprises, in terms of workloads, estimated costs, offered price, obtained orders or in discussion, and also typologies of negotiated products; (ii) the transfers of physical flows of products between the enterprises of the supply chain. The Java technology has been used to reach these aims. Java is an Open Source object-oriented programming language, that allows future extensions by other scientists too, besides this language supports client-server structure, which has used to build a web-based simulation. The architecture is multi-agent, the structure is realized by elementary models, which represent the behaviour of every entity of supply chain: the customer, the central manager and the enterprises. Demands of the costumer (part typology, number of pieces, due-date and maximum cost, the relative weight of these factors) are send to central agent: this agent defines the workings sequence, referencing to customer wishes, to part typology, to capacity of the enterprises, to scheduled jobs and failures that occur on their resources. To represent manufacturing reality, it has been schematised the power for every enterprise of the network to receive external orders to supply chain, too; particularly the developed logic allows to every enterprise to analyze if it’s better or not producing the workings taking availability to supply chain. Further developments regard the check of created tools and algorithms on industrial reality.