Evaluation of Communication in Multiagent Systems for Supply Chain Planning and Control

Multi Agent Systems (MAS) are particularly suitable for the modelling and analysis of the structure and dynamics of industrial complex systems. They are often used in supply chains as a support for decision-making and seem especially helpful to test and analyze different control design alternatives. However, the lack of evaluation methods of MAS is a major disincentive to their use in such real world applications. In this paper, a MAS evaluation method is first presented. It focuses on communication as it is one of the most important functional characteristics of distributed systems in general and MAS in particular. Then three multiagent models representing different supply chain control architectures (distributed, centralized and mixed) are presented, evaluated and compared according to the advanced method. Experimentation is carried out with reference to a test-case supply chain network in the domain of furniture manufacturing.