Distributed CIM and DAI: for a knowledge and multiagent approach

This paper concerns the building of models (“modeling”) of D-CIM (distributed CIM) systems and more precisely their generation and validation, anticipating new trends for manufacturing enterprises. The originality of our modeling rests on a knowledge-oriented approach through the use of multiagent systems (MAS) according to distributed artificial intelligence (DAI). The objective is to justify a modeling paradigm conciliating both major CIMOSA (CIM open system architecture) concepts and the MAS paradigm to support the modeling of D-CIM systems. In a first time we develop this new D-CIM paradigm. In a second time, we expose the needs for a methodological knowledge oriented approach. Next, we present a set of requirements that need to be fulfilled. After that, we develop the knowledge-oriented approach, presenting the steps that we have retained. Finally, we make an analysis confronting the proposed approach with the requirements to be fulfilled