Model based design of knowledge bases in multi agent systems for enabling automatic reconfiguration capabilities of material flow modules

Abstract

Dealing with the requirement to produce many different products at the same automated production system, flexible material flow modules, which are connected to a material flow system (MFS) are required. Therefore a definition of the architecture and interfaces of such modules is provided by a model based engineering approach and an appropriate meta model. Hence, a reduced engineering effort for setting up flexible MFS could be reached. Since the flexible composition of the modules should be performed without (re-)programming, an intelligent agent based control concept will be introduced. Based on the agents' knowledge, the procedure of online system configuration can be performed. Therefore, a main point are geometrical parameters for the online definition of interfaces for communication to neighboring modules. Additionally, the knowledge bases of the agents in the multi agent system can be derived automatically from the model. Furthermore, the communication concept, which has to consider the real-time requirements for synchronization of module interfaces is presented. An exemplary system shows how the reconfiguration triggered by connecting or disconnecting a module can be performed. Subsequently, the performance of the automatic reconfiguration procedure is evaluated regarding reconfiguration time.