Multi-agent based system architecture for agile manufacturing system and its learning algorithm with CMAC (Cerebellar Model Articulation Controller)

Abstract

As the world market is becoming increasingly dynamic and intense, manufacturing enterprises require responsive manufacturing systems which having ability to response rapidly to dynamic changes, to satisfied customer needs. Multi-agent system is used to model manufacturing system in this paper. A multi-agent system based system architecture for agile manufacturing systems which include agent architectures, communication interface between agents in the manufacturing system, agent data module, knowledge module for multi-agent system and sensory input and output controller are proposed. A agent learning model and its mapping algorithm for CMAC (Cerebellar Model Articulation Controller) are presented to make the agent system have self learning ability. Simulation result shows that the model and the solution are effective to the manufacturing system.