Agent-based Petri net models for AGV management in manufacturing systems

The problem of managing and controlling automatic guided vehicles (AGV) in manufacturing shop floor systems is addressed. In such systems, resources are connected through a network of paths that are limited-capacity shared resources as, in general, two or more AGVs cannot use the same section at the same time. The proposed approach makes use of the integration between a multi-agent system and Petri nets. The behaviour and the interactions among AGVs, path sections, and all other resources in the shop floor system are represented by means of Petri nets, whereas agents handle decisional activities. Two architectures are proposed: in the first, decisional agents are associated with path sections (path agents) whereas in the second scheme decisional agents are AGVs themselves (AGV agents)