Optimizing distributed production planning

Abstract

Production planning is one of the big challenges of artificial intelligence because of its complexity and the dynamic behavior of the different components of a manufacturing environment. We present a new organization scheme for a distributed planning method to coordinate the planning and scheduling of several autonomous agents as intelligent robots and of manufacturing cells on the task level. An implementation of a multi-agent system is described that has been implemented on a net of workstations in order to test the practical relevance of the approach. Each agent has its own knowledge-base and decision component.