Knowledge-based Multi-Agent Architecture for Dynamic Scheduling in Manufacturing Systems

The scheduling of production resources is one of the key features in the current competitive and dynamic manufacturing environment. The scheduling has to be flexible and able to cope with conflicts derived from the resources sharing among the production orders. We present a knowledge-based multi-agent approach combined with negotiation mechanisms for task allocation as a preliminary step towards optimized scheduling. We explained the scheduling strategy, where each agent performs local scheduling using dispatching rules and coordinates its actions with other agents optimizing at the same time global dynamic scheduling of the manufacturing system. An ontology is used to support "understanding" between agents during the communication, as well as to explicitly define the domain entities and relations between them.