Manufacturing Enterprise Integration is a complex and distributed task; hence it is an ideal candidate for application of distributed artificial intelligence in the form of multi-agent systems. We model various functional entities in a manufacturing enterprise as intelligent software agents, which negotiate with each other to reach scheduling and planning as well as operating decisions. The proposed system uses a variant of the popular contract net protocol for agent interactions, which includes a new bid selection and resource-based activity short-listing approach with concepts of agent ratings and monetary transactions for better modeling of the complex and dynamic environment. Since integration involves application of negotiation strategies at different organizational levels, we develop a multi-agent system based solution framework which can be applied to diverse problem domains. This framework helps in addressing the problems more efficiently by using a network of intelligent entities rather than a complex monolithic solver. Enterprise entities are classified into generic agent types with a standard ontology being followed for interaction at each level of the system. These agents may have conflicting individual objective functions and they might compete with other agents for the limited resources available in the system or collaborate with other agents to undertake a particular distributed task. We model information exchanges within and between enterprises as software agent interactions which are in sync with the goals of the entity they represent.
展开▼
