A cooperative problem solving framework for computer-aided process planning

Computer aided process planning is the link between design and manufacturing in a computer integrated manufacturing (CIM) environment. The process planning problem comes out as a rigid hierarchical structure of tasks where the lowest levels are well certified tasks, such as determination of machine and cutting parameters, while the top levels control, coordinate, and manage the entire system. Such a top down manner of coordination makes it a tightly coupled distributed decision making situation. The complexity of CIM system with hierarchical structures grows rapidly with the size of the system, making them very expensive. Also, assimilating a large array of knowledge sources to plan the activities is a major hurdle. Cooperative distributed problem solving (CDPS) is concealed with the cooperative solution of problems by a decentralized group of cooperating intelligent agents. CDPS techniques have good adaptive capability, make the system modular, are faster due to parallel operation of subsystems, possess greater reliability through redundancy, can over-ride limited resources, and can collect knowledge. The paper describes a conceptual framework of applying CDPS to CAPP for general purpose process modeling. The problem in process planning has been mapped onto a CDPS architecture based on multistage negotiation. The needed agents are defined and their roles in the architecture and the knowledge that each agent would need are described at the conceptual level. Potential benefits of the framework in solving process planning problems, in an intelligent and distributed manner are discussed.