An Agent-based Model Management Framework for Providing a Solution Method with Inter-model Relationship

초록

Increasing diversity of user requirements on model management forces a model agent to develop adaptive abilities building a realistic model and providing an adequate solution method. The relationship between models is important knowledge for the role of an agent, and it helps the agent effectively build a new model, adaptively adjust an existing model under a problem, and efficiently connect the new model into an adequate solution method. Since the generating process of the inter-model relationship is more difficult than the building a new model, however, the process mostly depends on the knowledge of operation research experts. Without the adequate scheme of the inter-model relationship, the burden of the management of an agent increases rapidly, and the quality of his services may worsen. This study goes beyond an existing view of the model management in using the knowledge of the relationship. We present a frame-based framework to manage the knowledge consisted of common and exclusive objects between models represented by a UNIK-OPT, a knowledge-assisted optimization modeler. The framework includes adaptive abilities which identify a model through a forward chaining method from the problem that a decision maker describes with a frame-based format and search a solution method appropriate to the model by using the knowledge of the inter-model relationship. We illustrate the framework with electronic commerce applications involving a delivery planning in the supply chain management (SCM) system.