MULTIAGENT PROBLEM-SOLVING AND OBJECT-ORIENTED DECISION-SUPPORT SYSTEMS FOR NATURAL-RESOURCE MANAGEMENT

Object-oriented programming and design is becoming the leading way of constructing Al software and decision support systems. However, there has been no clear methodology for handling the problem-solving process in object-oriented terms. Based on agent-oriented programming and con cepts of distributed AI, we present such a framework. We define the ne cessary concepts of agents, tasks, plans, and activities. Then we defi ne a process where an agent selects tasks, instantiates plans for prob lem solving, and executes activities. A programming framework based on IntelliCorp's Kappa(TM) environment is presented, where the user can analyze a domain and create an intelligent decision support applicatio n with his/her own agents and other objects. The present architecture is useful for constructing decision support systems, because there is a built-in problem analysis and task scheduling module. The system bec omes a model of an expert organization, and hence has direct relevance to object-oriented information engineering, a software engineering me thodology that builds systems based on enterprise models.