TOWARDS PROBLEM SOLVING METHODS IN MULTI-AGENT SYSTEMS
Paul Bogg, Ghassan Beydoun, Graham Cedric Low · Research Online (University of Wollongong) · 2009
Problem Solving Methods (PSM) are abstract structures that describe specific reasoning processes employed to solve a set of similar problems and have proved very effective at enhancing reuse and extensibility in developing knowledge-based systems. We envisage that off-the-shelf PSMs can similarly assist in the development of agent-oriented solutions using Multi-Agent Systems (MAS). A challenge towards the effective use of PSMs in MAS is that current approaches to formulating PSMs do not adequately address the complexity of problems to which agent-oriented systems are suited. Towards addressing this, this paper focuses on providing an approach to guide developers in adequately formulating PSMs for complex problem-solving where interactions are involved, such as in domains where negotiation and cooperation are essential for solving a problem.