TítolDescribing and configuring multi-agent systems at the knowledge level
Publication TypeConference Paper
Year of Publication2003
AuthorsGómez M, Plaza E
EditorAguiló I, Valverde L, M. Escrig T
Conference NameFrontiers in Artificial Intelligence and Applications
Volume100
EditorIOS Press
Paginació221-232
Resum

Cooperative Problem Solving is usually focused on the coordination and cooperation mechanisms of Multi-Agent Systems, leaving out the user and the problem requirements. This paper introduces the knowledge description level of ORCAS (Open, Reusable and Configurable multi-Agent Systems), a framework to develop MAS applications configurable on demand, according to problem requirements and user preferences. ORCAS introduces the idea of configuring a MAS application at two layers, the knowledge and the operational layers. During the knowledge layer a configuration of MAS components is found, including agent capabilities and domain knowledge; and after that the knowledge configuration is operationalized by a customized team of problem solving agents. This framework is based on applying a knowledge modeling approach to describe Multi-Agent Systems, specifically, it proposes to describe agent capabilities as Problem-Solving Methods.