Data mining

On the approach to rationality recognition in agents systems

M. A. Mikheenkova, V. K. Finn

All-Russian Institute for Scientific and Technical Information of Russian Academy of Sciences, Moscow

Abstract: The behaviour of an artificial agent and group of agents is formalized via the structured representation of an agent and his actions. The approach allows to analyze and to forecast the agent’s behaviour and to form agents’ classes on the basis of determinants of actions set. The apparatus can be used to analyze the rationality of both agents and multi-agent systems.

Keywords: agent, multi-agent systems, JSM-method, automated hypotheses generation, rationality.

Bibliographic databases:

•