The Representation of Events in Multiagent Domains

Michael P. Georgeff

The purpose of this paper is to construct a model of actions and events suited to reasoning about domains involving multiple agents or dynamic environments. A model is constructed that provides for simultaneous action, and the kind of facts necessary for reasoning about such actions are described. A model-based law of persistence is introduced to describe how actions affect the world. No frame axioms or syntactic frame rules are involved in the specification of any given action, thus allowing a proper model-theoretic semantics for the representation. Some serious deficiencies with existing approaches to reasoning about multiple agents are also identified. Finally, it is shown how the law of persistence, together with a notion of causality, makes it possible to retain a simple model of action while avoiding most of the difficulties associated with the frame problem.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.