Modeling belief and action in a multi-agent system

Saso, a rule-based shell for multiagent problems involving actions, beliefs and communication, is described. Saso is a Prolog-based forward-chaining system. Conventions for clear representations of actions, beliefs and goals are discussed. Saso has been applied in research aimed at developing robust, naturalistic coordination of autonomous agents. In particular, the control acts of simple conversations have been simulated. The system has also been applied to cooperative problem solving in a blocks world.<<ETX>>

[1]  David G. Novick,et al.  Control of mixed-initiative discourse through meta-locutionary acts: a computational model , 1988 .

[2]  S Fickas,et al.  Control knowledge in expert systems: relaxing restrictive assumptions , 1987 .

[3]  Dennis Merritt,et al.  Building Expert Systems in Prolog , 1989, Springer Compass International.

[4]  R. Fritzson,et al.  Adding forward chaining and truth maintenance to Prolog , 1989, [1989] Proceedings. The Fifth Conference on Artificial Intelligence Applications.

[5]  L. Suchman Plans and situated actions , 1987 .

[6]  Stephen Fickas Design issues in a Rule-Based System , 1985, ACM SIGPLAN Notices.

[7]  Stephen Fickas Design issues in a Rule-Based System , 1985, SLIPE '85.

[8]  Stuart C. Shapiro,et al.  A Model for Belief Revision , 1988, Artif. Intell..