On representing actions in multi-agent domains

Chitta Baral, Gregory Gelfond

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Scopus citations

Abstract

Reasoning about actions forms the foundation of prediction, planning, explanation, and diagnosis in a dynamic environment. Most of the research in this field has focused on domains with a single agent, albeit in a dynamic environment, with considerably less attention being paid to multi-agent domains. In a domain with multiple agents, interesting issues arise when one considers the knowledge of various agents about the world, as well about as each other's knowledge. This aspect of multi-agent domains has been studied in the field of dynamic epistemic logic. In this paper we review work by Baltag and Moss on multi-agent reasoning in the context of dynamic epistemic logic, extrapolate their work to the case where agents in a domain are classified into three types and suggest directions for combining ideas from dynamic epistemic logic and reasoning about actions and change in order to obtain a unified theory of multi-agent actions.

Original languageEnglish (US)
Title of host publicationLogic Programming, Knowledge Representation, and Nonmonotonic Reasoning - Essays Dedicated to Michael Gelfond on the Occasion of His 65th Birthday
Pages213-232
Number of pages20
DOIs
StatePublished - 2011
EventSymposium on Constructive Mathematics in Computer Science - Lexington, KY, United States
Duration: Oct 25 2010Oct 26 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume6565 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherSymposium on Constructive Mathematics in Computer Science
Country/TerritoryUnited States
CityLexington, KY
Period10/25/1010/26/10

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'On representing actions in multi-agent domains'. Together they form a unique fingerprint.

Cite this