On representing actions in multi-agent domains

  • Authors:
  • Chitta Baral;Gregory Gelfond

  • Affiliations:
  • Department of Computer Science and Engineering, Arizona State University;Department of Computer Science and Engineering, Arizona State University

  • Venue:
  • Logic programming, knowledge representation, and nonmonotonic reasoning
  • Year:
  • 2011

Quantified Score

Hi-index 0.01

Visualization

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.