Programming cognitive agents

  • Authors:

  • John-Jules Ch. Meyer

  • Affiliations:

  • Institute of Information and Computing Sciences, Utrecht University, The Netherlands

  • Venue:
  • MATES'05 Proceedings of the Third German conference on Multiagent System Technologies
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Although there is a lot of theory around about cognitive agents since the seminal work by researchers such as Bratman, Cohen & Levesque and Rao & Georgeff practice of programming ’truly’ cognitive agents is still in its infancy. Of course, several architectures have been proposed and even occasionally been implemented, and there is a prospect of many potential applications of agent-based systems, but is there a truly systematic way of programming agents with cognitive / mental attitudes such beliefs, desires, intentions, goals, plans, commitments, emotions...? We believe that for this dedicated agent-oriented languages are needed. A number of these have been developed in the last decade or so. But programming in them is still hard. Is there a methodology for agent-oriented programming? Can one structure agent programs better making use of cognitive notions? And how to verify that an agent program is correct? And how is this combined with programming multi-agent systems and agent societies where coordination of these autonomous agents and more generally social notions such as norms seem most important? In this paper a number of the issues related to programming cognitive (multi) agents will be discussed on the basis of work done in Utrecht around the agent language 3APL.