LNCS Homepage
CD ContentsAuthor IndexSearch

Cooperative Problem Solving Using an Agent-Based Market

David Cornforth1 and Michael Kirley2

1School of Environmental and Information Sciences, Charles Sturt University, Albury, NSW, Australia
dcornforth@csu.edu.au

2Department of Computer Science and Software Engineering, University of Melbourne, Melbourne, Vic, Australia
mkirley@cs.mu.oz.au

Abstract. A key problem in multi-agent systems research is identifying appropriate techniques to facilitate effective cooperation between agents. In this paper, we investigate the efficacy of a novel market-based aggregation technique in addressing this problem. An incremental transaction-based protocol is introduced where agents establish links by buying and selling from each other. Market transactions equate to agents coordinating their plans and sharing their resources to meet the global objective. An important contribution of this study is to clarify whether, in some circumstances, a market-based model leads to the effective formation of agent teams (or coalitions) and thus, solutions to the problem-solving task.

LNCS 3102, p. 60 ff.

Full article in PDF


lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2004