Using opponent models for efficient negotiation

ABSTRACT

Information about the opponent is essential to improve automated negotiation strategies for bilateral multi-issue negotiation. In this paper we propose a negotiation strategy that combines a Bayesian technique to learn the preferences of an opponent during bidding and a Tit-for-Tat-like strategy to avoid exploitation by the opponent. The learned opponent model is used to achieve two important goals in negotiation. It may be used to increase the efficiency of negotiation by searching for Pareto optimal bids and to avoid exploitation by making moves that mirror the move of the other party. The performance of the proposed negotiation strategy is analyzed in a tournament setup.

REFERENCES

• 1: Axelrod, R. 1984. The Evolution of Cooperation. Basic Books, Inc., Publishers, New York, USA.
• 2: Koen Hindriks , Catholijn M. Jonker , Dmytro Tykhonov, Analysis of Negotiation Dynamics, Proceedings of the 11th international workshop on Cooperative Information Agents XI, September 19-21, 2007, Delft, The Netherlands [doi>10.1007/978-3-540-75119-9_3]
• 3: Koen Hindriks , Dmytro Tykhonov, Opponent modelling in automated multi-issue negotiation using Bayesian learning, Proceedings of the 7th international joint conference on Autonomous agents and multiagent systems, May 12-16, 2008, Estoril, Portugal
• 4: Raiffa, H. 1982. The Art and Science of Negotiation, Harvard University Press.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.11 Distributed Artificial Intelligence
            Subjects: Intelligent agents

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.11 Distributed Artificial Intelligence
            Subjects: Multiagent systems

General Terms:
Algorithms, Economics, Experimentation, Performance, Theory

Keywords:
Bayesian learning, Tit-for-Tat, automated multi-issue negotiation, negotiation strategy, opponent modelling