Automated Bargaining Agents ( Preliminary Results )

Rapid commercialization of the Internet and development of the National Information Infrastructure are likely to change the nature of retailing and commerce in profound ways. These changes pose both challenges and opportunites for AI, such as how to design intelligent assistants that help humans cope with the increasing complexities of electronic commerce. In this paper, we present the rst steps towards creating automated bargaining agents | intelligent assistants that can reason about the relative supply and demand for goods and services and negotiate to reach a good deal. We rst review the economic and game theoretic background for such an endeavor. We then present a simple economic model of electronic commerce and describe several bargaining strategies. Preliminary experimental results suggest that sophisticated, adaptive strategies perform better than simple discounting approaches over a range of economic conditions. Many thanks to Jim Dana and Kathy Spier for discussions about game theory and economics. We would also like to thank Adam Carlson, Denise Draper, Oren Etzioni, Terrance Goan, Keith Golden, Nicholas Kushmerick, Neal Lesh, W. Derrick Weathersby, Mike Williamson, and Erh-Chun Yeh for suggesting bargaining strategies and providing helpful comments. This research was funded in part by O ce of Naval Research Grant N00014-94-1-0060 and by National Science Foundation Grant IRI-9303461. This paper has not already been accepted by and is not currently under review for a journal or another conference. Nor will it be submitted to one during IJCAI's review period.