Abstract

A fundamental problem in business-to-business exchanges is the efficient design of mechanisms to promote cooperation and coordination of various self-interested agents. We study the behavior of artificial agents in a bidding and contracting framework [Eur. J. Oper. Res. (2002); D.J. Wu, P. Kleindorfer, J.E. Zhang, Integrating Contracting and Spot Procurement with Capacity Options, Working Paper, Department of Operations and Information Management, The Wharton School, University of Pennsylvania, 2001]. In this framework, there is a long-term contract market as well as a backstop spot market. Seller agents bid their contract offers in terms of price and capacity via an electronic bulletin board, while Buyer agents decide how much to contract with Sellers and how much to shop from the spot market. This two-tiered market has been modeled [Eur. J. Oper. Res. (2002); D.J. Wu, P. Kleindorfer, J.E. Zhang, Integrating Contracting and Spot Procurement with Capacity Options, Working Paper, Department of Operations and Information Management, The Wharton School, University of Pennsylvania, 2001] as a von-Stackelberg game with Seller agents as leaders, and the necessary and sufficient conditions for the existence of market equilibrium are given. What happens if the resulting equilibrium is noncooperative and Pareto dominated by some nonequilibrium bidding? What happens if there are multiple equilibria, some Pareto dominated by others, then which will be selected? What happens when there is no equilibrium? The goal of this paper is to study equilibrium and disequilibrium behavior of artificial agents in such systems, and explore the efficient design of mechanisms to promote cooperation and coordination of self-interested artificial agents.

Author Keywords: Artificial agents; Bidding; Capacity options; (Dis)equilibrium behavior

Article Outline

1. Introduction

2. Literature review

2.1. Models and analysis

2.2. Artificial agents

2.3. Social trust

3. The bidding game

3.1. Preliminaries and notation

3.2. Interesting bidding behaviors

3.3. Normal form bidding game

3.4. Numerical examples

4. Myopic bidding

5. Nonmyopic bidding

5.1. Rule definitions

5.2. Experimental results

5.2.1. One agent learning

5.2.2. All agents learning

5.3. The emergence of trust

6. Conclusions and future research

Acknowledgements

References

Vitae