This paper is also available in PostScript

The Impact of Nested Agent Models in an Information Economy

José M. Vidal and Edmund H. Durfee
Artificial Intelligence Laboratory, University of Michigan
1101 Beal Avenue, Ann Arbor, MI 48109-2110
jmvidal@umich.edu

August 21, 1996

Abstract:

We present our approach to the problem of how an agent, within an economic Multi-Agent System, can determine when it should behave strategically (i.e. model the other agents), and when it should act as a simple price-taker. We provide a framework for the incremental implementation of modeling capabilities in agents. These agents were implemented and different populations simulated in order to learn more about their behavior and the merits of using agent models. Our results show, among other lessons, how savvy buyers can avoid being ``cheated'' by sellers, how price volatility can be used to quantitatively predict the benefits of deeper models, and how specific types of agent populations influence system behavior.

Topic Areas: Agent Modeling/Learning, Economic Societies of Agents.

• Introduction
• Description of the UMDL
  • Implications of the information economy.
• A Simplified Model of the UMDL
• Learning recursive models
  • Agents with no models
    • Buyers with no models.
    • Sellers with no models.
  • Agents with One-level Models
    • Buyers with one-level models.
    • Sellers with one-level models.
  • Agents with Two-level Models
    • Buyers with two-level models.
    • Sellers with two-level models.
• Tests
  • Lessons
    • Micro versus macro behaviors.
    • 0-level buyers and sellers.
    • 0-level buyers and sellers, plus one 1-level seller.
    • 1-level buyers and 0 and 1-level sellers.
    • 1-level buyers and 1 and 2-level sellers.
• Conclusions
• References