Computer Games

---

This work started in a graduate class on Intelligent Agents (CSCI 6504) that I taught at Dalhousie University in Fall 2000.

• Opponent Modeling
  • This project (performed with graduate students Joe MacInnes and Omid Banyasad) involved using recursive modling on opponent agents in an adversarial "Quake-like" game. We used deterministic finite automate (DFA) for modeling agents. Every agent assumed that the other agents use the same model as its own but without recursion. The objective of this work was to investigate if recursive modeling allows an agent to outperform its opponents that are using similar models.
  • Further Reading
    • J. MacInnes, O. Banyasad, and M. A. Upal. Watching you watching me: opponent modeling using FSMs, in Proceedings of the Fourteenth Canadian Conference on Artificial Intelligence, pages 361-364, Springer Verlag, New York, 2001.
• Intelligent Coach for A RoboSoccer Team
  • This work performed by my graduate student Anthony Yuen. It involved designing a coach for a Simulation League RoboSoccer team that used backpropagation neural network to learn to recognize formation of the opponent team. Lazarus was one of only two Canadian that made it into the RoboSoccer 2001 Final.
  • Further Reading
    • A. Yuen. Lazarus Team Description. RoboCup 2001, Lecture Notes in Artificial Intelligence, pages 522-525, Springer Verlag, Berlin, 2001.

  ---

  Home | Research & Publications | Machine Learning for Planning | Data Mining Evaluation | Social Simulation | Military Logistics Planning | Automated Software Engg