To create neuro-controlled game opponent from UCT-created data

ABSTRACT

Adaptive Game AI improves adaptability of opponent AI as well as the challenge level of the gameplay, as a result the entertainment of game is augmented. Opponent game AI is usually implemented by scripted rules in video games, but the most updated algorithm of UCT (Upper Confidence bound for Trees) which perform well in computer go can also be used to achieve excellent result to control non-player characters (NPCs) in video games. However, due to computational intensiveness of UCT, it is actually not suitable for Online Games. As it is already known that UCT can create near optimal control, so it is possible to create Neuro-Controlled Game Opponent by off-line learning from the UCT created sample data; finally Neuro-Controlled Game Opponent for Online Games from UCT-Created Data without worry about computational intensiveness is generated. And also if the optimization approach of Neuro-Evolution is applied to the above generated Neuro-Controller, the performance of the opponent AI is enhanced even further.

REFERENCES

• 1: Yannakakis, Giorgios N, AI in Computer Games: Generating Interesting Interactive Opponents by the use of Evolutionary Computation. PhD thesis, University of Edinburgh, 2005.
• 2: Georgios N. Yannakakis, John Levine, and John Hallam. An Evolutionary Approach for Interactive Computer Games. In Proceedings of the Congress on Evolutionary Computation (CEC-04), pages 986--993, June 2004.
• 3: Peter Bentley Article, LETTING STONES GO UNTURNED, Feb, 2008.
• 4: Simon M. Lucas, Computational Intelligence and Games: Challenges and Opportunities, International Journal of Automation and Computing, Pages 45--57, January, 2008.
• 5: P. Auer , N. Cesa-Bianchi , Y. Freund , R. E. Schapire, Gambling in a rigged casino: The adversarial multi-armed bandit problem, Proceedings of the 36th Annual Symposium on Foundations of Computer Science, p.322, October 23-25, 1995
• 6: Multi-armed bandit, Wikipedia, Retrieved from http://en.wikipedia.org/wiki/Multi-armed_bandit (22/08/08).
• 7: Peter Auer , Nicolò Cesa-Bianchi , Paul Fischer, Finite-time Analysis of the Multiarmed Bandit Problem, Machine Learning, v.47 n.2-3, p.235-256, May-June 2002 [doi>10.1023/A:1013689704352]
• 8: Levente Kocsis and Csaba Szepesvari. Bandit based Monte Carlo planning. In 15th European Conference on Machine Learning (ECML), pages 282--293, 2006.
• 9: Suoju He, Fan Xie, Yi Wang, Jin Meng, Hongtao Chen, Zhiqing Liu, Qiliang Zhu. Game Player Strategy Pattern Recognition and How UCT Algorithms Apply Pre-Knowledge of Player's Strategy to Improve Opponent AI. In the International Conference on Innovation in Software Engineering (ISE'2008), 2008.
• 10: Suoju He, Fan Xie, Yi Wang, Sai Luo, Yiwen Fu, Jiajian Yang, Zhiqing Liu, Qiliang Zhu. To Create Adaptive Game Opponent by Using UCT. In the International Conference on Intelligent Agents, Web Technologies and Internet Commerce -- IAWTIC'2008, 2008.
• 11: Thomas M. Mitchell, Machine Learning, McGraw-Hill Higher Education, 1997
• 12: Neuroevolution, Wikipedia, Retrieved from http://en.wikipedia.org/wiki/Neuroevolution (22/08/08).

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.8 Problem Solving, Control Methods, and Search
          Subjects: Control theory

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.8 Problem Solving, Control Methods, and Search
          Subjects: Heuristic methods; Graph and tree search strategies; Plan execution, formation, and generation

General Terms:
Algorithms, Design, Experimentation, Performance, Verification

Keywords:
UCT, adaptive game ai, dead end, neuro-controller, neuro-evolution