Konferensartikel

Gemini - learning cooperative behaviors without communicating

Masayuki Ohta
Tokyo Institute of Technology, Japan

Ladda ner artikelhttp://www.ep.liu.se/ecp_article/index.en.aspx?issue=004;article=008

Ingår i: RobocCup-99 Team Descriptions. Simulation League

Linköping Electronic Conference Proceedings 4:8, s. 36-39

Linköping Electronic Articles in Computer and Information Science 6:8, s. 36-39

Visa mer +

Publicerad: 1999-12-15

ISBN:

ISSN: 1650-3686 (tryckt), 1650-3740 (online)

Abstract

This paper describes the design of Gemini a client program for SoccerServer. The goal of Gemini is learning cooperative behaviors without direct communication in multi-agent environment. With recent implementation; it can select the best strategy against opponent; statistically.

Nyckelord

Inga nyckelord är tillgängliga

Referenser

[1] M.Ohta and T.Ando \Cooperative Reward in Reinforcement Learning" Proc. of 3rd JSAI RoboMech Symposia pages 7-11 April 1998.

[2] Grefenstette, John.J. \Credit Assignment in Rule Discoverry Systems Based on Genetic Algorithms" Machine Learning, Vol.3 pages 225-245 1988.

[3] Kaelbling L. P., Littman M. L. and Moore A. W. \Reinforcement Learning: A Survey" Journal of Arti cial Intelligence Research 4, pages 237-285 1996.

[4] Kimura H., Yamamura M. and Kobayashi S. \Reinforcement Learning in Partially Observable Markov Decision Processes: A Stochastic Gradient Method" Journal of Japanese Society for Arti cial Intelligence, Vol.11, No.5 papges 761-768 1996

Citeringar i Crossref