SUPERVISED LEARNING FOR AGENT POSITIONING BY USING SELF-ORGANIZING MAP

Kazuma Moriyasu, Takeshi Yoshikawa, Hidetoshi Nonaka

2010

Abstract

We propose a multi-agent cooperative method that helps each agent to cope with partial observation and reduces the number of teaching data. It learns cooperative actions between agents by using the Self-Organizing Map as supervised learning. Input Vectors of the Self-Organizing Map are the data that reflects the operator’s intention. We show that our proposed method can acquire cooperative actions between agents and reduce the number of teaching data by two evaluation experiments using the pursuit problem that is one of multi-agent system.

References

  1. Akiyama, H. and Noda, I. (2008). Triangulation based approximation model for agent positioning problem. Transactions of the Japanese Society for Artificial Intelligence, 23(4):255-267.
  2. Ito, A. and Kanabuchi, M. (2001). Speeding up multi-agent reinforcement learning by coarse-graining of perception -hunter game as an example-. The transactions of the Institute of Electronics, Information and Communication Engineers, J84-D-1(3):285-293.
  3. Katayama, K., Koshiishi, T., and Narihisa, H. (2004). Reinforcement learning agents with analytic hierarchy process: a case study of pursuit problem. Transactions of the Japanese Society for Artificial Intelligence, 19(4):279-291.
  4. Self-Organizing Maps(3rd ed.).
  5. Ono, N. and Fukumoto, K. (1997). A modular approach to multi-agent reinforcement learning. Springer-Verlag.
  6. Rumelhart, D. and McClelland, J. (1986). Learning internal representation by error propagation. Parallel Distributed Processing: Explorations in the Microstructure of Cognition, 1:318-362.
  7. Tateyama, T., Kawate, S., and Oguchi, T. (2004). A teaching method by using self-organizing map for reinforcement learning. Transactions of the Japan Society of Mechanical Engineers, 70:1722-1729.
Download


Paper Citation


in Harvard Style

Moriyasu K., Yoshikawa T. and Nonaka H. (2010). SUPERVISED LEARNING FOR AGENT POSITIONING BY USING SELF-ORGANIZING MAP . In Proceedings of the 12th International Conference on Enterprise Information Systems - Volume 2: ICEIS, ISBN 978-989-8425-05-8, pages 368-372. DOI: 10.5220/0003018603680372


in Bibtex Style

@conference{iceis10,
author={Kazuma Moriyasu and Takeshi Yoshikawa and Hidetoshi Nonaka},
title={SUPERVISED LEARNING FOR AGENT POSITIONING BY USING SELF-ORGANIZING MAP},
booktitle={Proceedings of the 12th International Conference on Enterprise Information Systems - Volume 2: ICEIS,},
year={2010},
pages={368-372},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0003018603680372},
isbn={978-989-8425-05-8},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 12th International Conference on Enterprise Information Systems - Volume 2: ICEIS,
TI - SUPERVISED LEARNING FOR AGENT POSITIONING BY USING SELF-ORGANIZING MAP
SN - 978-989-8425-05-8
AU - Moriyasu K.
AU - Yoshikawa T.
AU - Nonaka H.
PY - 2010
SP - 368
EP - 372
DO - 10.5220/0003018603680372