Presentation 1998/3/19
Multiagent reinforcement learning
Kei Fukuzawa, Bahman Kerumanshahi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper deals with determination of multiagent problem using the reinforcement learning. In this study, an actual soccer problem is taking into account with imitating the problem (two sides members, ground, etc.) by a simple model. In this model, ten agents are assigned confronted on the computer:five agents for self-team members and five agents for opposite-team members. From the obtained results of numerical computations, when a plenty of games is repeated, it has been confirmed that each of the agents of self-team have their own activities and have produced their own individual performances in the team, respectively. Furthermore, in this paper a learning method that becomes even more efficient than the conventional reinforcement learning is proposed. Ultimately, in order to verify the effectiveness and validity of the proposed method, and also for verification of learning efficiency, it is compared with conventional method and showed promising results.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) multiagent / reinforcenment learning / soccer
Paper #
Date of Issue

Conference Information
Committee NC
Conference Date 1998/3/19(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Multiagent reinforcement learning
Sub Title (in English)
Keyword(1) multiagent
Keyword(2) reinforcenment learning
Keyword(3) soccer
1st Author's Name Kei Fukuzawa
1st Author's Affiliation Tokyo Metropolitan University()
2nd Author's Name Bahman Kerumanshahi
2nd Author's Affiliation Tokyo Metropolitan University
Date 1998/3/19
Paper #
Volume (vol) vol.97
Number (no) 623
Page pp.pp.-
#Pages 5
Date of Issue