Presentation 2004/10/12
Adaptation of the Online Policy-Improving System by using a Mixture Model of Bayesian Networks to Dynamic Environments
Daisuke KITAKOSHI, Hiroyuki SHIOYA, Ryohei NAKANO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We have proposed a system improving reinforcement learning agents' policies by using a mixture model of Bayesian Networks (BNs). Each BN in the mixture model corresponds to a stochastic knowledge of an environment. In this paper, we introduce autonomous mechanisms for recognizing changes of environments, and for learning mixing rates of BNs in the mixture model, to the above system. Computer simulations in the agent navigation problem are carried out in order to discuss the adaptability of our online policy-improving system to dynamic environments and properties of the policy-improving procedure.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Mixture model of Bayesian Networks / Stochastic knowledge / Policy-improving system / Profit sharing
Paper # NC2004-71
Date of Issue

Conference Information
Committee NC
Conference Date 2004/10/12(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) Adaptation of the Online Policy-Improving System by using a Mixture Model of Bayesian Networks to Dynamic Environments
Sub Title (in English)
Keyword(1) Mixture model of Bayesian Networks
Keyword(2) Stochastic knowledge
Keyword(3) Policy-improving system
Keyword(4) Profit sharing
1st Author's Name Daisuke KITAKOSHI
1st Author's Affiliation Graduate School of Engineering, Nagoya Institute of Technology()
2nd Author's Name Hiroyuki SHIOYA
2nd Author's Affiliation Faculty of Engineering, Muroran Institute of Technology
3rd Author's Name Ryohei NAKANO
3rd Author's Affiliation Graduate School of Engineering, Nagoya Institute of Technology
Date 2004/10/12
Paper # NC2004-71
Volume (vol) vol.104
Number (no) 349
Page pp.pp.-
#Pages 6
Date of Issue