Fukuoka Institute of Technology REPOSITORY SYSTEM

HOME    repository info    mypage        Japanese    library    university    Feedback

Fukuoka Institute of Technology Repository >
02.University Bulletins & Report >
01.RESEARCH BULLETIN OF FUKUOKA INSTITUTE OF TECHNOLOGY >
Vol.44.No.1(67) >

 
Title :確率環境下でのモデルベース学習
Authors :田村 剛士
鶴岡 久
山口 明宏
Issue Date :Sep-2011
Abstract :Model based learning can reevaluate the utility of every state, according to a measure of urgency. Prioritized sweeping is a typical algorithm for efficient state updating. In a stochastic environment, a probability distribution can be used to represent the uncertainty of the Q-value caused by probabilistic state transitions or probabilistic rewards. The product of the confidence interval and the Bellman error is used to provide a measure for prioritizing,which takes account of the level of confidence and also yields a measure of urgency. The performance of this approach in the trap domain is examined and compared with that of the ordinary sweeping method. Experimental results indicate that the proposed approach results in a more effective exploration of the state than does the use of conventional sweeping methods.
Type Local :紀要論文
ISSN :02876620
Publisher :福岡工業大学
フクオカ コウギョウ ダイガク
URI :http://hdl.handle.net/11478/1289
citation :福岡工業大学研究論集
44
1
9
12
Citation :福岡工業大学研究論集 Vol.44 no.1 p.9 -12
Appears in Collections:Vol.44.No.1(67)

Files in This Item:

File Description SizeFormat
11478-1289_p9.pdf835KbAdobe PDFView/Open