Reinforcement Learning in Multi-dimensional State-action Space Using Random Tiling and Gibbs Sampling

• KIMURA Hajime

  Dept. of Marine Engineering, Graduate School of Engineering, Kyushu University

Bibliographic Information

Other Title

• ランダムタイリングとGibbs-samplingを用いた多次元状態-行動空間における強化学習
• ランダムタイリング ト Gibbs sampling オ モチイタ タジゲン ジョウタイ コウドウ クウカン ニ オケル キョウカ ガクシュウ

Abstract

In real-robot applications, learning controllers are often required to obtain control rules over high-dimensional continuous state-action space. Random tile-coding is a promising method to deal with high-dimensional state space for representing the state value function. However, there is no standard reinforcement learning scheme to deal with action selection in high-dimensional action space, especially the probability of action variables are mutually dependent. This paper introduces a new action selection scheme using random tile-coding and Gibbs sampling, and shows the Q-learning algorithm applying the proposed scheme. We demonstrate it through a Rod in maze problem and a redundant arm reaching task.

Journal

• Transactions of the Society of Instrument and Control Engineers

  Transactions of the Society of Instrument and Control Engineers 42 (12), 1336-1343, 2006

  The Society of Instrument and Control Engineers

Keywords

• reinforcement learning
• Q-learning
• random tile coding
• Gibbs sampling
• action selector

Details

• CRID: 1390001204503515136
• NII Article ID: 10018422317
• NII Book ID: AN00072392
• DOI: 10.9746/sicetr1965.42.1336
• ISSN: 18838189
• ISSN: 04534654
• ISSN: http://id.crossref.org/issn/04534654
• NDL BIB ID: 8625992
• Web Site: https://ndlsearch.ndl.go.jp/books/R000000004-I8625992
• Web Site: https://www.jstage.jst.go.jp/article/sicetr1965/42/12/42_12_1336/_pdf
• Data Source:
  • JaLC
  • NDL
  • Crossref
  • CiNii Articles
  • KAKEN
• Abstract License Flag: Disallowed