Presentation 2005/3/22
Reinforcement Learning under Dyanamic Emvironment : Sequential Monte Carlo with Sample Re-initialization
Akio TANAKA, Yohei NAKADA, Takashi MATSUMOTO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Reinforcement Learinng (RL) consists of an agent and its environment. The agent learns optimal or sub-optimal action sequence from the available reward. If an agent presumes static environment, while the actual environment changes, it has natural limitations in its performance. This study proposes a new RL algorithm which can cope with abrupt environmental changes within online Bayesian framework. Implementation is accomplished via "Sample Re-initialization" ("Sample Re-drawing") with Sequential Monte Carlo (SMC).
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Reinforcement Learning / Sequential Monte Carlo / Sample Re-initialization / Sample Re-drawing
Paper # NC2004-186
Date of Issue

Conference Information
Committee NC
Conference Date 2005/3/22(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) Reinforcement Learning under Dyanamic Emvironment : Sequential Monte Carlo with Sample Re-initialization
Sub Title (in English)
Keyword(1) Reinforcement Learning
Keyword(2) Sequential Monte Carlo
Keyword(3) Sample Re-initialization
Keyword(4) Sample Re-drawing
1st Author's Name Akio TANAKA
1st Author's Affiliation Graduate School of Science and Engineering, Waseda University:Core Research for Evolutional Science and Technology()
2nd Author's Name Yohei NAKADA
2nd Author's Affiliation Graduate School of Science and Engineering, Waseda University
3rd Author's Name Takashi MATSUMOTO
3rd Author's Affiliation Graduate School of Science and Engineering, Waseda University:Core Research for Evolutional Science and Technology
Date 2005/3/22
Paper # NC2004-186
Volume (vol) vol.104
Number (no) 759
Page pp.pp.-
#Pages 6
Date of Issue