Presentation 2003/11/14
Nonlinear control by learning the switching of multiple controllers
Masaya NISHIMURA, Junichiro YOSHIMOTO, Shin ISIHI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Reinforcement learning(RL) has recently been applied to contracting a controller for nonlinear systems. Merits of the RL are that an exact dynamics model of the controlled object is not required and adaptability to non-stationary environments. If the dynamics has strong nonlinearity and/or high dimensionality, however, the RL suffers from unstable learning and/or a huge number of learning episodes. For this reason, a naive RL method is not suitable for the tasks in a real world. In order to overcome the disadvantage of the RL, we propose a new RL scheme using multiple incomplete controllers, each of which is constructed based on a locally linear dynamics in a specific subspace. An inverted-pendulum task and an acrobot control task on computer simulation showed that our method was able to select an appropriate incomplete controller for each subspace, so that a fairly good nonlinear control was realized.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) reinforcement learning / acrobot / nonlinear control
Paper # NC2003-79
Date of Issue

Conference Information
Committee NC
Conference Date 2003/11/14(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) Nonlinear control by learning the switching of multiple controllers
Sub Title (in English)
Keyword(1) reinforcement learning
Keyword(2) acrobot
Keyword(3) nonlinear control
1st Author's Name Masaya NISHIMURA
1st Author's Affiliation Nara Institute of Science and Technology()
2nd Author's Name Junichiro YOSHIMOTO
2nd Author's Affiliation Nara Institute of Science and Technology:CREST, Japan Science and Technology Agency
3rd Author's Name Shin ISIHI
3rd Author's Affiliation Nara Institute of Science and Technology:CREST, Japan Science and Technology Agency
Date 2003/11/14
Paper # NC2003-79
Volume (vol) vol.103
Number (no) 465
Page pp.pp.-
#Pages 6
Date of Issue