Presentation	1999/7/19 Acceleration of Reinforcement Learning Process Based out Action Allocated Multi Agents Shuji Nakamura, Masaki Sano, Yasiji Sawada,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	This paper describes an algorithm to speed up Q-Learning which is often used in reinforcement learning. Usually this learning method needs much time and computational resources, because it evaluates state-action pairs in high dimensional space. We propose the algorithm that positively regard single agent as multi agents and divide decision for outputs (actions). It reduces the number of the iterations until converge, calculation time, and memory size since search space is reduced. Besides it, we evaluate this algorithm and show some applications.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	reinforcement learning / acceleration / multi agent / hierarchical
Paper #	NC99-34
Date of Issue

Conference Information
Committee	NC
Conference Date	1999/7/19(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)	Acceleration of Reinforcement Learning Process Based out Action Allocated Multi Agents
Sub Title (in English)
Keyword(1)	reinforcement learning
Keyword(2)	acceleration
Keyword(3)	multi agent
Keyword(4)	hierarchical
1st Author's Name	Shuji Nakamura
1st Author's Affiliation	Graduate School of Information Science, Tohoku Univ.()
2nd Author's Name	Masaki Sano
2nd Author's Affiliation	Research Institute of Electrical Communication, Tohoku Univ.
3rd Author's Name	Yasiji Sawada
3rd Author's Affiliation	Research Institute of Electrical Communication, Tohoku Univ.
Date	1999/7/19
Paper #	NC99-34
Volume (vol)	vol.99
Number (no)	193
Page	pp.pp.-
#Pages	8
Date of Issue