An improved reinforcement Q-learning method with BP neural networks in robot soccer

Shi Chao Wang, Zheng Xi Song, Hao Ding, Hao Bin Shi

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

9 Scopus citations

Abstract

In traditional reinforcement Q-Learning method, there exists two problems: difficulty of dividing the state information, complexity of extreme large dimension input. To solve these two problems, this paper proposed an improved reinforcement Q-Learning method with BP neutral network. In this method, the large Q table is replaced by a BP neural network. Continuous environmental information is the input. The Q value is the output. The Q value and weight of the network are also adjusted by the action rewards. This paper presents an algorithm for single agent's action selection. Simulation shows proposed method is more stable and applicable for the agent's strategy selection.

Original languageEnglish
Title of host publicationProceedings - 2011 4th International Symposium on Computational Intelligence and Design, ISCID 2011
Pages177-180
Number of pages4
DOIs
StatePublished - 2011
Event2011 4th International Symposium on Computational Intelligence and Design, ISCID 2011 - Hangzhou, China
Duration: 28 Oct 201130 Oct 2011

Publication series

NameProceedings - 2011 4th International Symposium on Computational Intelligence and Design, ISCID 2011
Volume1

Conference

Conference2011 4th International Symposium on Computational Intelligence and Design, ISCID 2011
Country/TerritoryChina
CityHangzhou
Period28/10/1130/10/11

Keywords

  • BP Neural Networks
  • Reinforcement Q-Learning
  • Robot Soccer

Fingerprint

Dive into the research topics of 'An improved reinforcement Q-learning method with BP neural networks in robot soccer'. Together they form a unique fingerprint.

Cite this