Title :
A Comparative Study of Parallel Reinforcement Learning Methods with a PC Cluster System
Author :
Kushida, Masayuki ; Takahashi, Kenichi ; Ueda, Hiroaki ; Miyahara, Tetsuhiro
Author_Institution :
Fac. of Inf. Sci., Hiroshima City Univ., Hiroshima
Abstract :
This paper presents a comparative study of three parallel implementation models for reinforcement learning. Two of them utilize Q-learning, and the other one utilizes fuzzy Q-learning for agent learning. In order to evaluate performance and validity of the three method, a PC (personal computer) cluster system consisting of 16 PCs connected via Gigabit ethernet has been built. For communications to deliver data among PCs, MPI (Message Passing Interface) is employed. Experimental results are compared with one another to show the performance and characteristics of the three methods.
Keywords :
fuzzy set theory; learning (artificial intelligence); message passing; parallel algorithms; workstation clusters; PC cluster system; agent learning; fuzzy Q-learning; gigabit Ethernet; message passing interface; parallel implementation model; parallel reinforcement learning; Computer simulation; Ethernet networks; Learning systems; Master-slave; Message passing; Personal communication networks; Process control;
Conference_Titel :
Intelligent Agent Technology, 2006. IAT '06. IEEE/WIC/ACM International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
0-7695-2748-5
