DocumentCode :
2442584
Title :
Reinforcement Learning with Inertial Exploration
Author :
Bergeron, Dany ; Desjardins, Charles ; Laumônier, Julien ; Chaib-Draa, Brahim
Author_Institution :
Laval Univ., Laval
fYear :
2007
fDate :
2-5 Nov. 2007
Firstpage :
277
Lastpage :
280
Abstract :
In the Q-learning framework, the exploration of large environment is influenced by the time credit assignment problem. In this context, abstraction techniques may be used. Thus, multi-step actions (MSA) Q-learning has been proposed to take advantage of the fact that few action switches are usually required in optimal policies. In this article, we propose the concept of inertial exploration, we apply a log-selection of the scales to MSA Q-learning and we go further by proposing a dynamic time scale approach. We demonstrate that the same improvement in learning speed can be achieved without the full scales set. This improvement is shown on the mountain car problem and on a more realistic application of vehicle control.
Keywords :
learning (artificial intelligence); abstraction technique; dynamic time scale approach; inertial exploration; log-selection; mountain car problem; multistep actions Q-learning; reinforcement learning; time credit assignment problem; vehicle control; Computer science; Intelligent agent; Machine learning; Machine learning algorithms; Optimal control; Software engineering; Switches; Vehicle dynamics; Vehicle safety; Vehicles;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Agent Technology, 2007. IAT '07. IEEE/WIC/ACM International Conference on
Conference_Location :
Fremont, CA
Print_ISBN :
978-0-7695-3027-7
Type :
conf
DOI :
10.1109/IAT.2007.74
Filename :
4407297
Link To Document :
بازگشت