An application of decision rules in reinforcement learning

Authors
Michalski, A
Citation
A. Michalski, An application of decision rules in reinforcement learning, CONTROL CYB, 29(4), 2000, pp. 989-996
Citations number
12
Categorie Soggetti
AI Robotics and Automatic Control
Journal title
CONTROL AND CYBERNETICS
ISSN journal
03248569 → ACNP
Volume
29
Issue
4
Year of publication
2000
Pages
989 - 996
Database
ISI
SICI code
0324-8569(2000)29:4<989:AAODRI>2.0.ZU;2-Z
Abstract
In this paper an application of decision rules to function representation i n reinforcement learning is described. Rules are generated incrementally by method based on rough set theory from instances recorded in state-action-Q -value memory. Simulation experiment investigating the performance of tile system and results achieved are reported.