Skip to main content

Temporal Difference and Policy Search Methods for Reinforcement Learning: An Empirical Comparison

Matthew E. Taylor‚ Shimon Whiteson and Peter Stone

Abstract

Reinforcement learning (RL) methods have become popular in recent years because of their ability to solve complex tasks with minimal feedback. Both genetic algorithms (GAs) and temporal difference (TD) methods have proven effective at solving difficult RL problems, but few rigorous comparisons have been conducted. Thus, no general guidelines describing the methods' relative strengths and weaknesses are available. This paper summarizes a detailed empirical comparison between a GA and a TD method in Keepaway, a standard RL benchmark domain based on robot soccer. The results from this study help isolate the factors critical to the performance of each learning method and yield insights into their general strengths and weaknesses.

Book Title
AAAI 2007: Proceedings of the Twenty−Second National Conference on Artificial Intelligence
Month
July
Note
(Nectar Track)
Pages
1675−1678
Year
2007