Reducing policy degradation in neuro-dynamic programming

Abstract

We focus on neuro-dynamic programming methods to learn state-action value functions and outline some of the inherent problems to be faced, when performing reinforcement learning in combination with function approximation. In an attempt to overcome some of these problems, we develop a reinforcement learning method that monitors the learning process, enables the learner to reflect whether it is better to cease learning, and thus obtains more stable learning results. © 2006 i6doc.com publication. All rights reserved.