Theoretical and Empirical Analysis of Reward Shaping in Reinforcement Learning

Research output: Contribution to conferencePaper

Author(s)

Department/unit(s)

Publication details

DatePublished - 2009
Original languageUndefined/Unknown

Abstract

Reinforcement learning suffers scalability problems due to the state space explosion and the temporal credit assignment problem. Knowledge-based approaches have received a significant attention in the area. Reward shaping is a particular approach to incorporate domain knowledge into reinforcement learning. Theoretical and empirical analysis of this paper reveals important properties of this principle, especially the influence of the reward type, MDP discount factor, and the way of evaluating the potential function on the performance.

Discover related content

Find related publications, people, projects, datasets and more using interactive charts.

View graph of relations