Abstract
This paper presents an analysis of criteria which measure policy optimality for average reward reinforcement learning. In previous works for undiscounted tasks, two criteria, gain-optimality and bias-optimality have been presented. The former is one to measure an average reward and the latter is one to evaluate transient actions. However, a limit factor in the definition of the gain-optimality makes real meaning of the criterion unclear, and what si worse, the performance function for the bias-optimality does not always converge. Thus, previous methods calculate an optimal policy with approximation approaches, that is, they don’t always acquire the optimal policy because of some finite errors. In addition, the theoretical proof of the convergence to the optimal policy is a difficult task. To eliminate ambiguity over these criteria, we show a necessary and sufficient condition of the gain-optimality: if and only if a policy is gain-optimal, it includes an optimal cycle-In other words, we only need to search a stationary cycle that has the highest average reward to find a gain optimal policy. We also make the performance function for the bias-optimality always converge by dividing it into two terms cycle-bias-value and path-bias-value. Finally, we build foundation of LC-learning, an algorithm for computing the bias optimal policy in a cyclic domain.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Leslie P. Kaelbling. Hierarchical learning in stochastic domains: Preliminary results. In Proceedings of the Tenth International Conference on Machine Learning (ICML-1993), pages 167–173, 1993.
Leslie P. Kaelbling. Learning to achieve goals. In Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence (IJCAI-1993), pages 1094–1099, 1993.
Leslie P. Kaelbling, Michael L. Littman, and Andrew P. Moore. Reinforcement learning: A survey. Journal of Artificial Intelligence Research, 4:237–285, 1996.
Taro Konda, Tensyo Shinjiro, and Tomohiro Yamaguchi. Lc-learning: In-stages model-based average reward reinforcement learning: Complexity and empirical results. In Proceedings of the Seventh Pacific Rim International Conference on Artificial Intelligence (PRICAI-2002), 2002.
Sridhar Mahadevan. An average-reward reinforcement learning algorithm for computing bias-optimal policies. In Proceedings of the Thirteenth AAAI (AAAI-1996), pages 875–880, 1996.
Sridhar Mahadevan. Average reward reinforcement learning: Foundations, algorithms, and empirical results. Machine Learning, 22(1-3):159–195, 1996.
Sridhar Mahadevan. Sensitive-discount optimality: Unifying average-reward and discounted reinforcement learning. In Proceedings of the Thirteenth International Conference on Machine Learning (ICML-1996), pages 328–336, 1996.
Andrew W. Moore and Christopher G. Atkeson. Prioritized sweeping: Reinforcement learning with less data and less time. Machine Learning, 13:103–130, 1993.
Martin L. Puterman. Markov Decision Processes: Discrete Dynamic Stochastic Programming, 92-93. John Wiley, 1994.
Stuart J. Russell and Peter Norvig. Artificial Intelligence: A Modern Approach. Prentice Hall, 1995.
Anton Schwartz. A reinforcement learning method for maximizing undiscounted rewards. In Proceedings of the Tenth International Conference on Machine Learning (ICML-1993), pages 298–305, 1993.
Richard S. Sutton and Andrew G. Barto. Reinforcement Learning: An Introduction. MIT Press, 1998.
Prasad Tadepalli and DoKyeong Ok. Model-based average reward reinforcement learning. Artificial Intelligence, 100(1-2):177–223, 1998.
Christopher J. Watkins and Peter Dayan. Q-learning. machine learning. Machine Learning, 3(8):279–292, 1992.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2002 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Konda, T., Yamaguchi, T. (2002). LC-Learning: Phased Method for Average Reward Reinforcement Learning —Analysis of Optimal Criteria —. In: Ishizuka, M., Sattar, A. (eds) PRICAI 2002: Trends in Artificial Intelligence. PRICAI 2002. Lecture Notes in Computer Science(), vol 2417. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45683-X_23
Download citation
DOI: https://doi.org/10.1007/3-540-45683-X_23
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-44038-3
Online ISBN: 978-3-540-45683-4
eBook Packages: Springer Book Archive