Algorithms and Bounds for Rollout Sampling Approximate Policy Iteration
Several approximate policy iteration schemes without value functions, which focus on policy representation using classifiers and address policy learning as a supervised learning problem, have been proposed recently. Finding good policies with such methods requires not only an appropriate classifier, but also reliable examples of best actions, covering the state space sufficiently. Up to this time, little work has been done on appropriate covering schemes and on methods for reducing the sample complexity of such methods, especially in continuous state spaces. This paper focuses on the simplest possible covering scheme (a discretized grid over the state space) and performs a sample-complexity comparison between the simplest (and previously commonly used) rollout sampling allocation strategy, which allocates samples equally at each state under consideration, and an almost as simple method, which allocates samples only as needed and requires significantly fewer samples.
KeywordsReinforcement Learning Policy Iteration Bandit Problem Policy Improvement Covering Scheme
Unable to display preview. Download preview PDF.
- 2.Auer, P., Ortner, R., Szepesvari, C.: Improved Rates for the Stochastic Continuum-Armed Bandit Problem. In: Bshouty, N.H., Gentile, C. (eds.) COLT 2007. LNCS, vol. 4539, pp. 454–468. Springer, Heidelberg (2007)Google Scholar
- 3.Bertsekas, D.: Dynamic programming and suboptimal control: From ADP to MPC. Fundamental Issues in Control, European Journal of Control 11(4-5) (2005); From 2005 CDC, Seville, SpainGoogle Scholar
- 4.Dimitrakakis, C., Lagoudakis, M.: Rollout sampling approximate policy iteration. Machine Learning 72(3) (September 2008)Google Scholar
- 6.Fern, A., Yoon, S., Givan, R.: Approximate policy iteration with a policy language bias. Advances in Neural Information Processing Systems 16(3) (2004)Google Scholar
- 9.Lagoudakis, M.G., Parr, R.: Reinforcement learning as classification: Leveraging modern classifiers. In: Proceedings of the 20th International Conference on Machine Learning (ICML), Washington, DC, USA, pp. 424–431 (August 2003)Google Scholar
- 10.Langford, J., Zadrozny, B.: Relating reinforcement learning performance to classification performance. In: Proceedings of the 22nd International Conference on Machine learning (ICML), Bonn, Germany, pp. 473–480 (2005)Google Scholar