Abstract
We consider planning in a Markovian decision problem, i.e., the problem of finding a good policy given access to a generative model of the environment. We propose to use fitted Q-iteration with penalized (or regularized) least-squares regression as the regression subroutine to address the problem of controlling model-complexity. The algorithm is presented in detail for the case when the function space is a reproducing-kernel Hilbert space underlying a user-chosen kernel function. We derive bounds on the quality of the solution and argue that data-dependent penalties can lead to almost optimal performance. A simple example is used to illustrate the benefits of using a penalized procedure.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Antos, A., Szepesvári, C., Munos, R.: Value-iteration based fitted policy iteration: learning with a single trajectory. In: IEEE ADPRL, pp. 330–337 (2007)
Antos, A., Munos, R., Szepesvári, C.: Fitted Q-iteration in continuous action-space MDPs. In: Advances in Neural Information Processing Systems 20, NIPS 2007 (in print, 2008)
Antos, A., Szepesvári, C., Munos, R.: Learning near-optimal policies with Bellman-residual minimization based fitted policy iteration and a single sample path. Machine Learning 71, 89–129 (2008)
Bertsekas, D.P., Shreve, S.E.: Stochastic Optimal Control (The Discrete Time Case). Academic Press, New York (1978)
Bunea, F., Tsybakov, A., Wegkamp, M.: Sparsity oracle inequalities for the lasso. Electronic Journal of Statistics 1, 169–194 (2007)
Engel, Y., Mannor, S., Meir, R.: The kernel recursive least squares algorithm. IEEE Transaction on Signal Processing 52(8), 2275–2285 (2004)
Engel, Y., Mannor, S., Meir, R.: Reinforcement learning with Gaussian processes. In: ICML 2005: Proceedings of the 22nd international conference on Machine learning, pp. 201–208. ACM, New York (2005)
Ernst, D., Geurts, P., Wehenkel, L.: Tree-based batch mode reinforcement learning. Journal of Machine Learning Research 6, 503–556 (2005)
Farahmand, A.M., Ghavamzadeh, M., Szepesvári, C., Mannor, S.: Regularized policy iteration. In: Advances in Neural Information Processing Systems 21, NIPS 2008 (to appear, 2008)
Györfi, L., Kohler, M., Krzyżak, A., Walk, H.: A distribution-free theory of nonparametric regression. Springer, New York (2002)
Jung, T., Polani, D.: Least squares SVM for least squares TD learning. In: ECAI, pp. 499–503 (2006)
Kearns, M., Mansour, Y., Ng, A.Y.: A sparse sampling algorithm for near-optimal planning in large Markovian decision processes. In: Proceedings of IJCAI 1999, pp. 1324–1331 (1999)
Lagoudakis, M.G., Parr, R.: Reinforcement learning as classification: Leveraging modern classifiers. In: ICML 2003, pp. 424–431 (2003)
Loth, M., Davy, M., Preux, P.: Sparse temporal difference learning using LASSO. In: IEEE International Symposium on Approximate Dynamic Programming and Reinforcement Learning (2007)
Mannor, S., Menache, I., Shimkin, N.: Basis function adaptation in temporal difference reinforcement learning. Annals of Operations Research 134, 215–238 (2005)
Munos, R., Szepesvári, C.: Finite-time bounds for fitted value iteration. Journal of Machine Learning Research 9, 815–857 (2008)
Ng, A.Y., Jordan, M.: PEGASUS: A policy search method for large MDPs and POMDPs. In: Proceedings of the 16th Conference in Uncertainty in Artificial Intelligence, pp. 406–415 (2000)
Ormoneit, D., Sen, S.: Kernel-based reinforcement learning. Machine Learning 49, 161–178 (2002)
Parr, R., Painter-Wakefield, C., Li, L., Littman, M.L.: Analyzing feature generation for value-function approximation. In: ICML, pp. 737–744 (2007)
Schölkopf, B., Smola, A.J.: Learning with Kernels. MIT Press, Cambridge (2002)
Srebro, N., Ben-David, S.: Learning bounds for support vector machines with learned kernels. In: Lugosi, G., Simon, H.U. (eds.) COLT 2006. LNCS, vol. 4005, pp. 169–183. Springer, Heidelberg (2006)
Xu, X., Hu, D., Lu, X.: Kernel-based least squares policy iteration for reinforcement learning. IEEE Trans. on Neural Networks 18, 973–992 (2007)
Zhou, D.-X.: Capacity of reproducing kernel spaces in learning theory. IEEE Transactions on Information Theory 49, 1743–1752 (2003)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Farahmand, A.m., Ghavamzadeh, M., Szepesvári, C., Mannor, S. (2008). Regularized Fitted Q-Iteration: Application to Planning. In: Girgin, S., Loth, M., Munos, R., Preux, P., Ryabko, D. (eds) Recent Advances in Reinforcement Learning. EWRL 2008. Lecture Notes in Computer Science(), vol 5323. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-89722-4_5
Download citation
DOI: https://doi.org/10.1007/978-3-540-89722-4_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-89721-7
Online ISBN: 978-3-540-89722-4
eBook Packages: Computer ScienceComputer Science (R0)