%0 Conference Paper %1 farahmand2008 %A Farahmand, A.m. %A Ghavamzadeh, M. %A Szepesvári, Cs. %A Mannor, S. %B EWRL %D 2008 %K approximation, function iteration learning, nonparametrics, planning, regularization, reinforcement theory, value %P 55--68 %R 10.1007/978-3-540-89722-4_5 %T Regularized Fitted Q-Iteration: Application to Planning %X 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.

@inproceedings{farahmand2008, 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.}, added-at = {2020-03-17T03:03:01.000+0100}, author = {Farahmand, A.{m}. and Ghavamzadeh, M. and Szepesv{\'a}ri, {Cs}. and Mannor, S.}, bdsk-url-1 = {http://dx.doi.org/10.1007/978-3-540-89722-4_5}, bibsource = {DBLP, http://dblp.uni-trier.de}, biburl = {https://www.bibsonomy.org/bibtex/2b7bba33a321d034a630545eebec58402/csaba}, booktitle = {EWRL}, crossref = {EWRL2008}, date-added = {2010-08-28 17:38:14 -0600}, date-modified = {2010-11-25 00:52:54 -0700}, doi = {10.1007/978-3-540-89722-4_5}, interhash = {5a32f86cdd262fb77fbe8175617d6793}, intrahash = {b7bba33a321d034a630545eebec58402}, keywords = {approximation, function iteration learning, nonparametrics, planning, regularization, reinforcement theory, value}, pages = {55--68}, pdf = {papers/RegFQI-Plan-EWRL08.pdf}, timestamp = {2020-03-17T03:03:01.000+0100}, title = {Regularized Fitted {Q}-Iteration: Application to Planning}, year = 2008 }