Bandit Based Monte-Carlo Planning Export

@incollection{@INPROCEEDINGSKocsis06banditbased, author = Levente Kocsis and Csaba Szepesvári, title = Bandit based Monte-Carlo Planning, booktitle = In: ECML-06. Number 4212 in LNCS, year = 2006, pages = 282--293, publisher = Springer, abstract = {For large state-space Markovian Decision Problems Monte-Carlo planning is one of the few viable approaches to find near-optimal solutions. In this paper we introduce a new algorithm, UCT, that applies bandit ideas to guide Monte-Carlo planning. In finite-horizon or discounted MDPs the algorithm is shown to be consistent and finite sample bounds are derived on the estimation error due to sampling. Experimental results show that in several domains, UCT is significantly more efficient than its alternatives.}, author = {Kocsis, Levente and Szepesv\'{a}ri, Csaba}, citeulike-article-id = {2480968}, citeulike-linkout-0 = {http://dx.doi.org/10.1007/11871842_29}, citeulike-linkout-1 = {http://www.springerlink.com/content/d232253353517276}, doi = {10.1007/11871842_29}, journal = {Machine Learning: ECML 2006}, keywords = {carlo, monte, search, uct}, pages = {282--293}, posted-at = {2009-11-06 09:26:15}, priority = {5}, title = {Bandit Based Monte-Carlo Planning}, url = {http://dx.doi.org/10.1007/11871842_29}, year = {2006} }

TY - CHAP ID - @INPROCEEDINGSKocsis06banditbased, author = Levente Kocsis and Csaba Szepesvári, title = Bandit based Monte-Carlo Planning, booktitle = In: ECML-06. Number 4212 in LNCS, year = 2006, pages = 282--293, publisher = Springer L3 - citeulike-article-id:2480968 N2 - For large state-space Markovian Decision Problems Monte-Carlo planning is one of the few viable approaches to find near-optimal solutions. In this paper we introduce a new algorithm, UCT, that applies bandit ideas to guide Monte-Carlo planning. In finite-horizon or discounted MDPs the algorithm is shown to be consistent and finite sample bounds are derived on the estimation error due to sampling. Experimental results show that in several domains, UCT is significantly more efficient than its alternatives. JF - Machine Learning: ECML 2006 EP - 293 TI - Bandit Based Monte-Carlo Planning SP - 282 KW - carlo KW - monte KW - search KW - uct AU - Kocsis, Levente AU - Szepesvári, Csaba PY - 2006/// UR - http://dx.doi.org/10.1007/11871842_29 ER -