Sunehag, PeterHutter, Marcus2015-08-172015-08-17978-3-642-35100-60302-9743http://hdl.handle.net/1885/14734We use optimism to introduce generic asymptotically optimal reinforcement learning agents. They achieve, with an arbitrary finite or compact class of environments, asymptotically optimal behavior. Furthermore, in the finite deterministic case we provide finite error bounds.© Springer-Verlag Berlin Heidelberg 2012. http://www.sherpa.ac.uk/romeo/issn/0302-9743/..."Author's post-print on any open access repository after 12 months after publication" from SHERPA/RoMEO site (as at 17/08/15)Reinforcement LearningOptimismOptimality201210.1007/978-3-642-35101-3_2