Friend-or-Foe Q-learning in General-Sum Games Export

@inproceedings{littman01a, abstract = {This paper describes an approach to reinforcement learning in multiagent multiagent general-sum games in which a learner is told to treat each other agent as a friend or foe. This Q-learning-style algorithm provides strong convergence guarantees compared to an existing Nash-equilibrium-based learning rule.}, author = {Littman, Michael L.}, booktitle = {Proceedings of the Eighteenth International Conference on Machine Learning}, citeulike-article-id = {5971456}, citeulike-linkout-0 = {http://jmvidal.cse.sc.edu/library/littman01a.pdf}, keywords = {2001, competitive, game-theory, learning, multiagent, reinforcement, reinforcement-learning}, pages = {322--328}, posted-at = {2009-10-19 15:52:29}, priority = {0}, publisher = {Morgan Kaufmann}, title = {Friend-or-Foe Q-learning in General-Sum Games}, url = {http://jmvidal.cse.sc.edu/library/littman01a.pdf}, year = {2001} }

TY - CONF ID - littman01a L3 - citeulike-article-id:5971456 N2 - This paper describes an approach to reinforcement learning in multiagent multiagent general-sum games in which a learner is told to treat each other agent as a friend or foe. This Q-learning-style algorithm provides strong convergence guarantees compared to an existing Nash-equilibrium-based learning rule. EP - 328 TI - Friend-or-Foe Q-learning in General-Sum Games PB - Morgan Kaufmann T2 - Proceedings of the Eighteenth International Conference on Machine Learning SP - 322 KW - 2001 KW - competitive KW - game-theory KW - learning KW - multiagent KW - reinforcement KW - reinforcement-learning AU - Littman, Michael PY - 2001/// UR - http://jmvidal.cse.sc.edu/library/littman01a.pdf ER -