Playing is believing: the role of beliefs in multi-agent learning

dc.contributor.author	Chang, Yu-Han
dc.contributor.author	Kaelbling, Leslie P.
dc.date.accessioned	2003-11-17T16:46:08Z
dc.date.available	2003-11-17T16:46:08Z
dc.date.issued	2003-01
dc.identifier.uri	http://hdl.handle.net/1721.1/3688
dc.description.abstract	We propose a new classification for multi-agent learning algorithms, with each league of players characterized by both their possible strategies and possible beliefs. Using this classification, we review the optimality of existing algorithms and discuss some insights that can be gained. We propose an incremental improvement to the existing algorithms that seems to achieve average payoffs that are at least the Nash equilibrium payoffs in the long-run against fair opponents.	en
dc.description.sponsorship	Singapore-MIT Alliance (SMA)	en
dc.format.extent	114175 bytes
dc.format.mimetype	application/pdf
dc.language.iso	en_US
dc.relation.ispartofseries	Computer Science (CS);
dc.subject	multi-agent learning algorithm	en
dc.subject	repeated games	en
dc.subject	belief	en
dc.subject	game theory	en
dc.subject	Matrix games	en
dc.subject	Nash equilibrium	en
dc.subject	Stochastic games	en
dc.subject	Reinforcement learning	en
dc.subject	PHC-Exploiter	en
dc.title	Playing is believing: the role of beliefs in multi-agent learning	en
dc.type	Article	en