Improved QMDP policy for partially observable Markov decision processes in large domains: Embedding exploration dynamics

Apostolikas, G; Tzafestas, S

dc.contributor.author	Apostolikas, G	en
dc.contributor.author	Tzafestas, S	en
dc.date.accessioned	2014-03-01T02:42:48Z
dc.date.available	2014-03-01T02:42:48Z
dc.date.issued	2004	en
dc.identifier.issn	1079-8587	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/31089
dc.relation.uri	http://www.scopus.com/inward/record.url?eid=2-s2.0-4344697161&partnerID=40&md5=e3fd8e8490f9ff67ea9d6f90b0bbd0e5	en
dc.subject	Action selection	en
dc.subject	POMDP	en
dc.subject	QMDP	en
dc.subject.classification	Automation & Control Systems	en
dc.subject.classification	Computer Science, Artificial Intelligence	en
dc.subject.other	Computational complexity	en
dc.subject.other	Costs	en
dc.subject.other	Embedded systems	en
dc.subject.other	Evolutionary algorithms	en
dc.subject.other	Feedback	en
dc.subject.other	Information retrieval	en
dc.subject.other	Learning systems	en
dc.subject.other	Statistical methods	en
dc.subject.other	Vectors	en
dc.subject.other	Action Selection	en
dc.subject.other	Optimal policy	en
dc.subject.other	Partially Observable Markov Decision Processes (POMDP)	en
dc.subject.other	QMDP	en
dc.subject.other	Markov processes	en
dc.title	Improved QMDP policy for partially observable Markov decision processes in large domains: Embedding exploration dynamics	en
heal.type	conferenceItem	en
heal.language	English	en
heal.publicationDate	2004	en
heal.abstract	Artificial Intelligence techniques were primarily focused on domains in which at each time the state of the world is known to the system. Such domains can be modeled as a Markov Decision Process (MDP). Action and planning policies for MDPs have been studied extensively and several efficient methods exist. However, in real world problems pieces of information useful for the process of action selection are often missing. The theory of Partially Observable Markov Decision Processes (POMDP's) covers the problem domain in which the full state of the environment is not directly perceivable by the agent. Current algorithms for the exact solution of POMDP's are only applicable to domains with a small number of states. To cope with more extended state spaces, a number of methods that achieve sub-optimal solutions exist and among these the Q(MDP) approach seems to be the best. We introduce a novel technique, called Explorative Q(MDP) (EQ(MDP)) which constitutes an important enhancement of the Q(MDP) method. To the best knowledge of the authors, EQ(MDP) is currently the most efficient method applicable to large POMDP domains.	en
heal.publisher	AUTOSOFT PRESS	en
heal.journalName	Intelligent Automation and Soft Computing	en
dc.identifier.isi	ISI:000223356100002	en
dc.identifier.volume	10	en
dc.identifier.issue	3	en
dc.identifier.spage	209	en
dc.identifier.epage	220	en