untitled
<OAI-PMH schemaLocation=http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd>
<responseDate>2018-01-15T15:37:53Z</responseDate>
<request identifier=oai:HAL:hal-00542922v1 verb=GetRecord metadataPrefix=oai_dc>http://api.archives-ouvertes.fr/oai/hal/</request>
<GetRecord>
<record>
<header>
<identifier>oai:HAL:hal-00542922v1</identifier>
<datestamp>2017-12-21</datestamp>
<setSpec>type:ART</setSpec>
<setSpec>subject:info</setSpec>
<setSpec>collection:BNRMI</setSpec>
<setSpec>collection:UNIV-AG</setSpec>
</header>
<metadata><dc>
<publisher>HAL CCSD</publisher>
<title lang=en>Building Accurate Strategies in Non Markovian Environments without Memory</title>
<creator>Gilles, Enée</creator>
<creator>Peroumalnaïk, Mathias</creator>
<contributor>Laboratoire de Mathématiques Informatique et Applications (LAMIA) ; Université des Antilles et de la Guyane (UAG)</contributor>
<description>International audience</description>
<source>Lecture notes in computer science</source>
<publisher>springer</publisher>
<identifier>hal-00542922</identifier>
<identifier>https://hal.archives-ouvertes.fr/hal-00542922</identifier>
<source>https://hal.archives-ouvertes.fr/hal-00542922</source>
<source>Lecture notes in computer science, springer, 2010, IWLCS (6471), pp.107-126. 〈10.1007/978-3-642-17508-4_8〉</source>
<identifier>DOI : 10.1007/978-3-642-17508-4_8</identifier>
<relation>info:eu-repo/semantics/altIdentifier/doi/10.1007/978-3-642-17508-4_8</relation>
<language>en</language>
<subject lang=en>APCS</subject>
<subject lang=en>XCS</subject>
<subject lang=en>classier systems</subject>
<subject lang=en>non-markovian multi-step environments</subject>
<subject lang=en>strategy</subject>
<subject>I.2.6 [Learning]: [Concept learning, knowledge acquisition]</subject>
<subject>[INFO.INFO-LG] Computer Science [cs]/Machine Learning [cs.LG]</subject>
<type>info:eu-repo/semantics/article</type>
<type>Journal articles</type>
<description lang=en>This paper focuses on the study of the behavior of a genetic algorithm based classier system, the Adapted Pittsburgh Classier System (A.P.C.S), on maze type environments containing aliasing squares. This type of environment is often used in reinforcement learning literature to assess the performances of learning methods when facing problems containing non markovian situations. Through this study, we discuss on the performance of the APCS upon two mazes (Woods 101 and Maze E2) and also on the eciency of an improvement of the APCS learning method inspired from the XCS: the covering mechanism. We manage to show that, without any memory mechanism, the APCS is able to build and to keep accurate strategies to produce regular sub-optimal solution to these maze problems. This statement is shown through a comparison between the results obtained by the XCS on two specic maze problems and those obtained by the APCS.</description>
<date>2010</date>
</dc>
</metadata>
</record>
</GetRecord>
</OAI-PMH>
