Improving MACS thanks to a comparison with 2TBNs

Thierry Gourdin 1 Olivier Sigaud 1 Pierre-Henri Wuillemin 2
1 Animatlab
LIP6 - Laboratoire d'Informatique de Paris 6
2 DECISION
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : Factored Markov Decision Processes is the theoretical framework underlying multi-step Learning Classifier Systems research. This framework is mostly used in the context of Two-stage Bayes Networks, a subset of Bayes Networks. In this paper, we compare the Learning Classifier Systems approach and the Bayes Networks approach to factored Markov Decision Problems. More specifically, we focus on a comparison between MACS, an Anticipatory Learning Classifier System, and Structured Policy Iteration, a general planning algorithm used in the context of Two-stage Bayes Networks. From that comparison, we define a new algorithm resulting from the adaptation of Structured Policy Iteration to the context of MACS. We conclude by calling for a closer communication between both research communities.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01501406
Contributor : Lip6 Publications <>
Submitted on : Tuesday, April 4, 2017 - 11:42:13 AM
Last modification on : Thursday, March 21, 2019 - 2:17:01 PM

Links full text

Identifiers

Citation

Thierry Gourdin, Olivier Sigaud, Pierre-Henri Wuillemin. Improving MACS thanks to a comparison with 2TBNs. GECCO 2004 - Genetic and Evolutionary Computation Conference, Jun 2004, Seattle, WA, United States. pp.810-823, ⟨10.1007/978-3-540-24855-2_95⟩. ⟨hal-01501406⟩

Share

Metrics

Record views

69