Supervision : Jean-Yves JAFFRAY
Co-supervision : GONZALES Christophe

Prise de décision multiattribut avec le modèle GAI

GAI networks are a graphical model, both compact and expressive, for representing the preferences of a Decision Maker in the context of Multiattribute Decision Making, i.e., in situations where the set of alternatives among which the Decision Maker has to make decisions are described as a set of attributes (or features). GAI network's graphical structures are exploited to develop efficient elicitation procedures (determination of the Decision Maker's preferences using questionnaires) as well as effective Decision Making algorithms (e.g., computing the preferred alternative or the k-best alternatives). The goal of this PhD thesis is twofold. First, it extends the aforementioned state-of-the-art Decision Making algorithms to be able to cope with dense GAI networks, i.e., with situations where the GAI network's treewidth is too high for these algorithms to complete in a reasonable amount of time. For this purpose, a new triangulation method has been developed which produces approximated GAI networks on which tailored inference mechanisms determine the alternatives that are actually optimal for the original GAI network. Second, we have proposed new inference algorithms for Multicriteria Decision Making. More precisely, new approaches for determining Pareto-optimal sets (exact and approximate with performance guarantee) and Lorenz-optimal sets have been developed. In addition, we have also proposed new algorithms for computing the optimal solutions in situations where criteria are aggregated using various operators like OWA (Ordered Weighted Average), Choquet integrals and Tchebycheff's norm.

Phd defence : 09/23/2010

Jury members :

GONZALES Christophe , Professeur à l'Université Pierre et Marie Curie (directeur de thèse)
LERAY Philippe , Professeur à l'Université de Nantes (rapporteur)
SABBADIN Régis , Chargé de recherche INRA (rapporteur)
MAUDET Nicolas , Maître de conférence à l'Université Paris Dauphine (examinateur)
MUNIER-KORDON Alix, Professeur à l'Université Pierre et Marie Curie (examinateur)
PERNY Patrice , Professeur à l'Université Pierre et Marie Curie (examinateur)

2009-2011 Publications

• 2011
  • Ch. Gonzales, P. Perny, J.‑Ph. Dubus : “Decision Making with Multiple Objectives using GAI networks”, Artificial Intelligence, vol. 175 (7-8), pp. 1153-1179, (Elsevier) (2011)
• 2010
  • J.‑Ph. Dubus : “Prise de décision multiattribut avec le modèle GAI”, thesis, phd defence 09/23/2010, supervision Jaffray, Jean-Yves, co-supervision : Gonzales, Christophe (2010)
• 2009
  • J.‑Ph. Dubus, Ch. Gonzales, P. Perny : “Choquet Optimization using {GAI} Networks for Multiagent/Multicriteria Decision-Making”, Algorithmic Decision Theory, vol. 5783, Lecture Notes in Computer Science, Venice, Italy, pp. 377-389, (Springer) (2009)
  • J.‑Ph. Dubus, Ch. Gonzales, P. Perny : “Fast Recommendations using {GAI} Models”, IJCAI'09, International Joint Conference on Artificial Intelligence, Pasadena, California, United States, pp. 1896-1901 (2009)
  • J.‑Ph. Dubus, Ch. Gonzales, P. Perny : “Multiobjective Optimization using {GAI} Models”, IJCAI'09 International Joint Conference on Artificial Intelligence, Pasadena, California, United States, pp. 1902-1907 (2009)
  • J.‑Ph. Dubus, Ch. Gonzales, P. Perny : “Choix multiattribut à l’aide de réseaux {GAI} de forte densité”, 10^e Congrès de la Société Française de Recherche Opérationnelle et d'Aide à la Décision (ROADEF 2009), Nancy, France, pp. 25-39 (2009)