Algorithmes et complexités génériques pour différents cadres de décision séquentielle dans l'incertain

Recently, a generic algebraic framework called PFU (for Plausibilities-FeasibilitiesUtilities) has been introduced in order to represent generic forms of sequential decision making under uncertainties. The PFU framework covers various existing probabilistic or possibilistic formalisms. This article introduces generic algorithms defined in the PFU framework which are therefore applicable for various decision models. Theoretical complexity results are also provided based on a parameter called the constrained induced-width. MOTS-CLÉS : cadre PFU, algorithmes génériques, élimination de variable, largeur induite.