Qualitative Decision under Uncertainty: Back to Expected Utility

Different qualitative models have been proposed for decision under uncertainty in Artificial Intelli­ gence, but they generally fail to satisfy the princi­ ple of strict Pareto dominance or principle of "ef­ ficiency", in contrast to the classical numerical cri­ terion — expected utility. In [Dubois and Prade, 1995J qualitative criteria based on possibility the­ ory have been proposed, that are appealing but inef­ ficient in the above sense. The question is whether it is possible to reconcile possibilistic criteria and efficiency. The present paper shows that the an­ swer is yes, and that it leads to special kinds of expected utilities. It is also shown that although nu­ merical, these expected utilities remain qualitative: they lead to two different decision procedures based on min, max and reverse operators only, generaliz­ ing the leximin and leximax orderings of vectors. DECISION THEORY 303

• Type: Conference
• Year: 2003
• Where: IJCAI
• Authors: Hélène Fargier, Régis Sabbadin
• Keywords: IJCAI 2003, IJCAI 2007, Possibility The­ Ory, Qualitative Criteria, Strict Pareto Dominance