Mardi 16 avril de 14:00 à 15:00
TITLE: Aggregating Incomplete Pairwise Preferences by Weight
ABSTRACT We introduce a model for the representation and aggregation of incomplete preferences, designing a family of rules based on the weight a pairwise ranking of two alternatives receives when reported within a preference set. Our model is relevant to multiagent scenarios where agents may satisfy minimal rationality requirements and violate not only completeness but also transitivity. We examine two original normative principles—restricted majoritarianism and splitting—that tie in directly with the incompleteness of preferences and characterise two weight rules with particularly simple weight functions. Interestingly, the rules thus characterised reduce to well-known rules for the special case of complete input profiles and suggest the definition of new ones in the standard voting framework.