Robust Voting Rules on the Interval Domain

Abstract: In this paper, we study voting rules on the interval domain, where the alternatives are arranged according to an externally given strict total order and voters report intervals of this order to indicate the alternatives they support. For this setting, we introduce and characterize the class of position-threshold rules, which compute a collective position of the voters with respect to every alternative and choose the left-most alternative whose collective position exceeds its threshold value. Our characterization of these rules mainly relies on reinforcement, a well-known population consistency condition, and robustness, a new axiom that restricts how the outcome is allowed to change when a voter removes the left-most or right-most alternative from his interval. Moreover, we characterize a generalization of the median rule to the interval domain, which selects the median of the endpoints of the voters' intervals.