A fuzzy information combination approach to multi-winner voting

Computational social choice (COMSOC) deals with the problem of how to aggregate individual preferences into a collective decision from an algorithmic perspective. As one of the most important themes in COMSOC, multi-winner voting is concerned with aggregating preferences from multiple voters over a set of candidates to return a <i>k</i>-element subset of candidates as the winning committee. The central challenge of this field lies in designing suitable and efficient multi-winner voting rules. To take this challenge, this paper attempts to apply information integration algorithms to the COMSOC domain by proposing multi-winner voting rules based on the Fagin algorithm which is widely used to improve the efficiency and precision of information aggregation. To identify the correctness of the proposed voting rule, we first show that it satisfies all Arrowian conditions except the Independence of Irrelevant Alternatives. Then we investigate its plausible properties and implications in the context of multi-winner voting and justify its appropriateness for multi-winner voting. This work is meant to be a first step in the application of techniques from information integration to multi-winner voting.