The possible winner problem with uncertain weights

Abstract

The original possible winner problem is: Given an unweighted election with partial preferences and a distinguished candidate c, can the preferences be extended to total ones such that c wins? We introduce a novel variant of this problem in which not some of the voters' preferences are uncertain but some of their weights. Not much has been known previously about the weighted possible winner problem. We present a general framework to study this problem, both for integer and rational weights, with and without upper bounds on the total weight to be distributed, and with and without ranges to choose the weights from. We study the complexity of these problems for important voting systems such as scoring rules, Copeland, ranked pairs, plurality with runoff, and (simplified) Bucklin and fallback voting.