Pragmatics of Multiwinner Voting: Algorithms and Preference Data Analysis

ongoing

This proposal is in the area of computational social choice, an area on the intersection of computer science and economics. We study multiwinner elections, with a focus on a pragmatic approach. Our goal is to provide a principled framework for applying multiwinner voting in various settings that may appear in real-life (ranging from small-scale elections in various institutions, through participatory budgeting settings, to applications directly within computer science). In particular, we are interested in: (a) designing new, fast algorithms for computing the outcomes of multiwinner voting rules (results of such rules are often NP-hard to compute), also for new languages of specifying preferences that are needed in practical settings; (b) obtaining algorithmic and mathematical understanding of preference data; and (c) providing algorithms for analyzing elections and their results. We are interested both in theoretical studies (designing new algorithms, analyzing computational complexity of election-related problems, establishing axiomatic features of multiwinner voting rules, etc.) and in experimental evaluations (finding out running times of algorithms, establishing their approximation ratios, evaluating properties of preference data, etc.).