Multiobjective Optimization for Politically Fair Districting: A Scalable Multilevel Approach

Abstract

Gerrymandering has been a fundamental issue in American democracy for more than two centuries, with significant implications for electoral representation. Traditional optimization models for political districting primarily model nonpolitical fairness metrics such as the compactness of districts. In “Multiobjective Optimization for Politically Fair Districting: A Scalable Multilevel Approach,” Swamy, King, and Jacobson propose optimization models that explicitly incorporate political fairness objectives using political data from past elections. These objectives model fundamental fairness principles such as vote-seat proportionality (efficiency gap), partisan (a)symmetry, and competitiveness. They propose a solution strategy, called the multilevel algorithm, that solves large instances of the problem using a series of matching-based graph contractions. A case study on congressional districting in Wisconsin demonstrates that district plans balance the interests of the voters and the political parties.