Optimization of Polling Stations Locations for National Election

Jadis Quek Hui Lin

Abstract

The efficient allocation of polling stations is critical to ensuring accessibility, minimizing voter congestion, and optimizing resource distribution during elections. This project focuses on optimizing polling station locations in a district of Singapore by leveraging advanced mathematical modelling and optimization techniques. Using Gurobi as the solver, we implement three key models—Set Covering, p-Centre, and p-Median—to determine the most effective placement of polling stations.

The Set Covering Model ensures that every voter is within an acceptable distance of at least one polling station while minimizing the number of stations required. The p-Centre Model aims to minimize the maximum travel distance for voters, ensuring equitable access to polling stations. Meanwhile, the p-Median Model optimizes locations by minimizing the total travel distance across all voters, thereby improving overall efficiency.

This approach involves data collection on voter distribution, geographical constraints, and existing infrastructure. The models are formulated as integer programming problems and solved using Gurobi, which provides optimal or near-optimal solutions within reasonable computational time. Results from the analysis indicate trade-offs between the number of polling stations, voter accessibility, and operational efficiency, offering valuable insights for policymakers.

This study demonstrates how mathematical optimization can enhance electoral logistics, ensuring fair and efficient voting processes. The findings can be adapted for broader applications, such as emergency facility placement and public service optimization.