Computing Pure-Strategy Nash Equilibria in a Two-Party Policy Competition: Existence and Algorithmic Approaches

A team of researchers from Academia Sinica and National Taiwan University has published groundbreaking work on modeling political competition using game theory. In their paper "Computing Pure-Strategy Nash Equilibria in a Two-Party Policy Competition," the authors formalize political competition as a two-player non-cooperative game where parties select policy vectors from compact Euclidean subsets. They introduce an affine isotonic function to model how a policy's winning probability increases with its total utility across voters, with each player's payoff defined as the expected utility received by its supporters.

The researchers first validated their isotonicity hypothesis through extensive voting simulations, then proved the existence of pure-strategy Nash equilibria in both one- and multi-dimensional policy spaces. Despite constructing a counterexample showing the game's non-monotonic nature, their experiments revealed that decentralized gradient-based algorithms typically converge rapidly to approximate equilibria. Most significantly, they developed a grid-based search algorithm that finds ε-approximate pure-strategy Nash equilibria in time polynomial in both input size and 1/ε, providing an efficient computational solution to what was previously considered a complex optimization problem.

This work extends Lin et al.'s 2021 research and represents a major advancement in computational political science. The mathematical framework allows for rigorous analysis of political strategy optimization, while the algorithmic approaches offer practical tools for understanding how parties might optimize their policy positions in competitive environments. The paper has been accepted as an extended abstract for AAMAS 2026, indicating its significance in the multi-agent systems community.