Distributed Methods for Computing Approximate Equilibria

TL;DR

This paper introduces a distributed algorithm for computing approximate Nash equilibria in bimatrix games, achieving improved bounds and efficiency in communication and query complexity compared to previous methods.

Contribution

It presents a novel distributed approach that independently solves LPs for each payoff matrix, enabling better bounds and efficiency in computing approximate Nash equilibria.

Findings

Computes a 0.6528-WSNE in polynomial time.

Uses poly-logarithmic communication for approximate equilibria.

Requires O(n log n) payoff queries for a 0.6528-WSNE.

Abstract

We present a new, distributed method to compute approximate Nash equilibria in bimatrix games. In contrast to previous approaches that analyze the two payoff matrices at the same time (for example, by solving a single LP that combines the two players payoffs), our algorithm first solves two independent LPs, each of which is derived from one of the two payoff matrices, and then compute approximate Nash equilibria using only limited communication between the players. Our method has several applications for improved bounds for efficient computations of approximate Nash equilibria in bimatrix games. First, it yields a best polynomial-time algorithm for computing \emph{approximate well-supported Nash equilibria (WSNE)}, which guarantees to find a 0.6528-WSNE in polynomial time. Furthermore, since our algorithm solves the two LPs separately, it can be used to improve upon the best known algorithms in the limited communication setting: the algorithm can be implemented to obtain a randomized expected-polynomial-time algorithm that uses poly-logarithmic communication and finds a 0.6528-WSNE. The algorithm can also be carried out to beat the best known bound in the query complexity setting, requiring $O(n \log n)$ payoff queries to compute a 0.6528-WSNE. Finally, our approach can also be adapted to provide the best known communication efficient algorithm for computing \emph{approximate Nash equilibria}: it uses poly-logarithmic communication to find a 0.382-approximate Nash equilibrium.