Exact Algorithms for Solving Stochastic Games

TL;DR

This paper presents algorithms capable of exactly solving classical two-player zero-sum stochastic games, including discounted, recursive, and undiscounted models, which are fundamental in game theory.

Contribution

The paper introduces novel algorithms specifically designed for the exact solution of various classical stochastic game models, advancing computational methods in game theory.

Findings

Algorithms successfully solve classical stochastic games

Exact solutions are obtained for discounted, recursive, and undiscounted models

The methods improve computational efficiency over previous approaches

Abstract

Shapley's discounted stochastic games, Everett's recursive games and Gillette's undiscounted stochastic games are classical models of game theory describing two-player zero-sum games of potentially infinite duration. We describe algorithms for exactly solving these games.