Computing Equilibrium Strategies in Infinite Games of Incomplete Information
This paper describes an algorithm for computing best-response strategies in a class of infinite games of incomplete information (of which various auctions are prime examples). Under certain conditions, this algorithm can be used to automatically compute Bayes-Nash equilibria. The predominant game-solving algorithms generally address finite games and approximate infinite games by discretizing the types and actions. It demonstrates that the class of games the method addresses cannot be approximated in this way with the current state-of-the-art game solver.