Existence and Approximation of Continuous Bayesian Nash Equilibria in Games with Continuous Type and Action Spaces

Shaoyan Guo,Huifu Xu,Liwei Zhang
DOI: https://doi.org/10.1137/19m1298032
IF: 2.763
2021-01-01
SIAM Journal on Optimization
Abstract:Meirowitz [Econom. Lett., 78 (2003), pp. 213-218] showed existence of the continuous Bayesian Nash equilibrium for Bayesian games with continuous type and action spaces under the condition that the best response strategies are equicontinuous. In this paper, we take a step forward by presenting some verifiable conditions for the required equicontinuity, namely, some growth conditions of the expected utility function of each player. Moreover, under some monotonicity conditions, we demonstrate uniqueness of a continuous Bayesian Nash equilibrium. We then move on to develop some computational approaches for finding an approximate continuous Bayesian Nash equilibrium. First, we restrict the response strategies to polynomial functions of certain degree over the type spaces, and consequently finding a polynomial Bayesian Nash equilibrium is reduced to solving a finite dimensional stochastic Nash equilibrium problem. Second, we apply the optimal quantization method to discretize the finite dimensional stochastic Nash equilibrium problem so that the discretized problem becomes an ordinary finite dimensional deterministic Nash equilibrium problem which can be readily solved by existing numerical methods in the literature. In the case when the discretization approach is carried out by sample average approximation, we show exponential rate of convergence with respect to sample size. Finally, we carry out numerical tests on the combined approaches through an illustration with rent-seeking games.
What problem does this paper attempt to address?