A Mean Field Approach for Discounted Zero-Sum Games in a Class of Systems of Interacting Objects
Carmen G. Higuera-Chan,J. Adolfo Minjárez-Sosa
DOI: https://doi.org/10.1007/s13235-021-00377-0
2021-02-17
Dynamic Games and Applications
Abstract:The paper deals with systems composed of a large number of N interacting objects (e.g., agents, particles) controlled by two players defining a stochastic zero-sum game. The objects can be classified according to a finite set of classes or categories over which they move randomly. Because N is too large, the game problem is studied following a mean field approach. That is, a zero-sum game model (mathcal {GM}_{N}), where the states are the proportions of objects in each class, is introduced. Then, letting (N
ightarrow infty ) (the mean field limit) we obtain a new game model (mathcal {GM}), independent on N, which is easier to analyze than (mathcal {GM}_{N}). Considering a discounted optimality criterion, our objective is to prove that an optimal pair of strategies in (mathcal {GM}) is an approximate optimal pair as (N
ightarrow infty ) in the original game model (mathcal {GM}_{N}).
mathematics, interdisciplinary applications