Information structures and values in zero-sum stochastic games

We consider a zero-sum stochastic game where two players have a common observation of a global state, and each player makes a private observation of its local state at every time step. This asymmetry of information among the players makes it difficult to the compute the equilibrium cost (called the value of the zero-sum game). To help us determine the value of such a game, we first consider a game with just the common observations and no local state. We argue that the value of this game with symmetric information can be computed using an information structure expansion followed by the methodology described in [1]. We then argue that the value of the asymmetric information game with global and local states is equal to the value of a virtual game with symmetric information. This allows us to use the results for the symmetric information game for computing the value of the game with global and local states.

[1]  Pravin Varaiya,et al.  Stochastic Systems: Estimation, Identification, and Adaptive Control , 1986 .

[2]  Jérôme Renault,et al.  The Value of Repeated Games with an Informed Controller , 2008, Math. Oper. Res..

[3]  Jeff S. Shamma,et al.  LP formulation of asymmetric zero-sum stochastic games , 2014, 53rd IEEE Conference on Decision and Control.

[4]  L. Shapley,et al.  Stochastic Games* , 1953, Proceedings of the National Academy of Sciences.

[5]  Yi Ouyang,et al.  Dynamic Games With Asymmetric Information: Common Information Based Perfect Bayesian Equilibria and Sequential Decomposition , 2015, IEEE Transactions on Automatic Control.

[6]  Jérôme Renault,et al.  The Value of Markov Chain Games with Lack of Information on One Side , 2006, Math. Oper. Res..

[7]  Fabien Gensbittel,et al.  The Value of Markov Chain Games with Incomplete Information on Both Sides , 2012, Math. Oper. Res..

[8]  Tamer Basar,et al.  Common Information Based Markov Perfect Equilibria for Stochastic Games With Asymmetric Information: Finite Games , 2014, IEEE Transactions on Automatic Control.

[9]  Jean-François Mertens,et al.  The value of two-person zero-sum repeated games with lack of information on both sides , 1971 .

[10]  Abhinav Sinha,et al.  A Systematic Process for Evaluating Structured Perfect Bayesian Equilibria in Dynamic Games With Asymmetric Information , 2016, IEEE Transactions on Automatic Control.

[11]  T. Başar On the saddle-point solution of a class of stochastic differential games , 1981, 1979 18th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.