A probabilistic numerical method for a class of mean field games

Abstract

The Mean Field Games PDEs system is at the heart of the Mean Field Games theory initiated by [J.-M. Lasry and P.-L. Lions, Jeux à champ moyen. I–le cas stationnaire, C. R. Math. 343 (2006) 619–625; J.-M. Lasry and P.-L. Lions, Jeux à champ moyen. II–horizon fini et contrôle optimal, C. R. Math. 343 (2006) 679–684; J.-M. Lasry and P.-L. Lions, Mean field games, Jpn. J. Math. 2 (2007) 229–260] which constitutes a seminal contribution to the modeling and analysis of games with a large number of players. We propose here a numerical method of resolution of such systems based on the construction of a discrete mean field game where the controlled state-variable is a Markov chain approximating the controlled stochastic differential equation [H. Kushner and P. G. Dupuis, Numerical Methods for Stochastic Control Problems in Continuous Time, Stochastic Modeling and Applied Probability, Vol. 24 (Springer Science & Business Media, 2013)]. In particular, existence and uniqueness properties of the discrete MFG are investigated with convergence results under adequate assumptions.