Special Session 107: 

Variational analysis of two convex optimization problems in duality applied to some Mean field game systems

Daniela Tonon
Paris Dauphine University
France
Co-Author(s):    
Abstract:
Mean field games (MFG) systems have been introduced to describe Nash equilibria in differential games with infinitely many players. In some simple cases, the model collapses into a system consisting on a backward Hamilton-Jacobi equation coupled with a forward Fokker-Plank equation. The starting point of the current study is that in some cases the MFG system can be understood as the optimality system of two convex optimization problems in duality. This leads to a variational analysis strategy to study the well-posedness of the PDE system. Following this methodology, we will discuss first the existence and uniqueness of weak solutions of some possibly degenerated Mean Field Games and then the existence of solutions of a modified problem prescribing the final distribution of the agents.