1: \begin{abstract}

2: We address  the numerical approximation of Mean Field Games  with 

3: local couplings. For power-like Hamiltonians, we consider   both the 

4: stationary system introduced in \cite{LasryLions06i,LasryLions07} and 

5: also a similar system involving density constraints in order to model 

6: hard congestion effects \cite{filippo,Mészáros2015}. For finite 

7: difference discretizations of the Mean Field Game system as in 

8: \cite{AchdouCapuzzo10}, we follow a variational approach. We prove 

9: that the aforementioned schemes can be obtained as the optimality 

10: system of suitably defined optimization problems. In order to prove 

11: the existence of solutions of the scheme with a variational argument, 

12: the monotonicity of the coupling term is not used, which allow us to 

13: recover general existence results proved in \cite{AchdouCapuzzo10}. 

14: Next, assuming next that the coupling term is monotone, the 

15: variational problem is cast as a convex optimization problem for 

16: which we study and compare  several proximal type methods. These 

17: algorithms have several interesting features, such as global 

18: convergence and stability with respect to the viscosity parameter, 

19: which can eventually be zero. We assess the performance of the 

20: methods via numerical experiments. 

21: %, and have good performance for density constrained systems. 

22: \end{abstract}

23: