A wavelet particle approximation for McKean-Vlasov and 2D-Navier-Stokes statistical solutions (Q2469500)

The paper is interested in a probabilistic approach to the Mc-Kean-Vlasov equations with random initial conditions. When these conditions are parametrized by a real random variable \(\theta\), the weak measure solutions of the equations are linked to the laws of nonlinear stochastic differential equations for which the drift coefficients are expressed as conditional expectations in the laws of the diffusions given \(\theta\). The unknown conditional expectations are then replaced by wavelet regression estimators, and an estimation of the rate of convergence of this stochastic particle approximation is obtained. The results are then carried over to 2D-Navier-Stokes equations. Numerical simulations of the random velocity vector field are performed in some test cases and the results are compared with those of two other methods.