A Micro-Macro Acceleration Method for the Monte Carlo Simulation of Stochastic Differential Equations

摘要: We present and analyze a micro-macro acceleration method for the Monte Carlo simulation of stochastic differential equations with separation between (fast) time scale individual trajectories (slow) macroscopic function interest. The algorithm combines short bursts path simulations extrapolation number state variables forward in time. new microscopic state, consistent extrapolated variables, is obtained by matching operator that minimizes perturbation caused extrapolation. provide proof convergence this method, absence statistical error, we various strategies matching, as an on probability measures. Numerical experiments show illustrate effects different approximations resulting error predictions.