An optimized Monte Carlo procedure and its application in rough surfaces scattering

In this paper we propose an optimized Monte Carlo procedure with the variance reduction technique incorporated. The proposed method makes use of some auxiliary variables which embed the correlation information intrinsic in the iteration process, hence the procedure can be systemized. The use of intrinsic information and the capability to systemize the procedure draw the essential difference of the proposed method with the control variable method in the Monte Carlo literature. The expression for the variance is derived, and an appropriate optimization problem is solved. The proposed method is applied to the analysis of electromagnetic scattering from rough surfaces.