Improved Forward Backward Method with Multiple Correction Vectors for Layered Random Rough Surfaces of Exponential Correlation Functions

摘要

The forward-backward method (FBM) is well-known for its robust convergence when applied to solve one rough interface problems. Challenges arise with the presence of multiple rough surfaces, resulting in a poor performance of the FBM. The improved forward-backward method with multiple correction vectors is proposed to enhance the convergence rate of the FBM. This is achieved by applying a residual minimisation step at the end of every iteration. In this paper, the application of the proposed method to exponential correlation surfaces is investigated. Numerical results are presented to demonstrate the accuracy and the efficiency of the proposed method as compared to the FBM.