Wavelet BEMのためのマルチグリッド反復法

抄録

Multigrid algorithm is developed for wavelet BEM. Since the wavelet bases have hierarchical structure, in the wavelet BEM using the wavelet bases as its functional bases, coefficient matrix and vectors include submatrix and subvector corresponding to each resolution level. Hence, one can readily apply the multigrid method to wavelet BEM. In this paper, V-, Sawtooth-, and FMV-cycles are considered. Numerical experiments show that these methods improve the convergence of iterative solution. As the iterative methods, Jacobi, Bi-CG, and GMRES methods are employed. It is found that the multigrid method can be an effective strategy for enhancement of convergence and stability of these iterative methods.