Further Study on (I+S_(max)) Preconditioning Gauss-Seidel Iterative Method
Linzhang Lu
Abstract
Linzhang Lu
Abstract
Hisashiki Kotakemori had proposed a preconditioner (I+S_(max)) for irreducibly diagonally dominant Z-matrix, which achieves better convergence rate than the classical Gauss-Seidel method and even better than Modified Gauss-Seidel method with preconditioner (I+S) under certain circumstances. We extend his convergence theorem to the case of H-matrix, and apply the preconditioner (I+S_(max)) to twice preconditioning for irreducible non-singular M-matrix, combining with another preconditioner (I+S). Numerical examples had been given to confirm that the convergence rate had been improved on considerably.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Hisashiki Kotakemori had proposed a preconditioner (I+S_(max)) for irreducibly diagonally dominant Z-matrix, which achieves better convergence rate than the classical Gauss-Seidel method and even better than Modified Gauss-Seidel method with preconditioner (I+S) under certain circumstances. We extend his convergence theorem to the case of H-matrix, and apply the preconditioner (I+S_(max)) to twice preconditioning for irreducible non-singular M-matrix, combining with another preconditioner (I+S). Numerical examples had been given to confirm that the convergence rate had been improved on considerably.
Key concepts: Preconditioner, Gauss–Seidel method, Mathematics, Rate of convergence, Convergence (economics), Applied mathematics, Matrix (chemical analysis), Iterative method