Closed-form multigrid smoothing factors for lexicographic Gauss-Seidel (Q2902192)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: Closed-form multigrid smoothing factors for lexicographic Gauss-Seidel |
scientific article; zbMATH DE number 6067204
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Closed-form multigrid smoothing factors for lexicographic Gauss-Seidel |
scientific article; zbMATH DE number 6067204 |
Statements
Closed-form multigrid smoothing factors for lexicographic Gauss-Seidel (English)
0 references
17 August 2012
0 references
multigrid
0 references
smoothing factor
0 references
local Fourier analysis
0 references
elliptic partial differential equations
0 references
lexicographic Gauss-Seidel method
0 references
Poisson equation
0 references
Möbius transformations
0 references
relaxation
0 references
anisotropic diffusion equations
0 references
convection-diffusion equations
0 references
The authors study the computation of smoothing factors on uniform meshes in arbitrary dimensions. Applying the complex analysis results on the maximum modulus principle and properties of Möbius transformations they are able to derive closed-form expressions for the smoothing factors of the lexicographic pointwise and block Gauss-Seidel method. In the case of the pointwise method the effect of a relaxation parameter is analysed as well. The results presented in the paper are applicable to various model problems, such as the Poisson and anisotropic diffusion equations and the convection-diffusion equations in arbitraty space dimensions. It is shown that the lexicographic Gauss-Seidel smoothing is efficient for equations with strong asymmetry such as the convection-diffusion equations. For the constant coefficient convection-diffusion equation with equal mesh Reynolds numbers it is shown for the upwind discretizations that the lexicographic Gauss-Seidel smoothing has a smaller factor than the red-black Gauss-Seidel smoothing.
0 references