The Finite Difference Based Fast Adaptive Composite Grid Method (Q4220439)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1226276
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | The Finite Difference Based Fast Adaptive Composite Grid Method |
scientific article; zbMATH DE number 1226276 |
Statements
23 November 1998
0 references
elliptic boundary value problems
0 references
finite difference discretization
0 references
fast adaptive composite grid method
0 references
convergence
0 references
iterative method
0 references
FAC method
0 references
numerical examples
0 references
The Finite Difference Based Fast Adaptive Composite Grid Method (English)
0 references
0.8973508477210999
0 references
0.8673183917999268
0 references
The fast adaptive composite grid method (the FAC method) which is studied in this paper is an iterative method for solving discrete problems resulting from finite difference discretizations on a composite grid. The approach here is to approximate the solution by means of several uniform grids with various meshsizes covering different parts of the domain. At least one grid should cover the entire domain. The meshsize of this global coarse grid is chosen in agreement with the smooth behaviour of the solution outside the high activity regions (i.e., the regions in which the solution varies much more rapidly than in the remaining part of the domain). In each iteration step of the FAC method a residual is computed on the composite grid, corrections are computed on the uniform subgrids and the composite grid approximation is corrected. In this way expensive computing, i.e., solving linear systems of equations, takes place on the uniform subgrids only. It is shown that the convergence behaviour of the FAC method depends mainly on the way in which the composite grid residuals are restricted on the interfaces between the uniform subgrids. An optimality criterion for the restriction operators is defined and it is shown that for a composite grid problem resulting from the uniform interface discretization (i.e., the grid points on the interfaces are treated as the grid points of the coarsest uniform grid adjacent to the interface), each restriction operator which equals the trivial injection on the interfaces satisfies this optimality criterion. It is also shown that for the composite problem resulting from non-uniform interface discretization one should not use trivial injection on the interface. The theoretical results are illustrated by two numerical examples.
0 references