Implicit Schemes and LU Decompositions (Q3316501)
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 3846786
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Implicit Schemes and LU Decompositions |
scientific article; zbMATH DE number 3846786 |
Statements
Implicit Schemes and LU Decompositions (English)
0 references
1981
0 references
implicit method
0 references
LU decomposition
0 references
well-conditioned
0 references
diagonally dominant
0 references
multidimensional backward Euler method
0 references
Newton-Raphson method
0 references
ADI method
0 references
0.7822672724723816
0 references
0.7773808836936951
0 references
0.773284375667572
0 references
0.7661519050598145
0 references
This paper analyzes implicit methods for hyperbolic equations by constructing LU factorizations. It is shown that the important requirement for a well-conditioned problem is that both L and U be diagonally dominant. It is shown that this is true for most standard schemes even though the original matrix is not diagonally dominant. Beginning with the L and U factors, new implicit methods are constructed. This methods are conditionally stable for all space dimensions. It is shown that the multidimensional backward Euler method is equivalent to the Newton-Raphson method for large time steps. In one-space dimension, the LU factorization is less efficient than standard methods. For two dimensions, the LU factorization is slightly more efficient than the ADI method. The main advantages of the LU decomposition are in three-space dimensions where this method requires only two sweeps through the grid and is unconditionally stable while ADI requires three sweeps and (in delta form) is unconditionally unstable.
0 references