High-order sparse factorization methods for elliptic boundary value problems
We are interested in solving the sparse linear systems, Av = b, that arise from finite difference or finite element approximations to partial differential equations. May iterative methods require solving an easier approximate equation, Pv = b, on each iteration. This is often called preconditioning or operator splitting (1,2,4,6-9,12-14). The methods we consider factor A approximately into the product of an upper and lower triangular matrix P is identical to LU approx. = A. These methods are called incomplete LU factorization methods and their convergence rate depends on how well P approximates A. We describe some new algorithms to generate accurate LU decompositions based on the continuity of the solution v.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 7046018
- Report Number(s):
- LA-UR-84-1717; CONF-8406113-3; ON: DE84012582
- Resource Relation:
- Conference: 5. IMACS international symposium, Bethlehem, PA, USA, 19 Jun 1984; Other Information: Portions are illegible in microfiche products
- Country of Publication:
- United States
- Language:
- English
Similar Records
Performance analysis of high quality parallel preconditioners applied to 3D finite element structural analysis
A fast, memory efficient and robust sparse preconditioner based on a multifrontal approach with applications to finite‐element matrices