A Block Constant Approximate Inverse for Preconditioning Large Linear Systems

From MaRDI portal





The authors propose a new class of approximate inverses for preconditioning large linear systems issued from the discretization of elliptic boundary value problems. The method proposed here is situated at the intersection of multipole, multigrid and sparse approximate inverse methods, and consists in approximating the inverse of a matrix \(A\) by a block constant matrix. An approximate inverse for preconditioning an iterative method is determined, following the multipole strategy, i.e. approximating the off-diagonal blocks of the inverse by low-rank matrices using the Green function. The block constant preconditioner is compared to a sparse approximate inverse with the same sparsity pattern as \(A\), and to the incomplete Cholesky factorization. A lot of numerical examples is performed which illustrate the proposed methods











This page was built for publication: A Block Constant Approximate Inverse for Preconditioning Large Linear Systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4443776)