On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems

We consider conjugate-gradient like methods for solving block symmetric indefinite linear systems that arise from saddle point problems or, in particular, regularizations thereof. Such methods require preconditioners that preserve certain sub-blocks from the original systems but allow considerable f...

Description complète

Détails bibliographiques
Auteurs principaux: Dollar, H, Gould, N, Schilders, W, Wathen, A
Format: Report
Publié: Unspecified 2005
_version_ 1826269816705515520
author Dollar, H
Gould, N
Schilders, W
Wathen, A
author_facet Dollar, H
Gould, N
Schilders, W
Wathen, A
author_sort Dollar, H
collection OXFORD
description We consider conjugate-gradient like methods for solving block symmetric indefinite linear systems that arise from saddle point problems or, in particular, regularizations thereof. Such methods require preconditioners that preserve certain sub-blocks from the original systems but allow considerable flexibility for the remaining blocks. We construct fourteen families of implicit factorizations that are capable of reproducing the required sub-blocks and (some) of the remainder. These generalize known implicit factorizations for the unregularized case. Improved eigenvalue clustering is possible if additionally some of the non-crucial blocks are reproduced. Numerical experiments confirm that these implicit-factorization preconditioners can be very effective in practice.
first_indexed 2024-03-06T21:31:04Z
format Report
id oxford-uuid:44b2c52a-0fa2-4a0d-b485-f4affd4d18f1
institution University of Oxford
last_indexed 2024-03-06T21:31:04Z
publishDate 2005
publisher Unspecified
record_format dspace
spelling oxford-uuid:44b2c52a-0fa2-4a0d-b485-f4affd4d18f12022-03-26T15:03:14ZOn iterative methods and implicit-factorization preconditioners for regularized saddle-point systemsReporthttp://purl.org/coar/resource_type/c_93fcuuid:44b2c52a-0fa2-4a0d-b485-f4affd4d18f1Mathematical Institute - ePrintsUnspecified2005Dollar, HGould, NSchilders, WWathen, AWe consider conjugate-gradient like methods for solving block symmetric indefinite linear systems that arise from saddle point problems or, in particular, regularizations thereof. Such methods require preconditioners that preserve certain sub-blocks from the original systems but allow considerable flexibility for the remaining blocks. We construct fourteen families of implicit factorizations that are capable of reproducing the required sub-blocks and (some) of the remainder. These generalize known implicit factorizations for the unregularized case. Improved eigenvalue clustering is possible if additionally some of the non-crucial blocks are reproduced. Numerical experiments confirm that these implicit-factorization preconditioners can be very effective in practice.
spellingShingle Dollar, H
Gould, N
Schilders, W
Wathen, A
On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
title On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
title_full On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
title_fullStr On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
title_full_unstemmed On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
title_short On iterative methods and implicit-factorization preconditioners for regularized saddle-point systems
title_sort on iterative methods and implicit factorization preconditioners for regularized saddle point systems
work_keys_str_mv AT dollarh oniterativemethodsandimplicitfactorizationpreconditionersforregularizedsaddlepointsystems
AT gouldn oniterativemethodsandimplicitfactorizationpreconditionersforregularizedsaddlepointsystems
AT schildersw oniterativemethodsandimplicitfactorizationpreconditionersforregularizedsaddlepointsystems
AT wathena oniterativemethodsandimplicitfactorizationpreconditionersforregularizedsaddlepointsystems