Preconditioning of LSQR and CGLS: Variants, Properties and Relations
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3A10511074" target="_blank" >RIV/00216208:11320/25:10511074 - isvavai.cz</a>
Result on the web
<a href="https://doi.org/10.1007/978-3-031-86173-4_42" target="_blank" >https://doi.org/10.1007/978-3-031-86173-4_42</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1007/978-3-031-86173-4_42" target="_blank" >10.1007/978-3-031-86173-4_42</a>
Alternative languages
Result language
angličtina
Original language name
Preconditioning of LSQR and CGLS: Variants, Properties and Relations
Original language description
LSQR and CGLS represent well known Krylov subspace methods for the solution of linear approximation problems including ill-posed ones. In order to accelerate convergence or impose solution constraints, various preconditioning strategies have been proposed. Basic LSQR and CGLS are mathematically equivalent, but relations among their preconditioned variants have been studied only partially. In this paper, we do not restrict to particular applications, but we algebraically describe underlying subspaces, optimality conditions, and relate quantities generated by selected preconditioned variants. Furthermore, motivated by available right preconditioned CGLS avoiding transformation of coordinates, we give a transformation free right preconditioned LSQR algorithm.
Czech name
—
Czech description
—
Classification
Type
D - Article in proceedings
CEP classification
—
OECD FORD branch
10102 - Applied mathematics
Result continuities
Project
—
Continuities
S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace
Others
Publication year
2025
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Article name in the collection
Lecture Notes in Computational Science and Engineering
ISBN
978-3-031-86172-7
ISSN
1439-7358
e-ISSN
2197-7100
Number of pages
10
Pages from-to
415-424
Publisher name
Springer Nature
Place of publication
Switzerland
Event location
Lisbon, portugal
Event date
Sep 4, 2023
Type of event by nationality
WRD - Celosvětová akce
UT code for WoS article
—