We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.NA

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Numerical Analysis

Title: A Novel Randomized XR-Based Preconditioned CholeskyQR Algorithm

Abstract: CholeskyQR is a simple and fast QR decomposition via Cholesky decomposition, while it has been considered highly sensitive to the condition number. In this paper, we provide a randomized preconditioner framework for CholeskyQR algorithm. Under this framework, two methods (randomized LU-CholeskyQR and randomized QR-CholeskyQR) are proposed and discussed. We prove the proposed preconditioners can effectively reduce the condition number, which is also demonstrated by numerical tests. Abundant numerical tests indicate our methods are more stable and faster than all the existing algorithms and have good scalability.
Comments: 23 pages, 11 figures, 6 tables
Subjects: Numerical Analysis (math.NA)
MSC classes: 65F25, 65Y05
Cite as: arXiv:2111.11148 [math.NA]
  (or arXiv:2111.11148v2 [math.NA] for this version)

Submission history

From: Ting Lin [view email]
[v1] Mon, 22 Nov 2021 12:20:32 GMT (248kb)
[v2] Wed, 24 Nov 2021 15:43:24 GMT (248kb)

Link back to: arXiv, form interface, contact.