References & Citations
Mathematics > Numerical Analysis
Title: A Novel Randomized XR-Based Preconditioned CholeskyQR Algorithm
(Submitted on 22 Nov 2021 (v1), last revised 24 Nov 2021 (this version, v2))
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.
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.