Structure-Preserving and Rank-Revealing QR-Factorizations

From MaRDI portal





The authors present a framework for the efficient implementation of rank- revealing \(QR\)-factorization (RRQR-factorization) algorithms, in particular for sparse matrices. They propose to compute an initial \(QR\)- factorization using a restricted column-pivoting strategy guarded by an incremental condition estimation, and then to apply the algorithm suggested by \textit{T. F. Chan} [Linear Algebra Appl. 88/89, 67-81 (1987; Zbl 0624.65025)] and \textit{L. V. Foster} [ibid. 74, 47-71 (1986; Zbl 0589.65031)] to this \(QR\)-factorization. The column exchanges are restricted such that they do not produce extra fill-in. Some numerical experiments illustrate that the proposed initial \(QR\)- factorization produces a good approximate \(RRQR\)-factorization.




Cited in
(30)








This page was built for publication: Structure-Preserving and Rank-Revealing QR-Factorizations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3982109)