Randomized methods for rank-deficient linear systems

Josef Sifuentes, Zydrunas Gimbutas, Leslie Greengard

Research output: Contribution to journalArticle

Abstract

We present a simple, accurate method for solving consistent, rank-deficient linear systems, with or without additional rank-completing constraints. Such problems arise in a variety of applications such as the computation of the eigenvectors of a matrix corresponding to a known eigenvalue. The method is based on elementary linear algebra combined with the observation that if the matrix is rank-k deficient, then a random rank-k perturbation yields a nonsingular matrix with probability close to 1.

Original languageEnglish (US)
Pages (from-to)177-188
Number of pages12
JournalElectronic Transactions on Numerical Analysis
Volume44
StatePublished - 2015

Fingerprint

Linear Systems
Nonsingular or invertible matrix
Linear algebra
Eigenvector
Eigenvalue
Perturbation

Keywords

  • Eigenvectors
  • Integral equations
  • Null space
  • Null vectors
  • Randomized algorithms
  • Rank-deficient systems

ASJC Scopus subject areas

  • Analysis

Cite this

Randomized methods for rank-deficient linear systems. / Sifuentes, Josef; Gimbutas, Zydrunas; Greengard, Leslie.

In: Electronic Transactions on Numerical Analysis, Vol. 44, 2015, p. 177-188.

Research output: Contribution to journalArticle

@article{ff2d6099786b4f88b96f460e34a86b22,
title = "Randomized methods for rank-deficient linear systems",
abstract = "We present a simple, accurate method for solving consistent, rank-deficient linear systems, with or without additional rank-completing constraints. Such problems arise in a variety of applications such as the computation of the eigenvectors of a matrix corresponding to a known eigenvalue. The method is based on elementary linear algebra combined with the observation that if the matrix is rank-k deficient, then a random rank-k perturbation yields a nonsingular matrix with probability close to 1.",
keywords = "Eigenvectors, Integral equations, Null space, Null vectors, Randomized algorithms, Rank-deficient systems",
author = "Josef Sifuentes and Zydrunas Gimbutas and Leslie Greengard",
year = "2015",
language = "English (US)",
volume = "44",
pages = "177--188",
journal = "Electronic Transactions on Numerical Analysis",
issn = "1068-9613",
publisher = "Kent State University",

}

TY - JOUR

T1 - Randomized methods for rank-deficient linear systems

AU - Sifuentes, Josef

AU - Gimbutas, Zydrunas

AU - Greengard, Leslie

PY - 2015

Y1 - 2015

N2 - We present a simple, accurate method for solving consistent, rank-deficient linear systems, with or without additional rank-completing constraints. Such problems arise in a variety of applications such as the computation of the eigenvectors of a matrix corresponding to a known eigenvalue. The method is based on elementary linear algebra combined with the observation that if the matrix is rank-k deficient, then a random rank-k perturbation yields a nonsingular matrix with probability close to 1.

AB - We present a simple, accurate method for solving consistent, rank-deficient linear systems, with or without additional rank-completing constraints. Such problems arise in a variety of applications such as the computation of the eigenvectors of a matrix corresponding to a known eigenvalue. The method is based on elementary linear algebra combined with the observation that if the matrix is rank-k deficient, then a random rank-k perturbation yields a nonsingular matrix with probability close to 1.

KW - Eigenvectors

KW - Integral equations

KW - Null space

KW - Null vectors

KW - Randomized algorithms

KW - Rank-deficient systems

UR - http://www.scopus.com/inward/record.url?scp=84937406396&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84937406396&partnerID=8YFLogxK

M3 - Article

VL - 44

SP - 177

EP - 188

JO - Electronic Transactions on Numerical Analysis

JF - Electronic Transactions on Numerical Analysis

SN - 1068-9613

ER -