Embedded techniques for choosing the parameter in Tikhonov regularization

S. Gazzola, P. Novati, M. R. Russo

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

This paper introduces a new strategy for setting the regularization parameter when solving large-scale discrete ill-posed linear problems by means of the Arnoldi-Tikhonov method. This new rule is essentially based on the discrepancy principle, although no initial knowledge of the norm of the error that affects the right-hand side is assumed; an increasingly more accurate approximation of this quantity is recovered during the Arnoldi algorithm. Some theoretical estimates are derived in order to motivate our approach. Many numerical experiments performed on classical test problems as well as image deblurring problems are presented.

Original languageEnglish
Pages (from-to)796-812
Number of pages17
JournalNumerical Linear Algebra with Applications
Volume21
Issue number6
Early online date3 Apr 2014
DOIs
Publication statusPublished - 1 Dec 2014

Fingerprint

Arnoldi
Tikhonov Regularization
Image Deblurring
Discrepancy Principle
Regularization Parameter
Test Problems
Experiments
Numerical Experiment
Norm
Approximation
Estimate
Strategy
Knowledge

Keywords

  • Arnoldi algorithm
  • Discrepancy principle
  • Linear discrete ill-posed problems
  • Tikhonov regularization

ASJC Scopus subject areas

  • Algebra and Number Theory
  • Applied Mathematics

Cite this

Embedded techniques for choosing the parameter in Tikhonov regularization. / Gazzola, S.; Novati, P.; Russo, M. R.

In: Numerical Linear Algebra with Applications, Vol. 21, No. 6, 01.12.2014, p. 796-812.

Research output: Contribution to journalArticle

@article{86880737822349718539e3c72f51608b,
title = "Embedded techniques for choosing the parameter in Tikhonov regularization",
abstract = "This paper introduces a new strategy for setting the regularization parameter when solving large-scale discrete ill-posed linear problems by means of the Arnoldi-Tikhonov method. This new rule is essentially based on the discrepancy principle, although no initial knowledge of the norm of the error that affects the right-hand side is assumed; an increasingly more accurate approximation of this quantity is recovered during the Arnoldi algorithm. Some theoretical estimates are derived in order to motivate our approach. Many numerical experiments performed on classical test problems as well as image deblurring problems are presented.",
keywords = "Arnoldi algorithm, Discrepancy principle, Linear discrete ill-posed problems, Tikhonov regularization",
author = "S. Gazzola and P. Novati and Russo, {M. R.}",
year = "2014",
month = "12",
day = "1",
doi = "10.1002/nla.1934",
language = "English",
volume = "21",
pages = "796--812",
journal = "Numerical Linear Algebra with Applications",
issn = "1070-5325",
publisher = "John Wiley and Sons Inc.",
number = "6",

}

TY - JOUR

T1 - Embedded techniques for choosing the parameter in Tikhonov regularization

AU - Gazzola, S.

AU - Novati, P.

AU - Russo, M. R.

PY - 2014/12/1

Y1 - 2014/12/1

N2 - This paper introduces a new strategy for setting the regularization parameter when solving large-scale discrete ill-posed linear problems by means of the Arnoldi-Tikhonov method. This new rule is essentially based on the discrepancy principle, although no initial knowledge of the norm of the error that affects the right-hand side is assumed; an increasingly more accurate approximation of this quantity is recovered during the Arnoldi algorithm. Some theoretical estimates are derived in order to motivate our approach. Many numerical experiments performed on classical test problems as well as image deblurring problems are presented.

AB - This paper introduces a new strategy for setting the regularization parameter when solving large-scale discrete ill-posed linear problems by means of the Arnoldi-Tikhonov method. This new rule is essentially based on the discrepancy principle, although no initial knowledge of the norm of the error that affects the right-hand side is assumed; an increasingly more accurate approximation of this quantity is recovered during the Arnoldi algorithm. Some theoretical estimates are derived in order to motivate our approach. Many numerical experiments performed on classical test problems as well as image deblurring problems are presented.

KW - Arnoldi algorithm

KW - Discrepancy principle

KW - Linear discrete ill-posed problems

KW - Tikhonov regularization

UR - http://www.scopus.com/inward/record.url?scp=84920444135&partnerID=8YFLogxK

UR - http://dx.doi.org/10.1002/nla.1934

U2 - 10.1002/nla.1934

DO - 10.1002/nla.1934

M3 - Article

VL - 21

SP - 796

EP - 812

JO - Numerical Linear Algebra with Applications

JF - Numerical Linear Algebra with Applications

SN - 1070-5325

IS - 6

ER -