Extra-updates criterion for the limited memory BFGS algorithm for large scale nonlinear optimization

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

This paper studies recent modifications of the limited memory BFGS (L-BFGS) method for solving large scale unconstrained optimization problems. Each modification technique attempts to improve the quality of the L-BFGS Hessian by employing (extra) updates in a certain sense. Because at some iterations these updates might be redundant or worsen the quality of this Hessian, this paper proposes an updates criterion to measure this quality. Hence, extra updates are employed only to improve the poor approximation of the L-BFGS Hessian. The presented numerical results illustrate the usefulness of this criterion and show that extra updates improve the performance of the L-BFGS method substantially.

Original languageEnglish
Pages (from-to)557-572
Number of pages16
JournalJournal of Complexity
Volume18
Issue number2
DOIs
Publication statusPublished - 2002

Fingerprint

Large-scale Optimization
Nonlinear Optimization
Update
Data storage equipment
Limited Memory Method
BFGS Method
Unconstrained Optimization
Optimization Problem
Iteration
Numerical Results
Approximation

Keywords

  • Large scale optimization
  • Limited memory BFGS method
  • Quasi-Newton methods

ASJC Scopus subject areas

  • Analysis
  • Computational Mathematics

Cite this

Extra-updates criterion for the limited memory BFGS algorithm for large scale nonlinear optimization. / Al-Baali, M.

In: Journal of Complexity, Vol. 18, No. 2, 2002, p. 557-572.

Research output: Contribution to journalArticle

@article{537a7c0f35a24c2598e95501d1186e3a,
title = "Extra-updates criterion for the limited memory BFGS algorithm for large scale nonlinear optimization",
abstract = "This paper studies recent modifications of the limited memory BFGS (L-BFGS) method for solving large scale unconstrained optimization problems. Each modification technique attempts to improve the quality of the L-BFGS Hessian by employing (extra) updates in a certain sense. Because at some iterations these updates might be redundant or worsen the quality of this Hessian, this paper proposes an updates criterion to measure this quality. Hence, extra updates are employed only to improve the poor approximation of the L-BFGS Hessian. The presented numerical results illustrate the usefulness of this criterion and show that extra updates improve the performance of the L-BFGS method substantially.",
keywords = "Large scale optimization, Limited memory BFGS method, Quasi-Newton methods",
author = "M. Al-Baali",
year = "2002",
doi = "10.1006/jcom.2001.0623",
language = "English",
volume = "18",
pages = "557--572",
journal = "Journal of Complexity",
issn = "0885-064X",
publisher = "Academic Press Inc.",
number = "2",

}

TY - JOUR

T1 - Extra-updates criterion for the limited memory BFGS algorithm for large scale nonlinear optimization

AU - Al-Baali, M.

PY - 2002

Y1 - 2002

N2 - This paper studies recent modifications of the limited memory BFGS (L-BFGS) method for solving large scale unconstrained optimization problems. Each modification technique attempts to improve the quality of the L-BFGS Hessian by employing (extra) updates in a certain sense. Because at some iterations these updates might be redundant or worsen the quality of this Hessian, this paper proposes an updates criterion to measure this quality. Hence, extra updates are employed only to improve the poor approximation of the L-BFGS Hessian. The presented numerical results illustrate the usefulness of this criterion and show that extra updates improve the performance of the L-BFGS method substantially.

AB - This paper studies recent modifications of the limited memory BFGS (L-BFGS) method for solving large scale unconstrained optimization problems. Each modification technique attempts to improve the quality of the L-BFGS Hessian by employing (extra) updates in a certain sense. Because at some iterations these updates might be redundant or worsen the quality of this Hessian, this paper proposes an updates criterion to measure this quality. Hence, extra updates are employed only to improve the poor approximation of the L-BFGS Hessian. The presented numerical results illustrate the usefulness of this criterion and show that extra updates improve the performance of the L-BFGS method substantially.

KW - Large scale optimization

KW - Limited memory BFGS method

KW - Quasi-Newton methods

UR - http://www.scopus.com/inward/record.url?scp=0036292396&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036292396&partnerID=8YFLogxK

U2 - 10.1006/jcom.2001.0623

DO - 10.1006/jcom.2001.0623

M3 - Article

VL - 18

SP - 557

EP - 572

JO - Journal of Complexity

JF - Journal of Complexity

SN - 0885-064X

IS - 2

ER -