Damped Techniques for the Limited Memory BFGS Method for Large-Scale Optimization

Mehiddin Al-Baali, Lucio Grandinetti, Ornella Pisacane

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

This paper is aimed to extend a certain damped technique, suitable for the Broyden–Fletcher–Goldfarb–Shanno (BFGS) method, to the limited memory BFGS method in the case of the large-scale unconstrained optimization. It is shown that the proposed technique maintains the global convergence property on uniformly convex functions for the limited memory BFGS method. Some numerical results are described to illustrate the important role of the damped technique. Since this technique enforces safely the positive definiteness property of the BFGS update for any value of the steplength, we also consider only the first Wolfe–Powell condition on the steplength. Then, as for the backtracking framework, only one gradient evaluation is performed on each iteration. It is reported that the proposed damped methods work much better than the limited memory BFGS method in several cases.

Original languageEnglish
Pages (from-to)688-699
Number of pages12
JournalJournal of Optimization Theory and Applications
Volume161
Issue number2
DOIs
Publication statusPublished - May 1 2014

Fingerprint

Limited Memory Method
Large-scale Optimization
Damped
Data storage equipment
Positive Definiteness
Uniformly Convex
Backtracking
Unconstrained Optimization
Global Convergence
Convergence Properties
Convex function
Update
Gradient
Iteration
Numerical Results
Large-scale optimization
Evaluation

Keywords

  • Damped technique
  • Large-scale optimization
  • Line search framework
  • The limited memory BFGS method

ASJC Scopus subject areas

  • Applied Mathematics
  • Control and Optimization
  • Management Science and Operations Research

Cite this

Damped Techniques for the Limited Memory BFGS Method for Large-Scale Optimization. / Al-Baali, Mehiddin; Grandinetti, Lucio; Pisacane, Ornella.

In: Journal of Optimization Theory and Applications, Vol. 161, No. 2, 01.05.2014, p. 688-699.

Research output: Contribution to journalArticle

@article{8a8472a12b344ef082dc0e26e7d052b1,
title = "Damped Techniques for the Limited Memory BFGS Method for Large-Scale Optimization",
abstract = "This paper is aimed to extend a certain damped technique, suitable for the Broyden–Fletcher–Goldfarb–Shanno (BFGS) method, to the limited memory BFGS method in the case of the large-scale unconstrained optimization. It is shown that the proposed technique maintains the global convergence property on uniformly convex functions for the limited memory BFGS method. Some numerical results are described to illustrate the important role of the damped technique. Since this technique enforces safely the positive definiteness property of the BFGS update for any value of the steplength, we also consider only the first Wolfe–Powell condition on the steplength. Then, as for the backtracking framework, only one gradient evaluation is performed on each iteration. It is reported that the proposed damped methods work much better than the limited memory BFGS method in several cases.",
keywords = "Damped technique, Large-scale optimization, Line search framework, The limited memory BFGS method",
author = "Mehiddin Al-Baali and Lucio Grandinetti and Ornella Pisacane",
year = "2014",
month = "5",
day = "1",
doi = "10.1007/s10957-013-0448-8",
language = "English",
volume = "161",
pages = "688--699",
journal = "Journal of Optimization Theory and Applications",
issn = "0022-3239",
publisher = "Springer New York",
number = "2",

}

TY - JOUR

T1 - Damped Techniques for the Limited Memory BFGS Method for Large-Scale Optimization

AU - Al-Baali, Mehiddin

AU - Grandinetti, Lucio

AU - Pisacane, Ornella

PY - 2014/5/1

Y1 - 2014/5/1

N2 - This paper is aimed to extend a certain damped technique, suitable for the Broyden–Fletcher–Goldfarb–Shanno (BFGS) method, to the limited memory BFGS method in the case of the large-scale unconstrained optimization. It is shown that the proposed technique maintains the global convergence property on uniformly convex functions for the limited memory BFGS method. Some numerical results are described to illustrate the important role of the damped technique. Since this technique enforces safely the positive definiteness property of the BFGS update for any value of the steplength, we also consider only the first Wolfe–Powell condition on the steplength. Then, as for the backtracking framework, only one gradient evaluation is performed on each iteration. It is reported that the proposed damped methods work much better than the limited memory BFGS method in several cases.

AB - This paper is aimed to extend a certain damped technique, suitable for the Broyden–Fletcher–Goldfarb–Shanno (BFGS) method, to the limited memory BFGS method in the case of the large-scale unconstrained optimization. It is shown that the proposed technique maintains the global convergence property on uniformly convex functions for the limited memory BFGS method. Some numerical results are described to illustrate the important role of the damped technique. Since this technique enforces safely the positive definiteness property of the BFGS update for any value of the steplength, we also consider only the first Wolfe–Powell condition on the steplength. Then, as for the backtracking framework, only one gradient evaluation is performed on each iteration. It is reported that the proposed damped methods work much better than the limited memory BFGS method in several cases.

KW - Damped technique

KW - Large-scale optimization

KW - Line search framework

KW - The limited memory BFGS method

UR - http://www.scopus.com/inward/record.url?scp=84944653600&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84944653600&partnerID=8YFLogxK

U2 - 10.1007/s10957-013-0448-8

DO - 10.1007/s10957-013-0448-8

M3 - Article

VL - 161

SP - 688

EP - 699

JO - Journal of Optimization Theory and Applications

JF - Journal of Optimization Theory and Applications

SN - 0022-3239

IS - 2

ER -