The effect of diagonal scaling on projected gradient methods for bound constrained quadratic programming problems

Jesse L. Barlow; Gerardo Toraldo

doi:10.1080/10556789508805613

The effect of diagonal scaling on projected gradient methods for bound constrained quadratic programming problems

Jesse L. Barlow, Gerardo Toraldo

Computer Science and Engineering

Research output: Contribution to journal › Article › peer-review

6 Scopus citations

Abstract

We examine the effect of diagonal scaling on the projected gradient method of Moré and Toraldo [SIAM J. Optimization 1, (1991), pp. 93-114] when the constraints are bound constraints and the quadratic form is positive definite. It is shown that scalings causes the method to visit fewer faces and to minimize the function more quickly. Numerical results are given with the journal bearing problem and some dense matrix bound constrained problems.

Original language	English (US)
Pages (from-to)	235-245
Number of pages	11
Journal	Optimization Methods and Software
Volume	5
Issue number	3
DOIs	https://doi.org/10.1080/10556789508805613
State	Published - Jan 1 1995

All Science Journal Classification (ASJC) codes

Software
Control and Optimization
Applied Mathematics

Access to Document

10.1080/10556789508805613

Cite this

@article{b08bd63f9d9349409e343b0b1141e074,

title = "The effect of diagonal scaling on projected gradient methods for bound constrained quadratic programming problems",

abstract = "We examine the effect of diagonal scaling on the projected gradient method of Mor{\'e} and Toraldo [SIAM J. Optimization 1, (1991), pp. 93-114] when the constraints are bound constraints and the quadratic form is positive definite. It is shown that scalings causes the method to visit fewer faces and to minimize the function more quickly. Numerical results are given with the journal bearing problem and some dense matrix bound constrained problems.",

author = "Barlow, {Jesse L.} and Gerardo Toraldo",

note = "Funding Information: s he research of Jesse L. Barlow was supported by the National Science Foundation 9000526 and contract no. CCR-9201612.",

year = "1995",

month = jan,

day = "1",

doi = "10.1080/10556789508805613",

language = "English (US)",

volume = "5",

pages = "235--245",

journal = "Optimization Methods and Software",

issn = "1055-6788",

publisher = "Taylor and Francis Ltd.",

number = "3",

}

TY - JOUR

T1 - The effect of diagonal scaling on projected gradient methods for bound constrained quadratic programming problems

AU - Barlow, Jesse L.

AU - Toraldo, Gerardo

N1 - Funding Information: s he research of Jesse L. Barlow was supported by the National Science Foundation 9000526 and contract no. CCR-9201612.

PY - 1995/1/1

Y1 - 1995/1/1

N2 - We examine the effect of diagonal scaling on the projected gradient method of Moré and Toraldo [SIAM J. Optimization 1, (1991), pp. 93-114] when the constraints are bound constraints and the quadratic form is positive definite. It is shown that scalings causes the method to visit fewer faces and to minimize the function more quickly. Numerical results are given with the journal bearing problem and some dense matrix bound constrained problems.

AB - We examine the effect of diagonal scaling on the projected gradient method of Moré and Toraldo [SIAM J. Optimization 1, (1991), pp. 93-114] when the constraints are bound constraints and the quadratic form is positive definite. It is shown that scalings causes the method to visit fewer faces and to minimize the function more quickly. Numerical results are given with the journal bearing problem and some dense matrix bound constrained problems.

UR - http://www.scopus.com/inward/record.url?scp=0029425799&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029425799&partnerID=8YFLogxK

U2 - 10.1080/10556789508805613

DO - 10.1080/10556789508805613

M3 - Article

AN - SCOPUS:0029425799

SN - 1055-6788

VL - 5

SP - 235

EP - 245

JO - Optimization Methods and Software

JF - Optimization Methods and Software

IS - 3

ER -

The effect of diagonal scaling on projected gradient methods for bound constrained quadratic programming problems

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this