Numerical experience with a recursive trust-region method for multilevel nonlinear bound-constrained optimization

Serge Gratton, Mélodie Mouffe, Annick Sartenaer, Philippe Toint, Dimitri Tomanos

Résultats de recherche: Contribution à un journal/une revueArticle

5 Downloads (Pure)

Résumé

We consider an implementation of the recursive multilevel trust-region algorithm proposed by Gratton et al. (A recursive trust-region method in infinity norm for bound-constrained nonlinear optimization, IMA J. Numer. Anal. 28(4) (2008), pp. 827-861) for bound-constrained nonlinear problems, and provide numerical experience on multilevel test problems. A suitable choice of the algorithm's parameters is identified on these problems, yielding a satisfactory compromise between reliability and efficiency. The resulting default algorithm is then compared with alternative optimization techniques such as mesh refinement and direct solution of the fine-level problem. It is also shown that its behaviour is similar to that of multigrid algorithms for linear systems.
langue originaleAnglais
Pages (de - à)359-386
Nombre de pages28
journalOptimization Methods and Software
Volume25
Numéro de publication3
Les DOIs
étatPublié - 1 juin 2010

Empreinte digitale

Trust Region Method
Recursive Method
Constrained optimization
Constrained Optimization
Trust Region Algorithm
Mesh Refinement
Nonlinear Optimization
Optimization Techniques
Test Problems
Nonlinear Problem
Linear Systems
Infinity
Linear systems
Norm
Alternatives
Experience

Citer ceci

@article{161f1abd902345f5a8833b79e945b363,
title = "Numerical experience with a recursive trust-region method for multilevel nonlinear bound-constrained optimization",
abstract = "We consider an implementation of the recursive multilevel trust-region algorithm proposed by Gratton et al. (A recursive trust-region method in infinity norm for bound-constrained nonlinear optimization, IMA J. Numer. Anal. 28(4) (2008), pp. 827-861) for bound-constrained nonlinear problems, and provide numerical experience on multilevel test problems. A suitable choice of the algorithm's parameters is identified on these problems, yielding a satisfactory compromise between reliability and efficiency. The resulting default algorithm is then compared with alternative optimization techniques such as mesh refinement and direct solution of the fine-level problem. It is also shown that its behaviour is similar to that of multigrid algorithms for linear systems.",
author = "Serge Gratton and M{\'e}lodie Mouffe and Annick Sartenaer and Philippe Toint and Dimitri Tomanos",
note = "Publication code : FP SB010/2008/10 ; QA 0002.2/001/08/10",
year = "2010",
month = "6",
day = "1",
doi = "10.1080/10556780903239295",
language = "English",
volume = "25",
pages = "359--386",
journal = "Optimization Methods and Software",
issn = "1055-6788",
publisher = "Taylor & Francis",
number = "3",

}

Numerical experience with a recursive trust-region method for multilevel nonlinear bound-constrained optimization. / Gratton, Serge; Mouffe, Mélodie; Sartenaer, Annick; Toint, Philippe; Tomanos, Dimitri.

Dans: Optimization Methods and Software, Vol 25, Numéro 3, 01.06.2010, p. 359-386.

Résultats de recherche: Contribution à un journal/une revueArticle

TY - JOUR

T1 - Numerical experience with a recursive trust-region method for multilevel nonlinear bound-constrained optimization

AU - Gratton, Serge

AU - Mouffe, Mélodie

AU - Sartenaer, Annick

AU - Toint, Philippe

AU - Tomanos, Dimitri

N1 - Publication code : FP SB010/2008/10 ; QA 0002.2/001/08/10

PY - 2010/6/1

Y1 - 2010/6/1

N2 - We consider an implementation of the recursive multilevel trust-region algorithm proposed by Gratton et al. (A recursive trust-region method in infinity norm for bound-constrained nonlinear optimization, IMA J. Numer. Anal. 28(4) (2008), pp. 827-861) for bound-constrained nonlinear problems, and provide numerical experience on multilevel test problems. A suitable choice of the algorithm's parameters is identified on these problems, yielding a satisfactory compromise between reliability and efficiency. The resulting default algorithm is then compared with alternative optimization techniques such as mesh refinement and direct solution of the fine-level problem. It is also shown that its behaviour is similar to that of multigrid algorithms for linear systems.

AB - We consider an implementation of the recursive multilevel trust-region algorithm proposed by Gratton et al. (A recursive trust-region method in infinity norm for bound-constrained nonlinear optimization, IMA J. Numer. Anal. 28(4) (2008), pp. 827-861) for bound-constrained nonlinear problems, and provide numerical experience on multilevel test problems. A suitable choice of the algorithm's parameters is identified on these problems, yielding a satisfactory compromise between reliability and efficiency. The resulting default algorithm is then compared with alternative optimization techniques such as mesh refinement and direct solution of the fine-level problem. It is also shown that its behaviour is similar to that of multigrid algorithms for linear systems.

UR - http://www.scopus.com/inward/record.url?scp=77949560755&partnerID=8YFLogxK

U2 - 10.1080/10556780903239295

DO - 10.1080/10556780903239295

M3 - Article

VL - 25

SP - 359

EP - 386

JO - Optimization Methods and Software

JF - Optimization Methods and Software

SN - 1055-6788

IS - 3

ER -