On the convergence of derivative-free methods for unconstrained optimization

Andy Conn; Katya Scheinberg; Philippe Toint

On the convergence of derivative-free methods for unconstrained optimization

Andy Conn, Katya Scheinberg, Philippe Toint

Research output: Contribution in Book/Catalog/Report/Conference proceeding › Chapter

Abstract

The purpose of this paper is to examine a broad class of derivative-free trust-region methods for unconstrained optimization inspired by the proposals of \citebb{Powe94b} and to derive a general framework in which reasonable global convergence results can be obtained. The developments make extensive use of an interpolation error bound derived by Sauer and Xu (1995) in the context of multivariate polynomial interpolation.

Original language	English
Title of host publication	Approximation Theory and Optimization: Tributes to M. J. D. Powell
Editors	A Iserles, M Buhmann
Place of Publication	Cambridge, UK
Publisher	Cambridge University Press
Pages	83-108
Number of pages	26
Publication status	Published - 1997

1 Article

BFO, a trainable derivative-free Brute Force Optimizer for nonlinear bound-constrained optimization and equilibrium computations with continuous and discrete variables
Porcelli, M. & Toint, P., 30 Jun 2017, In: Transactions of the American Methematical Society on Mathematical Software. 44, 1, 28 p., 6.
Research output: Contribution to journal › Article › peer-review

File
47 Downloads (Pure)

Cite this

@inbook{6fa543b8ccdc46ce823b8363a38d2a73,

title = "On the convergence of derivative-free methods for unconstrained optimization",

abstract = "The purpose of this paper is to examine a broad class of derivative-free trust-region methods for unconstrained optimization inspired by the proposals of \citebb{Powe94b} and to derive a general framework in which reasonable global convergence results can be obtained. The developments make extensive use of an interpolation error bound derived by Sauer and Xu (1995) in the context of multivariate polynomial interpolation.",

author = "Andy Conn and Katya Scheinberg and Philippe Toint",

note = "Publication editors : A. Iserles and M. Buhmann",

year = "1997",

language = "English",

pages = "83--108",

editor = "A Iserles and M Buhmann",

booktitle = "Approximation Theory and Optimization: Tributes to M. J. D. Powell",

publisher = "Cambridge University Press",

}

On the convergence of derivative-free methods for unconstrained optimization. / Conn, Andy; Scheinberg, Katya; Toint, Philippe.
Approximation Theory and Optimization: Tributes to M. J. D. Powell. ed. / A Iserles; M Buhmann. Cambridge, UK: Cambridge University Press, 1997. p. 83-108.

Research output: Contribution in Book/Catalog/Report/Conference proceeding › Chapter

TY - CHAP

T1 - On the convergence of derivative-free methods for unconstrained optimization

AU - Conn, Andy

AU - Scheinberg, Katya

AU - Toint, Philippe

N1 - Publication editors : A. Iserles and M. Buhmann

PY - 1997

Y1 - 1997

N2 - The purpose of this paper is to examine a broad class of derivative-free trust-region methods for unconstrained optimization inspired by the proposals of \citebb{Powe94b} and to derive a general framework in which reasonable global convergence results can be obtained. The developments make extensive use of an interpolation error bound derived by Sauer and Xu (1995) in the context of multivariate polynomial interpolation.

AB - The purpose of this paper is to examine a broad class of derivative-free trust-region methods for unconstrained optimization inspired by the proposals of \citebb{Powe94b} and to derive a general framework in which reasonable global convergence results can be obtained. The developments make extensive use of an interpolation error bound derived by Sauer and Xu (1995) in the context of multivariate polynomial interpolation.

M3 - Chapter

SP - 83

EP - 108

BT - Approximation Theory and Optimization: Tributes to M. J. D. Powell

A2 - Iserles, A

A2 - Buhmann, M

PB - Cambridge University Press

CY - Cambridge, UK

ER -

On the convergence of derivative-free methods for unconstrained optimization

Abstract

Fingerprint

Research output

BFO, a trainable derivative-free Brute Force Optimizer for nonlinear bound-constrained optimization and equilibrium computations with continuous and discrete variables

Cite this