2010•Computational Mathematics and Mathematical PhysicsRequires access

Structure of quasi-Newton minimization methods

А. М. Ветошкин

Open publisher page 0 citations

Abstract

A scheme for the design of quasi-Newton methods for unconstrained optimization problems is examined. A criterion for the positive definiteness of the quasi-Newton modification of the Hessian matrix is given. Quasi-Newton methods are described that cannot be placed within the classical scheme specified by the family of Broyden methods.

About this research paper

What this paper is about

A scheme for the design of quasi-Newton methods for unconstrained optimization problems is examined. A criterion for the positive definiteness of the quasi-Newton modification of the Hessian matrix is given. Quasi-Newton methods are described that cannot be placed within the classical scheme specified by the family of Broyden methods.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

A scheme for the design of quasi-Newton methods for unconstrained optimization problems is examined. A criterion for the positive definiteness of the quasi-Newton modification of the Hessian matrix is given. Quasi-Newton methods are described that cannot be placed within the classical scheme specified by the family of Broyden methods.

Key concepts: Hessian matrix, Quasi-Newton method, Newton's method, Mathematics, Positive definiteness, Minification, Scheme (mathematics), Applied mathematics

Related papers

Back to paper searchBrowse research topicsOriginal source
Structure of quasi-Newton minimization methods — Research Paper | ScholarLens