2014•Unpublished venueRequires access

Different Second Order Approximationsin a model-based SQP Trust-Region DFO Method

Anke Tröltzsch

Abstract

A trust-region SQP method for general nonlinear constrained optimization without derivatives is proposed. The trust-region step is computed by a Byrd-Omojokun-like approach. An active-set strategy is used to handle bound constraints and inequality constraints. The objective and constraint functions are approximated by local linear or quadratic interpolation. In this work, we want to examine different approximation techniques.

About this research paper

What this paper is about

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

Key concepts: Trust region, Sequential quadratic programming, Mathematical optimization, Interpolation (computer graphics), Constraint (computer-aided design), Mathematics, Nonlinear programming, Set (abstract data type)