An inexact Newton path-following interior-point algorithm for nonlinear programming
Abstract
We present a path-following algorithm for solving large scale and/or degenerate nonlinear programs. We use interior-point methodology, a linesearch strategy within a trust region globalization strategy, and the conjugate gradient method with Steihaug's termination tests. The algorithm's path-following strategy makes use of the quasicentral path, the modified augmented Lagrangian, and weighted proximity measures. Some promising numerical results are reported.
Subject Area
Mathematics
Recommended Citation
Hernandez, Jaime, "An inexact Newton path-following interior-point algorithm for nonlinear programming" (2004). ETD Collection for University of Texas, El Paso. AAIEP10559.
https://scholarworks.utep.edu/dissertations/AAIEP10559
COinS