Quasi Newton Methods
No Thumbnail Available
Date
2018-06-07
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
In this work, we investigate quasi Newton methods for solving unconstrained optimization
problems. We consider two di erent quasi-Newton update formulas, namely, Broyden-
Fletcher-Goldfarb and shanno (BFGS) update and Davidon-Fletcher-powell (DFP) up-
date. Line search method is used to nd the step length at each iteration. The methods
are tested on seven benchmark probelems and comparisons are made among Newton's
method, quasi-Newton methods (BFGS and DFP updates) and steepest descent method.
Also comparisons are made between the quasi-Newton methods (BFGS and DFP up-
dates). Finally conclusions are drawn upon the obtained results.
Description
Keywords
Convex Functions, Preliminaries