Newton Algorithms for Large-Scale Strictly Convex Separable Network Optimization

Aleksandar Donev (donevpa.msu.edu)
Phillip Duxbury (duxburypa.msu.edu)

Abstract: In this work we summarize the basic elements of primal and dual Newton algorithms for network optimization with continuously differentiable (strictly) convex arc cost functions. Both the basic mathematics and implementation are discussed, and hints to important tuning details are made. The exposition assumes that the reader posseses a significant level of prior knowledge in the field. The algorithms have been drawn from a very large pool of literature spanning over 20 years of research in the area. Please visit http://computation.pa.msu.edu/NO

Keywords: Network optimization, Convex programming

Category 1: Network Optimization

Category 2: Convex and Nonsmooth Optimization (Convex Optimization )

Citation: Working internal report, Computational Materials Science Group (Prof. Phil Duxbuty), Physics and Astronomy Department, Michigan State University, January 2001

Download: [Compressed Postscript]

Entry Submitted: 01/19/2001
Entry Accepted: 01/28/2001
Entry Last Modified: 01/19/2001

Modify/Update this entry