CISE Submissions

Item: REP-2004-353

Item:REP-2004-353

Title:A sparse proximal implementation of the LP dual active set algorithm

T. A. Davis
W. W. Hager

Abstract:
We present an implementation of the LP Dual Active Set Algorithm (LPDASA) based on a quadratic proximal approximation, a strategy for dropping inactive equations from the constraints, and recently developed algorithms for updating a sparse Cholesky factorization after a small rank change. Although our main focus is linear programming, the first and second-order proximal techniques that we develop are applicable to general concave-convex Lagrangians and to linear equality and inequality constraints. We use Netlib LP test problems to compare our proximal implementation of LPDASA to Simplex and Barrier algorithms, as implemented in CPLEX.

Supporting File:here

Select Departmental Reports

To select a single entry given the item reference number, enter the reference number in the Item: box. Other select options will then be ignored.

The other select options (Category:, Select by Time:, and Select by Author: can be used in any combination. However, only one of the Select by Author: options can be used. You can either select entries from a specific CISE faculty member, or from any (non faculty) CISE user, or you can enter in part or all of the author's name <-- ' --> (which is case insensitive). But only the FIRST one will be used. For many entries, the username may not have been entered, and in this case the first two options will not work for those entries, but selecting by last name will.

The Display Level: option tells how much information to display for each entry ranging from the lease (Index) to the most (Full Listing).

Item:

Category:

Select by Time:

Select by Author:

CISE Faculty Member:
CISE Member:
Name:

Display Level:

Index
Short Listing
Full Listing

Departmental Report Categories

Technical Reports
	Coordinator: Alexander M Thompson Coordinator: Dan H. Eicher Coordinator: John L. Kramer Jr.

To submit items for this page, please log in on the CISE submit page.