Exact regularization of linear programs
Michael P. Friedlander (mpfcs.ubc.ca)
Abstract: We show that linear programs (LPs) admit regularizations that either contract the original (primal) solution set or leave it unchanged. Any regularization function that is convex and has compact level sets is allowed--differentiability is not required. This is an extension of the result first described by Mangasarian and Meyer (SIAM J. Control Optim., 17(6), pp. 745-752, 1979). We show that there always exist positive values of the regularization parameter such that a solution of the regularized problem simultaneously minimizes the original LP and minimizes the regularization function over the original solution set. We illustrate the main result using the nondifferentiable L1 regularization function on a set of degenerate LPs. Numerical results demonstrate how such an approach yields sparse solutions from the application of an interior-point method.
Keywords: linear programming, regularization, degeneracy, sparse solutions
Category 1: Convex and Nonsmooth Optimization
Citation: Department of Computer Science Tech. Rep. TR-2005-31, December 2005, University of British Columbia.
Entry Submitted: 01/16/2006
Modify/Update this entry
|Visitors||Authors||More about us||Links|
Search, Browse the Repository
Give us feedback
|Optimization Journals, Sites, Societies|