Solving Regularized Linear Programs Using Barrier Methods and KKT Systems

        We discuss methods for solving the key linear equations (KKT systems) within primal-dual barrier methods for linear programming. To allow sparse indefinite Cholesky-type factorizations of the KKT systems, we perturb the problem slightly. Peturbations improve the stability of the Cholesky factorizations, but affect the efficiency of the cross-over to simplex (to obtain a basic solution to the original problem). We explore these effects by running OSL on the larger Netlib examples.

By: Michael A. Saunders (Stanford Univ.) and John A. Tomlin

Published in: RJ10064 in 1996

This Research Report is not available electronically. Please request a copy from the contact listed below. IBM employees should contact ITIRC for a copy.

Questions about this service can be mailed to reports@us.ibm.com .