Solving a class of LP problems with a primal-dual logarithmic barrier method

Abstract

Applying a higher order primal-dual logarithmic barrier method for solving large real-life linear programming problems is addressed in this paper. The efficiency of an interior point algorithm on these problems is compared with the one of the state-of-the-art simplex code MINOS version 5.3. Based on such experience, a wide class of LP problems is identified for which logarithmic barrier approach seems advantageous over the simplex one. Additionally, some practical rules for model builders are derived that should allow them to create problems that can easily be solved with logarithmic barrier algorithms.