An Optimal Method for Adjusting the Centering Parameter in the Wide-Neighborhood Primal-Dual Interior-Point Algorithm for Linear Programming

Abstract

In this paper we present a dynamic optimal method for adjusting the centering parameter in the wide-neighborhood primal-dual interior-point algorithms for linear programming, while the centering parameter is generally a constant in the classical wide-neighborhood primal-dual interior-point algorithms. The computational results show that the new method is more efficient.