A modified centered climbing algorithm for linear programming

In this paper, we propose a modified centered climbing algorithm (MCCA) for linear programs, which improves the centered climbing algorithm (CCA) developed for linear programs recently. MCCA implements a specific climbing scheme where a violated constraint is probed by means of the centered vector used by CCA. Computational comparison is made with CCA and the simplex method. Numerical tests show that, on average CPU time, MCCA runs faster than both CCA and the simplex method in terms of tested problems. In addition, a simple initialization technique is introduced.