A new steepest edge approximation for the simplex method for linear programming

The purpose of this paper is to present a new steepest edge (SE) approximation scheme for the simplex method. The major advantages are its simplicity of recurrences and implementation, low computational overhead (compared to both the exact SE method and the DEVEX approximation scheme), and surprisingly good performance. The paper contains a brief account of the exact SE algorithm, the new recurrences developed in the same framework and some discussion on the possible reasons for the method's apparent success. Finally, numerical experiments are presented to assess the practical value of the method. The results are very promising.