Abstract
An approximate solution method for solving the optimization problem which contains semi-fixed costs represented as a lower semi-continuous step function is developed. The fundamental idea of the algorithm is based on the simplex procedure of linear programming. We define the decrease in the objective function considering twice pivot calculations, and preparing two kinds of simplex tableau we propose the computational procedure to systematically obtain the approximate solution. Also some properties of the pivot calculations are theoretically analyzed. Finally some numerical examples are solved to illustrate the procedure and to test the effectiveness of the algorithm.