MULTIPLICATIVE INTERVAL VARIATION OF OBJECTIVE FUNCTION COEFFICIENTS IN LINEAR PROGRAMMING.

Traditional sensitivity analysis of linear programming objective function coefficients concerns itself with variations in single parameters. In a recent book, Gal developed a theoretical framework to determine the effect of mutliple variations in the parameters. In this paper, we develop a methodology to implement and extend the theoretical results in the earlier work for interval variations of objective function coefficients. This methodology uses necessary and sufficient conditions to determine all optimal solutions to the linear programming problem for objective function coefficients within given intervals. The resulting procedure has been coded for computer use and tested on various types of test problems. Computational results and an example are presented. [ABSTRACT FROM AUTHOR]