Sensitivity Analysis is an essential part in both the construction and interpretation of the results of linear programming models. Two main approaches have been developed in the literature over the years. Wendell’s tolerance sensitivity approach and Wagner’s variance based sensitivity approach. This work introduces an approach based on the integral decomposition of a multivariate function, which allows us to nest Wendell’s and Wagner’s approaches. Our method presents several advantages, such as, the possibility of determining tolerance region abating computational cost. Several insights on the structural dependence of the optimal profit on parameters of a linear programming problem are also obtained.
Homepage of Emanuele BorgonovoShort Bio of Emanuele Borgonovo