Matrix Sensitivity Analysis from an Interior Solution of a Linear Program

Abstract

This article considers the effect of changing matrix coefficients in a linear program after we have obtained an interior solution. Changes are restricted to where there remains an optimal solution to the perturbed problem (called “admissible”). Mills' minimax theorem provides one approach and has been used for similar sensitivity analysis from a basic optimum. Here we consider the effect on the optimal partition and how the analysis results relate to the classical approach that uses a basic solution.