Sensitivity Analysis of Additive Multiattribute Value Models

We present two simple computational procedures for sensitivity analysis of additive multiattribute value models that yield variations in attribute weights (scaling constants). For the first, or entropy-based, procedure, the point of departure is the concept of equal weights for all attributes. For the second, or least squares procedure, the point of departure is a set of arbitrary weights for attributes. In each procedure we either (a) calculate the “closest” set of weights that equates the multiattribute value for a pair of alternatives—that alternative whose overall value is optimal for the original attribute weights and any other specific nondominated alternative—or (b) calculate the “closest” set of weights required for the specific alternative to exceed the “optimal” alternative by a specified amount.