1 Sensitivity Analysis in a Generic Multi-Attribute

Abstract. This paper describes several possible sensitivity analyses associated with a generic multi-attribute Decision Support System, which is capable of considering all the steps in the Decision Analysis cycle and is aimed at aiding decision-makers in the choice of the most preferred alternative. The system evaluates the set of alternatives through an additive multi-attribute utility function that allows for imprecise assignments concerning utilities and weights and uncertainty about consequences in terms of intervals instead of single values. The first sensitivity analysis is related to the traditional approach that can be used to answer “what if” questions. A series of graphical displays are an excellent aid in the decision-making process, because they give useful and important insight into the final ranking. Then, other complementary procedures are considered that exploit the imprecise inputs, generating results that provide insights into the model recommendations. This leads to the computation of non-dominated and potentially optimal alternatives, a weight stability interval assessment and simulation techniques that utilize Monte Carlo simulation methods allowing simultaneous weight changes