Novel types of sensitivity analysis for additive MCDM methods

Three novel types of sensitivity analysis are presented. They determine the following: 1) the sensitivity of a ranking to specific changes in the evaluations of all alternatives on certain criteria; 2) the influence of specific changes in certain criterion-scores of an alternative; 3) the minimum modification of the weights required to make an alternative ranked first. The use of these instruments is demonstrated by the results of a simulation experiment. Results show that the first and second type of sensitivity analysis enable to apply multiple criteria decision making (MCDM) methods in dynamic circumstances. The third type of sensitivity analysis is demonstrated to be a tool to analyze the total weight space. © 1995.