Comparative Study of Methods for Estimating Interval Priority Weights Focusing on the Accuracy in Selecting the Best Alternative

The interval analytic hierarchy process (interval AHP) has been proposed to analyze the MCDM (MCDM) problem considering the vagueness of the decision maker’s evaluation. In the interval AHP, interval priority weights are estimated instead of crisp ones. Various methods have been proposed for estimating the interval priority weights from the crisp pairwise comparison matrix given by the decision maker. Up till now, the efficiency of various estimation methods of interval priority weights has been analyzed by numerical experiments, focusing on the accuracy of ranking alternatives. The evaluations have been done using three representative solutions to the interval priority weight estimation problem, which frequently has non-unique solutions. However, in real world decision making problems, choosing the best from a set of alternatives is more frequent. In this study, we compare various methods for estimating interval priority weights from the viewpoint to what extent the best alternative is accurately estimated. Our analysis extends beyond counting correct selections of the best alternative to the distribution of accuracy scores, for revealing more detailed tendencies. The results highlight the advantages of interval priority weight estimation over crisp one, demonstrating its effectiveness in identifying the best alternative.