A distance based multiple criteria sorting method without class thresholds

Abstract

In this paper, a new distance-based sorting method (DISWOTH) that does not require specifying class thresholds between preference-ordered classes is developed for multiple criteria sorting problems. The aim of the DISWOTH method is to assign each alternative to a class or a set of possible adjacent classes considering distance to class centroids. Distance of alternatives to centroids are used as a criteria aggregation function. A mathematical model is formulated to determine weights of criteria. Assignment is performed according to the weighted distance of each alternative to class centroids. Performance of the DISWOTH method is compared with the Preference DISaggregation analysis (PDIS) and the UTilites Additives DIScriminantes (UTADIS) methods from literature.