Nonlinear Scaled Preferences in Linguistic Multi-criteria Group Decision Making

Abstract

Even though the use of data-driven decision models has increased its popularity during recent years, the resolution of some decision making problems still relies on the use of the expertise of specialists in the corresponding area, leading to decision situations characterized by the uncertainty and vagueness of the available information which may also require to model hesitancy between multiple choices. Thus, new approaches have been defined to model decision makers’ indecision by means of complex linguistic expressions such as Extended Comparative Linguistic Expressions with Symbolic Translation (ELICIT), based on the 2-tuple linguistic model. Nevertheless, this approach, alike many other linguistic models, uses linear scales to model decision makers’ preferences. Recent studies show that humans do not measure the distances between values at different levels of the linear scale in the same way and better decisions are obtained when nonlinear scales are considered. Therefore, this chapter introduces a multi-criteria group decision making model based on fuzzy TOPSIS dealing with ELICIT information which considers the nonlinear scales provided by the recently defined extreme values amplifications. It provides flexibility to express decision makers’ preferences and guarantees more reliable results than those obtained with the classical linear scaled preferences.