A Linguistic 2-tuple Best-Worst Method

Nowadays, multi-criteria decision making (MCDM) problems are usually defined under changing contexts in which the emergence of uncertain information is common Under these circumstances, linguistic information and computing with words (CW) processes have been successfully applied for modelling such uncertainty and computing the decision results. Particularly, the linguistic 2-tuple model presents important benefits both readability and precision points of view. In the MCDM resolution process, the experts' preferences elicitation task plays a key role. The Best-Worst method (BWM) was proposed to solve some behavioral errors in similar MCDM methods and, consequently, reduces the number of pairwise comparisons and inconsistency in such a task. In the classical BWM, the experts use a numerical scale to provide their assessments. Lately, several BWM extensions under uncertain contexts have been proposed, but they still present drawbacks regarding the readability of the results. Hence, this contribution aims to introduce an BWM extension under a CW approach based on the linguistic 2-tuple model to model uncertainty, accomplish accurate computations and obtain understandable results. Furthermore, a novel consistency ratio to measure the experts' preferences consistency is proposed. Finally, the proposal is applied to an illustrative MCDM problem.