From numerical to heterogeneous linguistic best–worst method: Impacts of personalized individual semantics on consistency and consensus

This paper investigates the impacts of personalized individual semantics (PIS) on consistency and consensus in Best–Worst method (BWM) with heterogeneous linguistic preference information. Firstly, a PIS driven consistency measurement and information transformation method is presented to analyze the consistency level of linguistic preference information (i.e., best-to-others and others-to-worst vectors) of each decision maker, and this is conducted by maximizing the consistency level of additive preference relation that converted from linguistic preference information via personalized numerical scales. Secondly, a PIS driven maximum consensus optimization model within the BWM framework is designed to yield the maximum consensus level among the additive preference relations generated from heterogeneous linguistic preference information by means of personalized numerical scales. Thirdly, a PIS-based heterogeneous linguistic consensus reaching process is put forward to promote the consensus establishment among decision makers. Finally, the validity of the proposed BWM framework is verified by a case study, a sensitive analysis, and a comparison analysis.