Semantic differences and psychological behavior in multi-criteria group decision-making: Do they need consideration?

In Multi-Criteria Group Decision-Making (MCGDM), linguistic evaluation is widely used due to its flexibility. In particular, Personalized Individual Semantics (PIS), namely decision-makers hold heterogeneous understandings of the same word, has been extensively studied to elicit numeric meaning to the linguistic scale. However, few studies consider the need to tailor linguistics to the different indicators used, as well as the semantic difference in the same level linguistic that experts use when situations change. The traditional PIS models overlook the psychological aspect of the decision makers. To solve these problems, this paper proposes a method that considers semantic differences arising from the decision-makers and the indicators and incorporates the experts' psychological behavior. We apply this method to assess the alternatives for a smart refrigerator product service system. The results suggest that a large semantic difference, when evaluating the linguistic indicators, can influence the final outcome. We validate the flexibility and the effectiveness of the method by comparing against the base case of no semantic difference, using the Euclidean and Manhattan distance measures. •We proposed a method to obtain more accurate numeric meanings of the linguistic terms.•The method can tailor linguistics to the different indicators.•The method considers semantic differences arising from the decision-makers and the indicators.•The method incorporates the psychological behavior of the decision-makers.•The method uses more information, including alternative-preference matrix, criteria-preference matrix and decision matrix.