An endogenous and continual learning approach to personalize individual semantics to support linguistic consensus reaching

In linguistic group decision making, it is known that decision makers are individualized in understanding the meanings of words, i.e., decision makers have personalized individual semantics (PISs) in the representation of linguistic preferences. Since individuals influence each other mutually in the consensus reaching process, PISs will accordingly change. This suggests that there is an updating process of PISs for individuals. This paper proposes an endogenous and continual learning-based approach to update PISs in consensus reaching process by incorporating the endogenous consistency-driven PIS model and continual PIS learning based consensus model. Through this approach, individuals' PISs are endogenously updated and learned while ensuring an optimal level of consistency and an increased level of collective consensus during consensus reaching process. At the end of the study, numerical examples and some simulation and comparative analyses are presented to justify the effectiveness of proposed approach.