A cyclic dynamic trust-based consensus model for large-scale group decision making with probabilistic linguistic information

Abstract This paper investigates a consensus reaching process (CRP) considering dynamic trust in large-scale group decision making (LSGDM). In the traditional trust-based consensus model, it is assumed that the trust relationship generated by decision makers (DMs)’ previous knowledge remain unchanged during the whole decision process. However, this relationship will be dynamic rather than static especially in a social network with a new decision problem. This study explores the dynamic nature of trust through two stages. In the first stage, the trust degree will be functionally reformed by the conflict caused by DM’s opposite preferences. In the second stage, it will be effected by surroundings according to the “assimilation effect” in network. To handle the CRP with large-scale decision settings, a clustering technique is used to classify DMs with similar preference and preference accuracy. Based on the classifications, an optimization model is constructed to obtain the trust degrees between subgroups. The consensus measurements are investigated from similarity network within subgroups and min–max programming model between subgroups, respectively. Moreover, preference modification will effect trust in the aggregation and next iteration, the cyclic dynamic trust mechanism is established. The feasibility of the proposed model is verified by a numerical example. Comparisons declare the constructed consensus model’s universality without any essential conditions, as well as superiority with fully consideration of DM’s utility and centrality in network.