On disaggregation modeling of uncertainty group preferences considering trust networks

Preference disaggregation is a common method used in multi-criteria decision aid (MCDA) to derive preference models and reproduce the preference information of decision makers (DMs). Due to cognitive uncertainty, DMs may not be able to accurately, directly, and deterministically portray their preferences based on their own cognition, causing difficulties when trying to efficiently merging group information. To explore reasonable resolutions, this article addresses the uncertainty in the preference information provided by DMs and portrays this information by utilizing uncertainty theory. Furthermore, it integrates group information by constructing social networks and analyzing the similarity of decision results of DMs. Specifically, a group preference learning model is designed to aggregate the diverse individual preference information into group decision-making in accordance with trust-similarity centrality. Ultimately, by fully utilizing this integrated information, a group decision-making recommendation is produced. For missing values in the trust network, the product axiom is used as a propagation operator to maximize the propagated trust. To illustrate the proposed model, it is applied to rank European countries in terms of their university qualities. Finally, the proposed model is compared with the existing models that incorporate preference disaggregation or social networks, and the effectiveness of the proposed model in handling uncertainty and the advantages in interpretability are demonstrated.