Reliability-based multi-attribute large group decision making under probabilistic linguistic environment

This study proposes a more rational and effective multi-attribute large group decision making (MALGDM) method with probabilistic linguistic term set (PLTS) from reliability perspective. A reliability measure method is first proposed to compute the reliability degree of the PLTS, and then a normalization method is presented to normalize the ignorant PLTSs with respect to maximizing their reliability degrees. An efficient clustering method combining the opinion similarity of experts and the reliability degrees of the clusters formed is introduced. Moreover, an objective method of determining the similarity and reliability thresholds is presented. After classifying the large-scale experts, the consensus levels of clusters and the global consensus level are measured and the cluster that need to adjust information is identified based on its consensus level and reliability degree. Then, an optimization model to maximize the global consensus level and the global reliability degree is then built to obtain the evaluation values for improving the consensus levels and reliability degrees. The deviation between the expectation values of the evaluation values before and after adjustment is constrained by the parameter provided by the experts within the cluster that need adjustment. Finally, an application example of the selection of the hotel for isolating the entry personnel during the Covid-19 pandemic and some comparative analyses are provided to validate the proposed method.