A reliability-based consensus model and regret theory-based selection process for linguistic hesitant-Z multi-attribute group decision making

Abstract

In real-life decision making problems, reliability of information is one of the most concerning factors. Accordingly, to avoid the involvement of inconsistent information in a multi-attribute group decision making problem, this study aims to develop a consensus model by considering reliability of experts’ opinions in Z-number framework. To do so, first we reflect human uncertainty in the information system by linguistic hesitant-Z-number (LHZN) that has sufficient description power to express hesitancy, vagueness and reliability of evaluation information in a single form. Then, a novel normal cloud model is developed to handle LHZN information rationally without distorting its originality. Furthermore, we develop the Maclaurin symmetric mean operator based on normal cloud to take into account interrelationships among attributes. Thereafter, the weights of the experts are determined based on dynamic importance score of experts which consist of preference similarity score and reliability score. Afterwards, a group consensus degree is measured based on preference values and their reliability to identify the consensus level among experts. Then, to ameliorate the consensus efficiency, experts’ unreliable behaviours are identified and modified with reference to preference value and reliability. Lastly, a regret theory-based selection process is presented to detect the optimal alternative with respect to multiple attributes. The validity, superiority and advantages of the proposed model are exemplified by two case studies of sustainable zero-carbon measures prioritization for Indian transport sector, and failure mode and effect analysis, respectively.