Consensus reaching model with self-confidence-based dynamic weights and personalized adjustment constraints for multi-attribute group decision making

It is quite a remarkable fact that experts always hope that the final decision outcome can preserve their original assessments in practical multi-attribute group decision making (MAGDM) activities to a high degree, which can be reflected from their self-confidence in their own assessments. Meanwhile, experts may have different attitude towards the assessment modification at the levels of attribute and alternative in consensus reaching process (CRP), which is commonly characterized by their personalized adjustment constraints on the acceptance of the assessment’ modification suggestions recommended by moderators. Motivated by the facts outlined above, this study proposes a consensus reaching method with self-confidence-based dynamic weights and personalized adjustment constraints for the MAGDM problem. First, a self-confidence-based dynamic weight management method is proposed to accelerate the CRP. Next, considering that experts may have different levels of sensitivities/tolerances for their assessment modifications at the various attribute and alternative levels, a dynamic weight and personalized adjustment constraint-driven consensus model is proposed with the objective of minimizing the distance between experts’ initial assessments and the adjusted collective assessment. Meanwhile, an associated maximum consensus model is implemented to examine whether a predetermined consensus threshold can be achieved with the given assessments. Then, a resolution approach with an interactive CRP that can reconcile the goals that experts should be allocated enough attention and their original assessments should be preserved as much as possible is developed. Finally, three illustrative examples and a comparative study are conducted to show the validity and advantages of the proposal. Overall, this study exhibits three facets of originality: the dynamic weight management method is developed, the personalized assessment-modification willingness is formed, and the interactive consensus reaching algorithm is created.