A multi-stage hybrid consensus reaching model for multi-attribute large group decision-making: Integrating cardinal consensus and ordinal consensus
Xiangyu Zhong,Xuanhua Xu,Xuanpeng Yin
DOI: https://doi.org/10.1016/j.cie.2021.107443
2021-08-01
Abstract:<p>Based on the notion that cardinal consensus and ordinal consensus are of equal importance, a novel consensus reaching model for multi-attribute large group decision-making (MALGDM) is developed in this paper via the integration of cardinal consensus and ordinal consensus. First, an improved ordinal distance measure method is proposed that considers the importance of positions in the ordinal ranking. Next, a hybrid <em>k</em>-means based clustering method that reflects the integration of the cardinal distance and ordinal distance is presented to divide large-scale experts into several clusters. After measuring the consensus of clusters, including both the cardinal consensus and ordinal consensus, they are analyzed by a cardinal-ordinal consensus plot, and the clusters are classified into four categories based on the cardinal consensus threshold and ordinal consensus threshold. Then, by introducing the comprehensive consensus threshold, four consensus control rules are developed to control the consensus reaching process (CRP), and the clusters are subdivided into six categories. For clusters in different categories under different consensus control rules, a specific multi-stage hybrid feedback mechanism is proposed to improve the consensus. In addition, an objective method for the determination of the ordinal consensus threshold is presented. Finally, an application example that simulates the scene of the selection of a hotel for the centralized isolation of entry personnel during the COVID-19 epidemic and a comparative analysis are provided to confirm the validity and effectiveness of the proposed model.</p>
computer science, interdisciplinary applications,engineering, industrial