The Research Of The Probabilistic Linguistic Consensus Model Driven By Trust Relationship In Large-scale Group Decision-making

Group decision-making problems have changed dramatically due to rapid scientific and technological advances and continuous social progress.The requirement for methods or models that effectively deal with large-scale group decision-making(LSGDM)is increasing.In the increasingly complex decision-making environment,it is difficult for decision makers(DMs)to provide evaluation information with accurate real numbers,and they have a greater tendency to express evaluation information through fuzzy linguistic information.Among various forms of information,probabilistic linguistic term sets(PLTSs)can reflect DMs’ preferences for different linguistic terms,comprehensively reflect the uncertainty and fuzziness of DMs’ evaluation information,and have broad applicability.Meanwhile,with the continuous development of information technology and social media,DMs constantly interacting through social trust networks.This allows the opinions of DMs to influence each other in the decision-making process,increasing the complexity of LSGDM.In the decisionmaking process,rational use of trust relationships among DMs can not only help reduce the complexity of LSGDM and improve the efficiency of decision-making but also facilitate the large group to obtain acceptable group consensus opinion to facilitate consensus reaching.Therefore,studying the trust relationship-driven probabilistic linguistic large group consensus decision-making model is of great theoretical and practical significance.The research in this paper focuses on the following.(1)The research on the consensus model of LSGDM in a probabilistic linguistic context considers DMs’ hesitation degree driven by the trust relationship.PLTS is used as the information form for DM evaluation,and a consensus model for LSGDM that considers DMs’ hesitation degree and trust relationships is investigated.A clustering algorithm based on opinion similarity and trust relationships is proposed in response to the current problem of a single clustering indicator in LSGDM.A method for objectively determining clustering thresholds that consider the internal cohesion of subgroups is provided.In the complex decision-making environment,considering that the hesitation degree of DMs’ evaluation information affects the quality of decision results,a method for determining the hesitation degree of DMs in a probabilistic linguistic environment is proposed.Based on this,a weight determination method based on social trust networks and hesitation degree is further proposed.Finally,a large-scale group consensus decision-making method based on the social network DeGroot(SNDG)model is established,and the feasibility and superiority of the method are verified through case study analysis.(2)The research of the no-threshold consensus model of LSGDM is driven by trust relationships in a probabilistic linguistic context.Considering that the trust relationship between DMs is easily influenced by the decision environment and the DM’s behavior during the decision-making process,a trust relationship adjustment method based on the "similarity-attraction paradigm" is proposed to make the decision-making process more realistic.An identification mechanism with no consensus threshold and a feedback adjustment mechanism based on the SNDG model is established to solve the problem of consensus threshold setting in the consensus reaching process.The no-threshold consensus reaching process based on the SNDG model can reduce the subjectivity of the decision-making process and improve the rationality and scientific of the decision-making results.Based on the above methods,a large group decision-making method based on trust relationships and the no-threshold consensus model is established,and its feasibility and effectiveness are verified by case study analysis.(3)The research of the consensus model for LSGDM is driven by trusting relationships and considering the behaviour management of DMs.Considering that the manipulative behaviour of DMs in the LSGDM will violate the principles of fairness and truthfulness of the decision results,the method of identifying the manipulative behaviour of DMs based on trust relationships is established.Moreover,the management method of manipulative behaviour of DMs,including weight punishment,is proposed to reduce the influence of the manipulative behaviour of DMs on the decision results in the decision-making process.At the same time,the identification and management method of DMs’ overconfidence behaviour is proposed for the existence of DMs’ overconfidence behaviour in the SNDG model.On this basis,a consensus model for LSGDM that considers trust relationships and management of DMs’ behaviour is further developed,which effectively improves the reliability and objectivity of the decision-making results.