| The rapid development of information technology has brought about tremendous changes in human communication patterns.People have become closer,communication has become easier,and more decision makers(DMs)have been able to participate in decisionmaking problems.The group decision-making problem begins to change to the large group decision-making problem,so as to realize the decision-making goal more democratically and scientifically.In actual decision-making,there is usually a trust relationship among DMs,which is a key factor affecting the decision-making process.At the same time,in the decision-making process,DMs may take non-cooperative behaviors,which will greatly affect the efficiency and results of decision-making.Therefore,it is necessary to reasonably analyze the mutual trust between DMs and deal with their non-cooperative behavior in decision-making problems.Against this background,this paper proposes a large group decision-making method considering non-cooperative behavior under a dual-trust network.The main work is as follows:(1)In order to comprehensively reflect the trust relationships among DMs in the real environment,this paper proposes a dual-trust network based on familiarity trust and similarity trust,and uses it to guide the clustering process of decision-making groups.First,inspired by the friend-of-friend algorithm,a method that can objectively quantify the familiarity trust between DMs is proposed;based on the preference differences of DMs,similar trust is obtained,and then a dual-trust network between DMs is constructed by combining the two trust relationships above,which can comprehensively consider the interpersonal relationship and opinion similarity among DMs at the same time.On this basis,the hierarchical clustering algorithm is introduced,and the dual-trust relationship between DMs is used as the criteria of clustering,and a large group hierarchical clustering method under the dual-trust network is proposed,which can obtain ideal clustering results,namely members of the same decision-making unit are familiar with each other and have similar opinions.(2)Aiming at different decision-making stages in large group decision-making problems,this paper proposes an interactive consensus method and a supervisory consensus method respectively.Through the combination of the two consensus methods,a two-stage consensus method is proposed.First,considering the low level of collective consensus at the early stage of decision-making,the interactive consensus method is used to require all decision-making units to adjust their opinions according to the adjustment suggestions,in which the adjustment suggestions are generated based on the opinions of other decision-making units that they trust;when the collective consensus level is high enough,the supervisory consensus method is used to require the specific decision-making unit to adjust his opinions according to the adjustment suggestions,where the adjustment suggestions are provided and supervised by the rest of the decision-making units.On this basis,by introducing a control coefficient,the appropriate consensus method can be flexibly selected according to the collective consensus level.(3)Due to the fact that the actual performance and behavior of decision-making units need to be considered,this paper proposes different non-cooperative behavior identification and management methods for different decision-making stages.First,in the interactive consensus process,the decision-making units’ non-cooperative willingness and trust risk are used to classify their adjustment behavior,and their trust relationships and weights are rewarded or punished according to their behavior;in the supervisory consensus process,through the decision-making unit’s subjective adjustment willingness and adjustment recommendations to identify and manage its adjustment behavior.This paper intends to study the large group decision-making problem and decisionmaking behavior under the trust network environment.The effectiveness of the proposed method is verified by example,simulation and comparative analysis,which can enrich the research on the theory and method of large group decision-making in complex situations on the theoretical level,and provide tool support for the selection of community e-commerce platform and other large group decision-making problems on the application level. |
