Research On Sorting Consensus Based On The Interaction Of Group Opinions Under Hierarchical Social Network

In order to satisfy the increase of decision-making complexity and requirements,group decision making can cope with challenges and tackle decision-making problems through cooperation and competition,which shows an overall advantage beyond the individual decision-making.Group decision making using collective intelligence is widely applied in emergency management,e-commerce,recommendation management and so on.With respect to group decision making,due to the differences in professional fields,knowledge backgrounds,preferences,habits and judgment levels of decision makers,they may lead to the inconsistency or even decision conflict,which makes it difficult to implement the group result.Thus,the effectively usage of group information to make scientific decisions is an important research problem.So far,the researches on group consensus have made remarkable progress.Considering the social relationship of decision makers,scholars explored the group consensus under social network by introducing the decision makers’ trust information.Remarkably,in the group decision making scene of the organizational structure,there are both decision makers as leaderships and participants as employees in the department.The social network shows a hierarchical characteristic,which impacts group consensus reaching.Therefore,based on the hierarchical social network,this research deeply discusses four core issues of group consensus.Firstly,there is the heterogeneity of participants in hierarchical social network.Decision makers and participants are involved in group decision making.The behavioral characteristics of participants can reflect the influence of decision makers,which helps to reduce the manipulation behavior of traditional group decision making in evaluating the influence of decision makers.The network hierarchy may impact the adjustment and interaction of decision makers,and affect the group consensus reaching.Secondly,there is a decision demand based on the sorting result consensus.Traditional consensus decision-making models often rely on opinion consensus to get the group results.However,opinion consensus does not fully guarantee the result consensus.Then,the sorting result consensus oriented decision goal will affect the adjustment mechanism and the process of reaching consensus.Thirdly,the hierarchical social network may be overlapping.In the cross-departmental collaboration decision-making activities,it may be faced with multiple departments to participate in the decision-making situation and exist the cross-departmental decision makers.The overlap of the hierarchical social network will impact the information transmission and interaction among decision makers,which changes the speed and direction of consensus.Finally,the hierarchical social network may exist higher-order structures.Because of the connection mode of hierarchical social network,the existence of higher-order structure is inevitable.Based on the higherorder structure,decision makers produce higher-order opinion interaction which is different from pairwise interaction,and changes the consensus reaching process.The above problems are discussed in this study,the main research works are shown as follows:(1)This dissertation develops a group sorting consensus method based on the influence of opinion leader under hierarchical social network.Firstly,this dissertation builds a hierarchical social network based on the characteristic of organizational structure.Inspired by the idea of three-way decision,the method learns the participants’ following behavior to the decision makers,and utilizes the participants’ following behavior to calculate the decision makers’ influence and identify opinion leader.Considering the dynamic adjustment cost and budget,based on the opinion leader and optimization theory,the minimum adjustment and maximum closeness model under the cost constraint is constructed,and the opinion leader is used to guide the decision maker’s adjustment and promote the group consensus.(2)This dissertation explores a group sorting consensus method based on sorting result viewpoint in hierarchical social network.Our research constructs a sorting consensus index from the perspective of sorting result.The asymmetric misclassification cost function is designed to describe the decision makers’ adjustment preferences in different situations.The opinion leader’s and the group results’ dominant set in hierarchical social network are determined as the sorting target to guide the preference adjustment of decision maker.The intra-class distance between classes is introduced to restrict the decision makers’ adjustments.An adjustment model based on sorting target guidance considering asymmetric misclassification cost under distance restriction is established to guide the decision maker’s adjustment,which improves group sorting consensus and achieves weak opinion consensus.(3)This dissertation investigates a group sorting consensus method based on the influence of externality in multi-hierarchical social network with cross-departmental cooperation.The coordination degree of cross-departmental decision makers and the social network relationship are used to determine the influence of decision makers,identify cross-departmental opinion leaders,and construct a minimum adjustment and maximum closeness model based on opinion leaders.This method analyzes the externality influence of the decision maker’s adjustment behavior on the participant’s following behavior,designs the externality influence identification rule,updates the decision maker’s weight and establishes the termination condition,and improve the adjustment model and group consensus.(4)This dissertation studies a group sorting consensus method based on higher-order opinion interaction in multi-hierarchical social network.This method learns the trait of simple complex,and utilizes the trust information of social network to identify the higherorder structure in the multi-hierarchical social network.Inspired by the simplex dynamics model,higher-order opinion interaction considering the enhancement mechanism and competition mechanism are designed.This method constructs the local stability and global stability of the network structure and establishes an adjustment model based on the higherorder opinion interaction and the network stability,which fulfills group consensus.To sum up,this dissertation investigates the impact of the hierarchical social network’s structure character for the decision makers’ influence,opinion interaction and the group consensus reaching.Our research enriches the application situation of group decision making,and provides effective solutions and theoretical support for the realization of group decision sorting consensus.