Group decision-making (GDM) in large-group social network environment (LGSNE) has attracted considerable attention in the field of decision science. Social relationships exist among decision-makers, and individual decisions are often influenced by others they are connected with. Opinions among large-scale decision-makers can easily be controversial and conflicting. Reaching consensus is necessary, but it requires the adjustment of some individual opinions. Due to differences in self-interest and perception, some decision-makers are noncooperative with regard to adjusting their opinions to promote consensus. This may delay consensus convergence and ultimately affect decision quality. This study proposes a two-dimensional consensus convergence model considering noncooperative behaviors. We first describe the characteristics of GDM problems in LGSNE. Two measurement attributes – trust relationship and opinion similarity – are identified as important factors throughout the decision-making process. Then, we propose a hierarchical clustering method based on the trust–similarity measure. A weight-determining method for clusters is presented that considers the internal and external features of a cluster. Based on these, a two-dimensional consensus convergence process is designed to reduce opinion differences and manage noncooperative behaviors. Finally, a numerical experiment is used to illustrate the feasibility and efficacy of the proposed model, and comparative analysis reveals its features and advantages.

大型团体社交网络环境（LGSNE）中的团体决策（GDM）在决策科学领域引起了相当大的关注。决策者之间存在社会关系，个人决策通常会受到与其相关的其他人的影响。大型决策者之间的观点很容易引起争议和冲突。达成共识是必要的，但它需要调整一些个人意见。由于自身利益和看法的差异，一些决策者在调整意见以促进共识方面不合作。这可能会延迟共识收敛，并最终影响决策质量。这项研究提出了一种考虑非合作行为的二维共识收敛模型。我们首先描述LGSNE中GDM问题的特征。在整个决策过程中，两个度量属性（信任关系和观点相似性）被确定为重要因素。然后，我们提出了一种基于信任相似度度量的分层聚类方法。提出了一种考虑集群内部和外部特征的集群权重确定方法。基于这些，设计了一个二维共识收敛过程，以减少意见分歧并管理非合作行为。最后，通过数值实验说明了该模型的可行性和有效性，并通过比较分析揭示了该模型的特点和优点。我们提出了一种基于信任相似度度量的分层聚类方法。提出了一种考虑集群内部和外部特征的集群权重确定方法。基于这些，设计了一个二维共识收敛过程，以减少意见分歧并管理非合作行为。最后，通过数值实验说明了该模型的可行性和有效性，并通过比较分析揭示了该模型的特点和优点。我们提出了一种基于信任相似度度量的分层聚类方法。提出了一种考虑集群内部和外部特征的集群权重确定方法。基于这些，设计了一个二维共识收敛过程，以减少意见分歧并管理非合作行为。最后，通过数值实验说明了该模型的可行性和有效性，并通过比较分析揭示了该模型的特点和优点。