In the era of social media, the issue of large-scale group decision-making (LSGDM) is becoming increasingly prominent. The complexity of large group interactions increases rapidly with the expansion of the group size. Current research has often used cluster analysis to reduce the dimensionality of LSGDM, but the decision agents following dimensionality reduction are not clearly defined, which hinders the practical application of LSGDM methods. This paper studies the LSGDM through the context of community representatives which are determined by combining strong and weak ties. The participation of such representatives in subsequent decision-making can reduce negotiation costs and efficiently form a consensus from both global and local perspectives. To identify these community representatives, this study improves the traditional Laplacian centrality, calculates the local and global centrality of individuals, and defines metrics for measuring representatives. The opinion dynamics model is used to verify the effectiveness of the LSGDM model with community representatives. The karate club network is used to illustrate the application of the proposed LSGDM model. Simulation and comparative analysis results show that the proposed model takes less time to reach consensus than traditional ones. Furthermore, there is no complete linear relationship between the representative size and the time taken to reach group consensus.

中文翻译：

---

社区代表参与的大规模群体决策：强弱关系结合的视角

社交媒体时代，大规模群体决策（LSGDM）问题日益突出。随着群体规模的扩大，大群体交互的复杂性迅速增加。目前的研究经常使用聚类分析来降低LSGDM的维数，但是降维后的决策主体并没有明确的定义，这阻碍了LSGDM方法的实际应用。本文通过强弱联系确定的社区代表背景来研究LSGDM。这些代表参与后续决策可以降低谈判成本，并从全球和本地角度有效形成共识。为了识别这些社区代表，本研究改进了传统的拉普拉斯中心性，计算个体的局部和全局中心性，并定义了衡量代表的指标。意见动态模型用于与社区代表验证LSGDM模型的有效性。空手道俱乐部网络用于说明所提出的 LSGDM 模型的应用。仿真和比较分析结果表明，该模型比传统模型达成共识所需的时间更短。此外，代表人数与达成群体共识所需的时间之间不存在完全的线性关系。