Scholarlink

<p>With the rapid growth of social media, decision-making groups are expanding in scale and facing increasingly complex problems. Clustering decision-makers (DMs) into subgroups is commonly adopted to reduce complexity. However, DMs from different subgroups often exhibit divergent opinions due to diverse social experiences and educational backgrounds, making the consensus-reaching process (CRP) a critical challenge in large-scale group decision-making (LGDM). Additionally, accurately characterizing DMs’ preference information remains a key research focus. This study proposes a multi-attribute LGDM (MALGDM) method to address scenarios in which DMs express preferences via uncertain linguistic variables and where non-cooperative, leadership, and delegation behaviors emerge during the CRP in social networks. First, an improved social network analysis (SNA) method is applied to cluster DMs, addressing real-world scenarios where nodes may belong to multiple subgroups simultaneously or exist as isolated nodes. Second, a management mechanism for non-cooperative, leadership, and delegation behaviors is designed based on the Uninorm operator and social trust networks to facilitate group consensus level (GCL). Third, comprehensive weight is determined by integrating social network centrality, trust, and similarity. Finally, a three-dimensional uncertain linguistic density (T<sub>2</sub>ULD<sub>WAA</sub>) operator is proposed to aggregate information, and alternatives are ranked using score functions. An example shows that the proposed approach accelerates the CRP while ensuring decision outcomes’ accuracy, fairness, and reliability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Multi-Attribute Large-Scale Group Decision-Making Method Considering Decision-Makers’ Behaviors In Social Networks

  • 期刊论文

  • 发表日期：

    2026年4月4日

查看全文
  • Gang Chen,
  • Anqiong Lang,
  • Xun Han

摘要

With the rapid growth of social media, decision-making groups are expanding in scale and facing increasingly complex problems. Clustering decision-makers (DMs) into subgroups is commonly adopted to reduce complexity. However, DMs from different subgroups often exhibit divergent opinions due to diverse social experiences and educational backgrounds, making the consensus-reaching process (CRP) a critical challenge in large-scale group decision-making (LGDM). Additionally, accurately characterizing DMs’ preference information remains a key research focus. This study proposes a multi-attribute LGDM (MALGDM) method to address scenarios in which DMs express preferences via uncertain linguistic variables and where non-cooperative, leadership, and delegation behaviors emerge during the CRP in social networks. First, an improved social network analysis (SNA) method is applied to cluster DMs, addressing real-world scenarios where nodes may belong to multiple subgroups simultaneously or exist as isolated nodes. Second, a management mechanism for non-cooperative, leadership, and delegation behaviors is designed based on the Uninorm operator and social trust networks to facilitate group consensus level (GCL). Third, comprehensive weight is determined by integrating social network centrality, trust, and similarity. Finally, a three-dimensional uncertain linguistic density (T2ULDWAA) operator is proposed to aggregate information, and alternatives are ranked using score functions. An example shows that the proposed approach accelerates the CRP while ensuring decision outcomes’ accuracy, fairness, and reliability.