报告题目:Dynamic Clustering for Consensus Reaching Model in LSGDM Dealing with Overlapping Communities
报告摘要:This talk introduces a novel dynamic clustering-based consensus model for Large-Scale Group Decision Making (LSGDM) that explicitly manages overlapping communities within social networks of decision makers. It begins by outlining the limitations of traditional CRPs — static trust structures, disjoint communities, and limited scalability — which reduce realism and efficiency in large, socially connected groups.
The proposed model formalizes LSGDM as a networked problem, where decision makers are nodes linked by evolving trust relationships. It employs a LINE-based Label Propagation (LNLP) algorithm to detect overlapping communities, enabling DMs to belong to multiple influence zones. Consensus is computed hierarchically at evaluation, cluster, and group levels using a hybrid centrality measure that integrates both local and global network properties.
A key innovation is the overlapping community-driven feedback mechanism, where decision makers in intersecting clusters act as mediators, guiding adjustments across communities. The process iteratively updates trust degrees, reclusters the network, and recalculates consensus until a satisfactory global consensus degree (GCD) is achieved.
Validated through a case study, the model demonstrates improved consensus levels, interpretability, and adaptability. The talk concludes by emphasizing that this dynamic, socially-aware CRP enhances scalability and realism in decision networks and outlines future extensions toward fuzzy, hesitant, and linguistic environments, promising richer uncertainty management and broader applicability in intelligent decision support systems.
报告人简介:Luis Martinez Lopez 教授是西班牙哈恩大学计算机科学系教授, 主要研究方向为模糊决策、词计算、多准则决策等,目前担任 International Journal of Computational Intelligence Systems 主编,Information Sciences、Expert System with Applications、Knowledge-based System、IEEE Transactions on fuzzy systems、Information Fusion 等期刊副主编,IFSA 会士(IFSA Fellow),IEEE senior member,欧洲模糊逻辑与技术学会(ESFLT)会士。发表 270 余篇文章,2017-2024 年连续入选全球高被引学者榜单。