Bi-level Consensus in Large-Scale Group Decision-Making: Integrating Structural Holes and Community Dynamics

Large-scale group decision-making (LSGDM) is deeply influenced by social network structures, where relational patterns shape how individuals relate to each other by exchanging information and adjusting opinions during consensus-reaching processes (CRPs). However, the strategic role of structural hole spanners—individuals who bridge communities and enable cross-cluster communication—has often been overlooked. To address this gap, we propose a bi-level consensus framework that integrates structural hole theory with community dynamics in LSGDM. First, a community bridge spanner detection (CBSD) algorithm is introduced to identify structural hole spanners while considering overlapping communities. Second, a bi-level feedback mechanism is established: local influencers resolve intra-community conflicts to strengthen internal consensus, while spanners mediate inter-community disagreements through trust-aware opinion aggregation. Third, a fuzzy incremental reward-based trust update (FIR-TU) mechanism models trust evolution and community adjustment throughout iterative CRPs. A supplier selection case study under dual hesitant fuzzy sets validates the framework. Comparative experiments demonstrate that the proposed method enhances consensus-reaching efficiency and more effectively captures structural influence patterns through spanner-driven decision modeling.