A Semantic–Preference Similarity Fusion Approach for Large-Scale Group Decision Making
Large-scale group decision-making (LGDM) often relies on unified semantic scales, single-perspective clustering, and mean-based feedback adjustment, which can distort experts’ intended preferences and lead to ranking instability during the consensus-reaching process. To overcome these limitations, this paper proposes a multi-stage semantic–preference fusion framework that integrates personalized individual semantics with order-preserving consensus regulation.First, a consistency-driven personalized individual semantic modeling mechanism is developed to calibrate each expert’s numerical scale function, enhancing the additive consistency of fuzzy preference matrices while strictly preserving ordinal constraints. Second, a semantic–preference joint clustering strategy is introduced to reconstruct subgroup structures from dual perspectives by incorporating semantic merging, subgroup size reallocation, and preference migration, together with a hierarchical weighting scheme for aggregating intra-group and inter-group information. Third, a multi-level consensus measurement system is established at the individual, subgroup, and group levels, and an anchor-guided nonlinear feedback mechanism is designed to regulate bounded preference adjustments for low-consensus experts while minimizing the risk of order violations. Finally, numerical examples and comparative experiments demonstrate that the proposed method effectively improves consensus and ranking stability, thereby providing an interpretable and convergent framework for large-scale decision-making.
Authors
- Geng Xiu-li
- Lin Ze-xi
Institutions
- University of Shanghai for Science and Technology (CN)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-19
- DOI
- https://doi.org/10.5281/zenodo.24359163
- Primary Topic
- Multi-Criteria Decision Making
- Type
- article
- Field-Weighted Citation Impact
- 0.00