Three-way decision-based consensus feedback mechanism considering conflict level and social network for large-scale group decision-making

Large-scale group decision-making (LSGDM) has recently gained significant attention, with most studies viewing decision makers’ (DMs) behaviors as either supportive or non-supportive. However, as decision-making complexity increases, DMs encounter diverse situations, which introduces greater uncertainty into the decision process. The three-way decision theory provides a new perspective to address this problem by dividing the universal set into three regions: positive, boundary and negative, thereby reducing decision-making risks. Building on this, we propose a three-way decision-based consensus feedback mechanism (3WD-CFM) considering conflict level and social network for LSGDM, in which DMs are classified into three behaviors of support, uncertainty, and non-support by introducing a quantitative loss function, mitigating reliance on subjective prior knowledge. A targeted feedback strategy is developed that prioritizes non-supportive DMs and adapts adjustment rules according to behavioral classifications to enhance consensus efficiency. To jointly optimize global consensus and local subgroup harmony, a dual-layered control mechanism is introduced, incorporating conflict monitoring as a parallel indicator. Furthermore, a collaborative representation approach is employed to uncover latent relational structures among DMs, based on which a reputation-based weighting mechanism is constructed, balancing individual preferences with structural interdependencies. Finally, the validity and robustness of the 3WD-CFM are proved by numerical experiments and parameter analysis, and its superiority is highlighted through several comparisons.