Reflexive trust Index for assessing internal consistency in Z-number-based information

Decision-making under uncertainty often relies on information that is both imprecise and associated with varying levels of reliability. Z-numbers provide a unified framework for modeling such dual uncertainty through the pair Z = (A, B), where A represents a fuzzy constraint and B expresses the associated reliability. However, in most existing Z-number-based approaches, reliability is primarily interpreted through the marginal properties of the B component, while the internal structural consistency between A and B remains insufficiently quantified. In this paper, a Reflexive Trust Index is proposed to assess the internal consistency of Z-number-based information. The proposed index integrates three complementary components: informativeness compatibility, semantic consistency, and a reflexive peak-confidence ratio capturing the relationship between confidence and informative strength. The resulting indicator is normalized to the interval [0,1] and provides an interpretable measure of the coherence between the informational and reliability components of a Z-number. Numerical case studies demonstrate that the proposed index can effectively distinguish between Z-numbers with similar nominal confidence levels by revealing differences in their internal structure and consistency. In particular, the results show that not only the magnitude but also the distribution of the confidence component plays a crucial role in decision-making. The proposed approach serves as a practical diagnostic tool that enhances the interpretability and robustness of Z-number-based decision-making models without replacing existing methodologies. Its effectiveness is illustrated through scenario-based validation and numerical case studies, demonstrating its ability to distinguish structurally consistent and inconsistent Z-number configurations.