A DBN‐Based Risk and Reliability Information Mapping Method for Adaptive Decision Support in Safety‐Critical Operations

To support decision‐making under dynamic risk, uncertainty, and high information load in safety‐critical operations, this study proposes a dynamic bayesian network (DBN)‐based risk and reliability information mapping method. The method establishes a five‐layer framework that connects dynamic risk inference, semantic state representation, mode determination, strategy regulation, and trustworthy decision‐support output. DBN posterior results are transformed into decision‐oriented semantic signals, including risk severity, variation trend, escalation probability, inference certainty, dominant risk chain, and task stage. These signals determine the current support mode and regulate the presentation of recommended actions, key evidence, procedural guidance, verification items, and confidence cues. Strategy regulation mechanisms further stabilize support under changing uncertainty and risk‐state transitions. A human‐in‐the‐loop experiment compared predefined Baseline, Baseline‐Fixed, and Adaptive conditions. Based on Holm‐adjusted p ‐values, Adaptive achieved significantly better performance than both comparison conditions across all six predefined indicators ( p < 0.05). The main improvements included faster response and critical‐information localization, higher action accuracy and critical‐item hit rate, and lower decision omission and fewer incorrect operations. The results demonstrate that DBN‐based dynamic risk information can be systematically transformed into adaptive and reliability‐oriented decision support for safety‐critical operations.