Shen Jinting, Si Zeyang, Li Hongyu, Ma Ning, He Wei
Harbin Normal University, Harbin, 150025, China.
Sci Rep. 2025 Jan 27;15(1):3356. doi: 10.1038/s41598-025-86851-w.
The health status of aerospace equipment directly affects the operational capability of the entire system. Belief rule base (BRB) is an effective method for assessing health status that combines expert knowledge and historical data. However, in the actual assessment, the data provided by experts only form the basic framework of the model. Therefore, the BRB model with joint optimization of structure and parameters (BRB-SPO) is proposed to achieve a balance between the model's accuracy and complexity. First, to balance complexity and accuracy of the model, parameter structure backward stepwise selection method (BSS) and full factorial design (FFD) are introduced. BSS constructs the optimal parameter set, while FFD determines the best parameter values for the model. Subsequently, the constructed model is deduced using the evidential reasoning (ER) calculation procedure, the other parameters are optimized using the projection covariance matrix adaptive evolution strategy (P-CMA-ES). Finally, the practicality of the proposed method is validated through two examples.
航空航天装备的健康状态直接影响整个系统的作战能力。置信规则库（BRB）是一种结合专家知识和历史数据评估健康状态的有效方法。然而，在实际评估中，专家提供的数据仅构成模型的基本框架。因此，提出了结构与参数联合优化的BRB模型（BRB-SPO），以实现模型准确性和复杂性之间的平衡。首先，为平衡模型的复杂性和准确性，引入了参数结构反向逐步选择方法（BSS）和全因子设计（FFD）。BSS构建最优参数集，而FFD确定模型的最佳参数值。随后，使用证据推理（ER）计算程序推导构建的模型，使用投影协方差矩阵自适应进化策略（P-CMA-ES）优化其他参数。最后，通过两个实例验证了所提方法的实用性。
