【摘要】 针对BP求解方法存在应用局限性和计算误差较大的缺陷，提出了一种基于冠豪猪优化算法(CPO)与BP神经网络相结合的可靠性分析方法。利用CPO结构简单、能广泛探索搜索空间的能力等优点，改善BP神经网络陷入局部极小值的问题。通过多间隙机构可靠度分析表明CPO-BP精度优于传统BP。与传统的Monte Carlo法相比，所需的样本数量较少，得出的可靠度的误差范围小，可以节约大量的计算量与计算时间。更多还原

【Abstract】 A reliability analysis method based on the combination of Crested Porcupine Optimizer(CPO) and BP neural network is proposed to address the limitations of BP solving method in application and large calculation errors. By utilizing the advantages of simple CPO structure and the ability to explore search space extensively, the problem of BP neural network getting stuck in local minima can be improved. The reliability analysis of the multi-gap mechanism shows that CPO-BP has better accuracy than traditional BP. Compared with the traditional Monte Carlo method, it requires fewer samples and has a smaller range of reliability errors, which can save a lot of computation and time.更多还原