一种基于混合遗传算法优化的截断重要抽样法

Truncated Importance Sampling Method based on Optimization of Mixed Genetic Algorithm

针对并联系统失效概率的计算,提出一种基于混合遗传算法的截断重要抽样法。所提算法中,基于十进制编码的混合遗传算法被用来寻找并联系统最可能失效点x*和系统近似可靠度指标β。以β为半径建立以坐标原点为球心的截球,并以x*为抽样中心构造重要抽样概率密度函数,从而建立针对并联系统可靠性分析的β球截断重要抽样法。通过算例分析,比较了几种不同并联系统失效概率计算的方法,结果表明本文方法比连续顺序近似法、一次二阶矩法具有更高的计算精度,比蒙特卡洛法具有更快的收敛速度,尤其是针对小失效概率问题;与β球截断抽样法和重要抽样法相比,计算效率也进一步提高。

A truncated importance sampling method is employed for the failure probability of parallel system with multiple failure modes.The mixed genetic optimization algorithm is chosen to search for the the most probable failure point in the failure domain,and solve the approximate reliability index β of the parallel system.A β-sphere truncated importance sampling method for the failure probability of parallel system is consisted of β-sphere truncation and importance sampling probability function.Comparing with the successive sequential approximation approach and FORM,the presented method performs much more efficiently than Monte-Carlo method with higher accuracy,especially for the small failure probability.