制导精度一体化试验的Bayesian样本量计算方法

针对制导精度一体化试验的样本量计算(SSD)问题,分析了应用经典样本量计算方法和标准幂先验的Bayesian样本量计算方法存在的问题和矛盾;为了解决验前样本量很大时,基于标准幂先验得到的设计先验与Bayesian平均后验方差准则之间的矛盾,综合考虑仿真可信度与验前样本量的影响,提出了试验的设计效应指标,给出一种基于试验的设计效应等价确定修正幂先验指数的方法,并用于构造Bayesian样本量计算的设计先验;以兴趣参数Bayesian估计的平均后验方差为输出精度,分别设计了在试验经费约束和评估精度要求约束下的一体化试验方案样本量优化计算方程,并通过示例分析证明了所提Bayesian样本量计算方法的有效性。

Bayesian sample size determination for integrated test of missile hit accuracy

Sample size determination (SSD) methods for integrated test of missile hit accuracy are analyzed, which reveals the problem and contradiction in classical method and Bayesian method using standard power prior for design. In order to solve the contradiction between standard power prior for design and average posterior variance criterion of Bayesian SSD while prior sample size is very large, design effect of experiment is proposed with a comprehensive consideration on simulation test credibility and prior sample size. Thus, a modified power exponent for design prior elicitation based on design effect equivalence of experiment is given. Taking Bayesian average posterior variance for parameter of interest as the output precision, we get the optimization equations for SSD of integrated test scheme under both test cost constraint and required posterior precision constraint. In the end, the effectiveness of our proposed Bayesian SSD method for integrated test of missile hit accuracy is illustrated with two examples.