基于等效样本和实仿融合的现场小子样试验制导精度评估方法

Guidance accuracy evaluation method of field small sample test based on equivalent sample and real simulation fusion

Bayes方法通过融合仿真先验信息可以提高现场小子样试验评估精度,但Bayes方法直接将仿真先验分布方差作为度量仿真先验信息的精度参数参与计算后验加权系数,存在大样本仿真数据容易“淹没”现场小子样试验数据、估计结果不是最优估计值等现实问题。针对该问题,提出了一种基于等效样本和实仿融合的现场小子样试验制导精度评估方法。在仿真可信度C、统计理论和仿真评估精度等条件约束下,使用仿真可信度C引起的评估误差E1和仿真抽样引起的评估误差E2共同度量仿真先验精度。建立仿真可信度C和评估精度之间的关系模型,将E1折算成相同评估精度下的现场试验样本量。使用等效样本表征E1,仿真抽样样本表征E2共同参与计算后验加权系数,基于Bayes框架完成了对后验分布参数的精确估计。通过仿真实验验证了本文方法的有效性和正确性。

Bayesian method can improve the evaluation accuracy of field small sample test by fusing simulation prior information.However,Bayesian method directly takes the variance of simulation prior distribution as the precision parameter to measure the simulation prior information and participates in the calculation of posterior weighting coefficient,which leads to some practical problems easily such as big simulation data is easy to submerge the field small data and the estimation result is not the optimal estimation value.In order to solve the above problem,the guidance accuracy evaluation method of field small sample test based on equivalent sample and real simulation fusion is proposed.Under the constrains of simulation credibility degree C,statistic theory and simulation estimation error and so on,the evaluation accuracy E1 caused by simulation credibility degree C and the evaluation accuracy E2 caused by simulation samples are used to measure the simulation prior information error.The relationship model between credibility degree C and the evaluation accuracy is established,and E1 is converted into field test sample size under the same evaluation accuracy.The equivalent sample is used to represent E1,and the simulation sample is used to represent E2.The posterior distribution parameters are estimated accurately based on Bayesian framework.Simulation experiments results indicate the effectiveness and correctness of the proposed method.