Objective To take full advantages of the producer's risks and allocate risks of both the producer and the customer in ship-to-air missile kill probability hypothesis testing reasonably based on established risk and test sample size. Methods The most powerful test which was named as stochastic test was induced by likelihood function. A Ber-noulli test was taken through when necessary after the completion of the test to determine whether to accept the null hy-pothesis or not. Results Compared with the classical hypothesis testing method, stochastic test method can ensure that the actual producer's risk does not exceed established producer's risk limit, while reducing risks of the customer. Conclusion This method can make good use of risks of the producer and reduce risks of the customer without increasing test samples. It has very good engineering practice significances.%目的 充分有效利用生产方风险,在既定的生产方风险和有限的试验样本条件下,合理分配舰空导弹杀伤概率假设检验中的双方风险.方法 通过似然函数推导检验问题的最优势检验,设计一种随机化检验方法,在检验完成后且必要时进行一次Bernoulli试验,以确定是否接受原假设.结果 计算结果显示,相对经典假设检验方法,随机化检验方法可以保证实际生产方风险不超过既定的生产方风险上限,同时可使军方风险更小.结论 该方法能够合理利用生产方风险,在不增加试验样本的条件下,降低军方风险,具有很好的工程实践意义.
展开▼
