总体方差未知时k法单侧计量检验方案的改进

Improvement on K Method Sampling Inspection Scheme by Variables with a Single Specification Limit under Unknown Total Variance

k法是产品计量抽样检验中的常用方法之一。就总体方差未知时的k法抽验方案,Wallis W.A.于1947年给出了一个沿用至今的以正态分布近似t分布的结果,我国有学者分别于上世纪和2014年提出了一些改进,但其中存在参数设定不够灵活且过程繁琐的缺陷,文章基于精确公式给出了灵活且精度更高的迭代解,有助于改进单侧计量抽验的样本量和临界标准的确定。同时,对精确公式给出了更合理的理论解释。

Quality products do not go without inspection. K method is one of the commonly used methods in product sampling inspection by variables. As for K method sampling inspection scheme by variables under the unknown total variance, Wallis W.A.gave the result of using normal distribution to approximate the t distribution in 1947, which has been used up to now. Some Chinese scholars made some improvements in 2014 and last century, but there exist such defects as inflexibility and cumbersome process in the parameter setting. This paper relies on the accurate formula to give a more flexible and more accurate iterative solution to help improve the sample size in sampling inspection by variables with a single specification limit and the establishment of the critical standard. Meanwhile, the paper offers a more reasonable theoretical explanation about the exact formula.