基于样本广义方差的改进|S|^(1/2)控制图  

An Improved |S|^(1/2) Control Chart Based on Sample Generalized Variance

作　　者：李颖[1] 师义民[1] 赵选民[1] 王彩玲[1] 

机构地区：[1]西北工业大学理学院数学系,陕西西安710072

出　　处：《数学的实践与认识》2007年第15期82-88,共7页Mathematics in Practice and Theory

基　　金：国家自然科学基金(70471057);航空科学基金(02J53079);陕西省自然科学基金(N5CS0002)

摘　　要：以常用的广义方差为基础,利用矩估计的方法,从一组样本推广到多组独立样本,得出具有无偏控制限的改进|S|^(1/2)控制图,并对比说明了矩估计方法的简洁性和准确性.与标准|S|^(1/2)控制图相比,其具有更小的平均运行链长(ARL).结合实例,采用数值模拟的方法说明该控制图对过程中方差的漂移有较强的检测能力.Based on the sample generalized variance, this paper uses moment estimate to construct an improved √｜S｜ control chart which have unbiased control limits in the case of independent multi-samples. It is shown that the method of moment estimate is concise and accurate. The improved √｜S｜ control chart has smaller ARL than the standard √｜S｜ control chart. An example accompanied by simulation is used to illustrate that this chart is more sensitive to shifts in the process variability.

关 键 词：√|S|控制图 样本广义方差 无偏估计 过程方差 

分 类 号：O213.1[理学—概率论与数理统计]