监控不合格品率的MEWMA-p控制图设计  

MEWMA-p Chart to Monitor Percent Defective

作　　者：朱永忠[1] 丁辉 ZHU Yongzhong;DING Hui(College of Science,Hohai University,Nanjing 211100,China)

机构地区：[1]河海大学理学院,江苏南京211100

出　　处：《工业工程》2023年第1期130-135,181,共7页Industrial Engineering Journal

基　　金：国家自然科学基金资助项目(11771120)。

摘　　要：传统Shewhart-p控制图只对单一属性的不合格品率进行监控,在过程发生偏移时有一定的滞后性。为提高不合格品率控制图的精度,提出一种多元指数加权移动平均不合格品率(multivariate exponentially weighted moving average p, MEWMA-p)控制图。该控制图将多个属性的不合格品率应用于多元指数加权移动平均控制图,可同时对多个属性进行监控,并且对于小范围的偏移更加敏感。对比分析同等偏移程度下指数加权移动平均不合格品率(exponentially weighted moving average p, EWMA-p)控制图与MEWMA-p控制图的平均运行长度(average run length,ARL)结果,并通过模拟仿真说明该方法的有效性。The traditional Shewhart-p control chart only monitors the percent defective of a single property and has a certain lag in the process offset. A multivariate exponentially weighted moving average p(MEWMA-p) control chart is proposed to improve the accuracy of the percent defective control chart. The control chart applies the percent defective of multiple properties to the multivariate exponentially weighted moving average(MEWMA) control chart, which can monitor multiple properties at the same time and is more sensitive to small ranges of shift. The average run length(ARL) table of exponentially weighted moving average p(EWMA-p) control chart and MEWMA-p control chart is studied and compared at the same offset degree, and the validity of this method is verified by simulation.

关 键 词：休哈特不合格品率控制图 多元指数加权移动平均控制图 平均运行长度 

分 类 号：O212[理学—概率论与数理统计]