监控一系列均值漂移的广义累积和控制图
new generalized CUSUM control chart for detecting a range of mean shifts
本文提出监控一系列均值漂移的广义累积和(GCUSUM)控制图, 不依赖于事先设定的漂移大小.同时为了简化运算提出了对GCUSUM的修正(MGCUSUM)并研究了参数影响. 通过蒙特卡洛模拟比较GCUSUM和广义的指数滑动平均(GEWMA),自适应累积和(ACUSUM)以及双累积和(DCUSUM), 结果显示所提出的控制图对各种漂移都很有效且稳健. 最后通过一个化学工厂的实际数据例子来说明了所提方法的应用并提供了平均运行步长(ARL)的表现.
In this paper, a generalized cumulative sum (GCUSUM) control chart,which does not depend on the reference value nor prespecified shiftsize, is developed for detecting a range of mean shifts. Amodification to the GCUSUM (MGCUSUM) control chart is also presentedto ease the computation and the parameter effect is studied indetail. We compare the GCUSUM with generalized exponentiallyweighted moving average (GEWMA), adaptive cumulative sum (ACUSUM)and dual cumulative sum (DCUSUM) based on Monte-Carlo simulation.Our simulation results show that the proposed chart provides quiteeffective and robust detecting ability for various of shifts withdifferent magnitude. The application of our proposed method isillustrated by a real data example from chemical process control andthe average run length (ARL) performance is also presented.
罗陨照、王兆军、周勤
自动化基础理论计算技术、计算机技术
数理统计平均运行步长累积和统计过程控制
Mathematical StatisticsAverage run lengthCumulative sumStatisticalprocess control.
罗陨照,王兆军,周勤.监控一系列均值漂移的广义累积和控制图[EB/OL].(2014-11-18)[2025-08-02].http://www.paper.edu.cn/releasepaper/content/201411-318.点此复制
评论