测量误差对方差控制图的性能影响分析  

Effects of Measurement Error on Control Chart for Monitoring Process Variability

作　　者：刘双 何川[2] 刘静 张久军[3] LIU Shuang;HE Chuan;LIU Jing;ZHANG Jiu-jun(School of Mathematics and Information Science,Anshan Normal University,Anshan 114007,China;College of Sciences,Northeastern University,Shenyang 118019,China;Department of Mathematics,Liaoning University,Shenyang 110036,China)

机构地区：[1]鞍山师范学院数学与信息科学学院,辽宁鞍山114007 [2]东北大学理学院,辽宁沈阳118019 [3]辽宁大学数学院,辽宁沈阳110036

出　　处：《数理统计与管理》2020年第6期990-999,共10页Journal of Applied Statistics and Management

基　　金：辽宁省自然科学科技基金(2019ZD-0193,2020-MS-139);辽宁省百千万人才项目;文化和旅游部信息化专项项目(xxhfzzx201904)。

摘　　要：在控制图的应用中,多数统计过程控制方法都是基于数据的精确测量,然而在生产与服务环境中,数据中存在测量误差是不可避免的。本文研究测量误差对Zhang等[1]提出的用于检测方差的控制图的性能影响。结果表明,当测量误差存在时,控制图的检测性能会受到一定的影响。文章最后给出采取多次测量方法来降低测量误差的影响。In the application of control charts,most of research works in statistical process monitoring assumes that the measurements are accurate.However,an exact measurement is a rare phenomenon in any manufacturing and service environment and the measurement error does exist.In this paper,the effects of measurement error on the performance of the control chart proposed by Zhang et al.[1]for monitoring process variance are studied in the presence of measurement error.The simulation results showed that the chart's performance can be affected when the measurement error exist.We also consider multiple measurements to reduce the effect of measurement error on the chart’s performance.

关 键 词：统计过程控制 测量误差 平均运行长度 

分 类 号：O212[理学—概率论与数理统计] O213.1[理学—数学]