基于最新差异信息的重启EWMA控制图  

A Restarting EWMA Control Chart Based on the Latest-Difference-Information of Sample

作　　者：直雪瑶 宋学力[1] 王凯明[1] 张亚洁 

机构地区：[1]长安大学理学院,陕西 西安

出　　处：《应用数学进展》2023年第11期4733-4742,共10页Advances in Applied Mathematics

摘　　要：在统计过程控制中,随机性会延迟对过程失控的诊断,所以挖掘与失控相关的非随机信息,尽早检测出失控信号是控制图设计亟待解决的关键问题。本文通过在样本数据中设计重启机制以消除反向信息的影响,以及在图统计量中加入样本最新差异信息以量化过程最新变化的影响,设计了改进型EWMA均值控制图(ReMEWMA),实现了过程均值单侧漂移的检测。实验表明,与现有的MEWMA、EWMA和IEWMA控制图相比,ReMEWMA均值检测控制图在仿真数据和真实过程数据实验中均具有更小的失控平均运行长度(ARL1)。In statistical process control, randomness would delay the identification of out-of-control(OC) sig-nals, therefore, mining the OC-related non-random information to discover OC signals earlier is the key issue when designing a control chart. In this paper, an improved EWMA Mean control chart (ReMEWMA) is designed, in which a restart mechanism is set to the sample data to eliminate the in-fluence of reverse information, and the latest difference-information of the sample is weighted-ly-introduced into the graph statistics to quantify the influence of the latest changes of the process, and at last the detection of the potential one-sided-drifting on process mean is accomplished. Nu-merical experiments show that compared with the existing MEWMA, EWMA and IEWMA methods, the ReMEWMA mean-detection-control-chart presents a smaller average run length of out-of-control (ARL1) on both simulated and real process data.

关 键 词：统计过程控制 指数加权移动平均 单边控制图 平均运行长度 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]