自适应指数加权移动平均指数控制图设计  

Design of an Adaptive Exponential Weighted Moving Average Exponential Control Chart

作　　者：平安 孙金生[1] 胡雪龙[2] 薛丽[3] Ping An;Sun Jinsheng;Hu Xuelong;Xue Li(School of Automation,Nanjing University of Science and Technology,Nanjing 210094,China;Institute of High Quality Development Evaluation,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;School of Information Management,Zhengzhou University of Aeronautics,Zhengzhou 450046,China)

机构地区：[1]南京理工大学自动化学院,南京210094 [2]南京邮电大学高质量发展评价研究院,南京210003 [3]郑州航空工业管理学院信息管理学院,郑州450046

出　　处：《统计与决策》2021年第22期27-31,共5页Statistics & Decision

基　　金：国家自然科学基金资助项目(71802110,71701188);江苏高校哲学社会科学重点研究基地(信息产业融合创新与应急管理研究中心)资助项目;江苏省哲学社会科学优秀创新团队项目((2017ZSTD022))。

摘　　要：现代制造过程中,某些产品的不合格率非常低,这类过程通常称为高质量过程,此时过程中相邻不合格产品之间的时间间隔服从指数分布。为了监控这类高质量过程,文章提出一种自适应指数加权移动平均(AEWMA)指数控制图。仿真结果表明,此控制图的平均运行链长(ARL)与Shewhart控制图、指数加权移动平均(EWMA)控制图相比具有优越性,尤其对整个区间内的偏移具有较好的检测性能。In modern manufacturing,the defect rate for some products is very low.Such processes are often referred to as high quality processes where the time interval between adjacent nonconforming products follows an exponential distribution.In order to monitor this kind of high-quality process,this paper puts forward an adaptive exponential weighted moving average(AEWMA)exponential control chart.Simulation results show that the average running length(ARL)of the control chart is superior to Shewhart control chart and exponential weighted moving average(EWMA)control chart.Especially,it has good detection performance for the offset within the whole range.

关 键 词：自适应指数加权移动平均 指数分布 单边控制图 平均运行链长 

分 类 号：O213.1[理学—概率论与数理统计]