监测Lévy稳定过程均值变点的平均运行时间估计  

The Average Run Lengths of Control Charts for Stable Lévy Processes

作　　者：张晨[1] 韩东[1] 宗福季[2] 

机构地区：[1]上海交通大学数学系,上海200240 [2]香港科技大学工业工程与工程管理系

出　　处：《应用概率统计》2007年第4期384-394,共11页Chinese Journal of Applied Probability and Statistics

基　　金：This work was supported by the Key Project of Science and Technology Committee of Shanghai(03JC14050);RGC Competitive Earmarked Research Grants HKUST6011/01E.

摘　　要：平均运行长度(时间)(ARL)是判断一控制图监测变点效果好坏的一个重要工具.本文主要研究Lévy稳定过程的均值变点监测问题.我们给出了三个控制图,即EWMA,GEWMA和GLR的ARL估计,并通过数值模拟比较了4个控制图监测均值变点的效果和差异.As we know that the average run length （ARL） is an extensively used measure in statistical process control （SPC） for evaluating and comparing the detection performance of various control charts. In this paper we not only present the asymptotic estimation of the ARL for the exponentially weighted moving average （EWMA）, generalized EWMA （GEWMA） and generalized likelihood ratio （GLR） control charts but also compare the detection performance by the numerical simulation among the four charts： EWMA, GEWMA, GLR and CUSUM in detecting the mean change of a stable Levy process.

关 键 词：变点监测 平均运行长度 Levy稳定过程 EWMA GLR GEWMA CUSUM. 

分 类 号：O213.1[理学—概率论与数理统计]