具有可变样本容量的非正态EWMA控制图

摘要

为了提高控制图的监控效率，本文研究非正态分布下，E WMA控制图的可变样本容量设计问题。首先利用 Burr分布近似各种非正态分布，构造可变样本容量的非正态 EWMA控制图；其次运用马！科夫链法计算可变样本容量非正态 E WMA 控制图的平均运行长度；然后与传统的非正态 E WMA 控制图进行比较得出：当过程中出现小波动时，可变样本容量的非正态 E WMA控制图能够更快地发现过程中的异常波动，具有较小的平均运行长度，其监控效率明显优于传统的非正态 E WMA控制图。%In order to improve the monitoring efficiency of control charts,the EWMA control chart under non-normal distribution with variable sample size (VSS )is constructed in this paper;The Markov chain method is applied to calculate the average run length(ARL)of the VSS EWMA control charts.The computing results show that the VSS EWMA control chart under non-normal distribution is the more efficient in detecting shifts than the traditional EWMA control chart,and has a shorter the average run length to find the abnormal fluctuation.