An EWMA Control Chart With Practical Importance Of Data

In many applications,if the parameter difference before and after the change point is small,the alarm signal of the control chart may not be needed.Traditional hypothesis testing is highly sensitive to some unnecessary minor changes,and the high false alarm rate greatly reduces the monitoring effect of control chart.In order to solve the frequent alarm problem caused by minor changes of parameters before and after the change point,this paper studies how to design the change detection with practical importance.The methodology is applied to multivariate exponential weighted moving average(MEWMA)control chart and multivariate exponential weighted moving covariance(MEWMC)control chart,and the control limits under different tolerances and the average running length(ARL)under different shifts are obtained by numerical simulation.The results show that the control charts with practical importance designed in this paper can not only avoid unnecessary alarms in the monitoring process,but also quickly detect changes with practical importance.