New Quality Control Chart to Quickly Detect the Changes of Process Average

Abstract

The objective of this article is to propose a new control chart-improved exponentially weighted moving average (IEWMA) control chart-to fast detect the mean shifts of process when quality characteristic data are normally distributed. This chart still has robust property even though its controls limits are created from data with outliers. The efficiency inspection of IEWMA control chart is managed 504 situations for the simulation data. Moreover, the four control charts, namely, exponentially weighted moving average (EWMA), robust exponentially weighted moving average (REWMA), median mean absolute deviation (MDMAD), and average control charts, are compared the performances with IEWMA control chart. All charts are constructed by using data set in two cases, i.e., the first case that data are not include outliers and the second case that data are composed of outliers. It is found that in the case of non-outliers in the data, the three charts-IEWMA, EWMA and REWMA control charts-tend to have the most capability for process average shift detection for all sample sizes and all levels of the process average changes. For the case of outliers in the data, the IEWMA control chart tends to have the most efficiency for all sample sizes, especially for the tiny process shifts from the target.