Enhancing the performance of EWMA charts

Control charts are extensively used in processes and are very helpful in determining the special cause variations so that a timely action may be taken to eliminate them. One of the charting procedures is the Shewhart-type control charts, which are used mainly to detect large shifts. Two alternatives to the Shewhart-type control charts are the cumulative (CUSUM) control charts and the exponentially weighted moving average (EWMA) control charts that are specially designed to detect small and moderately sustained changes in quality. Enhancing the ability of design structures of control charts is always desirable and one may do it in different ways. In this article, we propose two runs rules schemes to be applied on EWMA control charts and evaluate their performance in terms of the Average Run Length (ARL). Comparisons of the proposed schemes are made with some existing representative CUSUM and EWMA-type counterparts used for small and moderate shifts, including the classical EWMA, the classical CUSUM, the fast initial response CUSUM and EWMA, the weighted CUSUM, the double CUSUM, the distribution-free CUSUM and the runs rules schemes-based CUSUM. The findings of the study reveal that the proposed schemes are able to perform better than all the other schemes under investigation. Copyright © 2010 John Wiley & Sons, Ltd.