There are two major types of variations in processes that affect the product characteristics: one is special cause variation and theother is common cause variation. A process is considered in control in the presence of only common cause ariations, but thepresence of special cause variations brings it out of control. Control charts are famous tools to differentiate between these two states of a process (Shewhart1 ). Shewhart control charts are mostly used to detect large shifts in location and/or dispersion parameters. On the other hand, the exponentially weighted moving average (EWMA) control chart and the cumulative sum (CUSUM) control chart are popular for small to moderate shifts (cf. Roberts2 and Page,3 respectively).
There is a variety of literature on these types of charts for efficient monitoring of process parameters and improving the quality of the process outputs. In order to enhance the detection abilities of different kinds of charts, researchers have suggested certain modifications in the literature. Lucas4 proposed a combined Shewhart-CUSUM quality control scheme for efficient detection of small and large shifts. Lucas and Saccucci5 recommended a combined Shewhart-EWMA control chart for improved performance. Lucas and Crosier6 proposed fast initial response (FIR) CUSUM charts that provide a head start to the CUSUM statistics, and similarly, Steiner7 proposed FIR EWMA. Yashchin8 proposed the weighted CUSUM in which he assigned weights to the past information in CUSUM statistics. Riaz et al.9 used the idea of runs rules to enhance the performance of the CUSUM control charts for small to large shifts. Riaz et al.10 implemented different runs rules schemes, and Mehmood et al.11 used a variety of ranked set strategies to enhance performance of Shewhart charts. Abbas et al.12 applied the runs rules idea for the EWMA charts. Recently, Abbas et al.13 introduced the design structure of a mixed EWMA-CUSUM (MEC) control chart for improved monitoring of the process parameters. In the said MEC chart, the EWMA statistic is used as the input for the CUSUM structure. In this study, we propose a reverse version of this mixing, that is, a mixed CUSUM-EWMA (MCE) control chart. In this new setup, the CUSUM statistic will serve the input for the EWMA structure.
The organization of the rest of this study is as follows: Section 2 describes the classical CUSUM and classical EWMA control charts, the MEC of Abbas et al.,13 and the proposed scheme denoted by MCE; section 3 contains the explanations of the different measures that will be used to evaluate the performance, the comparisons using these performance measures, and some graphical presentations; in section 4, an example with a real data set is given for the practical aspects of the proposed scheme; finally, we conclude the article with a Section 5.