Charts for Exponentially Weighted Moving Average Classical and Robust: A Comparative Study

Abstract

This research discusses the comparison between two control charts classical & robust for both types (EWMA-SMQ) and (EWMA-SM) for the exponentially weighted moving average, which is showed that the robust (EWMA-SMQ) chart for the process is a superior alternative to the (EWMA-SM) chart when the outliers are present in the data. Generally, the (EWMA-SMQ) chart enables easier detection of outliers in the subgroups and is also more sensitive to other forms of out-of-control situations when outliers are present. Hence, the (EWMA-SMQ) becomes a preferred alternative to be taken and applied in quality control. The data used in this research represent the weights of the Al-Sabah Iraqi newspaper published by the Iraqi media networks. Twenty-five samples were taken and each sample consisted of five observations. These samples were taken at different production times, as the average weight of the newspaper was approximately (150) g. Through the application of the classical control chart (EWMA), exit four points for the upper and lower control limits, and the application of (EWMA-SM), (4) points are out of the upper and lower limits of the control; the control limits have been extended from the top and bottom sides of the (EWMA-SM) chart, making it less sensitive to the diagnosis of the shift in the mean process. The (EWMA-SMQ) chart detected extra points out of control. There were (7) points out of the upper and lower limits for the control, tight control limits of the (EWMA-SMQ) chart from the top and bottom sides. Due to this narrowness, three additional points were detected. Therefore (EWMA-SMQ) chart is more robust than the (EWMA-SM) control chart. This research aims to make a comparative study between the classical and robust for both types (EWMA-SM) & (EWMA-SMQ) for the exponentially weighted moving average control chart. The most important result of the research that the (EWMA-SMQ) chart is more suitable than the (EWMA-SM) & classical (EWMA) when outliers are present in the data. This gives importance to it to control the quality of the product. By this robust chart, the outliers must be detected, investigated and the special cause removed if possible. The presence of outliers will reduce the sensitivity of a control chart. The (EWMA-SMQ) control chart is more sensitive to out-of-control conditions when outliers are present in data. The limits computed from the estimate of the interquartile ranges for the (EWMA-SMQ) chart are less influenced by outliers than the (EWMA-SM) chart where the limits are computed based on the sample ranges. Thus the (EWMA-SMQ) chart is more robust than the (EWMA-SM) chart.