Development and Performance Evaluation of a Mixed DEWMA–MA Control Chart for Detecting Shifts in Standard Deviation

Abstract

This study introduces a Mixed Double Exponentially Weighted Moving Average–Moving Average (DEWMA–MAS) control chart for monitoring process variability, with a focus on detecting shifts in standard deviation. By combining the long-term memory of the
DEWMAS chart with the short-term smoothing of the MAS chart, the proposed method improves sensitivity to small and moderate variability changes. Simulation results, evaluated using ARL, SRL, and MRL metrics, show that larger moving average windows enhance detection of moderate shifts, while smaller 𝜆 values are more effective for small shifts. Compared with conventional S, MAS, and DEWMAS charts, the DEWMA–MAS consistently demonstrates superior performance, particularly for moderate shifts, and remains competitive for large shifts. A soft-drink filling case study further validates its practical advantages, confirming the DEWMA–MAS chart as a more effective and reliable tool for modern industrial variability monitoring.