Dynamic Process Dispersion Monitoring Through Bayesian‐Driven Smoothing Adjustments

ABSTRACT This study introduces a Bayesian adaptive exponentially weighted moving average (AEWMA) control chart designed to monitor process dispersion in case of normal distribution by integrating various loss functions. It exhibits robust performance in identifying shifts in process dispersion across different scales. Evaluation involves Monte Carlo simulations to calculate run length characteristics and a comprehensive comparative analysis against existing charts. Our findings emphasize the increased sensitivity of the Bayesian AEWMA control chart to shifts of different magnitudes. Furthermore, an experiment was performed in the context of another field, specifically semiconductor manufacturing, to compare the performance of the proposed Bayesian control chart using different loss functions. It showed that the suggested chart was much better than the existing control charts in terms of observing out‐of‐control signals. In summary, this article develops an innovative approach with various loss function approaches and improves the accuracy and efficiency depending on the dispersion change in the Phase II process, thus it is a valuable contribution to the further development of the quality control and monitoring process.