A comparison of EWMA-type charts under linear drifts in Poisson means

Various control charts have been recommended to detect step shift in the process mean, however, relatively less attention has been paid to monitor the linear drifts in Poisson mean. In this paper, we evaluate the ability of four EWMA-type control charts in detecting increases in the Poisson rate by using the steady-state average run length, which is the most commonly used measure in statistical process control (SPC) or health surveillance. The simulation study shows that for small drift rates, both the one-sided upper EWMA chart (referred as EZ chart) and the adaptive EWMA (AEWMA) chart with large γ value from Huber´s score function perform better than the other EWMA-type control charts; for large drift rates, the AEWMA chart with small γ value seems to behave the best amongst the four EWMA-type control charts for detecting Poisson mean drifts.