In this paper the EWMA and the X-bar control charts are considered for monitoring processes in which the observations can be represented as a first order autoregressive model. The charts are designed taking the serial correlation into account and the sampling strategy is set based on the rational subgroup concept. The control charts properties are studied and compared. Numerical results show that the positive correlation within-subgroup has a significant impact on the charts performance. The EWMA chart is substantially more efficient than the X-bar chart in detecting process disturbances, especially when the mean shifts are of small magnitude.
Autocorrelation; average run length; exponentially weighted moving average control chart; first order autoregressive model; statistical process control
