model is CUSUM chart that gives the lowest ARL1, and for MA(1) model is EWMA model, and is the best choice compared with other models. Shewhart chart is the most efficient chart design for ARMA(1,1) model.
4.2. Relative efficiency
The relative efficiency of the models have been estimated from Table V by taking the relative distance of ARL from the reference case
(i.e., in-control case) by estimating the ratio ARL1 . The best model is that it gives us the lowest ratio. For the cases of weak ARL0
autocorrelation and strong autocorrelation, these have been plotted in Figures 7–9 and Figures 10–12, respectively.
From the previous results that we gained by numerical comparisons and the used figures, the control charts’ efficiency with
respect to different models is summarized in Table VI.
4.3. Robustness
To investigate the robust of a control chart under time series processes, at first, estimate the control limits from a white noise process, then estimate the ARL of the model under different control charts. The more robust control chart is that gives ARLs closer to that we obtained under the white nose state (Figures 13–15).