Control charts and feedback adjustments for a jump disturbance model

We compare the performance of repeated process adjustments to postponed (bounded) adjustments for a model other than the widely used, normal-theory integrated moving average model. We generalize to an interesting, yet practical, alternative jump disturbance model. This model allows the process mean to remain constant for a random time period before a disturbance changes it (and includes a parameter that specifies the rate of mean changes in the process). Because the postponed adjustment can be based on a control chart, these strategies are included. Known analytical results are summarized and simulation is used to extend the performance analysis of different control strategies. We argue that for many process-operating conditions adjustments can be postponed, but that the magnitude of the adjustment (the diligent reaction) is a critical component for success.