Discussion

The short run problems is ill-posed since with few data, it is not possible to have as small false alarm rate and as efficient shift detection properties as in the known parameter case. There are inherently trade-offs between false alarm rate, and the shift detection capabilities. The real issue is that different methods for short run statistical process control (SPC) will provide different trade-offs. This paper focuses on Q charts and in general on statistical process control methods for short run processes.