ABSTRACT
This chapter discusses the exponentially weighted moving average chart which is similar to CUSUM but is designed so older measurements have less influence on the test statistic. Full adjustment is perhaps the simplest form of proportional control and, as shown by the funnel experiment, it increases rather than decreases the variation in the system. The explicit purpose of the damping factor G is to avoid overadjustment but, in the absence of anything but common cause variation, proportional control will nonetheless increase overall variation. The details of this kind of adjustment scheme are beyond the scope of this book but the valuable takeaway is that it is indeed possible to adjust systematic forms of time series-related variation out of a process if the time series behavior can be quantified. Proportional control can be combined with integral and derivative control to obtain proportional-integral-derivative control.
