A control chart for monitoring process mean based on attribute inspection

This article proposes a new control chart, namely the MON chart, which employs attribute inspection (inspecting whether units are conforming or nonconforming) to monitor the mean value of a variable characteristic x. A unit is classified as nonconforming if the value of x falls beyond a fixed warning limit. A sample is regarded as suspect if more than m out of n units (referred to as MON) in the sample are nonconforming. A MON chart produces an out-of-control signal when the interval between two suspect samples is smaller than a control limit. The MON chart is distinctively advantageous owing to its simplicity in implementation. In particular, the MON chart uses attribute inspection and eliminates the need for any computation. In addition, the MON chart makes use of information not only about the magnitude of x, but also the interval between two suspect samples. Therefore, it always outperforms the X chart and often excels the CUSUM chart on the basis of same inspection cost. Furthermore, the MON chart performs more uniformly over a wide range of mean shift than other charts.