Economic Design of CUSUM Control Charts

Abstract

In statistical process control, control charts are one tool for monitoring the control status of a process. One such type of chart is the cumulative sum (CUSUM) chart which has advantages over other styles of control chart. A study of the economic design of CUSUM control charts is undertaken via a comparative study of long-run hourly cost (LRHC) and a computational search algorithm is used to minimize LRHC for a CUSUM chart using nine parameters confined to their respective feasible parameter spaces as defined by the chart designer. Savings over similarly designed two stage Xbar charts are discovered and presented.