Track: Statistical Process Control

Abstract

Different types of run rules are proposed and applied to the control chart to identify unnatural variations as early as possible. Two opposite views are present in literature regarding pattern identification by run rules; one view in favour and the other view criticize the application of run rules to pattern identification. In this paper Nelson’s run rules are investigated in detail for sensitivity and identification of control chart patterns. The goal is to gain deeper insights on the opposite views about the run rules. The rules are applied individually and in combinations to fully developed patterns. The results confirm that most of the run rules individually do not identify specific pattern. However, the runs rules can be used to identify specific patterns when applied in proper combination. The results suggest that proper combinations of Nelson’s run rules can lead to effective identification of shift, trend, cycle, and systematic control chart patterns.