Robust Control Charts for Process Dispersion Based on Jackknife Estimates

Control chart is a powerful technique that can be of immense use in an industrial process in many ways like detecting assignable causes of variation from a target value, or giving an early warning on quality deterioration, or identifying outliers, or a continuous monitoring of the process. There are many parametric and nonparametric methods of developing control limits, in the literature. This article provides two general nonparametric procedures to develop dispersion control charts. First one is a general method to develop robust control limits using jackknife samples and the latter, is based on confidence intervals using jackknife sampling methods. Performance comparison of these limits is studied on simulated data. Average Run Length is also considered for comparison.