A reliable procedure for testing linear regulators with one-sided specification limits based on multiple samples

摘要

Process capability indices C_(PU) and C_(PL) are developed in the manufacturing industry to provide quantitative measures on process potential and performance, for processes with one-sided specification limits. Statistical properties of the estimators of C_(PU) and C_(PL) have been investigated extensively, but only restricted to cases with single samples. In this paper, we consider the estimation and capability testing of C_(PU) and C_(PL) based on multiple samples. We show that the proposed estimator of C_(PU) and C_(PL) are indeed the uniformly minimum variance unbiased estimators (UMVUEs). A simple procedure based on a hypothesis testing using the UMVUE is developed for the practitioners to use for judging whether a stable process meets the preset capability requirement.