Accuracy Analysis of the Estimated Process Yield Based on Spk

Process yield has been the most basic and common criterion used in the manufacturing industry as a base for measuring process performance. Boyles considered a measurement formula called S_pk, which establishes the relationship between the manufacturing specification and the actual process performance, providing an exact (rather than approximate) measure of process yield. Unfortunately, the sampling distribution and the associated statistical properties of S_pk are analytically intractable. In this paper, we consider the natural estimator of the measure S_pk. We investigate the accuracy of the natural estimator of S_pk computationally, using a simulation technique to find the relative bias and the relative mean square error for some commonly used quality requirements. Extensive simulation results are provided and analyzed, which are useful to the engineers for factory applications in measuring process performance. Copyright © 2004 John Wiley & Sons, Ltd.