By Panagiotis Tsiamyrtzis
University of Minnesota
August 31, 2000
Standard methods for Statistical Quality Control (SQC) require a substantial data gathering to establish a pattern against which further samples may be compared. Clearly, short production runs would severely affect the performance of such methods.
In this work we consider SQC processes where the total number of stages is small (short runs). The case where significant changes need to be detected as soon as they occur and not once the process is terminated, makes the short run problem even more complicated. In this research a statistical model to handle short runs is proposed and is also shown to generalize popular methods used in other fields of Statistics for detecting shifts in a parameter.
The theoretical development is based on a Bayesian updating procedure of a mixture of Normal distributions.