Quality Evaluation Plan Using Adaptive Kalman Filtering

An important function of the Bell Laboratories Quality Assurance Center and the Western Electric Quality Assurance Directorate is to 2081 audit the quality of the products manufactured, and the services provided by the Western Electric Company to determine if the intended quality standards are met. This is achieved by dividing the products and services into some 3000 homogeneous classes. A small sample is taken from each class during each period (there are eight rating periods in a year). Based on this data, an inference is made in each period regarding the compliance of each class to the quality standard. Until the sixth period of 1980, the £-rate system, evolved from the work of Dodge and others,1 was used to rate the product quality. Starting with the seventh period of 1980, the Quality Measurement Plan (QMP) was implemented. The QMP, developed by A. B. Hoadley,2 is based on an empirical Bayes model of the audit-sampling process. It uses the current and the preceding five periods of data. It represents a considerable improvement in the statistical power for detecting substandard quality as compared with the old rules based on the trate. However, QMP ignores the time order of the observations, so it is less sensitive to drifts in the process mean. The Quality Evaluation Plan (QEP) has been designed to take into account the time order of the data and to be more sensitive to drifts in the process mean. The object of this paper is to present the Quality Evaluation Plan, which uses the entire time series of data on a given class to determine if that class meets the quality standard.