OPTIMIZATION OF MEASUREMENT STRATEGY

摘要

Laboratory measurements are an integral part of process and quality management. Quite commonly, variables measured are statistically dependent and thus measurement of one variable is an indirect measurement of the others. Also, measurements are uncertain and thus there always is uncertainty about the quality. In this paper we formulate the optimal measurement scheduling problem of laboratory measurements. In particular, the objective is to minimize a weighted sum of measurement costs and uncertainty about whether the product confirms with given quality constraints or not. This problem is considered in the case when the joint probability density of variables is Gaussian, and when uncertainty evolves according to Ornstein-Uhlenbeck process.