ONE-SIDED BAYESIAN S2 CONTROL CHARTS FOR THE CONTROL OF PROCESS DISPERSION IN FINITE PRODUCTION RUNS

Abstract

In the industrial setting characterized by short run processes, adaptive Bayesian control charts have been demonstrated to be an efficient means to perform the on-line process monitoring. Up to now, Bayesian charts have been implemented on manufacturing processes to monitor attribute variables and the sample mean. In this paper the Bayesian approach is extended to the control of the process data dispersion: several adaptive schemes of one-sided Bayesian charts monitoring the sample variance S2 have been designed. The design of each Bayesian S2 chart is performed to achieve an economic goal: the minimization of the total quality cost incurred during the production run. The quality economic performance of the Bayesian S2 charts is compared with that of a static Shewhart S2 chart and two other strategies which do not require sampling from the process. A comprehensive sensitivity analysis has been carried out to investigate the positive effects on costs deriving from the adoption of the different adaptive policies. Finally, some practical guidelines are suggested to help decision makers to select the best performing strategy to be employed within their own industrial environment.