Optimisation of fully adaptive Bayesian charts for infinite-horizon processes

This article presents a model for the economic design and optimisation of an adaptive -chart used to monitor the process mean in infinite runs. Assignable causes may randomly affect the mean of the process by shifting it from its target value to an undesirable level. The proposed model allows the determination of the scheme parameters that minimise the total expected quality cost of the procedure. The monitoring mechanism of the process employs a Bayesian chart and at each sampling instance the economically optimum design parameters of the chart are estimated using the Bayes theorem. The effectiveness of the proposed model is estimated through economic comparisons with simpler variable-parameter charts and even simpler fixed-parameter charts. It is shown that significant economic improvement can be achieved through the use of the new model.