Optimal Economic Design of X¯-Control Charts When Cost Model Parameters are Not Precisely Known

Abstract We suggest a method for using cost models for designing control charts when precise estimates of the cost model parameters are not available. Our method seeks control chart designs which are not sensitive to the estimates of the various cost model parameters by incorporating a measure of the imprecision of the parameter estimates into the objective function. Using Duncan's single assignable cause model for designing X¯-charts as an example, we compare our method to the case where only point estimates of the parameters are used. The basic procedure can be generalized and applied not only to other economic models for designing X¯-control charts (such as models with different process failure mechanisms and models with multiple assignable causes) but to cost models for designing other types of control charts as well as for cost models used to design economical acceptance sampling plans.