A optimal project of exponentially weighted moving averages charts for statistical process control

A method is presented for the design of EWMA schemes for statistical process control. The solutions obtained optimize the ratio between the speed with which the scheme detects a shift of a given magnitude and the resources allocated to inspection. The final product is a set of tables which give directly the optimal parameters, making the use of the method immediate. This simplicity, together with the indication of optimal values for the whole set of implementation parameters, is the major advantage of the method over previous work. Additional contributions of the work are: the consideration of user constraints on the sample size or sampling frequency; and a performance comparison with Shewhart schemes (in which each scheme is considered with its optimal set of parameters), enabling the user to choose the better scheme for each specific situation.