A study on a multivariate EWMA controller

Many manufacturing processes have, by nature, Multiple Input and Multiple Output (MIMO) variables. However, despite its frequent use in practice, MIMO process feedback control has not been fully investigated in the literature. Under a linear MIMO model, this paper proposes a Multivariate Exponentially Weighted Moving Average (MEWMA) controller. The stability conditions of this controller and the feasible region of its associated discount factor are derived. A discount factor which minimizes the total Mean Square Error (MSE) of the process output is obtained under the assumption that the process disturbance is either a white noise series or an IMA(1,1) series. An example is given to demonstrate the applicability of the proposed method. Some related issues are also discussed.