Tuning MPC for desired closed-loop performance for SISO systems

Model Predictive Control (MPC) is widely used in the process industries. A successful implementation of MPC involves the setting of a considerable number of parameters which must be appropriately tuned. Despite a number of publications on MPC tuning, there is a lack of a systematic approach that relates MPC tuning to linear control theory. In this paper, a systematic tuning of the prediction horizon, the control horizon and the penalty weights is discussed such that desired closed-loop pole and zero locations result as long as the constraints are not active. We verify our approach on two challenging examples.