ABSTRACT
This chapter gives a series of examples and demonstrates how Model Predictive Control (MPC) might be tuned quickly and effectively. One of the advantages of MPC is the ability to incorporate constraints into the optimisation. MPC however, because it can include knowledge of the constraints within the optimisation, can be far more intelligent and hence give large improvements in performance where constraints are active. In a similar vein, conventionally designed MPC controllers of unstable systems are conditionally stabilising in that the output horizon must lie between an upper and a lower limit. The basic difficulty with MPC of unstable systems is that the optimisation of is ill-conditioned for large ny and ill posed for small ny. MPC works on the basis that the part of the predictions that are ignored should be largely in steady state; this is clearly not the case for unstable systems.
