Analytic Model Predictive Controller in Simple Symbolic Form

Abstract

Paper deals with an analytic solution of Model Predictive Controller in simple symbolic form. Process is approximated with a first order dynamical model. Special choice of prediction and control horizons is considered, so the symbolic solution is still applicable, and the controller has interesting "predictive" feature in case of known future set-point course. Such a controller can be used in simple devices like PLCs or microcontrollers without need of matrix operations. Its advantage is that the controller reacts to the process model parameters and penalty parameter change so the control can be very fast and efficient even in adaptive manner.