Model predictive control
Model Predictive Control, or MPC, is an advanced method of process control that has been in use in the process industries such as chemical plants and oil refineries since the 1980s. It improves on standard feedback control by predicting how a process such as distillation will react to inputs such as heat input. This means that feedback can be relied on much less since the effects of inputs will be mostly known ahead of time. Feedback is still used to correct for model inaccuracies, since no mathematical model is perfect and therefore cannot completely predict future values. The controller relies on an empirical model of a process obtained by plant testing to predict the future behavior of dependent variables of a dynamical system based on past moves of the independent variables. It usually relies on linear models of the process. The major commercial suppliers of MPC software in the US are Aspentech and Honeywell.