Model based predictive control using neural network and fuzzy logic

The application of neural network structures to model-based predictive control (MPC) is studied. First, basic structures of MPC algorithms are reviewed. Next, many techniques using neural network modeling to improve structural or computational properties of MPC algorithms are presented and discussed, from a neural network model of a process in standard MPC structures to modeling parts or entire MPC controllers with neural networks. Finally, a simulation example and conclusions are given.