Constrained predictive control using neural network

Efficient control algorithms are important for industrial processes with constraints. The presence of constraints often results in nonlinear optimization problems which are difficult to solve online. A constrained neural predictive controller is presented, in which the internal dynamics of the neural network will drive the states of neurons to their steady states just corresponding to the constrained predictive control law. The structure of the computer control system with the neural predictive controller is also investigated.<
>