Adaptive control with NeuCOP, the neural control and optimization package

NeuCOP is the neural control and optimization package that has been jointly developed by Texaco and NeuralWare. It is a state-of-the-art, multivariable, adaptive controller that combines the nonlinear modeling power of neural networks with nonlinear optimization algorithms. The NeuCOP controller belongs to the general class of model predictive controllers. One novel feature of the NeuCOP controller is its use of a nonlinear, neural network process model. We describe the identification subsystem that has been developed. More specifically, we address the issue of system re-identification, after the system is put online. The re-identification process allows the model to adapt to changing process conditions