Nonlinear Adaptive Control Of Mineral Processes

The modelling of phenomena by nonlinear dynamic Systems Theory representations require the specification and identification of suitable models of the phenomenon and the availability of good optimization algorithms to determine the optimal controls. Stochastic disturbances and difficulties of monitoring the process lead invariably, otherwise, to suboptimal performance. It will be shown that an algorithm can be formulated which will simultaneously identify and optimize such a system. The properties of the algorithm will be derived and its convergence to an optimal control proved. Some results of actual applications will be presented and it will be shown how this approach generalizes easily to optimal control of the actual processing system. These results generalize those of optimal learning and input conditioning.