Support vector machine based U-model generalized predictive controller for nonlinear dynamic plants

In this study, least-squares support vector machine (LS-SVM) is integrated into the generalized predictive controller based on U-model (a control-oriented prototype) for a wide range of nonlinear dynamics plants. Mathematically the final solution of the controller output is converted into resolving a polynomial equation in terms of current controller output, LS-SVM is used to obtain the current controller output, based on strong nonlinear function fitting ability, which dramatically reduces the difficulties encountered in nonlinear control system online implementation. For demonstration of the effectiveness of the design procedure, this method is applied in continuously stirred tank reactor.