Simulation Study of the Generalized Predictive Control Algorithm Based on BP MFN´s Model

For the present, there are still difficulties to construct many steps predictive models and its control laws for the strong nonlinear systems, though the generalized predictive control (GPC) algorithm had been obtained well control effect to the linear or weak nonlinear systems. For solving the questions, based on the high nonlinear mapping property of BP multilayer feedforward network (MFN) model, the control method which is GPC´s algorithm with the artificial neural network technologies is studied. The GPC´s control structure scheme based on BP MFN´s model for the nonlinear system is designed. The GPC´s control principle, control algorithm and setting different parameters of the GPC criterion function are analysed. Finally, simulation researches of GPC´s algorithm for the nonlinear system based on the BP MFN´s model have done. Simulation results show, the GPC´s control structure scheme based on BP MFN´s model is effective and GPC´s algorithm based on BP MFN´s model is successful