T-S model-based predictive control for nonlinear systems based on min-max optimization

In order to cope with change or Sudden change of the model parameter of nonlinear process under different circumstances, a new fuzzy predictive control is proposed in this paper based on fuzzy T-S modeling and min-max optimization. A series of models are applied in prediction horizon in the replacement of single prediction model. Considering the unknown effect of future nonlinear charaeteristics, the open optimization problem in predictive control is solved through a feedback min-max optimization to satisfy the requests of nonlinear systems. To relieve computation burden, the traditional quadratic cost function is replaced with a new one, which makes it possible that the optimization problem can be solved with linear programming. The robust property of controlled nonlinear system is improved due to the including Sudden change of the model parameter in the design of controller. The efficiency of the proposed algorithm is proven in simulation.