LS-SVM predictive control based on PSO for nonlinear systems

For the predictive control of nonlinear systems, we present a single-step predictive control algorithm based on model learning and particle swarm optimization (PSO). The method utilizes least square support vector machine (LSSVM) to estimate the model of a nonlinear system and forecast the output value, reducing the error in output feedback and error correction. The control values are obtained by the rolling optimization of PSO. This method can be used to design effective controllers for nonlinear systems with unknown mathematical models. For univariate and multivariate nonlinear systems, simulation results show that the predictive control algorithm is effective and has an excellent adaptive ability and robustness.