Generalized Predictive Control of Nonlinear System Based on Mind Evolutionary Algorithm

In order to apply the thought of generalized predictive control to nonlinear system successfully, this paper presents a way based on Mind Evolutionary Algorithm (MEA) in which nonlinear system is divided into nonlinear subsystem and linear subsystem. MEA optimizes and draws into the output of nonlinear subsystem. Linear subsystem uses CARIMA as its model and uses DLS as its on-line learning algorithm. The error of the nonlinear system model is compensated by itself feedback gain. A simulated study has proved its validity.