Due to fast receding horizon speed and global optimization of small-world optimization algorithm with real-coding (RSW), a new nonlinear multi-objective predictive controller was presented based on RSW and neural network (NN) identification trained by BP. NN model which was obtained by off-line identification was used to predict the present and future output of the plant, and RSW was applied to receding horizon control. Finally, an application to 500MW unit load control system with multi-objective optimization was given, and simulation results indicated the effectiveness of this new approach.
A Receding Horizon Multi-Objective Planner for Autonomous Surface Vehicles in Urban Waterways