一种有约束的神经网络预测控制方法

针对非线性系统的控制问题，提出了一种采用黄金分割法的神经网络预测控制算法。该算法以神经网络作为预测模型，以黄金分割法用于优化控制器，其中以系统输入的约束条件作为黄金分割法的动态搜索区间。该算法解决了控制量的范围和变化速度受约束的情况下，未知非线性系统的预测控制问题，通过仿真研究证明了该算法计算速度、稳定性和杭扰动能力。

A Strategy of Neural Network Predictive Control with Constraints

For the control of nonlinear system, a neural network predictive control strategy based on Golden Section Algorithm is presented. The neural network is used as predictive model. Golden Section is used in the optimization of controller. The variable searching region is set by the constraints of system inputs. The strategy can be used to unknown nonlinear systems with control variable constraints of range and rate. Computer simulation results demonstrate the calculate speed, stability and the ability to restrain disturb.