摘要
提出了一种基于BP神经网络预报的动态矩阵预测控制新算法 ,在该算法中 ,先用BP神经网络辨识对象模型 ,同时预测对象的未来输出 ,然后用动态矩阵控制算法进行滚动优化和反馈校正。该方法解决了非线性、时变对象难以建模及控制的问题 。
The paper proposes a new DMC algorithm using BP neural networks as predictive model.In this algorithm,BP neural networks is employed to identify model of the controlled plant and produce predictive output.Receding horizon optimization and revising feedback are used in DMC algorithm.In the method,the problem that nonlinear and time-variety plant are difficulty to build model is solved.Simulation results demonstrate the feasibility of the proposed algorithm.
关键词
神经网络
预测控制
动态矩阵控制
模型预测
neural network
predictive control
dynamic matrix control
model prediction
