摘要
为了使传统的BP神经网络预测控制的收敛速度更快、准确率更高,提出一种改进的人工鱼群算法。分别用BP神经网络、PSO-BP神经网络和IAFSA-BP神经网络来优化预测控制系统的建模部分和滚动优化部分,并进行仿真试验,结果表明:IAFSA-BP神经网络优化后的预测模型精度更高,并且滚动优化部分的响应速度加快,控制系统更稳定。
For purpose of obtaining a faster convergence rate and higher accuracy of the traditional BP neural network predictive control, an improved artificial fish swarm algorithm was proposed. Adopting BP neural network, PSO-BP neural network and IAFSA-BP neural network to optimize the modeling part and rolling optimization part of the neural network predictive control system respectively and then having it simulated to show that, the predictive model optimized by IAFSA-BP neural network has higher accuracy and the rolling optimization part’s responding speed is faster along with the more stable control system.
作者
黄丽华
李俊丽
HUANG Li-hua;LI Jun-li(Faculty of Information Engineering and Automation, Kunming University of Science and Technology)
出处
《化工自动化及仪表》
CAS
2019年第8期610-614,共5页
Control and Instruments in Chemical Industry
关键词
BP神经网络
预测控制
优化
改进人工鱼群算法
极值寻优
BP neural network
predictive control
optimization
improved artificial fish swarm algorithm
extremum optimization