摘要
针对工业生产过程中的多变量耦合系统难以实现解耦的问题,建立了一种改进的规划算法的RBF神经网络逆系统,构造了多变量神经网络控制器,用来对多变量耦合系统进行解耦控制。对一组给定的二变量耦合系统进行了仿真,结果表明:基于改进的进化规划算法的RBF神经网络逆系统的解耦控制不仅超调量小、响应速度快、控制精度高,而且具有很强的鲁棒性和自适应能力。该解耦控制使得解耦后的多变量系统具备良好的动、静态特性,达到了理想的控制效果。
Since it＇s very difficult to decouple for multivariable coupling system in the industrial process control field, an inverse system of RBF neutral network was set up based on an improved evolutionary programming algorithm. Then, the multivariable neutral network controller was constructed, which can be used to decouple control for the multivariable coupling system. A double variable coupling system was simulated in computer. The simulation results show that RBF neutral network inverse system based on an improved evolutionary programming algorithm works well with small overshoot, quick response, high control precision, good robustness and adaptive ability. This decoupled multivariable control system has better static and dynamic performance and attains the ideal control effect.
出处
《交通信息与安全》
2009年第2期59-62,共4页
Journal of Transport Information and Safety
关键词
解耦控制
RBF神经网络
逆系统
decoupling control, radial basis function neural network
inverse system
