摘要
提出了采用改进的遗传算法优化比例-积分-微分(PID)神经网络解耦控制器的连接权值,从而实现PID控制器参数的优化及非线性多变量系统的解耦控制.改进的遗传算法优于基本遗传算法,它使寻优过程中的计算量减少,计算效率提高,收敛速度加快.将优化后的PID神经网络解耦控制器应用于统一混沌系统的控制中,仿真实验收到良好的控制效果,证明了PID控制器应用于统一混沌系统控制的有效性.
An improved genetic algorithm （IGA） was proposed. It can optimize the proportional-integral-derivative （PID） neural network decoupling controller＇s connecting weight value, so that it makes the PID controller＇s parameser to be optimized and realizes the decoupling control of multivariate nonlinearity systems. The IGA is superior to the elementary genetic algorithm. In the PID controller＇s parameter optimization, the IGA uses less calculations, is more efficient, and faster in convergence. When the optimized PID controller was applied to unified chaotic systems, good control results were obtained by simulation experimentation, so it was proved that the PID controller when applied to unified chaotic systems was effective.
出处
《物理学报》
SCIE
EI
CAS
CSCD
北大核心
2007年第5期2493-2497,共5页
Acta Physica Sinica
关键词
混沌系统
遗传算法
比例-积分-微分
神经网络
chaotic system, genetic algorithm, proportional-integral-derivative, neural network
