摘要
回声状态网络储备池完全随机生成,数据预测时其参数的设置缺乏合理性,因此将果蝇优化算法应用于回声状态网络储备池参数的优化中,通过果蝇优化算法的自适应参数寻优提高回声状态网络的数据预测能力。变风量空调系统是一个多变量、强耦合和非线性的系统,为了实现变风量空调系统的智能控制,从而优化生产工艺,构建了基于果蝇优化的回声状态网络内模控制系统的正模型和逆模型。仿真实验结果表明:所提出的内模控制系统具有良好的跟踪性和抗干扰性。
The echo state networks （ESN） reservoir pool is completely random. The parameter setting is not reasonable when the algorithm is applied to data prediction. Therefore, the fruit fly optimization algorithm （FFOA） is applied to optimize the parameters of echo state networks. The predictive ability of echo state networks is improved by adaptive parameters optimization of the fruit fly optimization algorithm. Variable air volume system （VAV） is a muhivariable, strongly coupled and nonlinear system. To realize the intelligent control of VAV and optimize the production process, the positive model and inverse model of internal model control system based on ESN optimized by FFOA are constructed. Finally, the simulation results show that the proposed internal model control system has good tracking and disturbance rejection performance.
出处
《重庆理工大学学报（自然科学）》
CAS
2017年第6期120-126,153,共8页
Journal of Chongqing University of Technology：Natural Science
基金
国家自然科学基金资助项目(61502065)
重庆市科委基础科学与前沿技术研究重点项目(cstc2015jcyj BX0127)
重庆市教委科研项目(yjg143090)
关键词
回声状态网络
储备池
果蝇优化算法
内模控制
变风量空调系统
echo state network
reservoir pool
fruit fly optimization algorithm
internal modelcontrol
variable air volume system
