摘要
在基于速率反馈的网络流量控制系统中,信元在网络中的传播时延,特别是网络流量控制中的非线性大滞后,会带来极大的网络拥塞和数据丢失。针对流量控制系统中存在的非线性大滞后和不确定性,提出了一种新的基于智能前馈控制策略的组合控制器。前馈控制部分由两个神经网络分别实现逆模型辩识与直接逆控制,可以在线调整网络权值;基本控制器由PID和Fuzzy PID实现分段控制,根据不同误差变化范围调整控制组合。仿真表明本方案能使信元发送速率快速响应网络变化,特别对于大滞后对象,控制的适应性和鲁棒性更好。
The propagation delay of cells, especially the nonlinear large-lag of communication may create congestion and loss of data in rate-based flow control in networks. Proposed in this paper is a kind of feed-forward combination controller, which can better overcome the adverse effect caused by the large time delay and its indeterminacy. The feed-forward control part consists of two neural networks,one realizes identification of inverse model,and the other is used for inverse control. The basic control part is made up of PID and fuzzy PID controller which can self-switch between them according to various stages of output error so as to adjust combination of flow control in network. The simulation shows that the scheme can make source rates respond to the changes of network status rapidly, avoid the congestion. It has much better adaptability and robustness in large time delay control system.
出处
《贵州工业大学学报（自然科学版）》
CAS
2004年第5期99-102,共4页
Journal of Guizhou University of Technology(Natural Science Edition)
关键词
信元
流量控制
滞后
逆模型辩识
神经网络
cell
flow control
lag
identification of inverse model
neural network
