一种改进的主动队列管理算法

An Enhanced Active Queue Management Algorithm

传统的主动队列管理算法(AQM:Active Queue Management)存在响应时间较长等问题,PID(Proportional Integral Differentia1)主动队列管理算法对此作出了一定改进,然而在时延较大时也不能使队列长度收敛到期望值。本文利用BP神经网络自适应控制的特点,针对无线信道(TCP Westwood)提出了一种基于BP神经网络整定的PID主动队列管理算法。

The AQM（Active Queue Management） algorithm have too long respond time and some other problems. Though it makes some enhancement, the PID （Proportional Integral Differential） active queue management algorithm can not make the queue length converge to expectation with longer delay. Because of self-adapt control of BP（backpropagation） neural network, this issue bring forward a BP neural network based PID active queue management algorithm on wireless channel （TCP Westwood）.