摘要
利用神经网络的方法,研究了Internet网络路由器中的拥塞控制问题.根据过去相邻两个时刻缓存器中队列长度值的变化量来预测下一时刻路由器中队列长度值,及时调整控制增益的大小以防止拥塞的发生,该方法可以使路由器中队列长度稳定在一个期望值附近.仿真表明,该控制方法可以有效地保证网络系统中信息的平稳传输.
The congestion control in network Router is studied by Neural Network. Based on the buffer length of border two nodes ,the quese＇s length of next time in Router is predicted. Then the control gain can be turned timely to avoid network congestion. This method enables the queue length to converge at the desired steadystate value. Simulation results show that the algorithm ensure the stability of the network transmission.
出处
《鲁东大学学报（自然科学版）》
2007年第1期31-33,共3页
Journal of Ludong University:Natural Science Edition
关键词
拥塞控制
神经网络
预测控制
鲁棒性
congestion control
Neural Network
predictive control
robustness
