摘要
提出了异步转移模式ATM网络可用位速率ABR业务一种基于神经网络的流量控制方法 .采用基于径向基神经网络的流量控制算法可以实现在线学习 ,自适应根据流量大小的变化和网络的拥塞状况调整神经网络的模型参数 .仿真结果表明与传统的静态反馈方法相比 。
In this paper, we present a novel neural networks approach to ABR flow control in ATM networks and propose a feedback flow control algorithm based on the RBF_ANN model, which is an on_line learning ANN model and can modify the ANN model parameters to adapt the change of source rates and congestion. The results of the simulation suggest that our approach performs better than traditional st_ atic feedback control both on resource utilization and cell loss ratio.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2001年第7期51-54,共4页
Journal of South China University of Technology(Natural Science Edition)
基金
国家自然科学基金资助项目 (6 9972 0 15 )
