改进的网络拥塞控制策略算法研究  

Simulation and Application on Wireless Network Congestion Control

作　　者：朱春[1] 王赛云[1] 

机构地区：[1]浙江万里学院计算机与信息学院,浙江宁波315100

出　　处：《计算机仿真》2011年第12期141-144,共4页Computer Simulation

摘　　要：研究网络拥塞优化控制问题。针对网络承载量的不断增加,使得网络传输效率降低。传统网络拥塞控制算法要求系统根据无线网络的容量实时编号,动态调整TCP拥塞窗口的大小,难于建立准确的数学模型,从而导致网络带宽利用率低,网络拥塞严重。为了降低网络拥塞的概率,提出了一种改进的无线TCP拥塞控制算法。算法主要是集中解决在TCP拥塞窗口的大小调整问题上,首先利用BP神经网络对参数进行训练,有效地解决了TCP拥塞窗口大小的调整,从而实现了拥塞避免、快速重传和快速恢复机制,改善网络性能。实验结果表明改进的算法提高了网络平均吞吐量,带宽利用率更高,有效避免了网络拥塞。Study the optimization problem of network congestion control.The carrying capacity for network is growing,making the low efficiency of network.Traditional network congestion control algorithm requires real-time capacity numbering of wireless networks and the dynamic adjustment of TCP congestion window size,and it is difficult to establish accurate mathematical model,leading to low utilization of network bandwidth,and worse network congestion.In order to reduce the probability of network congestion,an improved wireless TCP congestion control algorithm was proposed.The original algorithm based on BP neural network was used for training the parameters,and the congestion control algorithm predicted by the network state can effectively distinguish between network congestion and packet loss network random errors,which avoided TCP congestion control algorithm congestion and realized the fast retransmit and fast recovery mechanisms.The experimental results show that the improved algorithm can increase the average network throughput and bandwidth utilization,and avoid network congestion.

关 键 词：拥塞控制 网络测量 神经网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]