Congestion control of HighSpeed TCP using the partial collaborative learning automaton team mode

Currently, the communication network is the foundation of social life. The current Internet is the development of the communication network. In the Internet, a wide common transport layer protocol is TCP Reno. However, the bandwidth of TCP Reno has been reduced in the high bandwidth. As one of mitigation measures, HighSpeed TCP have been proposed. But the bandwidth of HighSpeed TCP is assumed that recommended parameters are provided to 10Gbps. Through experiment know that reduces utilization of bandwidth when the bandwidth is changed. In this paper, we proposed a new method, called as Partial Collaborative Learning Automaton Team Model, into the congestion control process of HighSpeed TCP, aim to adjust the important parameter adaptively. Compare to traditional method, the simulation results have shown that, if the variation of the bandwidth of the backbone can not be predicted, our method has some better performances.