论文部分内容阅读
为了精确研究 TCP(transmission control protocol)算法 ,指导其他端到端流量控制算法以及路由器中的分布式流量控制算法的设计和实现 ,该文从非合作博弈的角度出发 ,提出了一种基于进化博弈理论的 TCP算法模型。使用该模型 ,针对 Reno和 Vegas两个版本的 TCP算法进行了描述、仿真和比较。分析和仿真结果表明 :TCP Reno使用 3个策略进行流量控制 ,而 Vegas是一个 5策略的流量控制算法 ,因此 ,Vegas能够比 Reno更好地适应复杂的网络环境 ,获得更好的性能。另外 ,该文从进化博弈理论的角度提出了一些能够提高 TCP性能的方法和改进方向
In order to accurately study the transmission control protocol (TCP) algorithm and guide the design and implementation of other end-to-end flow control algorithms and distributed flow control algorithms in routers, this paper proposes a game based on evolutionary game Theory of TCP algorithm model. Using this model, we describe, simulate and compare the TCP algorithms of Reno and Vegas. The analysis and simulation results show that TCP Reno uses 3 strategies for traffic control, and Vegas is a 5-strategy traffic control algorithm. Therefore, Vegas can better adapt to complex network environment and achieve better performance than Reno. In addition, the paper proposes some methods and ways to improve TCP performance from the perspective of evolutionary game theory