【作者】 向少华；

【导师】 胥布工；

【作者基本信息】 华南理工大学 ， 控制理论与控制工程， 2006， 博士

【摘要】 随着Internet的飞速发展，用户数量迅速增加，新的网络应用不断涌现，使得网络流量急剧增加，由此引发的网络拥塞已经成为制约网络发展和应用的瓶颈问题。拥塞易造成传输延迟和吞吐量等QoS性能指标下降，导致网络性能下降、网络资源利用率降低，从而无法提供有效的QoS保证。因此，有效地解决拥塞问题对于提高网络性能具有重要意义，而网络拥塞控制研究也成为当前计算机网络和控制理论交叉领域研究的一个热点课题，不仅具有重要的理论价值，同时具有广泛的应用研究背景和意义。 本文的课题主要来源于国家自然科学基金重点项目(60334010)；国家自然科学基金项目(60474047)；高等学校博士学科点专项基金项目(20030561013)；广东省自然科学基金博士启动项目(04300046)。通过对Internet网络拥塞控制领域最新研究成果系统的分析和总结，利用基于控制理论和优化理论的TCP／AQM对偶模型和TCP动态流体模型，结合成熟的先进控制技术，本文设计了两种基于端主机的新的Vegas源端算法和五种基于路由端的新的鲁棒AQM链路算法，所设计的新算法在提高网络的性能，提高网络资源的吞吐量和利用率，以及适应网络环境动态变化的鲁棒性等方面都有较大的改进，同时体现了较好的公平性，从而能够提供有效的QoS保证，并通过仿真实验验证了新算法的性能。本文主要研究工作的内容和创新点包括以下几个方面： 1．基于优化理论和对偶模型，提出了一种改进的Vegas算法。仿真结果表明，采用改进算法合理地设置Vegas的参数，在保持较高的利用率和极低的丢包率的同时，明显地改善了Vegas和Reno之间竞争的公平性。 2．基于源端算法在稳定状态下的循环模型，详细分析了TCP Vegas算法的不足，设计了一种新的加强型Vegas(NE-Vegas)算法。所设计的算法能够有效的解决Vegas连接在网络路由改变造成的吞吐量持续下降问题，同时有效的改善Vegas与Reno之间竞争的公平性。通过仿真实验验证了NE-Vegas算法的高效性和公平性。 3．充分利用模糊控制理论在处理不确定性问题上的优越性，设计了一种基于模糊自适应PI的主动队列管理(AQM)算法，通过仿真说明，所设计的算法能在保证队列控制的稳定性基础上，加快队列收敛速度，更多还原

【Abstract】 With the rapid development of Internet and the sudden increase of computer network users, various network applications appear almost every day. And the network congestion led by the sharp increase of net flux has been a bottleneck problem, which restricts the development and application of network. Congestion often results in the decline of a series of QoS (Quality of Service) parameters such as transmission delay and throughput. What’s more, the network performance and resource utilization are also affected seriously which can not provide an effective guarantee of QoS index. Therefore, it is important to solve the congestion problem effectively. And the research about network congestion problem has also been a hot subject in both computer network and control theory, which makes great sense in science as well as practical application.This work was supported by National Natural Science Foundation of China Key Project 60334010, National Natural Science Foundation of China Project 60474047, SRFDP Project 20030561013, and Guangdong Province Natural Science Foundation of China PH.D Project 04300046. The purpose of this paper is to investigate the new arithmetic theoretically and empirically to improve the network performance. Firstly, it concludes the development status of the network congestion control research and reviews the relative literature. Then by using the control theory and optimal principle of TCP/AQM duality model as well as TCP dynamic fluid model, two kinds of source algorithms based on end hosts and five kinds of link algorithms on AQM are designed via advanced control techniques. The network performance and the throughput of network resource are upgraded much by the proposed algorithms. Simulation results also show that the new algorithms can not only improve the robustness of various changes in adapting network environment, but also appear to be a rather good fairness, which can afford the effective QoS guarantees.In general, the main contents and primary innovations of this dissertation can be summarized as follows:1. The fairness between Vegas and Reno is considered in this paper更多还原