【作者】 李琳；

【导师】 井元伟；

【作者基本信息】 东北大学 ， 控制理论与控制工程， 2008， 硕士

【摘要】 科学技术高速发展,新型网络应用不断涌现,用户数量迅速增加,由此引发的网络佣塞成为制约网络发展和应用的瓶颈问题,如何更好地预防和控制网络拥塞成为近年来网络研究领域的热点。仅依靠基于源端的TCP拥塞控制已不能满足服务质量要求,基于路由器的拥塞控制机制逐渐成为研究的重点。主动队列管理算法(AQM)是基于路由器的拥塞控制的重要组成部分。本文首先介绍了拥塞控制的定义、产生的原因和TCP/IP拥塞控制机制;着重介绍了IP拥塞控制机制中的主动队列管理算法,并指出它在拥塞控制中的重要作用。主动队列管理算法解决路由器满队列问题的方法,是在队列满之前就对到达的数据包进行丢弃,这样端节点便能在队列溢出之前做出反应,从而避免真正拥塞的发生。随机早期检测(RED)算法是著名的AQM算法,而基于控制理论的拥塞控制成为新的研究热点。近几年,智能控制理论在网络拥塞控制中的应用取得了很大的进展。本文在研究主动队列管理算法的基础上,考虑了网络环境的复杂性和网络模型的不确定性,将智能控制中的模糊控制与常规PID控制结合起来,采用模糊增益调整PID控制算法作为主动队列管理算法,对网络拥塞进行控制。模糊增益调整PID控制器,既具有模糊控制灵活而适应性强的优点,又具有PID控制精度高的特点,能达到传统拥塞控制所不能达到的控制效果。为了解决模糊增益调整PID控制算法设计复杂、耗时的问题,本文研究了模糊/PID混合控制算法,将具有PD控制功能的二维单输出模糊控制与PI控制相结合,并将变速积分引入PI控制器,解决因固定的积分常数在某些调节时刻过大而造成超调过大的问题。针对模糊增益调整PID控制算法和模糊/PID混合控制算法基于Matlab进行仿真。通过仿真结果表明,两种模糊PID算法能有效地将队列长度稳定在目标值附近,具有较强的稳定性和鲁棒性,适应多变的网络环境,而后者的控制效果更优越。最后,对全文进行了概括性总结,在此基础上,对以后的工作进行展望。更多还原

【Abstract】 Nowadays, with the development of science and technology, the network congestion become a bottleneck problem, which restricts the development and application of network. Therefore,the congestion control has become an active issue in the field of network research. The TCP congestion control implemented in source node can not feed the quality of service, therefore, congestion control which is based on router gradually becomes the main direction in cunent research. Active queue management (AQM) is the significant constituent.The thesis introduces the concept and meaning of congestion and congestion control, and the mechanism of congestion control of TCP/IP is introduced. Then the AQM algorithm of IP congestion control mechanism is emphatically introduced, and the paper indicates that the AQM is very important in congestion control. The AQM algorithm drops the data-packages before the queue achieving full, so that the terminal node can response early, and thus the congestion can be avoided truly. RED is the most famous AQM algorithm, and congestion control based on control theory becomes a new research focus.In recent years, the research of intelligent control theory in congestion control has made great progress. Based on the analysis of AQM algorithm, considering the complexity of the network and the uncertainty of the network model, the thesis brings forward that combines the fuzzy control and the classical control algorithm PID, to process the congestion control, as the AQM algorithm. The self-tuning gain fuzzy-PID control algorithm which uses the output of fuzzy-controller to tune the parameters of PID in this paper, not only has the advantages of great precision of PID control, but also has the strong points of agility and adaptive of fuzzy control. The fuzzy-PID control can get the good control effect which the traditional control can’t achieve.To solve the problems of design-complexity and time-consuming, compositive control algorithm of fuzzy/PID was studied. Combine the planar single-output fuzzy control which has the capability of PD control with the PI control. And the gearshift integral is inducted in PI control to solve the problems caused by great integral.The simulation is processed in Matlab. Comparing the self-tuning gain fuzzy-PID control algorithm with the compositive control algorithm of fuzzy/PID, the simulation results show that the algorithm can stabilize the queue length to target value. Both of them are applicable to the complex and various network environments and have good stabilization and robustness. And the later algorithm has better performance. Lastly, the research work is concluded. The further research direction is pointed out, and the prospect of solving congestion control is looked forward.更多还原