基于遗传算法的网络拥塞控制研究

发布时间：2018-05-02 17:55

本文选题：网络拥塞控制 + 主动队列管理　；参考：《江西理工大学》2014年硕士论文

【摘要】：随着互联网Internet的飞速发展，网络多媒体业务日趋多样化，互联网上的用户和应用都在急剧增加，网络拥塞成为制约网络发展和应用的瓶颈。网络拥塞控制因而至关重要，，作为改善网络系统性能、提高服务质量的主要手段，对网络拥塞控制问题的研究具有重大的理论意义和应用价值。本文分别从局部和全局的角度对网络拥塞控制方法进行研究分析。针对网络系统的单一节点，主动队列管理（AQM）是一种网络拥塞控制效果较好且广泛使用的一种方法，是Internet拥塞控制领域的研究热点。本文在目前存在的AQM算法的基础上，在算法设计上引入智能优化算法，设计了新的AQM算法，从而明显地改善队列的性能。从网络系统的全局出发，针对网络中的关键节点和链路实施新的路由算法，通过网络路径优化，从而使得网络负载均衡分配，降低丢包率，提高链路吞吐量，减小时延，获得一个高QoS保证的网络环境。本文的主要工作包括以下两个方面：（1）基于队列长度和链路速率相对变化率，设计一种带有参数优化的模糊神经网络控制器的拥塞控制方法。该算法引入队列长度和期望队列长度以及链路速率与链路容量的相对误差量作为网络拥塞指示，通过改进后的遗传算法定时对模糊神经网络控制器进行参数优化，实现对网络拥塞的有效控制。在大时滞环境和突发流情况下，该算法的稳定性和控制效果都比较令人满意。（2）分析了QoS的技术特征、执行过程以及现有路由算法的优缺点，从路由优化的角度，基于移动Ad hoc网络（MANETs），设计一种运用在几何路由中的基于遗传算法的路由优化算法，目标是在满足带宽、时延、费用等多项QoS指标的基础上，路径时延最小化，负载尽量分布在有宽裕空闲资源的链路上，便于今后接纳更多的请求，达到提高网络吞吐量的目的，进而避免拥塞。
[Abstract]:With the rapid development of Internet Internet, network multimedia services are becoming more and more diversified, and the number of users and applications on the Internet is increasing rapidly. Network congestion has become a bottleneck restricting the development and application of network. Therefore, network congestion control is of great importance. As the main means to improve the performance of network system and improve the quality of service, the research on network congestion control has great theoretical significance and application value. In this paper, the network congestion control methods are studied and analyzed from the local and global perspectives. For a single node in a network system, active queue management (AQM) is an effective and widely used method for network congestion control. It is a hot research topic in the field of Internet congestion control. In this paper, based on the existing AQM algorithm, an intelligent optimization algorithm is introduced into the algorithm design, and a new AQM algorithm is designed, which obviously improves the performance of the queue. Based on the overall situation of the network system, a new routing algorithm is implemented for the key nodes and links in the network. Through the network path optimization, the network load distribution is balanced, the packet loss rate is reduced, the link throughput is improved, and the delay is reduced. Obtain a high QoS guaranteed network environment. The main work of this paper includes the following two aspects: 1) based on the relative change rate of queue length and link rate, a congestion control method of fuzzy neural network controller with parameter optimization is designed. The algorithm introduces queue length and expected queue length as well as the relative error between link rate and link capacity as network congestion indication, and optimizes the parameters of fuzzy neural network controller by the improved genetic algorithm. The effective control of network congestion is realized. The stability and control effect of the algorithm are satisfactory in the case of large time delay and sudden flow. This paper analyzes the technical characteristics, execution process and advantages and disadvantages of existing routing algorithms of QoS. From the point of view of routing optimization, a genetic algorithm based on genetic algorithm for geometric routing is designed based on mobile Ad hoc networks. The goal is to minimize path delay on the basis of satisfying bandwidth, delay, cost and other QoS indicators, and to distribute the load on links with spare resources as much as possible, so that more requests can be accepted in the future, so that the throughput of the network can be improved. Thus avoiding congestion.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06;TP18

【参考文献】