【作者】 郝琨；

【导师】 金志刚；

【作者基本信息】 天津大学 ， 计算机应用技术， 2010， 博士

【摘要】 网络编码打破了通信网络中传统的信息处理方式,网络节点对信息进行编码等处理,使得信息传输速率达到网络的最大流,提升网络的吞吐量、节约传输带宽和均衡负载等。但是,网络编码系统中的编解码操作,必将给网络带来额外的开销。因此,减少网络编码操作的额外开销,对网络编码的传输过程进行优化,对推动网络编码的实际应用有重要的意义。论文以网络编码理论及其优化机制为基础,对网络编码优化及其在不同领域下的应用进行了研究。研究成果包括以下几方面:基于遗传算法的网络编码优化:针对网络编码的单目标优化建立了优化模型,并在模型基础上设计了基于简单遗传算法的最小化编码节点的优化算法MCN。仿真结果表明,算法有效地减少了参与编码的节点数量。针对网络编码的多目标优化问题,综合考虑编码消耗和链路消耗问题,设计了基于NSGA-II算法的网络编码的多目标优化算法MOONC。仿真结果表明,该算法能有效得到Pareto解,并降低了网络的整体开销。结合机会路由的部分网络编码优化:针对无线网络中广播通信能量效率、网络编码延迟问题,设计了部分网络编码优化算法。考虑当前链路状况和节点的能量状况,提出了基于部分网络编码的能量约束机会路由协议PNCOR。仿真结果表明,部分网络编码与机会路由相结合,在降低网络延迟、提高网络吞吐量方面均有显著的优势。同传统的路由机制相比,吞吐量提高了22%,投递率提高了23%左右,网络延迟降低25%左右,提高了无线网络传输性能。基于确定性网络编码的文件分发:以P2P文件分发系统为研究背景,设计了一种新的网络编码文件分发模型及算法DLNCCD。该模型在保证编码解码成功的前提下,采用确定性编码方法降低了网络编码的复杂度。同时,将该编码方法应用在特定的网络拓扑结构上,在保证达到最大的网络编码增益的前提下,有效地解决了拓扑带来的线性相关编码块的问题。仿真结果表明,以该模型建立的文件分发系统同传统的文件分发系统相比吞吐量提高20%左右,平均下载时间节省了20%左右,提高了文件分发系统的性能。更多还原

【Abstract】 Network coding has reformed the traditional information processing in communication networks. It has lots of advantages, such as increasing the information transmission rate to the maximum network flow rate, improving the network throughput, saving the bandwidth and balancing load. But, the encoding and decoding operation of the nodes in network coding system will bring overhead to the network. Hence, it will be extremely important to reduce the overhead of the network coding operation and optimize the network coding transmission for promoting the practical application of the network coding. This paper studies on the optimization of the network coding and its application in different field based on the network coding theory and optimization mechanism. The research achievements include:Genetic algorithm based network coding optimization. Aiming at the single objective optimization of the network coding, the paper proposes the network coding optimization model. The paper designs the optimization algorithm MCN for minimization of the coding nodes based on simple genetic algorithm. The simulation results show that this algorithm reduces the network coding nodes. Aiming at the multi-objective optimization of the network coding, after considering the encoding overhead and link overhead, this paper proposes multi-objective optimization algorithm MOONC for network coding based on NSGA-II algorithm. The simulation results show that the algorithm can computer the Pareto optimal solution effectively and reduce the network overhead.Partial network coding optimization combined with opportunistic routing. The paper proposes the partial network coding optimization algorithm to improve the radio communication energy efficiency and decrease the network coding delays in wireless network. Moreover the energy constraint opportunistic routing protocol PNCOR is also proposed based on the partial network coding after considering the current link status and the energy status of nodes. The simulation results show that the partial network coding has advantages of reducing the network delay and increasing network throughput. It can improve throughput about 22%, delivery rate about 23% and reduce the delay about 25% after combining with the opportunistic routing and improve the transmission performance. File distribution strategy based on deterministic network coding. Based on the research background of P2P file distribution system, this paper designs a new kind of network coding file distribution model and algorithm DLNCCD. This model could reduce the complexity of network coding through using the deterministic coding method, when it guarantees the success of encoding and decoding. Meanwhile, this network coding method could be applied to the specific network topology, which could ensure the maximum network coding gain and could solve the linearly related coding block problems brought by topology. The simulation and results show that the throughput of the file distribution system with network coding based on this model can improve about 20% and the average download time can reduce about 20% and improve the file distribution performance.更多还原