传送网路由规划优化建模研究

发布时间：2018-04-14 18:14

本文选题：传送网 + 路由　；参考：《电子科技大学》2014年硕士论文

【摘要】：网络业务量的爆炸式增长,推动了传送网的不断发展。现在的传送网正朝着高速率、大容量、智能化、多业务能力的方向发展,这对网络规划和网络管理提出了更高的要求,原有的路由算法已难以满足未来传送网中要求的高效集中管理。传送网的网络规划与优化中,需要考虑光网络中的很多特性,其中影响路由规划的主要有光信号的传输损耗、波长一致性、OEO(Optical-Electrical-Optical)资源分配、波长资源分配等,这些特性同时给传送网中相关算法的设计提出了很大的挑战。在有限的网络资源下,如何优化传送网网络资源的分配,如何优化路由规划,关系到传送网的服务质量。现有的网络优化方案较少重视OEO资源的稀缺性,或对光信号的传输损耗约束表述不合理,并且尚缺乏针对重路由问题的研究,同时为提高网络生存性而提出的OEO资源预留、波长预留的保护策略还缺乏有效的验证。本文由传送网路由规划中的几个重要问题出发,研究传送网路由规划问题,针对不同场景进行ILP(Integer Linear Programming,整数线性规划)建模。首先,研究不同的网络优化建模方法,即Node-Link模型和Link-Path模型,分析不同场景中建模方法的选择策略,为之后的建模方案设计提供指导。然后,在OEO资源分配、OEO资源预留、重路由计算的场景下分析不同网络拓扑对路由规划结果的影响,由此总结传送网中传输损耗、关键资源数目、OEO节点数目、OEO资源放置位置等因素对传送网网络性能的影响。同时,在上述场景中比较不同算法策略对优化结果的影响,如OEO资源预留问题中的OEO资源共享、重路由问题中的资源重用和波长预留等。并且,对本文所提出的重路由问题的ILP模型进行扩展,使其可以应用于多种传送网路由计算场景。最后,对复杂的重路由模型进行拉格朗日分解,降低模型复杂度的同时使该问题得以在算法层面上解决。除此以外,ILP模型的计算结果可以为相应问题的算法验证提供数据依据。通过上述工作,使用数学模型分析传送网拓扑,总结影响网络性能的网络特征,为传送网网络规划提供理论依据。同时,验证资源预留策略对网络生存性的提升,并给相关问题的算法设计提供新的思路。
[Abstract]:The explosive growth of network traffic promotes the continuous development of transport network.The current transport network is developing towards the direction of high speed, large capacity, intelligence and multi-service capability, which puts forward higher requirements for network planning and network management.The original routing algorithm is unable to meet the requirements of the future transport network efficient centralized management.In the network planning and optimization of transport networks, many characteristics of optical networks need to be considered, among which the transmission loss of optical signals, wavelength consistency, and wavelength resource allocation are the main factors affecting routing planning.These characteristics also pose a great challenge to the design of related algorithms in transport networks.Under the limited network resources, how to optimize the distribution of transport network resources and how to optimize the routing planning are related to the quality of service of the transport network.The existing network optimization schemes pay little attention to the scarcity of OEO resources, or the unreasonable expression of optical signal transmission loss constraints, and the lack of research on rerouting problem, and the proposed OEO resource reservation to improve the network survivability.The protection strategy of wavelength reservation is still lack of effective verification.Based on several important problems in transport network routing planning, this paper studies the transport network routing planning problem and models ILP(Integer Linear programming (integer linear programming) for different scenarios.Firstly, different network optimization modeling methods, namely Node-Link model and Link-Path model, are studied, and the selection strategies of modeling methods in different scenarios are analyzed to provide guidance for later modeling scheme design.Then, in the scenario of OEO resource allocation and rerouting calculation, the influence of different network topologies on routing planning results is analyzed, and the transmission loss in transport network is summarized.The number of key resources and the number of OEO nodes and the location of OEO resources affect the network performance.At the same time, the effects of different algorithm strategies on the optimization results are compared in the above scenarios, such as OEO resource sharing in OEO resource reservation problem, resource reuse and wavelength reservation in rerouting problem, and so on.Furthermore, the ILP model of rerouting problem proposed in this paper is extended to be applied to various transport network routing computing scenarios.Finally, the complex rerouting model is decomposed by Lagrange decomposition, which reduces the complexity of the model and solves the problem at the algorithm level.In addition, the calculation results of ILP model can provide data basis for the algorithm verification of the corresponding problems.Through the above work, using mathematical model to analyze the transport network topology, summarize the network characteristics that affect the network performance, and provide a theoretical basis for transport network planning.At the same time, it verifies that the resource reservation strategy improves the survivability of the network, and provides a new idea for the algorithm design of related problems.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN915.0

【参考文献】