光无线混合宽带接入网络优化设计与绿色路由算法研究

Hybrid Optical-Wireless Broadband Access Networks (HOWBAN) can provide very high throughput in a cost-effective manner; and it is a promising architecture to provide a convenient broadband access service for users. Although network planning and routing in HOWBAN have become a research focus, a few research activities are reported on energy saving for HOWBAN. A reasonable configuration of renewable energy and traditional energy, i.e., mixing energy, can avoid the degradation of network performance while reducing carbon emissions. The introduction of mixing energy will bring about new changes for the network planing and routing in HOWBAN, which is required a further research. With the purpose of energy saving and based on more general assumptions than previous work such as random distribution of users' bandwidth requirements and ONU supporting different wavelengths, we investigate the network planning of HOWBAN and its solution by combinatorial optimization theory and particle swarm optimization algorithm respectively while considering the characteristics of the optical access network and the wireless access network, their physical limitations and interactions, renewable energy sources, sleep mode of node, the failure of network component and other factors. We propose some new concepts including differentiated reliable routing (OLT-disjoint routing, ONU-disjoint routing and channel-disjoint routing), energy factor and segment protection. Based on cross-layer design, graph coloring theory and wavelength plane model, we also propose inter-domain and inner-domain joint channel assignment and routing algorithms. With a consideration of the impact of green routing protocol, we use genetic algorithm (GA) to re-optimize network resources. Finally, we develop a simulation platform and a testbed to analyze and evaluate our models and protocols in term of network deployment cost, transmission delay, reliability, energy consumption and to collect running data of HOWBAN, which is beneficial to provide a theoretical and methodological guidance for the further research and application.

光无线混合宽带接入网极具应用前景，其优化设计和路由选择是相互影响的两个关键问题。合理使用混合能源（可再生能源和传统能源），可避免网络性能下降，同时降低传统能源消耗和减少碳排放量。引入混合能源将对解决上述问题带来新的变化，有待深入研究。项目以节能减排为目标，将光接入和无线接入作为有机整体，综合考虑各自的物理限制条件和相互影响、混合能源使用、节点休眠、网络元件失效等多种因素，基于更符合实际的一般性假设，利用组合优化理论和微粒群优化算法设计网络部署模型并研究求解方法；提出区分可靠路由（OLT分离、ONU分离及信道分离）、能量因子和支路保护等概念；基于跨层设计思想，借鉴图着色理论和波长平面图模型提出域内/域间联合信道分配和选路算法；联合绿色路由机制，利用遗传算法进一步优化资源配置；构建仿真和硬件验证平台，采用理论分析和实证相结合的方法，分析与评测网络关键指标，为进一步研究和应用提供理论和方法指导。

光无线混合宽带接入网结合了光纤接入和无线接入二者的优势，可用低廉的成本为用户提供满足QoS（Quality of Service）要求的灵活、便捷的宽带接入服务，是一种极具应用前景的“最后一公里”解决方案。它既不同于单纯的光接入网，也不同于传统的无线网状网，是两种网络的无缝连接，具有自身的特殊性。项目针对光无线混合宽带接入网前端无线网络和后端光网络所涉及的网络部署问题、绿色路由和QoS问题、生存性问题、动态资源配置问题以及仿真实验平台搭建等关键问题进行研究，设计高性价比的网络部署方案和有效的绿色节能协议。主要研究成果包括：（1）构建综合考虑多种限制因素的网络部署优化模型；（2）提出利用可再生能源的绿色路由算法；（3）提出基于混合能源的绿色共享保护算法；（4）提出支持QoS的动态带宽分配算法；（5）提出能量感知的多线程轮询算法；（6）构建网络仿真验证平台评测网络关键指标。项目相关研究成果可为光无线混合宽带接入网的进一步发展和应用提供理论和方法指导。项目执行期间，项目组已发表学术论文61篇，其中期刊论文42篇，会议论文19篇，已被SCI收录4篇，EI收录28篇，申请国家发明专利4项，获得计算机软件著作权登记证书1项，培养研究生22名。