基于蚁群优化的自组网负载均衡路由技术的研究

【作者】 刘仁婷；

【导师】 郭伟；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2006， 硕士

【摘要】 移动自组网在使用上有很多优点,如可以快速部署,具有自组织功能,非常适合于军用战术环境和民用应急通信。由于自组网的特殊组织形式,固定网络和有中心无线网络的很多协议都无法直接被它采用,因而需要设计专门适用于自组网的协议。其中,路由协议是自组网设计、研究的主要技术难点。在进行路由协议的设计时,考虑到自组网的自身特点和诸多限制,负载均衡的概念被大量的引入。本论文主要研究基于负载均衡的无线自组网关键技术与算法,期望通过在路由选择时考虑网络中各节点的负载和拥塞情况,对网络进行负载均衡,减轻自组网的瓶颈及拥塞问题,使网络保持连续、高效和稳定的运行,优化网络的综合性能。另一方面,蚁群优化算法是一种通用的分布式随机优化方法,能够大大提高系统的可靠性、鲁棒性以及增强通信网适应传输故障、突发业务等意外事件的能力,已经开始应用于通信网中进行路由的组合优化问题的求解。本文基于蚁群优化算法的优点,提出了基于蚁群优化的负载均衡路由协议(AOLR),设计针对蚁群优化算法特有的信息素浓度的更新规则,将蚁群优化和跨层优化方法结合起来,设计信息素浓度的挥发以及针对不同的数据分组的更新模式。同时,为了进一步的解决自组网中的负载均衡问题,在AOLR中引入了准入控制规则、拥塞反向抑制规则、捷径增强规则以及噪声选路规则,通过对节点负载状态的动态感知增加节点对负载的动态适应能力。本文在详细阐述和分析了基于蚁群优化的负载均衡路由协议(AOLR)的各种关键技术及规则后,介绍了AOLR协议的系统建模和仿真的整体框架,以及各个模块实现的功能和工作流程。通过对基于蚁群优化的负载均衡路由协议系统的仿真,从数据分组成功递交率、平均端到端时延、平均路由建立时延、路由开销等性能的仿真数据和曲线上,分别验证了该系统中的几项关键技术和算法。从仿真结果可以看出,该系统较好的改善了自组网中中普遍存在的拥塞问题、捷径问题和引入的路由开销问题等。同时,较之现有的负载均衡路由协议,在数据分组的成功递交率,平均端到端时延以及路由开销等性能上也有显著的提高。更多还原

【Abstract】 With several advantages, such as fast deployment and self-organization, Mobile Ad hoc Networks(MANETs) are suitable for applications involved in tactical battlefield and civilian emergency. However, due to their special form of organization, many protocols used in networks with pre-configured and centrally controlled infrastructure, cannot be directly applied to Mobile Ad-hoc Networks. Thus protocols need to be carefully designed for this kind of networks. Among them, routing algorithms are the main difficulty of design and research on self-organized networks. Considering the special characters and several limitations of MANETs in the design of routing protocols for MANETs, the idea of load balancing is largely introduced. This dissertation mainly focuses on the key techniques and algorithms involved in routing protocols of MANETs based on load balancing. The traffic load and congestion status of nodes in MANETs will play an important role in route selection, in hope of balancing load, alleviating the bottleneck and congestion problems, making the network stable and efficient, and optimizing the overall performance of the network.On the other side, the ant colony optimizing algorithm is a general distributed random optimization method, which can largely improve the stability, robustness of the system and increase the adaptability to unexpected situations, such as transmission failure and emergent traffic. It has already been applied to the solutions of combine optimization in communication networks. Based on the advantages of this algorithm, this dissertation proposed the Ant-colony Optimization based Load-balancing Routing protocol for MANETs(AOLR), which combines the algorithm of ant colony optimization with the idea of cross-layer design, in order to provide the mechanism for the evaporation of the pheromone strength and the its update mode triggered by different types of packets. Meanwhile, to further solve the load balancing problem of MANETs, AOLR introduced several mechanisms, such as the admission control mechanism, backward congestion restrain mechanism, shortcut enforcement mechanism, and the noise-encouraged route-selection mechanism, in order to dynamically perceive the load of nodes and increase the adaptability to it.This thesis elucidates and analyses the key techniques and rules of AOLR in detail, and introduces its simulation modeling of the integrated system infrastructure and each function modules. Through systematic simulation of AOLR in the performance of packet delivery rate, average end-to-end delay, average route set-up delay, routing overhead, and etc, the key techniques and algorithms are testified and can provide reference for practical system design and application. Simulation results show that, the system can improve several problems involved in MANETs, such as the congestion problem, the shortcut problem and the routing overhead problems. Meanwhile, compared to exist load balancing routing protocols, our system shows better performance of packet delivery ratio, end-to-end delay and routing overhead.更多还原