无线Mesh网络联合功率控制与信道分配研究

发布时间：2018-03-09 11:42

本文选题：无线Mesh网络　切入点：功率控制　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：无线Mesh网络(Wireless Mesh Network,WMN)作为提供高速互联网络服务的低成本解决方案,可以为其覆盖区域内的移动用户提供网络接入,在下一代宽带无线通信系统中具有重要的应用价值。Mesh路由器需转发Mesh客户端和其他Mesh路由器的数据业务,可以根据通信需求采用不同的发射功率和传输信道进行并发传输,因而带来了如何通过功率控制和信道分配来降低网络干扰的问题。网络干扰是制约WMN性能的主要因素,功率控制和信道分配共同决定了网络的干扰分布。当Mesh路由器的发射功率和传输信道被合理配置时,可以有效避免WMN链路冲突,并获得更高的数据传输速率。目前针对WMN联合功率控制与信道分配方法的研究,没有很好的考虑功率控制和信道分配子问题之间的交互反馈求解,同时对网络负载与干扰分布的衡量不够准确,未能准确捕捉网络的实际状况,容易产生网络瓶颈链路,造成网络拥塞。因此,设计合理有效的联合功率控制与信道分配优化模型及优化算法,可以有效改善WMN整体性能。本文针对WMN链路负载分布不均的问题,择优选择网关节点以均衡网关负载。通过定义链路权重,构建网络加权吞吐量最大化的联合资源分配模型。为求解所构建的联合资源分配模型,提出一种基于Q学习和差分进化的联合功率控制与信道分配算法(Q learning and Differential evolution based Joint Power control and Channel assignment Algorithm,QDJPCA)。QDJPCA算法采用基于多重变异和自适应交叉因子的差分进化算法进行信道分配,同时获取功率控制的反馈结果作为每次迭代的输入参数;针对每次迭代产生的信道分配结果,采用基于状态聚类和状态修正的Q学习算法实现功率控制;借助循环嵌套和反馈机制实现功率控制与信道分配子问题之间的交互反馈求解,以获得有效的联合资源分配结果。网络加权吞吐量最大化的联合资源分配模型仅考虑了网络吞吐量性能的优化,忽略了网络瓶颈链路和网络拥塞对网络性能的影响。为准确捕捉网络负载与干扰分布以解决网络拥塞问题,本文构建了网络拥塞避免系数最小化的联合资源分配模型。针对所构建的联合资源分配模型,在QDJPCA算法的基础上,提出一种面向拥塞控制的联合功率控制与信道分配算法(Congestion Control oriented Joint Power control and Channel assignment Algorithm,CCJPCA)。通过引入负载模型与干扰模型,衡量通信链路的干扰范围内各干扰链路对同一信道的竞争程度,进而更加准确的捕捉网络干扰分布。同时,定义网络拥塞避免系数来反映网络拥塞程度,以有效保护WMN瓶颈链路。CCJPCA算法通过混合编码策略实现了链路功率与信道变量的共同进化,利用Q学习算法的回报机制来实现变异策略的自适应选择,从而保证网络资源的合理配置。通过NS-3网络仿真平台,对QDJPCA和CCJPCA算法的性能进行了仿真验证。仿真结果表明,本文提出的联合资源分配模型及算法,能够优先为网络瓶颈链路分配网络资源,实现WMN性能改善。
[Abstract]:Wireless Mesh network (Wireless Mesh Network, WMN) as a low cost to provide high-speed Internet service solutions can provide network access for its coverage of the mobile users in the region, has important application value to the.Mesh router forwarding Mesh client and other Mesh router data services in the next generation broadband wireless communication system, can according to the communication needs of different transmission power and transmission channel for concurrent transmission, so how to reduce network interference brought by power control and channel allocation problem. The network interference is the main factor to restrict the performance of WMN, power control and channel allocation determines the distribution of network interference. When the transmit power of Mesh router and the transmission channel by reasonable configuration, can effectively avoid the WMN link conflict, and obtain high data rate. The combined power of WMN Research on control and channel allocation method, there is no interaction between good consideration of power control and channel allocation problem solving and feedback, to the network load and the distribution of interference measurement is not accurate enough, failed to accurately capture the actual situation of the network, the network is easy to produce the bottle Necklace Road, caused by network congestion. Therefore, the design of reasonable and effective joint power control and channel allocation optimization model and optimization algorithm, can effectively improve the overall performance of WMN. The WMN link load uneven distribution problems, choose the gateway node to gateway load equilibrium. By defining the link weight, joint resource allocation model of weighted network throughput maximization. Joint resource allocation model for solving and put forward a Q learning and differential evolution of joint power control and channel allocation algorithm based on Q (learning and Differential evolution based Joint Power control and Channel assignment Algorithm, QDJPCA).QDJPCA algorithm using adaptive crossover mutation and multiple factor based differential evolution algorithm for channel allocation, power control and obtain feedback results as input parameters for each iteration; channel assignment result for each iteration, using state clustering and modified Q learning algorithm to achieve the state power based on the interaction between control; using nested loop and feedback mechanism to realize the power control and channel allocation problem is solved to obtain feedback, joint resource allocation results effectively. The joint resource allocation model of network weighted throughput maximization only consider the optimization of network throughput performance, ignoring the influence of network bottleneck link and network congestion on network performance to accurately capture the network load and interference distribution to solve the network congestion problems, this paper. The construction of network congestion avoidance joint resource allocation model. According to the coefficient minimization of joint resource allocation model, based on the QDJPCA algorithm, proposed a congestion control oriented joint power control and channel allocation algorithm (Congestion Control oriented Joint Power control and Channel assignment Algorithm, CCJPCA). By introducing the load model with interference on the same channel model, the competition degree of each link interference interference range measure communication link, capture network interference and thus a more accurate distribution. At the same time, the definition of network congestion avoidance coefficient to reflect the degree of network congestion and to protect the WMN bottleneck of.CCJPCA algorithm by hybrid encoding method to achieve the common evolutionary link power and channel variables the use of Q learning algorithm to achieve the return mechanism of adaptive mutation strategy, so as to ensure network information The reasonable allocation of source. Through NS-3 network simulation platform, the performance of QDJPCA and CCJPCA algorithm is simulated and verified. The simulation results show that the joint resource allocation model and algorithm proposed in this paper can prioritize network resources for network bottleneck link and achieve WMN performance improvement.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5

【相似文献】