移动Ad Hoc网络分簇算法的研究

发布时间：2018-03-06 10:48

本文选题：Ad　切入点：Hoc　出处：《安徽大学》2013年硕士论文　论文类型：学位论文

【摘要】：由于移动Ad Hoc网络具有自组织、无中心和动态拓扑的特点,因此可以满足人们随时随地通信的需求,然而节点能量的受限和经常变化的拓扑结构成为设计Ad Hoc网络协议的重大障碍,特别是网络规模较大时其稳定性难以保持。Ad Hoc网络通过分簇算法来实现网络分层,以支持高效的资源管理和路由策略,分层结构可以有效的满足一定的Qos需求,负载平衡因子和簇稳定性是评价分簇算法性能的主要标准。本文首先介绍了Ad Hoc网络和分簇算法的相关知识和概念,并介绍了一个不借助GPS的辅助来描述簇移动性的方法。接着对TCA分簇算法进行了改进,提出了一种基于距离变化的加权分簇算法MSED,该算法将网络层上处于同一子网的节点尽量分在一个簇中,并通过一段时间内节点与其邻居节点间距离的变化来预测节点的稳定性,与TCA算法相比,MSED算法考虑了侦测周期期间节点入簇情况,并改进了权值计算公式。NS2仿真结果证明,同TCA算法相比,MSED算法得到的簇稳定性较高,在节点间传输有效距离相同或者最大移动速度相同时,其网络统治集更新频率和节点入簇频率均较低。常用的分簇算法没有很好的保持网络的稳定性,稳定的分簇算法可以减少网络的计算和通信开销,为了提高网络的稳定性,本文最后在AOW分簇算法的基础上,提出了一种基于链接率的自适应按需加权分簇算法AOWLR,该算法借助GPS的辅助,将节点的运动轨迹抽象为直线型和圆弧形两种,并引入邻居节点的平均链接率预测值作为衡量簇稳定性的一个重要标准。NS2仿真结果证明,同AOW相比,AOWLR算法的负载平衡因子较高,连通统治集(CDS)更新频率和节点充当簇头的公平性指数(HFI)较低。
[Abstract]:Because the mobile Ad Hoc network has the characteristics of self-organization, no center and dynamic topology, it can meet the needs of people communicating anytime, anywhere. However, the limitation of node energy and the frequently changing topology have become a major obstacle to the design of Ad Hoc network protocols. Especially, it is difficult to maintain the stability of Ad Hoc network by clustering algorithm when the network size is large. In order to support efficient resource management and routing strategy, hierarchical architecture can effectively meet certain Qos requirements. Load balancing factor and cluster stability are the main criteria to evaluate the performance of clustering algorithm. This paper first introduces the knowledge and concept of Ad Hoc network and clustering algorithm, and introduces a method to describe cluster mobility without the help of GPS, and then improves the TCA clustering algorithm. In this paper, a weighted clustering algorithm based on distance variation is proposed. The algorithm divides the nodes in the same subnet into a cluster as far as possible, and predicts the stability of the nodes by the distance between the nodes and their neighbors over a period of time. Compared with the TCA algorithm, the TCA algorithm takes into account the clustering situation during the detection period, and improves the weight calculation formula. NS2 simulation results show that compared with the TCA algorithm, the cluster stability of the MMSED algorithm is higher than that of the TCA algorithm. When the effective distance between nodes is the same or the maximum moving speed is the same, the updating frequency of the dominant set and the frequency of node joining are lower. The commonly used clustering algorithms do not keep the stability of the network very well. The stable clustering algorithm can reduce the computation and communication overhead of the network. In order to improve the stability of the network, this paper finally based on the AOW clustering algorithm. An adaptive on-demand weighted clustering algorithm based on link rate is proposed. With the aid of GPS, the motion trajectory of nodes is abstracted into straight line and circular arc. As an important criterion of cluster stability, NS2 simulation results show that the load balancing factor of AOWLR algorithm is higher than that of AOW, and the prediction value of average link rate of neighbor nodes is used as an important criterion to measure the stability of cluster. The frequency of CDS) renewal and the fairness index (HFI) of nodes acting as cluster heads are lower.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TN929.5

【参考文献】