VANET随机部署环境下基于改进型共享最近邻密度峰聚类的快速分簇算法

Fast Clustering Algorithm Using Improved Shared-Nearest-Neighbor-based Density Peaks Clustering in Random Deployment Environment of VANET

摘要:

针对车辆高速移动场景下，网络拓扑变化过大导致网络分簇结果不稳定的问题，提出一种基于改进型共享最近邻密度峰聚类的快速成簇算法SNNCA（Shared Nearest Neighbor Clustering Algorithm）。通过综合考虑节点的链路生存周期和移动相似性，提出一种全新的节点连接稳定程度评估指标，并将该评估指标应用于节点共享最近邻的计算过程，以组织网络节点为划分合理的多跳簇结构。为适应网络环境的动态变化，提出一种簇维护策略，其中每个层级的簇成员承担着维护下一层级簇成员的任务，该策略能够对簇成员进行批量分离或合并，从而实现了算法的分布式快速收敛。根据随机部署场景中进行的仿真实验结果显示，相比其他较新算法，SNNCA算法降低了74%的簇数量，并且簇成员的平均存活时间增加了近1倍，表现出更好的网络稳定性和健壮性。

Abstract:

For the issue of unstable network clustering results due to excessive network topology changes in high-speed vehicle movement scenarios, a fast clustering algorithm called SNNCA (Shared Nearest Neighbor Clustering Algorithm) using improved shared-nearest-neighbor-based density peaks clustering is proposed. By comprehensively considering the node's link survival period and movement similarity, a novel node connection stability evaluation metric is proposed. The metric is utilized in the shared nearest neighbor calculation process of the node to organize the network nodes into reasonable multi-hop cluster structure. To adapt to the dynamic changes of the network environment, a cluster maintenance strategy is introduced, where each level of cluster members takes on the task of maintaining the next level of cluster members, and this strategy can perform batch separation or merging of cluster members, achieving distributed and rapid convergence of the algorithm. According to the simulation results of the random deployment scenario, SNNCA algorithm reduces 74% of cluster numbers compared to other newer algorithms, and the average survival time of cluster members increases by nearly 1 time, demonstrating better network stability and robustness.