Clustering is an efficient technique for achieving high scalability on the Internet of Vehicles (IoV). However, the latency and overhead generated from forming and maintaining clusters are common barriers to the mass adoption of this technique. To this end, we propose an efficient clustering scheme for the IoV. Leveraging reinforcement learning, our scheme can quickly form network condition-aware clusters. In addition, our reinforcement learning-based clustering scheme (RLBC) assures dynamic and cooperative maintenance for clusters. The effectiveness of our scheme is evaluated through extensive simulations. The simulation results show that the RLBC outperforms a previously developed approach and allows for more persistent cluster heads with higher durations and stable connections with their members.

集群是一种在车联网 (IoV) 上实现高可扩展性的有效技术。然而，形成和维护集群产生的延迟和开销是大规模采用这种技术的常见障碍。为此，我们提出了一种有效的车联网聚类方案。利用强化学习，我们的方案可以快速形成网络状态感知集群。此外，我们基于强化学习的集群方案 (RLBC) 确保集群的动态和协作维护。我们的方案的有效性是通过广泛的模拟来评估的。模拟结果表明，RLBC 优于先前开发的方法，并允许具有更长持续时间的更持久的簇头和与其成员的稳定连接。