An Adaptive Clustering Approach for Dynamic Service Provisioning in Vehicular Cloud Networks

The integration of Internet of Things (IoT) and connected car technologies enables resource sharing and collaboration within the vehicular ecosystem. Through IoT connectivity, vehicles can dynamically share and optimize the utilization of resources. However, Vehicular Cloud Networks (VCNs) face unique challenges due to their dynamic topology changes caused by high vehicle mobility and unpredictability. Clustering is a key mechanism used in vehicular networks to address such issues. Clustering groups nearby vehicles into clusters to improve routing efficiency, reduce broadcast storms, and enable distributed coordination. When applied to VCNs, clustering can efficiently match resource requests to resource-rich vehicle candidates. This paper proposes an adaptive clustering approach that forms clusters based on vehicle mobility parameters and cloud resource profiles. A similarity score is used to group vehicles with similar attributes, and a connectivity-based method selects stable cluster heads. An adaptive joining policy dynamically adjusts clustering thresholds in response to changes in network traffic. Simulation results show that the proposed approach achieves more stable clusters and lower overhead compared to a baseline clustering algorithm. It also balances cluster count and service quality under varying network conditions through adaptive threshold adjustment.