<![CDATA[This study proposes a Deep Reinforcement-Driven Clustering and Routing Protocol (DRCRP) to enhance energy efficiency and routing stability in smart vehicular networks. The protocol integrates an Actor–Critic deep reinforcement learning framework with Proximal Policy Optimization (PPO) to enable adaptive decision-making in dynamic Internet of Vehicles (IoV) environments. Through continuous learning, DRCRP adjusts cluster head selection and routing paths according to real-time vehicular mobility, residual energy, and link quality. Simulation experiments conducted using NS-2 and VanetMobiSim show that DRCRP achieves superior performance compared to benchmark algorithms such as AI-EECR, GWO-CH, and DMCNF. Quantitatively, the proposed model improved the Packet Delivery Ratio (PDR) by up to 4.3%, reduced End-to-End Delay by 18–22%, and lowered Energy Consumption by 12–16%. Moreover, DRCRP effectively minimized communication overhead and extended cluster head and member lifetimes, confirming its ability to balance reliability and energy efficiency. These results demonstrate the capability of reinforcement learning-based architectures to support intelligent, sustainable, and scalable vehicular communication systems under complex mobility conditions]]>
Copyrights © 2025
