Deep Reinforcement Learning Empowered Edge Collaborative Caching Scheme for Internet of Vehicles

With the development of internet vehicles, traditional centralized content caching mode transmits through core network, which causes a large delay and cannot meet demands for delay-sensitive services. To solve these problems, on basis vehicle we propose an edge collaborative scheme. Road side unit (RSU) mobile computing (MEC) are used to collect information, predict cache popular content, thereby provide low-latency delivery However, storage capacity single RSU severely limits performance handle intensive requests at same time. Through sharing, can relieve burden servers. Therefore, integrate build MEC-assisted (MVECC) scheme, so as realize among cloud, vehicle. MVECC uses deep reinforcement learning what needs be cached RSU, enables RSUs more content. In addition, also introduces mobility-aware replacement scheme network reduce redundant improving efficiency, allows dynamically replace in response mobility vehicles. The simulation results show that proposed improve terms energy cost hit rate.