基于深度强化学习的基站协作与服务缓存研究OA

Edge caching technology,which stores content in proximity to end users,can effectively reduce transmission redundancy and service latency,making it one of the key enabling technologies for mobile edge networks.Given the limited cache capacity of individual base stations,cooperative edge caching can significantly improve storage utilization through resource sharing among base stations.To address the joint optimization problem of base station cooperation and service caching in mobile edge networks,this study proposes an efficient dynamic decision-making algorithm that integrates Lyapunov optimization with the deep deterministic policy gradient(DDPG)algorithm.The proposed approach transforms the latency minimization problem under long-term cost constraints into a time-slot-level optimization task.By leveraging DDPG to handle dynamic optimization in high-dimensional and nonlinear environments,the algorithm achieves efficient allocation and coordination of system resources.Simulation results demonstrate that,compared with existing approaches,the proposed algorithm is more effective in reducing overall system latency while satisfying cost constraints.