Edge Collaborative Assisted Caching Content Placement Optimization Strategy Based on DDSG

To improve the pre-fetching accuracy of mobile autonomous vehicles in highly dynamic network, we study the cache content update and placement optimization in edge collaborative caching systems, and propose an edge collaborative assisted cache content placement optimization scheme. This scheme leverages the mobility of intelligent driving vehicles as an edge node to create an effective cache content placement strategy, and solve the cooperation problem between roadside units (RSUs) that dynamically adapt to vehicle requests and intelligent driving vehicles. Specifically, by jointly optimizing vehicle scheduling, cache content task unloading ratio, and content placement decisions, the problem of minimizing cache task processing delay is formulated. Considering the non-convexity of the problem, high-dimensional state space and non-convexity of continuous action space, a cache content placement optimization algorithm based on deep deterministic policy gradient (DDPG) is proposed. Using this algorithm, the optimal cache content placement strategy can be obtained in an uncontrollable dynamic environment. Simulation results show that the proposed algorithm can converge quickly and has significant improvements in processing delay compared to other baseline algorithms.