Cache Strategy for Joint Content Recommendation and D2D Collaboration

Device to Device(D2D) caching employs direct communication technology to facilitate content sharing among adjacent users, effectively harnessing the storage resource of numerous terminal devices. Integrating user request diversity, resource limitation and content recommendation on user request probability, a joint optimization strategy for recommendation and collaborative cache is proposed to maximize the cache hit rate in D2D network. Firstly, the concept of regional user request probability is introduced, and the impact of content recommendation on user request probability is analyzed. And then, an initial regional user request model based on multi-level round-by-round iteration heuristic and long short-term memory algorithm is proposed to predict content popularity within collaborative clusters. Finally, combining recommendation and collaboration, the regional user request model is reconstructed, and an optimization algorithm is proposed to address the problem of content caching in D2D networks. Compared with other studies under different recommendation acceptance probabilities and cache capacities, the results show that the proposed algorithm improves the local cache hit rate by at least 24% and 9%, respectively.