Multifactor Recommendation-based Video Caching Strategy in Mobile Edge Computing
Qian Liu, Ying Nian Wu, Qilie Liu · 2022
Mobile edge computing (MEC) can provide users with high-quality video services by placing computing and storage capacity close to the user. Transferring large size video files usually consumes more time and energy, and the emergence of edge caching can effectively solve this problem. Current caching schemes generally consider the impact of only one or two factors among attributes such as ratings and reviews, without considering the impact of multiple factors together on the recommender system. In this paper, we propose a video caching strategy based on multi-factor recommendation (VCSMFR) to solve the above problem. First, the video file ratings and corresponding rankings are obtained by a recommendation algorithm that fuses multi-factor data (e.g., reviews, directors, and actors). Then, an optimized particle swarm algorithm is used to make caching decisions for files stored on the edge MEC server to solve the problem that traditional particle swarm algorithms are prone to local convergence. Simulation results show that the recommendation algorithm proposed in this paper can analyze user and video information more carefully by fusing multiple factors, and improve the cache hit rate by 11% over the traditional caching scheme. In addition, the greedy algorithm is introduced into the optimized particle swarm algorithm, which improves the local search ability as well as the convergence of the algorithm and achieves faster caching decisions.