Deep Reinforcement Learning for Video Caching and Updating in Mobile Edge Computing Systems

Using mobile edge computing (MEC) servers to cache and update video resources can not only save network bandwidth and computing resources, but also adjust the bit rate of video playback according to real-time network conditions, providing users with ideal video quality of experience ($\mathbf{Q o E}$). In this work, we propose a novel video caching and updating optimization scheme, which enables the MEC server to provide video caching and updating services for users. To achieve this, we design a video caching and update optimization algorithm based on deep reinforcement learning adaptive bitrate selection, which can perceive environmental state information and determine its own caching strategy based on environmental feedback. We perform simulations and show that our algorithm can significantly reduce bandwidth and computing resources in the edge network and better improve video quality of service when compared with several existing algorithm.