Reinforcement Learning-Based HTTP Adaptive Streaming Scheme in Edge Computing Environments

Abstract

Dynamic Adaptive Streaming over HTTP (DASH) is a representative scheme to improve Quality of Experience (QoE) of video streaming users. However, the QoE fairness is degraded when multiple users stream videos over the same network. The reasons of this problem are a lack of coordination among clients and bitrate adaptation with the fixed heuristics. In this paper, we propose an HTTP adaptive streaming scheme with reinforcement learning in edge computing environments. The proposed scheme generates a policy based on reinforcement learning to optimize QoE fairness and utilizes edge computing to assist policy generation. We evaluated the proposed scheme through simulations under various network conditions. The experimental results show that the proposed scheme improves QoE fairness of multiple clients compared with the existing scheme.