Reinforcement learning - based adaptation and scheduling methods for multi-source DASH

Dynamic adaptive streaming over HTTP (DASH) has been widely used in video recently. In DASH, the client downloads chunks order from a server. The rate adaptation function at enhances user?s quality-of-experience (QoE) by choosing suitable quality level for each chunk to download based on network condition. Today networks such as content delivery networks, edge caching contentcentric etc. usually replicate contents multiple cache nodes. We study sources this work. multi-source streaming, may arrive out of due different conditions paths. Hence, guarantee high QoE, needs not only adaptation, but also scheduling. Reinforcement learning (RL) emerged state-of-the-art control method various fields recent years. This paper proposes two algorithms sources: RL-based with greedy scheduling (RLAGS) and (RLAS). build simulation environment training evaluation. efficiency proposed is proved via extensive simulations real-trace data.