Deep Reinforcement Learning with Importance Weighted A3C for QoE enhancement in Video Delivery Services

Adaptive bitrate (ABR) algorithms are used to adapt the video bitrate based on the network conditions to improve the overall video quality of experience (QoE). Recently, reinforcement learning (RL) and asynchronous advantage actor-critic (A3C) methods have been used to generate adaptive bit rate algorithms and they have been shown to improve the overall QoE as compared to fixed rule ABR algorithms. However, a common issue in the A3C methods is the lag between behaviour policy and target policy. As a result, the behaviour and the target policies are no longer synchronized which results in suboptimal updates. In this work, we present ALISA: An Actor-Learner Architecture with Importance Sampling for efficient learning in ABR algorithms. ALISA incorporates importance sampling weights to give more weightage to relevant experience to address the lag issues with the existing A3C methods. We present the design and implementation of ALISA, and compare its performance to state-of-the-art video rate adaptation algorithms including vanilla A3C implemented in the Pensieve framework and other fixed-rule schedulers like BB, BOLA, and RB. Our results show that ALISA improves average QoE by up to 25%-48% higher average QoE than Pensieve, and even more when compared to fixed-rule schedulers.