Deep Q-Learning Based Algorithm for Dynamic Adaptive Streaming over HTTP

EasyChair Preprint no. 780

Abstract

As mobile technology getting more advanced, watching streaming video becomes an indispensable action in our daily life. How to maintain or provide a good quality of Experience (QoE) for the video user during bad network condition is an important and challenging issue. Therefore, we propose a Deep Q-learning based algorithm to substitute the adaptive bitrate algorithm (ABR). Our algorithm aims to learn and improve the Quality of Experience (QoE) for the video user in overall network condition. The result shows that our Deep Q-learning algorithm is able to learn and select the segment with good quality which gives the user the better (QoE).

Keyphrases: Adaptive bitrate algorithm, Deep Q-Learning, Quality of Experience

@Booklet{EasyChair:780,
  author = {Goay Fuh Yang and Wei-Tsong Lee and Hsin-Wen Wei},
  title = {Deep Q-Learning Based Algorithm for Dynamic Adaptive Streaming over HTTP},
  howpublished = {EasyChair Preprint no. 780},

  year = {EasyChair, 2019}}