Evaluation of Q-Learning approach for HTTP adaptive streaming
Virginia Martín, Julian M. Cabrera, Narciso N. Garcia · 2016
We propose a Q-Learning-based algorithm for an HTTP Adaptive Streaming (HAS) Client that maximizes the perceived quality, taking into account the relation between the estimated bandwidth and the qualities and penalizing the freezes. The results will show that it produces an optimal control as other approaches do, but keeping the adaptiveness.