Simulation code used in the paper "Online learning adaptation strategy for DASH clients", presented at ACM MMSys 2016 (https://dl.acm.org/citation.cfm?id=2910603). The MATLAB code has been cleaned up and commented before upload, but the functional parts are exactly the ones used in the article simulations.
The code includes a Reinforcement Learning client for DASH video streaming, as well as the simple rate-based heuristic used as a baseline. The repository also contains the main simulation code