This repo contains the implementation of the DQN algorithm and two variants, the first with the introduction of the dueling network architecture and the second one with the application of Prioritized Experience Replay.
The algorithms were tested on the Flappy Bird environments.
More information can be found in the report.