## Playing Flappy Bird Using Deep Reinforcement Learning (Based on Deep Q Learning DQN)
## Include NIPS 2013 version and Nature Version DQN
I rewrite the code from another repo and make it much simpler and easier to understand Deep Q Network Algorithm from DeepMind
The code of DQN is only 160 lines long.
To run the code, just type python FlappyBirdDQN.py
Since the DQN code is a unique class, you can use it to play other games.
## About the code
As a reinforcement learning problem, we knows we need to obtain observations and output actions, and the 'brain' do the processing work.
Therefore, you can easily understand the BrainDQN.py code. There are three interfaces:
1. getInitState() for initialization
2. getAction()
3. setPerception(nextObservation,action,reward,terminal)
the game interface just need to be able to feed the action to the game and output observation,reward,terminal
## Disclaimer
This work is based on the repo: [yenchenlin1994/DeepLearningFlappyBird](https://github.com/yenchenlin1994/DeepLearningFlappyBird.git)
DRL-FlappyBird-master.zip （38个子文件） 
DRL-FlappyBird-master saved_networks 
network-dqn-50000.meta 72KB network-dqn-50000 13.72MB checkpoint 283B FlappyBirdDQN.py 1KB assets sprites 
4.png 3KB redbird-midflap.png 3KB 3.png 3KB base.png 664B 7.png 3KB 5.png 3KB 1.png 3KB 2.png 3KB background-black.png 4KB 6.png 3KB 8.png 3KB redbird-downflap.png 3KB 0.png 3KB 9.png 3KB pipe-green.png 5KB redbird-upflap.png 3KB audio swoosh.ogg 13KB point.ogg 13KB hit.ogg 15KB wing.ogg 8KB wing.wav 29KB point.wav 173KB die.ogg 17KB die.wav 190KB hit.wav 94KB swoosh.wav 346KB README.md 1KB game wrapped_flappy_bird.py 8KB flappy_bird_utils.py 3KB __pycache__ wrapped_flappy_bird.cpython-36.pyc 5KB flappy_bird_utils.cpython-36.pyc 2KB __pycache__ BrainDQN_Nature.cpython-36.pyc 6KB BrainDQN_Nature.py 7KB BrainDQN_NIPS.py 5KB
评论1