AA-bot

Simple Anti-Air game with RL-agent

Requirements

pip install -r requirements.txt

Run this to check environment

python3 src/check_env.py

Training

set variables TIMESTEPS, EPOCHS in src/train_PPO.py

set variable UNIX_SERVER=True in src/aaEnv.py to enable training on UNIX server

python3 src/train_PPO.py

see training results in tensorboard

tensorboard --logdir=logs

Usage

Play game:

python3 src/play.py

Press A and D to turn left and right. Space to shot.

Test agent: set variable TIMESTEPS in src/test_PPO.py

python3 src/test_PPO.py

Project Organization

├── LICENSE
│
├── README.md          <- README for developers
│
├── docs               <- Documentation
│
├── logs               <- Training logs for tensorboard
│
├── models             <- Trained models
│
├── references         <- Data dictionaries, manuals, and all other explanatory materials.
│
├── requirements.txt   <- The requirements file for reproducing the analysis environment, e.g.
│                         generated with `pip freeze > requirements.txt`
└── src                <- Source code for use in this project.