This project is implementation of multiple AI agents based on different Reinforcement Learning methods to OpenAI Gymnasium Lunar-Lander environment which is classic rocket landing trajectory optimization problem.
- Documentation: https://lunarlander-gym.readthedocs.io.
Q-Learning Agent | Actor-critic Agent | |
---|---|---|
Training episodes | 3000 | 3000 |
Reward | 198.51 | 284.86 |
Output Models | link | link |
Demo | ![]() |
![]() |
RandomAgent | Gradient Policy Agent | |
---|---|---|
Training episodes | 0 | 10,000 |
Reward | -70.46 | 49.07 |
Output Models | link | link |
Demo | ![]() |
![]() |
The sources for lunarlander_gym can be downloaded from the Github repo
_.
-
Clone the repository
$ git clone git://github.com/ehsan2754/lunarlander_gym
-
Once you have a copy of the source, you can install it with:
$ sudo apt update && sudo apt upgrade $ sudo apt install make $ pip install -r requirements_dev.txt $ sudo make install
-
Now you can just immidiately use it:
$ lunarlander-gym -h usage: lunarlander_gym [-h] -m M options: -h, --help show this help message and exit -m M, --method M Specifies the Reinforcement Agent method { 0 -> Random, 1 -> Gradient based optimization, 2 -> Q-Learning Agent 3 -> Actor- critic }