Inspiration
We are optimistic about the potential of AI in games, and AI can help us in games.
What it does
Try to code RL algorithms play the Lunar Lander in Gymnasium.
How we built it
Install Gymnasium library and play with it first, then try to build a policy to auto play it then try join some structure make our agent play it well.
Challenges we ran into
We want to join implement some policy from pervious paper about RL, but our agent does not work as what we think, it must have some mistake in our design.
Accomplishments that we're proud of
The agent sometimes works very well and look very cool.
What we learned
Basic concept of RL and confirmed some of our previous ideas is right.
What's next for Lunar Lander
Improve the algorithm and try to generalize it to a wider range of games.
Log in or sign up for Devpost to join the conversation.