/grid-world-rl

Value iteration, policy iteration, and Q-Learning in a grid-world MDP.

Primary LanguagePythonMIT LicenseMIT

grid-world-rl

Implementations of MDP value iteration, MDP policy iteration, and Q-Learning in a toy grid-world setting.

TODO

The policy iteration implementation is suboptimal, as it does not use the closed-form solution. Pull requests are welcome.