rust_qlearn

A rust implementation of the Q-Learning algorithm

This implementation requires a finite, discrete state space.

In the botworld example provided, the 2D botworld state space is converted into a 1D state space for the QLearner using the discretize() function.

To run an example training session, clone this repository, cd into it, and run the following:

export RUST_LOG_LEVEL=4
cargo test -- --nocapture

You will see output like the following indicated a successful test:

#### RUNNING QLEARNER TEST ####
Start location: (9, 0)
Goal location: (0, 9)
5000 iterations, avg reward: -49.0656
#### QLEARNER TEST COMPLETE ####