dodopool/RL-Taxi-Driver

An RL agent trained via DQNs for maximizing rewards in the given environment.

RL-Taxi-Driver

A Reinforcement Learning agent trained via DQNs (Deep Q-Networks) for maximizing rewards in the given environment.

A short description of the files is as follows:

1. Report.pdf: Includes details on the architecture used and the results
2. Converged_1900.h5: A converged model, this can be used in the notebook named RL Model Testing.ipynb
3. TM_lower.npy: Contains the time matrix in numpy format, which tells time taken to go from one place to another at a given day and time.
4. RL Model Testing.ipynb: This notebook can be used for testing the model.
5. RL Model Training.ipynb: This notebook is used for training and plotting results from the model.