Dispatching ambulances in the Netherlands with Deep Reinforcement Learning

To run the code:

• Step 1: Make sure the environment is set up correctly

from Environment import Environment
env = Environment()
env.import_data()

• Step 2: Configure hyperparameters. In deep_q_learning_main.py you can configure number of episodes and their length. In Learner.py you can configure the batch size and in Memory.py the memory size. In Model.py you can configure the learning rate, and in QNet.py you can configure the layer nodes.
• Step 3: Run the deep_q_learning_main.py

If you have an already made version of the model, you can load it using the shelve module.

Project Report

The report for this project can be found here.

Information

Group project created in the context of TU Delft's CS4320TU Applied AI Project.

Team 12:

• Francesca Drummer
• Thomas Georgiou
• Athanasios Theocharis
• Silvana van der Voort