How can I use deep reinforcement learning for vehicle rerouting in SUMO?

Question

I want to use deep reinforcement learning for vehicle rerouting in SUMO, but I don't know how to start training the model.

I've already created road network and vehicle routing in SUMO-XML files (mymap.net.xml and mymap.rou.xml). Currently, I'm trying to train the model on Jupyter Notebook, importing TraCI library to control the SUMO simulator and allow for a reinforcement learning approach. However, I'm still confused in training step.

1. Do I need any traffic data to train my agent to take actions in the environment?

2. How can I train based on these SUMO-XML files I created?

3. Is it possible to run the simulation on Windows? or I need to change to Ubuntu instead?

I would appreciate if someone could guide me. Thank you in advance.

Answer 1

There is a platform called Flow, sounds like its what you're looking for

https://github.com/flow-project/flow