voting_traffic

This repository contains the code needed to run the simulations detailed in our paper, Democratizing traffic control in smart cities.

Setup

To run this, please install this fork of CityFlow: https://github.com/mbkorecki/CityFlow

The following packages and versions are needed:

geopandas==0.12.1
gym==0.21
torch>=1.13
dill==0.3

Installation can be done using

pip install -r requirements.txt

Training

Training scripts may be found in intersection_switching/run_train.py. Use this to train the models needed.

Running experiments

Training scripts may be found in intersection_switching/run_vote.py.

Citing

If you have found any part of this work useful or relevant to your research, please cite our article:

@article{Korecki2024,
  title = {Democratizing Traffic Control in Smart Cities},
  doi = {10.1016/j.trc.2024.104511},
  journal = {Transportation Research Part C: Emerging Technologies},
  author = {Korecki,  Marcin and Dailisan,  Damian and Yang,  Joshua and Helbing,  Dirk},
  year = {2024},
  vol = {160},
  pages = {104511}
}