Please run the cells in notebooks/osmnx_generator.ipynbto enable generation of graph inputs from OSMnx -- this may take a few minutes.

Code for all experiments can be found in the /notebooks/ directory. Experiment files follow the naming convention "exp_[experiment name]". GNN and greedy-t hyperparameter tuning can be found in /notebooks/gnn_tuner.ipynb and /notebooks/greedy_tuner.ipynb, respectively.

/gnn_library/ contains all information relevant to defining, training, testing, and saving GNN models. Several pre-trained models can be loaded from `/trained_models/'. The experiments contain many examples of loading these models.

Finally, instance_generator.py is used for the generation of OBBM instances of different graph configurations, torch_converter.py gives code for converting instances into PyG representations, evaluate.py contains code for estimating average competitive ratio for learned models, and algorithms.py contains implementations of baseline algorithms.