This model is describe at the following paper:
F. Geyer, G. Carle, "Learning and generating distributed routing protocols using graph-based deep learning", Proceedings of ACM SIGCOMM BigDAMA, 2018.

For this example you can find the directory data containing a very small subset of the dataset that we used, so that the user can get a general idea of what the dataset looks like --and maybe even reproduce it. Notice however that this dataset is enough to obtain an accurate model.

In addition to this dataset, we have included the rest of files properly filled, and thus it requires only to be executed. In fact, to train the corresponding GNN model to implement the GQNN, simply run:

    python main.py

This command will create the GNN specified in model_description.yaml file. We have also included the global_variables file, even though right now is not used by the model_description file. To learn to use them, we refer the user to (https://ignnition.org/doc/global_variables/).

If you want to execute any other functionality that is not train and validate, simply change the main.py file, to specify the new functionality. Visit (https://ignnition.org/doc/train_and_evaluate/) for more information.