Genetic algorithms for generation of new mobility networks

This project aims to generate new mobility networks using a Bio-Inspired approach. NSGA II is employed as the main evolutionary algorithm to evolve the networks.

Left: An example of a mobility network from MoGAN Right: The Pareto front of the degree distribution and the flux distribution of the networks generated by the evolutionary algorithm

This project is built on top of two works:

• Evolving Weighted and Directed Graphs with Constrained Properties
• Generating Mobility Networks with Generative Adversarial Networks

An overview of the project can be found in the report which is available in the repository.

Installation

First of all, clone our repository and right after that initialize the submodules:

git clone https://github.com/MattiaRigon/BioMobilityNetworks
git submodule update --init --recursive

In this way, MoGAN repository will be cloned and ready to be used. Then you can install the requirements:

pip install -r requirements.txt

Datasets

The datasets used in order to initialize the population are the ones used by the authors of MoGAN. They are available on the path GAN-flow/[dataset_name]/v_train.txt. The dataset names are:

• BikeNYC
• BikeCHI
• TaxiNYC
• TaxiCHI

For further information about the datasets, please refer to the MoGAN repository.

Usage

Firstly, you need to make sure that the dataset you want to use has a corresponding .pkl file in the data/preprocessed folder. If it does not exist, you can create it by running the following command:

python data/data_loader.py <dataset_name>

where <dataset_name> is the name of the dataset you want to use.

Evolution

Then you can run the main script by giving a configuration file as input following the format of the other config file in the config folder. For example:

python main.py config/config_taxi_nyc.json

Evaluation

In order to evaluate the results of the evolution, you can run the following command:

python analysis.py <path_to_experiment>

where <path_to_experiment> is the path to the folder containing the results of the evolution (it is defined in the configuration file).

Running multiple experiments

If you want to run multiple experiments, you can use the run_all.py script. You can define the experiments passing a folder containing the configs for the different experiments. For example:

python run_all.py config/

where the config folder contains the configuration files for the different experiments.

Custom optimization

If you want to run an optimization of different metrics or with different constraints, you can create a new property and define it in the graph_evolution/eval_functions.py file. Then you can follow the implementation in the reference_metrics.py file in order to get the target values for the new property. Finally, you can define the new property in the configuration file and run the optimization.

Authors

• Amir Gheser
• Mattia Rigon
• Matteo Mascherin