Equilibrium-Traffic-Networks

This document describes the datasets generated for and used in the study "A hybrid deep-learning-metaheuristic framework for bi-level network design problems" by Bahman Madadi and Gonçalo H. de Almeida Correia, published in Expert Systems with Applications. The datasets are generated and used to train and evaluate models for solving the User Equilibrium (UE) problem on three transportation networks (Sioux-Falls, Eastern-Massachusetts, and Anaheim) from the well-known "transport networks for research" repository. You can just download the datasets from this repository with the comprehensive metadata.

Dataset citation: Madadi, Bahman (2024). Equilibrium-Traffic-Networks. figshare. Dataset. https://doi.org/10.6084/m9.figshare.27889251.v1

Metadata

Network	Nodes	Edges	OD Pairs	Train Samples	Val Samples	Test Samples	Dataset Size	Solvers	Algorithm
SiouxFalls	24	76	576	18000	1000	1000	20,000	Aeq, Ipp	BFW
Eastern-Massachusetts	74	258	5476	4000	500	500	5,000	Aeq, Ipp	BFW
Anaheim	416	914	1444	4000	500	500	5,000	Aeq, Ipp	BFW

Features and Data Fields

Field	Type	Description
Node Features	Array	Represent origin-destination (OD) demand matrices for travel. Each OD pair specifies travel demand between zones.
Edge Features	Array	Include: Free-flow travel time (FFTT) and Capacity.
Edge Labels	Array	Optimal link flows derived from solving the DUE problem.
Number of Links	Int	Number of links in the network.
Number of Nodes	Int	Number of nodes in the network.
Number of OD Pairs	Int	Number of origin-destination pairs in the network.
Train Split	Int	Number of samples in the training set.
Validation Split	Int	Number of samples in the validation set.
Test Split	Int	Number of samples in the test set.
Dataset Size	Int	Total number of samples in the dataset.
Solvers	String	Solvers used for generating the dataset.
Algorithm	String	Algorithm used for generating the dataset.

Datasets

The datasets are generated using the scripts data_due_generate.py and data_dataset_prep.py from the GitHub repository. Each dataset corresponds to a different transportation network and contains solved instances of the DUE problem.

Dataset Structure

Each dataset is stored as a pickle file and contains three splits: train, validation, and test. Each split is a list of DGLGraph objects with node and edge features, along with edge labels.

• Node Features: Represent origin-destination (OD) demand matrices for travel. Each OD pair specifies travel demand between zones. Stored in the feat field of the DGLGraph.
• Edge Features: Include Free-flow travel time (FFTT) and Capacity. Stored in the feat field of the DGLGraph.
• Edge Labels: Optimal link flows derived from solving the DUE problem. A list of labels for each edge in the DGLGraph.

Available Networks

• SiouxFalls
• Eastern-Massachusetts
• Anaheim

Usage

To load a dataset, use the following code:

import pickle
from data_dataset_prep import DUEDatasetDGL

case = 'SiouxFalls'  # or 'Eastern-Massachusetts', 'Anaheim'
data_dir = 'DatasetsDUE'

with open(f'{data_dir}/{case}/{case}.pkl', 'rb') as f:
    train, val, test = pickle.load(f)

# Example: Accessing the first graph and its edge labels in the training set
graph, edge_labels = train[0]
print(graph)
print(edge_labels)

Data Generation Steps

1. Define Parameters: Set the parameters for dataset generation in the parameters() function.
2. Solve DUE Problem: Use the data_due_generate.py script to solve the DUE problem for each network.
3. Store Results: Save the results as CSV files and clean up the data.
4. Create DGL Dataset: Convert the data into DGL format using the data_dataset_prep.py script and save it as pickle files.

Scripts

data_due_generate.py (from GitHub repository)

This script generates the datasets by solving the DUE problem for each network in the benchmark networks. The parameters for dataset generation are defined in the parameters() function.

data_dataset_prep.py (from GitHub repository)

This script contains the classes and functions for preparing the datasets and converting them into DGL format.

References

• A hybrid deep-learning-metaheuristic framework for bi-level network design problems
• preprint
• GitHub Repository: HDLMF_GIN-GA
• Dataset repository

Citation

If you use these datasets in your research, please cite the following paper and dataset:

Paper citation:

Madadi B, de Almeida Correia GH. A hybrid deep-learning-metaheuristic framework for bi-level network design problems. Expert Systems with Applications. 2024 Jun 1;243:122814. https://doi.org/10.1016/j.eswa.2023.122814

Dataset citation:

Madadi, Bahman (2024). Equilibrium-Traffic-Networks. figshare. Dataset. https://doi.org/10.6084/m9.figshare.27889251.v1

---

Metadata

{
  "datasets": [
    {
      "name": "SiouxFalls",
      "description": "DGL dataset for the SiouxFalls transportation network with solved instances of the DUE problem.",
      "num_samples": 20000,
      "features": {
        "node_features": "Represent origin-destination (OD) demand matrices for travel. Each OD pair specifies travel demand between zones.",
        "edge_features": "Include Free-flow travel time (FFTT) and Capacity.",
        "edge_labels": "Optimal link flows derived from solving the DUE problem."
      },
      "splits": ["train", "val", "test"]
    },
    {
      "name": "Eastern-Massachusetts",
      "description": "DGL dataset for the Eastern-Massachusetts transportation network with solved instances of the DUE problem.",
      "num_samples": 5000,
      "features": {
        "node_features": "Represent origin-destination (OD) demand matrices for travel. Each OD pair specifies travel demand between zones.",
        "edge_features": "Include Free-flow travel time (FFTT) and Capacity.",
        "edge_labels": "Optimal link flows derived from solving the DUE problem."
      },
      "splits": ["train", "val", "test"]
    },
    {
      "name": "Anaheim",
      "description": "DGL dataset for the Anaheim transportation network with solved instances of the DUE problem.",
      "num_samples": 5000,
      "features": {
        "node_features": "Represent origin-destination (OD) demand matrices for travel. Each OD pair specifies travel demand between zones.",
        "edge_features": "Include Free-flow travel time (FFTT) and Capacity.",
        "edge_labels": "Optimal link flows derived from solving the DUE problem."
      },
      "splits": ["train", "val", "test"]
    }
  ],
  "references": [
    {
      "title": "A hybrid deep-learning-metaheuristic framework for bi-level network design problems",
      "doi": "10.1016/j.eswa.2023.122814",
      "url": "https://doi.org/10.1016/j.eswa.2023.122814",
      "authors": ["Bahman Madadi", "Gonçalo H. de Almeida Correia"],
      "journal": "Expert Systems with Applications",
      "year": 2024,
      "volume": 243,
      "pages": "122814"
    },
    {
      "title": "GitHub Repository: HDLMF_GIN-GA",
      "url": "https://github.com/bahmanmdd/HDLMF_GIN-GA"
    }
  ]
}