Pith. sign in

REVIEW 4 major objections 5 minor 39 references

Learning traffic flows: Graph Neural Networks for Metamodelling Traffic Assignment

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims a message-passing graph neural network with edge features can learn stochastic-user-equilibrium link flows and stay accurate when capacities or speed limits move out of distribution.

desk verdict A careful, honest benchmark of a GatedGCN metamodel for SUE on Sioux Falls, but the abstract overclaims out-of-distribution robustness because the demand-shift experiments fail. read the letter →

arxiv 2505.11230 v1 pith:MP2HET37 submitted 2025-05-16 cs.LG

Oskar Bohn Lassen , Serio Agriesti , Mohamed Eldafrawi , Daniele Gammelli , Guido Cantelmo , Guido Gentile , Francisco Camara Pereira This is my paper · ORCID
classification cs.LG
keywords trafficassignmentstochasticuserequilibriumgraphneuralnetworksmessage-passingout-of-distributiongeneralizationedgefeaturesflowpredictionmetamodel
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that a message-passing graph neural network can act as a metamodel for the Stochastic User Equilibrium traffic assignment problem: given a network with origin-destination demand, link capacities, speed limits, and free-flow times, it predicts equilibrium link flows without running the iterative simulator. The authors argue that by mirroring the simulator's message-passing structure and by feeding edge attributes into the network, the model should generalize to inputs outside its training range. They report that on the Sioux Falls network the proposed GatedGCN outperforms an MLP, a node-only GCN, and a mean-flow baseline in-distribution, and keeps low error when capacities and speed limits are moved up to 25% out of range. If this holds, traffic planners could replace repeated stochastic user equilibrium simulations with a fast forward pass for scenario analysis, which matters because conventional assignment is computationally expensive for large networks.

What carries the argument

The central object is a message-passing neural network built from GatedGCN layers, with node features holding the origin-destination demand rows (padded with zeros for non-centroid nodes) and edge features holding free-flow travel time, speed limit, and capacity. Each GatedGCN layer computes a message from the sender's and receiver's node embeddings plus the connecting edge's embedding, sums incoming messages at every node, and updates the node state; six such layers propagate demand and network attributes across the graph. The decoder concatenates the final embeddings of a link's endpoints with the edge features and sends them through an MLP to output the predicted flow on that link. This exchange-and-update structure is meant to mimic the iterative travel-cost information exchange in a traffic simulator, and it is the mechanism by which capacity and speed-limit perturbations enter the prediction.

What would settle it

Train the same GatedGCN on a training set whose OD matrices are drawn to vary total network demand substantially rather than with roughly uniform per-parameter marginals, then evaluate on the 70% OD-pair perturbation used in Fig. 4; the central claim would be contradicted if the model still falls below the mean baseline under that distribution shift.

Watch

Extended reading notes

Core claim

The paper claims that a GatedGCN message-passing network with edge features (free-flow travel time, speed limit, capacity) and node features formed by the origin-destination matrix can learn the Stochastic User Equilibrium mapping from network inputs to link flows. The key evidence is the out-of-distribution experiments: when 10–90% of links have capacities or speed limits moved up to 25% beyond the training range, the GatedGCN keeps mean absolute error low (below 0.4 even at 90% capacity changes) and beats all baselines for capacity perturbations, while the MLP stays competitive for speed limits and in-distribution. The paper's own boundary condition is demand: with 70% of OD pairs perturbed, the GatedGCN falls below the mean baseline and the MLP is best, which the authors attribute to the limited variation in total demand created by their Latin Hypercube sampling. Thus the strong claim is that edge features make the GNN robust to out-of-distribution changes in the network's physical attributes, not necessarily to all network-input shifts.

Load-bearing premise

The training set, built by sampling each parameter individually with Latin Hypercube Sampling, must be diverse enough in total demand for the model to extrapolate to out-of-distribution demand shifts, and the paper's own results show this assumption fails when 70% of OD pairs are perturbed.

Editorial extensions

If this is right

  • If the central claim holds, a trained message-passing metamodel can serve as a drop-in surrogate for SUE simulation when evaluating capacity or speed-limit scenarios, cutting the time per scenario from iterative assignment to a single network forward pass.
  • Edge attributes become the mechanism of extrapolation: the node-only GCN baseline degrades sharply on capacity changes, so the practical recipe is to include free-flow time, speed, capacity, and the demand matrix explicitly in the model.
  • The in-distribution result (MAE 0.029, R² 0.95) indicates that learning-based metamodeling has reached practical accuracy on a mid-size benchmark, supporting its use in repeated evaluation loops.
  • The demand-shift failure at 70% OD perturbation implies the surrogate is reliable for infrastructure-side changes but should not be trusted for large demand shocks without retraining or richer training data.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next test, which the paper does not run, is to train on datasets with deliberately wide variation in total OD demand; the paper's own diagnosis suggests that demand OOD behavior would improve substantially, isolating data generation rather than architecture as the bottleneck.
  • If the architecture generalizes to larger networks as the authors hypothesize, the model could be embedded as a differentiable layer in network design and capacity-expansion optimization, enabling gradient-based search over capacity and speed decisions without calling a simulator.
  • The same edge-feature message-passing recipe could be transferred to dynamic traffic assignment or multimodal assignment by adding temporal or mode-specific edge attributes, though the paper only studies static SUE.
  • Treating centroids separately from junctions, as the paper does, is a precondition for applying the learned model to real cities where OD data exists at zone level; a test on a second real network would confirm whether this design choice is sufficient.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a message-passing graph neural network (GatedGCN) as a metamodel for Stochastic User Equilibrium traffic assignment, trained on 10,000 synthetic Sioux Falls scenarios generated with Latin Hypercube Sampling and labeled by PTV Visum. The authors compare the GatedGCN against an MLP, a GCN, and a mean-flow baseline on in-distribution prediction and on out-of-distribution (OOD) perturbations of speed limits, capacities, and OD demand. The main claims are that the GatedGCN incorporates edge features effectively and achieves high OOD performance on changing capacities and speed limits, while the broader abstract and conclusion claim general OOD robustness. The in-distribution results show the GatedGCN marginally outperforms the MLP, and the OOD capacity/speed experiments show the GatedGCN is competitive or superior; however, under OD-demand perturbations the GatedGCN degrades sharply and is outperformed by the MLP and even the mean baseline at high perturbation levels.

Significance. If the capacity- and speed-OOD results hold, the paper makes a useful contribution to fast traffic-assignment metamodeling: it demonstrates a concrete architecture that uses edge features and transfers to inputs outside the training range for two practically relevant perturbation types, and it provides an honest comparison against an MLP baseline that is often neglected in GNN-for-traffic work. The experimental design, including external simulator ground truth, held-out OOD evaluation, and a deliberately challenging mean baseline, is a strength. However, the central 'out-of-distribution' claim as stated in the abstract and conclusion is only partially supported because the OOD demand-shift experiments fail for the proposed model, and the paper's own limitation section acknowledges the root cause. The theoretical framing of SUE in Eq. (1) is also not standard and would need correction. Overall, the empirical core is plausible but the manuscript overstates the generality of its OOD findings.

major comments (4)
  1. [Section VI, Abstract] The general out-of-distribution claim in the abstract and conclusion is not supported by the full set of experiments. Figure 4 shows that when 70% of OD pairs are perturbed, the GatedGCN's MAE exceeds that of the mean-flow baseline, and the MLP is the best model from 20% perturbation onward. The authors themselves attribute this to the limited variation in total demand caused by the LHS sampling strategy (Section IV-D.3 and Section V). Since OD demand is a first-class input for traffic assignment and for the stated real-time planning application, the capacity/speed OOD results are only a partial demonstration. The claims in the abstract ('outside the domain on which it was trained') and conclusion ('high performance out-of-distribution on changing capacities and speed limits') should be narrowed accordingly, or the demand-shift experiments should be revisited with a training distribution designed to support extrapolation in total demand.
  2. [Section II-A] The mathematical formulation of SUE in Eq. (1) is not a standard SUE objective and contains an undefined term. The expression 'Var(f_e)' is not defined: variance over what randomness or population, and how does it depend on the decision variables f_e? Moreover, minimizing the sum of link travel-time integrals plus a variance term does not correspond to the standard logit-based SUE formulation (e.g., Sheffi's expected perceived travel cost or the entropy-like objective). Since the paper states this is the problem being approximated by the GNN, the theoretical background should either be corrected to a standard SUE formulation or explicitly presented as a stylized variant with all terms defined.
  3. [Section IV-B] The training losses differ across models: the GatedGCN is trained with an L1 loss (Eq. (10)), while the MLP baseline is trained with MSE (Section IV-B). Table I then compares all models on MAE, R², MSE, and RMSE on the same test set. A model trained to minimize L1 is expected to have a lower MAE, and a model trained with MSE is expected to have a lower MSE, so this setup confounds architectural differences with loss-function differences. At minimum, the authors should train all models with the same loss, or report results under both losses, before drawing conclusions about which architecture is better in-distribution and under OOD shifts.
  4. [Section IV] No error bars, confidence intervals, or multiple-seed results are reported anywhere in the paper. The in-distribution differences between GatedGCN (MAE 0.02899) and MLP (MAE 0.03077) are small, and the OOD curves in Figures 2-4 appear to be single runs; without repeated runs or confidence intervals, the ranking claims, especially the claim that GatedGCN outperforms MLP under speed perturbations, are not established as statistically reliable.
minor comments (5)
  1. [Section III] The message-passing notation is inconsistent: Eq. (5) defines m_{ij}^{(l)}, while Eq. (6) sums m_{i->j}^{(l)}; Eq. (7) writes the update for node i using an aggregated message denoted `m ^ l j` that appears to be ^ m_j^{(l)} but with a subscript j rather than i. These equations should be harmonized so that the update for node i uses the aggregated message at node i.
  2. [Fig. 1] The caption of Figure 1 ends with 'GCN and mean comparison' and appears to be truncated; it should be completed with a full description of the plot.
  3. [Section IV-B] The R² metric used in Table I is not defined; it should be specified whether it is the coefficient of determination computed per edge, pooled across edges, or averaged across samples, since R² is not invariant to the aggregation unit for graph-structured regression.
  4. [Section V] The Limitations section is candid and useful, but the abstract and conclusion currently contradict it by making a broad OOD claim. The authors should align the high-level claims with the limitations they themselves identify.
  5. [Throughout] There are several typographical and grammatical errors, e.g., 'which would instead by captured' in the Introduction and 'excel by mimicking' in Section II-B; a careful proofreading pass is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: learned flows are benchmarked against an external simulator, not derived from the inputs.

full rationale

The proposed GatedGCN is a supervised regression model mapping node OD demand and edge features (free-flow time, speed, capacity) to SUE link flows. Ground truth comes from PTV Visum, an external simulator; the loss (Eq. 10) measures error against these simulated flows, and in- and out-of-distribution evaluations are on held-out scenarios with independently recomputed SUE assignments. No model parameter is fitted to the evaluation outputs, and no prediction is equivalent by construction to an input feature: link flow is not a component of the input representation in Eqs. (2)-(3). The architectural description 'mimic the algorithmic structure' is motivating rhetoric, not a derivation step, and the only self-citations (e.g., refs. [3,4,6,8-10,12-14]) are background algorithm references, not load-bearing evidence for the empirical claim. The limitation stated in Section V about LHS-generated demand variation weakens the general out-of-distribution claim, but that is an external-validity or correctness concern, not circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim depends on the hand-chosen synthetic data ranges, the OOD perturbation levels, and the architecture hyperparameters. The main domain assumption is that PTV Visum's SUE is a valid ground truth. No invented physical entities are introduced.

free parameters (4)
  • OOD perturbation magnitude = up to 25% beyond original range
    Hand-chosen in Section IV-A2; the OOD results would likely differ if this magnitude changed.
  • Fraction of modified links/nodes = 10%-90%
    Hand-chosen perturbation levels; results in Figs. 2-4 depend on this.
  • Input feature ranges (OD, speed, capacity) = 0-1500 veh, 45-80 km/h, 4000-26000
    Chosen in Section IV-A to be realistic; the synthetic training distribution is defined by these ranges.
  • Number of hidden units and layers = 64 hidden units, 6 layers
    Chosen in Section IV-B; 6 layers are chosen to match the network diameter, an ad hoc design choice.
assumptions (4)
  • domain assumption PTV Visum's SUE assignment produces correct ground-truth equilibrium flows
    The entire supervised learning pipeline uses PTV Visum outputs as target labels (Section IV-A); if these differ from true SUE, the model learns a biased target.
  • domain assumption The LHS-generated synthetic scenarios are representative of realistic traffic conditions
    The uniform ranges and independent sampling of OD, speed, and capacity may not reflect real-world correlations; the authors acknowledge limited total demand variation in Section V.
  • ad hoc to paper A 6-layer GatedGCN with messages passing along graph edges can represent the SUE mapping
    The choice of 6 layers is justified by the diameter of Sioux Falls, but there is no theoretical guarantee that message passing of this depth approximates the SUE fixed point; this is a learned hypothesis.
  • standard math Standard supervised learning assumptions: training and validation sets are drawn from the same distribution
    The models are trained on 60% of ID data and validated on 20% (Section IV-A1); this is standard but not explicitly discussed for OOD use.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning traffic flows: Graph Neural Networks for Metamodelling Traffic Assignment." pith.science (2026). https://pith.science/paper/MP2HET37

@misc{pith2026250511230,
  author       = {Pith},
  title        = {Pith review of: Learning traffic flows: Graph Neural Networks for Metamodelling Traffic Assignment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MP2HET37}},
  note         = {Machine review of arXiv:2505.11230}
}
read the original abstract

The Traffic Assignment Problem is a fundamental, yet computationally expensive, task in transportation modeling, especially for large-scale networks. Traditional methods require iterative simulations to reach equilibrium, making real-time or large-scale scenario analysis challenging. In this paper, we propose a learning-based approach using Message-Passing Neural Networks as a metamodel to approximate the equilibrium flow of the Stochastic User Equilibrium assignment. Our model is designed to mimic the algorithmic structure used in conventional traffic simulators allowing it to better capture the underlying process rather than just the data. We benchmark it against other conventional deep learning techniques and evaluate the model's robustness by testing its ability to predict traffic flows on input data outside the domain on which it was trained. This approach offers a promising solution for accelerating out-of-distribution scenario assessments, reducing computational costs in large-scale transportation planning, and enabling real-time decision-making.

Figures

Figures reproduced from arXiv: 2505.11230 by the authors.

Figure 1
Figure 1. Mean absolute error (MAE) of all models per edge. The [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 4
Figure 4. Mean absolute error across all models when changing x% [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 3
Figure 3. Mean absolute error across all models when changing x% [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 35 canonical work pages

  1. [1]

    Sheffi,Urban transportation networks

    Y . Sheffi,Urban transportation networks. Prentice-Hall, Englewood Cliffs, NJ, 1985, vol. 6

    work page 1985

  2. [2]

    Large- scale multimodal transportation network models and algorithms-part i: The combined mode split and traf- fic assignment problem,

    Y . Fan, J. Ding, H. Liu, Y . Wang, and J. Long, “Large- scale multimodal transportation network models and algorithms-part i: The combined mode split and traf- fic assignment problem,”Transportation Research Part E: Logistics and Transportation Review, vol. 164, p. 102832, 2022

    work page 2022

  3. [3]

    Local user cost equilibrium: a bush-based algorithm for traffic assignment,

    G. Gentile, “Local user cost equilibrium: a bush-based algorithm for traffic assignment,”Transportmetrica A: Transport Science, vol. 10, no. 1, pp. 15–54, 2014

    work page 2014

  4. [4]

    Reduced gradient algorithm for user equilibrium traf- fic assignment problem,

    A. Babazadeh, B. Javani, G. Gentile, and M. Florian, “Reduced gradient algorithm for user equilibrium traf- fic assignment problem,”Transportmetrica A: Transport Science, vol. 16, no. 3, pp. 1111–1135, 2020

    work page 2020

  5. [5]

    An algorithm for the stochastic user equilibrium problem,

    O. Damberg, J. T. Lundgren, and M. Patriksson, “An algorithm for the stochastic user equilibrium problem,” Transportation Research Part B: Methodological, vol. 30, no. 2, pp. 115–131, 1996

    work page 1996

  6. [6]

    New formulations of the stochastic user equilibrium with logit route choice as an extension of the deterministic model,

    G. Gentile, “New formulations of the stochastic user equilibrium with logit route choice as an extension of the deterministic model,”Transportation science, vol. 52, no. 6, pp. 1531–1547, 2018

    work page 2018

  7. [7]

    A general formulation for multi-modal dynamic traffic assignment considering multi-class vehicles, public transit and parking,

    X. Pi, W. Ma, and Z. S. Qian, “A general formulation for multi-modal dynamic traffic assignment considering multi-class vehicles, public transit and parking,”Trans- portation Research Part C: Emerging Technologies, vol. 104, pp. 369–389, 2019

    work page 2019

  8. [8]

    Spillback congestion in dynamic traffic assignment: a macroscopic flow model with time-varying bottlenecks,

    G. Gentile, L. Meschini, and N. Papola, “Spillback congestion in dynamic traffic assignment: a macroscopic flow model with time-varying bottlenecks,”Transporta- tion Research Part B: Methodological, vol. 41, no. 10, pp. 1114–1138, 2007

    work page 2007

Show all 39 references
  1. [9]

    A dynamic traffic assignment model for highly congested urban networks,

    M. E. Ben-Akiva, S. Gao, Z. Wei, and Y . Wen, “A dynamic traffic assignment model for highly congested urban networks,”Transportation research part C: emerg- ing technologies, vol. 24, pp. 62–82, 2012

    2012

  2. [10]

    The general link transmission model for dynamic network loading and a comparison with the due algorithm,

    G. Gentileet al., “The general link transmission model for dynamic network loading and a comparison with the due algorithm,”New developments in transport planning: advances in Dynamic Traffic Assignment, vol. 178, p. 153, 2010

    2010

  3. [11]

    Hierarchical travel demand estimation using multiple data sources: A forward and backward propagation algorithmic frame- work on a layered computational graph,

    X. Wu, J. Guo, K. Xian, and X. Zhou, “Hierarchical travel demand estimation using multiple data sources: A forward and backward propagation algorithmic frame- work on a layered computational graph,”Transportation Research Part C: Emerging Technologies, vol. 96, pp. 321–346, 2018

    2018

  4. [12]

    Using traffic data of various types in the estimation of dynamic od matrices,

    B. Kostic and G. Gentile, “Using traffic data of various types in the estimation of dynamic od matrices,” in2015 International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS). IEEE, 2015, pp. 66–73

    2015

  5. [13]

    Methodology for o-d matrix estimation using the re- vealed paths of floating car data on large-scale networks,

    A. Mitra, A. Attanasi, L. Meschini, and G. Gentile, “Methodology for o-d matrix estimation using the re- vealed paths of floating car data on large-scale networks,” IET Intelligent Transport Systems, vol. 14, no. 12, pp. 1704–1711, 2020

    2020

  6. [14]

    A practical approach to assignment-free dynamic origin–destination matrix estimation problem,

    X. Ros-Roca, L. Montero, J. Barcel ´o, K. N ¨okel, and G. Gentile, “A practical approach to assignment-free dynamic origin–destination matrix estimation problem,” Transportation Research Part C: Emerging Technologies, vol. 134, p. 103477, 2022

    2022

  7. [15]

    Deep learning for intelligent transportation systems: A survey of emerging trends,

    M. Veres and M. Moussa, “Deep learning for intelligent transportation systems: A survey of emerging trends,” IEEE Transactions on Intelligent transportation systems, vol. 21, no. 8, pp. 3152–3168, 2019

    2019

  8. [16]

    A graph neural network (gnn)-based approach for real-time es- timation of traffic speed in sustainable smart cities,

    A. Sharma, P. Nikashina, V . Gavrilenko, A. Tselykh, A. Bozhenyuk, M. Masud, and H. Meshref, “A graph neural network (gnn)-based approach for real-time es- timation of traffic speed in sustainable smart cities,” Sustainability, vol. 15, no. 15, 2023

    2023

  9. [17]

    Dynamic temporal position observant graph neural network for traffic forecasting,

    L. Waikhom, R. Patgiri, and L. D. Singh, “Dynamic temporal position observant graph neural network for traffic forecasting,”Applied Intelligence, vol. 53, no. 20, pp. 23 166–23 178, 2023

    2023

  10. [18]

    Co-prediction of multiple transportation demands based on deep spatio-temporal neural network,

    J. Ye, L. Sun, B. Du, Y . Fu, X. Tong, and H. Xiong, “Co-prediction of multiple transportation demands based on deep spatio-temporal neural network,” inProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 2019, pp. 305–313

    2019

  11. [19]

    Graph neural networks for intelligent transportation systems: A survey,

    S. Rahmani, A. Baghbani, N. Bouguila, and Z. Patterson, “Graph neural networks for intelligent transportation systems: A survey,”IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 8, pp. 8846–8885, 2023

    2023

  12. [20]

    Graph neural network for traffic forecasting: The research progress,

    W. Jiang, J. Luo, M. He, and W. Gu, “Graph neural network for traffic forecasting: The research progress,” ISPRS International Journal of Geo-Information, vol. 12, no. 3, 2023

    2023

  13. [21]

    Dynamic multi-graph neural network for traffic flow prediction incorporating traffic accidents,

    Y . Ye, Y . Xiao, Y . Zhou, S. Li, Y . Zang, and Y . Zhang, “Dynamic multi-graph neural network for traffic flow prediction incorporating traffic accidents,”Expert Sys- tems with Applications, vol. 234, 2023

    2023

  14. [22]

    Traffic flow matrix-based graph neural network with attention mechanism for traffic flow prediction,

    J. Chen, L. Zheng, Y . Hu, W. Wang, H. Zhang, and X. Hu, “Traffic flow matrix-based graph neural network with attention mechanism for traffic flow prediction,” Information Fusion, vol. 104, 2023

    2023

  15. [23]

    Multi-view dynamic graph convolution neural network for traffic flow prediction,

    X. Huang, Y . Ye, X. Yang, and L. Xiong, “Multi-view dynamic graph convolution neural network for traffic flow prediction,”Expert Systems with Applications, vol. 222, 2023

    2023

  16. [24]

    Improving traffic density forecasting in intelligent transportation systems using gated graph neural networks,

    R. H. Khan, J. Miah, S. Y . Arafat, M. M. Syeed, and D. M. Ca, “Improving traffic density forecasting in intelligent transportation systems using gated graph neural networks,” in2023 15th International Conference on Innovations in Information Technology (IIT), 2023, pp. 104–109

    2023

  17. [25]

    Graph neural network rein- forcement learning for autonomous mobility-on-demand systems,

    D. Gammelli, K. Yang, J. Harrison, F. Rodrigues, F. C. Pereira, and M. Pavone, “Graph neural network rein- forcement learning for autonomous mobility-on-demand systems,” in2021 60th IEEE Conference on Decision and Control (CDC), 2021, pp. 2996–3003

    2021

  18. [26]

    Domain adversarial graph neural network with cross-city graph structure learning for traffic prediction,

    X. Ouyang, Y . Yang, Y . Zhang, W. Zhou, J. Wan, and S. Du, “Domain adversarial graph neural network with cross-city graph structure learning for traffic prediction,” Knowledge-Based Systems, vol. 278, 2023

    2023

  19. [27]

    Data-driven traffic assign- ment: A novel approach for learning traffic flow patterns using graph convolutional neural network,

    R. Rahman and S. Hasan, “Data-driven traffic assign- ment: A novel approach for learning traffic flow patterns using graph convolutional neural network,”Data Science for Transportation, vol. 5, no. 2, p. 11, 2023

    2023

  20. [28]

    End-to-end heterogeneous graph neural networks for traffic assignment,

    T. Liu and H. Meidani, “End-to-end heterogeneous graph neural networks for traffic assignment,”Transportation Research Part C, 2024

    2024

  21. [29]

    Wardropnet: Traffic flow predictions via equilibrium- augmented learning,

    K. Jungel, D. Paccagnan, A. Parmentier, and M. Schiffer, “Wardropnet: Traffic flow predictions via equilibrium- augmented learning,” arXiv:2410.06656

    arXiv

  22. [30]

    A deep- learning approach for network traffic assignment with in- complete data,

    Y . Su, W. Fan, J. Puchinger, and M. Shen, “A deep- learning approach for network traffic assignment with in- complete data,” inTransportation Research Board 100th Annual Meeting, 2021

    2021

  23. [31]

    Deep learning-based dynamic traffic assignment with incom- plete origin–destination data,

    W. Fan, Z. Tang, P. Ye, F. Xiao, and J. Zhang, “Deep learning-based dynamic traffic assignment with incom- plete origin–destination data,”Transportation Research Record, vol. 2677, no. 3, pp. 1340–1356, 2023

    2023

  24. [32]

    Knowledge fusion enhanced graph neural network for traffic flow prediction,

    S. Wang, Y . Zhang, Y . Hu, and B. Yin, “Knowledge fusion enhanced graph neural network for traffic flow prediction,”Physica A: Statistical Mechanics and its Applications, vol. 623, 2023

    2023

  25. [33]

    PTV Visum,

    PTV-Group, “PTV Visum,” 2022

    2022

  26. [34]

    Recipe for a General, Powerful, Scalable Graph Transformer,

    L. Ramp ´aˇsek, M. Galkin, V . Dwivedi, A. Luu, G. Wolf, and D. Beaini, “Recipe for a General, Powerful, Scalable Graph Transformer,”Advances in Neural Information Processing Systems, vol. 35, 2022

    2022

  27. [35]

    and Laurent, T., Residual Gated Graph ConvNets

    Bresson, X. and Laurent, T., Residual Gated Graph ConvNets. https://arxiv.org/abs/1711.07553

    arXiv

  28. [36]

    Transportation networks for research,

    Transportation Networks for Research Core Team, “Transportation networks for research,” 2023, accessed: September 27, 2023

    2023

  29. [37]

    Latin hypercube sampling as a tool in uncertainty analysis of computer models,

    M. D. McKay, “Latin hypercube sampling as a tool in uncertainty analysis of computer models,” inProceedings of the 24th Conference on Winter Simulation, ser. WSC ’92. New York, NY , USA: Association for Computing Machinery, 1992, p. 557–564. [Online]. Available: https://doi.or...

    1992

  30. [38]

    Pytorch: An imperative style, high-performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. K ¨opf, E. Z. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala, “Pytorch: An imperative style, hi...

    1912 arXiv

  31. [39]

    Fast graph representation learning with pytorch geometric,

    M. Fey and J. E. Lenssen, “Fast graph representation learning with pytorch geometric,” 2019

    2019

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.