Open-Source METANET Calibration for Reproducible Freeway Traffic Macroscopic Simulation

Cathy Wu; Monica Chan; Shreyaa Raghavan

arxiv: 2605.23042 · v1 · pith:KN6MWRV3new · submitted 2026-05-21 · 📡 eess.SY · cs.SY

Open-Source METANET Calibration for Reproducible Freeway Traffic Macroscopic Simulation

Monica Chan , Shreyaa Raghavan , Cathy Wu This is my paper

Pith reviewed 2026-05-25 05:24 UTC · model grok-4.3

classification 📡 eess.SY cs.SY

keywords METANET calibrationmacroscopic traffic simulationfreeway traffic modelingnonlinear programmingIPOPT solverramp flow estimationI-24 MOTIONPeMS data

0 comments

The pith

An open-source nonlinear program calibrates METANET parameters to reproduce observed freeway traffic patterns from I-24 and PeMS data.

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

The paper establishes an open-source calibration, simulation, and visualization tool for the METANET second-order macroscopic traffic flow model. Calibration takes the form of a nonlinear program solved with the interior-point method IPOPT while jointly estimating ramp flows. When applied to trajectory data from the I-24 MOTION network and loop-detector data from the PeMS system, the resulting models match measured traffic behavior, including stop-and-go congestion waves, across varied network geometries and conditions. A sympathetic reader would care because the absence of such open tools has prevented easy reproduction or extension of calibration results to new freeway networks and control applications.

Core claim

This work provides an open-source METANET calibration, simulation, and data visualization tool. The calibration is formulated as a nonlinear program solved via the interior-point method IPOPT, with joint ramp flow estimation. Models calibrated using this method reproduce observed traffic patterns from I-24 MOTION and PeMS datasets across diverse network geometries and traffic conditions including complex stop-and-go congestion waves.

What carries the argument

The nonlinear program formulation for METANET parameter estimation solved via IPOPT with joint ramp flow estimation, which fits model parameters directly to measured freeway data.

If this is right

Calibrated METANET models can be used for simulation in ramp metering and variable speed limit control applications.
Open-source code allows other researchers to reproduce and extend the calibration to additional networks.
Growing volumes of sensor data can be turned into validated models that support real-world traffic control.
The approach works on both trajectory-based and loop-detector data sources.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same calibration pipeline could be adapted for online parameter updates as new sensor streams arrive.
Community use of the released code may produce standardized benchmark calibrations for other second-order traffic models.
Validation on additional datasets outside California could test whether the method remains effective under different driver populations or infrastructure types.

Load-bearing premise

The chosen nonlinear program solved by IPOPT with joint ramp flow estimation produces parameters that generalize beyond the specific I-24 and PeMS segments used for fitting.

What would settle it

Applying the calibrated parameters to a new freeway segment or later time period and measuring large discrepancies in simulated versus observed densities and flows would show that the calibration does not produce generalizable results.

Figures

Figures reproduced from arXiv: 2605.23042 by Cathy Wu, Monica Chan, Shreyaa Raghavan.

**Figure 2.** Figure 2: Speed time-space diagrams of the validation set of networks. a. is [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Synthetic bottleneck network. Segments 1–10 have 4 lanes; a lane [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

**Figure 5.** Figure 5: Macroscopic representation of the I-5 main corridor used for [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗

**Figure 6.** Figure 6: a) Original PeMS loop detector speed data and b) reconstructed [PITH_FULL_IMAGE:figures/full_fig_p004_6.png] view at source ↗

**Figure 7.** Figure 7: Results compared to ground truth data for a) synthetic scenario, b) I-24 on 11-30-2022, c) I-24 on 11-29-2022, and d) I-5. Each subplot compares [PITH_FULL_IMAGE:figures/full_fig_p005_7.png] view at source ↗

**Figure 8.** Figure 8: Parameter Recovery: Synthetic Scenario. Relative percent error between estimated and ground-truth METANET parameters. The synthetic scenario contains no on- or off-ramps, producing near-spatially-homogeneous dynamics; under these conditions SI-TV achieves lower MAPE than SV-TV (Table III). All real-world scenarios include ramps, inducing spatial variation in traffic state that SI-TV’s uniform parameterizat… view at source ↗

read the original abstract

METANET is a widely used second-order macroscopic traffic flow model for freeway networks, supporting applications across traffic simulation, ramp metering, and variable speed limit control. The predictive accuracy of any traffic model, however, hinges on careful calibration to real-world conditions. Despite its widespread use, there have not been open-source tools for calibrating METANET's parameters. Without open-source calibration, results cannot be easily reproduced or extended to other networks. This work provides an open-source METANET calibration, simulation, and data visualization tool. The calibration is formulated as a nonlinear program (NLP) solved via the interior-point method (IPOPT), with joint ramp flow estimation. We validate our calibration on real-world freeway data from two widely used traffic monitoring systems: Interstate-24 MObility Technology Interstate Observation Network (I-24 MOTION), one of the largest open-road trajectory instruments in the country, and loop detector data from the Caltrans Performance Measurement System (PeMS), which spans nearly 40,000 detectors across California freeways and serves as a standard benchmark in traffic research. Models calibrated using our method are able to reproduce these datasets' observed traffic patterns across diverse network geometries and traffic conditions including complex stop-and-go congestion waves. As large-scale traffic monitoring infrastructure continues to expand, open-source calibration tools are essential for translating growing volumes of sensor data into validated models that can support real-world traffic control. The complete code is publicly available at https://github.com/woxsao/metanet-calibration to support reproducible research in freeway traffic modeling and control.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The open-source METANET calibration code is the actual deliverable here, but joint ramp estimation leaves open whether the model parameters or the extra variables are reproducing the data.

read the letter

The paper's main point is the release of public code for calibrating METANET parameters via a nonlinear program solved with IPOPT, including joint estimation of ramp flows. This is the first such open implementation, and the validation on I-24 MOTION trajectories plus PeMS loop detectors shows the calibrated models can match observed speed and density patterns, including stop-and-go waves on different network layouts. The GitHub link gives others a concrete starting point for their own freeway simulations without having to build the solver from scratch. That directly addresses the reproducibility gap noted in the abstract. The work is practical for traffic engineers who already use METANET and want to fit it to new sensor data. The formulation itself is straightforward and the choice of two independent real-world datasets is reasonable. The soft spots sit in the validation details. The abstract supplies no numerical error measures and no head-to-head comparison against earlier calibration methods. More critically, treating ramp flows as decision variables alongside the METANET parameters (tau, kappa, eta, etc.) gives the optimizer extra freedom. If the ramps are noisy or unobserved, mismatch can be absorbed there instead of forcing the second-order dynamics to fit. Nothing in the description breaks out the relative contribution of the ramp variables or tests a version with ramps fixed to measurements. That concern from the stress-test note stands on the given information and weakens the claim that the calibrated parameters are what enable reproduction. This paper is for applied users who need a ready tool rather than readers hunting for new traffic physics. It is coherent on its own terms and deserves referee time so the implementation and any missing checks on ramp influence can be examined.

Referee Report

2 major / 1 minor

Summary. The manuscript presents an open-source Python tool for calibrating the METANET second-order macroscopic traffic model. Calibration is formulated as a nonlinear program solved via IPOPT with joint estimation of ramp flows as decision variables. Validation is reported on I-24 MOTION trajectory data and PeMS loop-detector data, with the central claim that the resulting models reproduce observed speed and density patterns, including complex stop-and-go waves, across different network geometries and conditions.

Significance. An open-source, reproducible calibration pipeline for METANET would be a useful contribution given the model's widespread use in simulation and control. Public release of the code at the cited GitHub repository is a clear strength that directly supports the reproducibility goal stated in the abstract. Use of two independent, large-scale real-world datasets (I-24 MOTION and PeMS) adds potential value if the quantitative performance can be demonstrated.

major comments (2)

[Abstract] Abstract: the validation statement that calibrated models 'reproduce these datasets' observed traffic patterns' is presented without any quantitative error metrics (RMSE, MAE, or similar on speed/density), without comparison to prior METANET calibration methods, and without reporting of IPOPT convergence behavior or parameter sensitivity. These omissions are load-bearing for the central claim of successful reproduction.
[NLP formulation (Abstract)] NLP formulation (Abstract): ramp flows are treated as jointly estimated decision variables together with the METANET parameters (tau, kappa, eta, etc.). No baseline experiment is described in which ramp flows are fixed to measured values (where available) or to zero; therefore it is impossible to determine whether reproduction performance is attributable to the second-order dynamics or to the extra degrees of freedom in the ramp estimates. This directly affects the claim that the calibrated METANET parameters enable the observed reproduction.

minor comments (1)

[Abstract] The GitHub link is supplied, which is helpful; however, the abstract does not indicate whether the released code includes the exact NLP formulation, solver settings, and data preprocessing scripts used for the reported results.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight important aspects for strengthening the validation claims. We agree that quantitative metrics and a ramp-flow baseline are needed to support the central claims and will revise the manuscript accordingly.

read point-by-point responses

Referee: [Abstract] Abstract: the validation statement that calibrated models 'reproduce these datasets' observed traffic patterns' is presented without any quantitative error metrics (RMSE, MAE, or similar on speed/density), without comparison to prior METANET calibration methods, and without reporting of IPOPT convergence behavior or parameter sensitivity. These omissions are load-bearing for the central claim of successful reproduction.

Authors: We agree that the abstract should report quantitative metrics to support the reproduction claim. The current abstract is qualitative; we will revise it to include key error metrics (RMSE/MAE on speed and density) from the validation experiments on both datasets. We will also add statements on IPOPT convergence status and a reference to the parameter sensitivity results already present in the full manuscript. A concise comparison to prior METANET calibration literature will be incorporated into the abstract or the opening of the results section. revision: yes
Referee: [NLP formulation (Abstract)] NLP formulation (Abstract): ramp flows are treated as jointly estimated decision variables together with the METANET parameters (tau, kappa, eta, etc.). No baseline experiment is described in which ramp flows are fixed to measured values (where available) or to zero; therefore it is impossible to determine whether reproduction performance is attributable to the second-order dynamics or to the extra degrees of freedom in the ramp estimates. This directly affects the claim that the calibrated METANET parameters enable the observed reproduction.

Authors: This is a fair and substantive point. Joint ramp estimation was chosen to accommodate networks where ramp measurements are incomplete or unavailable. However, to isolate the contribution of the METANET dynamics, we will add a new ablation experiment in the revised results section. In this experiment ramp flows will be fixed to measured values on segments where they exist (I-24) or set to zero otherwise, and reproduction quality will be compared directly against the joint-estimation case. revision: yes

Circularity Check

0 steps flagged

No significant circularity; calibration is an external-data optimization task

full rationale

The paper formulates METANET calibration as a nonlinear program solved via IPOPT with joint ramp flow estimation and validates reproduction of observed speed/density patterns on independent I-24 MOTION and PeMS datasets. No load-bearing steps reduce any reported reproduction performance to a fitted quantity by construction, nor do any self-citations, uniqueness theorems, or ansatzes appear in the provided text. The central claim is that the open-source NLP tool produces parameters capable of matching external sensor data; this is a standard fitting procedure against held-out measurements rather than a self-referential derivation. The joint ramp estimation is an explicit modeling choice within the optimization but does not create a definitional loop or rename a known result. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on the standard METANET second-order model equations (treated as given), the assumption that IPOPT reliably solves the resulting NLP, and the representativeness of the two chosen datasets. No new entities are postulated.

free parameters (2)

METANET model parameters (tau, kappa, eta, etc.)
These are the quantities being fitted by the NLP; their specific values are not reported in the abstract.
Ramp flow estimates
Jointly estimated within the same optimization; treated as decision variables rather than measured inputs.

axioms (2)

domain assumption METANET second-order macroscopic equations accurately describe freeway dynamics when parameters are chosen appropriately
Invoked implicitly by using the model for calibration and validation (abstract opening paragraph).
domain assumption Interior-point method (IPOPT) finds a global or sufficiently good local optimum for the formulated NLP
Stated as the solver choice without further justification in the abstract.

pith-pipeline@v0.9.0 · 5819 in / 1454 out tokens · 19161 ms · 2026-05-25T05:24:56.213106+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

15 extracted references · 6 canonical work pages

[1]

Metanet: a macroscopic simula- tion program for motorway networks,

A. Messmer and M. Papageorgiou, “Metanet: a macroscopic simula- tion program for motorway networks,”Traffic Engineering & Control, vol. 31, p. 466–470, Jan. 1990

1990
[2]

Hegyi,Model predictive control for integrating traffic control mea- sures

A. Hegyi,Model predictive control for integrating traffic control mea- sures. Delft: The Netherlands Trail Research School, 2004

2004
[3]

Reinforcement learning with model predictive control for highway ramp metering,

F. Airaldi, B. D. Schutter, and A. Dabiri, “Reinforcement learning with model predictive control for highway ramp metering,” no. arXiv:2311.08820, Feb. 2025, arXiv:2311.08820 [eess]. [Online]. Available: http://arxiv.org/abs/2311.08820

work page arXiv 2025
[4]

Traffic flow model validation using metanet, adol-c and rprop,

A. Poole and A. Kotsialos, “Traffic flow model validation using metanet, adol-c and rprop,”IFAC-PapersOnLine, vol. 49, no. 3, p. 291–296, 2016

2016
[5]

A parameter identification algorithm for the metanet model with a limited number of loop detectors,

J. R. D. Frejo, E. F. Camacho, and R. Horowitz, “A parameter identification algorithm for the metanet model with a limited number of loop detectors,” in2012 IEEE 51st IEEE Conference on Decision and Control (CDC), Dec. 2012, p. 6983–6988. [Online]. Available: https://ieeexplore.ieee.org/document/6426671/

work page arXiv 2012
[6]

Calibration of second order traffic models using continuous cross entropy method,

D. Ngoduy and M. Maher, “Calibration of second order traffic models using continuous cross entropy method,”Transportation Research Part C: Emerging Technologies, vol. 24, p. 102–121, Oct. 2012

2012
[7]

Filippoairaldi/sym-metanet,

F. Airaldi, “Filippoairaldi/sym-metanet,” Nov. 2025. [Online]. Available: https://github.com/FilippoAiraldi/sym-metanet

2025
[8]

Imputation of ramp flow data for freeway traffic simulation,

A. Muralidharan and R. Horowitz, “Imputation of ramp flow data for freeway traffic simulation,”Transportation Research Record: Journal of the Transportation Research Board, vol. 2099, no. 1, p. 58–64, Jan. 2009

2099
[9]

Cross- modal reconstruction pretraining for ramp flow prediction at highway interchanges,

Y . Li, J. Chen, Z. Li, C. Gao, Y . Li, C. Zhang, and C. Dong, “Cross- modal reconstruction pretraining for ramp flow prediction at highway interchanges,” no. arXiv:2510.03381, Nov. 2025, arXiv:2510.03381 [cs]. [Online]. Available: http://arxiv.org/abs/2510.03381

work page arXiv 2025
[10]

M. L. Bynum, G. A. Hackebeil, W. E. Hart, C. D. Laird, B. L. Nicholson, J. D. Siirola, J.-P. Watson, and D. L. Woodruff, Introduction. Cham: Springer International Publishing, 2021, p. 1–11. [Online]. Available: https://doi.org/10.1007/978-3-030-68928-5 1

work page doi:10.1007/978-3-030-68928-5 2021
[11]

On the implementation of an interior- point filter line-search algorithm for large-scale nonlinear programming,

A. W ¨achter and L. T. Biegler, “On the implementation of an interior- point filter line-search algorithm for large-scale nonlinear programming,” Mathematical Programming, vol. 106, no. 1, p. 25–57, Mar. 2006

2006
[12]

I-24 motion: An instrument for freeway traffic science,

D. Gloudemans, Y . Wang, J. Ji, G. Zachar, W. Barbour, and D. B. Work, “I-24 motion: An instrument for freeway traffic science,” no. arXiv:2301.11198, Jan. 2023, arXiv:2301.11198 [eess]. [Online]. Available: http://arxiv.org/abs/2301.11198

work page arXiv 2023
[13]

Discussion of traffic stream measurements and definitions,

L. Edie, “Discussion of traffic stream measurements and definitions,” in 2nd International Symposium on the Theory of Traffic Flow. London: OECD (Organization for Economic Cooperation and Development), 1963

1963
[14]

Freeway performance measurement system (pems),

C. Chen, “Freeway performance measurement system (pems),” 2003. [Online]. Available: https://escholarship.org/uc/item/6j93p90t

2003
[15]

Calibrating adaptive smoothing methods for freeway traffic reconstruction,

J. Ji, D. Gloudemans, G. Zach ´ar, M. Nice, W. Barbour, and D. B. Work, “Calibrating adaptive smoothing methods for freeway traffic reconstruction,” no. arXiv:2602.02072, Feb. 2026, arXiv:2602.02072 [cs]. [Online]. Available: http://arxiv.org/abs/2602.02072

work page arXiv 2026

[1] [1]

Metanet: a macroscopic simula- tion program for motorway networks,

A. Messmer and M. Papageorgiou, “Metanet: a macroscopic simula- tion program for motorway networks,”Traffic Engineering & Control, vol. 31, p. 466–470, Jan. 1990

1990

[2] [2]

Hegyi,Model predictive control for integrating traffic control mea- sures

A. Hegyi,Model predictive control for integrating traffic control mea- sures. Delft: The Netherlands Trail Research School, 2004

2004

[3] [3]

Reinforcement learning with model predictive control for highway ramp metering,

F. Airaldi, B. D. Schutter, and A. Dabiri, “Reinforcement learning with model predictive control for highway ramp metering,” no. arXiv:2311.08820, Feb. 2025, arXiv:2311.08820 [eess]. [Online]. Available: http://arxiv.org/abs/2311.08820

work page arXiv 2025

[4] [4]

Traffic flow model validation using metanet, adol-c and rprop,

A. Poole and A. Kotsialos, “Traffic flow model validation using metanet, adol-c and rprop,”IFAC-PapersOnLine, vol. 49, no. 3, p. 291–296, 2016

2016

[5] [5]

A parameter identification algorithm for the metanet model with a limited number of loop detectors,

J. R. D. Frejo, E. F. Camacho, and R. Horowitz, “A parameter identification algorithm for the metanet model with a limited number of loop detectors,” in2012 IEEE 51st IEEE Conference on Decision and Control (CDC), Dec. 2012, p. 6983–6988. [Online]. Available: https://ieeexplore.ieee.org/document/6426671/

work page arXiv 2012

[6] [6]

Calibration of second order traffic models using continuous cross entropy method,

D. Ngoduy and M. Maher, “Calibration of second order traffic models using continuous cross entropy method,”Transportation Research Part C: Emerging Technologies, vol. 24, p. 102–121, Oct. 2012

2012

[7] [7]

Filippoairaldi/sym-metanet,

F. Airaldi, “Filippoairaldi/sym-metanet,” Nov. 2025. [Online]. Available: https://github.com/FilippoAiraldi/sym-metanet

2025

[8] [8]

Imputation of ramp flow data for freeway traffic simulation,

A. Muralidharan and R. Horowitz, “Imputation of ramp flow data for freeway traffic simulation,”Transportation Research Record: Journal of the Transportation Research Board, vol. 2099, no. 1, p. 58–64, Jan. 2009

2099

[9] [9]

Cross- modal reconstruction pretraining for ramp flow prediction at highway interchanges,

Y . Li, J. Chen, Z. Li, C. Gao, Y . Li, C. Zhang, and C. Dong, “Cross- modal reconstruction pretraining for ramp flow prediction at highway interchanges,” no. arXiv:2510.03381, Nov. 2025, arXiv:2510.03381 [cs]. [Online]. Available: http://arxiv.org/abs/2510.03381

work page arXiv 2025

[10] [10]

M. L. Bynum, G. A. Hackebeil, W. E. Hart, C. D. Laird, B. L. Nicholson, J. D. Siirola, J.-P. Watson, and D. L. Woodruff, Introduction. Cham: Springer International Publishing, 2021, p. 1–11. [Online]. Available: https://doi.org/10.1007/978-3-030-68928-5 1

work page doi:10.1007/978-3-030-68928-5 2021

[11] [11]

On the implementation of an interior- point filter line-search algorithm for large-scale nonlinear programming,

A. W ¨achter and L. T. Biegler, “On the implementation of an interior- point filter line-search algorithm for large-scale nonlinear programming,” Mathematical Programming, vol. 106, no. 1, p. 25–57, Mar. 2006

2006

[12] [12]

I-24 motion: An instrument for freeway traffic science,

D. Gloudemans, Y . Wang, J. Ji, G. Zachar, W. Barbour, and D. B. Work, “I-24 motion: An instrument for freeway traffic science,” no. arXiv:2301.11198, Jan. 2023, arXiv:2301.11198 [eess]. [Online]. Available: http://arxiv.org/abs/2301.11198

work page arXiv 2023

[13] [13]

Discussion of traffic stream measurements and definitions,

L. Edie, “Discussion of traffic stream measurements and definitions,” in 2nd International Symposium on the Theory of Traffic Flow. London: OECD (Organization for Economic Cooperation and Development), 1963

1963

[14] [14]

Freeway performance measurement system (pems),

C. Chen, “Freeway performance measurement system (pems),” 2003. [Online]. Available: https://escholarship.org/uc/item/6j93p90t

2003

[15] [15]

Calibrating adaptive smoothing methods for freeway traffic reconstruction,

J. Ji, D. Gloudemans, G. Zach ´ar, M. Nice, W. Barbour, and D. B. Work, “Calibrating adaptive smoothing methods for freeway traffic reconstruction,” no. arXiv:2602.02072, Feb. 2026, arXiv:2602.02072 [cs]. [Online]. Available: http://arxiv.org/abs/2602.02072

work page arXiv 2026