Calibration of CORSIM models under saturated traffic flow conditions.

Paz, Alexander; Molano-Paz, Victor Hugo · 2013 · ROSA P / University of Nevada, Las Vegas. Transportation Research Center

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of calibrating microscopic traffic flow simulation models, specifically CORSIM, under saturated traffic conditions. Previous calibration methodologies often considered only a subset of model parameters, relied on single performance measures, or pre-calibrated demand patterns, leading to convergence and stability issues. This research proposes a comprehensive methodology to simultaneously calibrate all model parameters—including global and local settings, driver behavior, vehicle performance, and demand patterns (turning volumes)—using multiple performance measures, specifically link counts and speeds. The goal is to improve model accuracy and reduce user effort by optimizing the entire parameter set concurrently. The methodology employs a Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm to minimize the difference between actual and simulated network states. The calibration problem is formulated as a mathematical programming task that minimizes a normalized root mean square (NRMS) objective function, which aggregates the normalized differences in link counts and speeds across all links and time periods. The SPSA algorithm was selected for its computational efficiency, requiring only two simulation evaluations per iteration to update all parameters. A custom Java-based software tool was developed to implement this framework, featuring a graphical user interface and layered architecture to handle input validation, calculation, and model updates. Calibration criteria were based on Federal Highway Administration guidelines, requiring a GEH statistic below 5 for at least 85% of links and a total volume difference under 5%. The proposed methodology was validated through three experiments using CORSIM models. The first experiment involved an arterial network on Pyramid Highway in Reno, NV, using link counts and speeds. Calibration reduced the NRMS from 0.042 to 0.010, and the GEH statistic fell below 5 for 100% of the links, satisfying all criteria. The second experiment calibrated a mixed network of freeway ramps and arterials on I-75 in Miami, FL. While the NRMS decreased from 0.270 to 0.245, the GEH statistic improved significantly for freeway ramps (99.6% of links below 5), though arterial improvements were less pronounced due to data availability. The third experiment used a time-dependent arterial network from McTrans sample data, dividing the simulation into four 15-minute periods. The NRMS dropped from 0.51 to 0.09, and the GEH statistic met the threshold for nearly all links across all time periods, demonstrating the method's ability to handle time-varying demand with static parameters. The significance of this work lies in its demonstration that simultaneous calibration of all model parameters and demand patterns is feasible and effective. By using the SPSA algorithm, the study achieves better parameter estimates with fewer simulation runs compared to previous methods, reducing computational costs and improving convergence stability. The results confirm that the methodology can accurately reproduce complex traffic conditions, including mixed freeway-arterial systems and time-dependent flows, thereby enhancing the reliability of microscopic simulation models for transportation planning and analysis.

Key finding

The proposed simultaneous calibration methodology significantly improved model accuracy, reducing the normalized root mean square error and achieving GEH statistics below 5 for 100% of links in the first experiment and 99.6% of freeway ramp links in the second experiment.

Methodology

modeling

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• traffic density

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Theoretical Contribution: computational model