A dynamic traffic assignment model for highly congested urban networks

Moshe Ben‐Akiva; Gao, Song; Wei, Zheng; Yang, Wen · 2012 · OpenAlex-citations

DOI: 10.1016/j.trc.2012.02.006

archive: archived pipeline: cataloged verified

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

Summary

This paper addresses the challenge of accurately modeling highly congested urban traffic networks, specifically focusing on replicating realistic traffic dynamics such as long queues and spillbacks. The authors identify that conventional Dynamic Traffic Assignment (DTA) models often fail in complex environments due to unrealistic gridlocks, excessive congestion on specific routes, and inadequate handling of network topology. Motivated by the need to evaluate demand management and traffic control strategies in dense urban areas, the study presents enhancements to the DynaMIT-P mesoscopic traffic simulation system, validated through a case study of Beijing, China. The methodology involves enhancing the DynaMIT-P model to address four specific modeling challenges: route choice bias, lane-based queuing, short link dynamics, and non-motorized traffic impacts. The researchers calibrated all demand and supply parameters simultaneously using sensor counts and floating car travel time data. Key methodological adjustments included replacing the standard Multinomial Logit (MNL) route choice model with the Path-size Logit (PSL) model to account for overlapping routes. Additionally, the supply model was modified to use explicit lane groups rather than segment-level queuing to prevent artificial blocking of through traffic by turning lanes. The model also incorporated a modified treatment of acceptance capacity for short links—segments comparable in length to a vehicle—and utilized dynamic road segment capacities to capture the interference of bicycles and pedestrians on auto traffic, despite not explicitly modeling non-motorized users. The results demonstrate that these combined enhancements successfully resolved the unrealistic gridlocks and excessive congestion observed in earlier simulation attempts. The Path-size Logit model corrected the bias toward expressway links caused by overlapping paths, leading to more realistic flow distributions. The implementation of lane groups eliminated artificial bottlenecks where left-turn queues incorrectly blocked through and right-turn traffic. Furthermore, adjusting the minimum speed parameters and acceptance capacity calculations for short links prevented vehicles from moving unrealistically slowly or queuing unnecessarily. The calibrated model accurately replicated traffic characteristics in Beijing, validating the necessity of synthesizing solutions on both the demand and supply sides of the DTA model. The significance of this work lies in providing a robust framework for applying DTA models to real-world, highly congested urban networks. The study concludes that no single model feature is sufficient to handle such complexity; rather, a synthesis of advanced route choice modeling, explicit lane-group representations, and careful handling of short links and non-motorized traffic impacts is crucial. These findings imply that for accurate transportation planning and policy evaluation in dense urban environments, simulation-based DTA models must move beyond simplified assumptions to incorporate detailed network topology and behavioral corrections.

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.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

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

• traffic density