Demand-Oriented Approach to Estimate the Operational Performance of Urban Networks With and Without Traffic Information Provision

Tsekeris, Theodore; Stathopoulos, Antony · 2020 · Crossref

DOI: 10.18757/ejtir.2005.5.2.4390

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 estimating the operational performance of extended urban transport networks under varying conditions of traffic information availability. Existing dynamic traffic assignment (DTA) models often rely on fixed, historical demand data and fail to account for how users adjust their travel behavior—specifically route choices and departure times—in response to real-time route guidance. The authors propose a demand-oriented, three-stage simulation approach to evaluate network performance with and without traffic information provision, emphasizing the dynamic adjustment of origin-destination (O-D) demand. The methodology consists of three stages. First, the current network state is estimated without traffic information by combining dynamic O-D matrix estimation with DTA. This uses an entropy maximization model adjusted by real-time link traffic counts to derive the most recent O-D matrix. Second, the network performance under real-time information provision is simulated using a Dynamic Network Assignment (DNA) model. This model simulates users’ en-route diversion behavior by updating route choices at specific control intervals based on prevailing traffic conditions. Third, the future network state is predicted by hypothesizing that users adjust both their departure times and route choices to the new traffic conditions generated by the guidance. This involves predicting a new dynamic O-D matrix based on the simulated link flows from the DNA model. The approach was implemented on the inner road network of Athens, Greece, using morning peak data from February 2000. The results demonstrate that the frequency of information updating significantly impacts network performance and demand patterns. Shorter control intervals (30 and 60 seconds) led to more accurate predictions of traffic conditions, as evidenced by lower GEH error statistics. Loading the predicted O-D matrix (reflecting user adjustment) resulted in significantly lower vehicle-hours traveled (VHT) and higher average travel speeds compared to using target or estimated matrices, particularly at high update frequencies. While vehicle-kilometers traveled (VKT) increased slightly due to path spreading, the overall congestion levels decreased. Furthermore, high-frequency information provision released latent demand, increasing the total O-D demand size serviced within the study period, whereas longer intervals (300 and 900 seconds) led to a decrease in satisfied demand as users shifted trips outside the peak period. Statistical analysis confirmed that shorter intervals caused significant changes in the temporal distribution of O-D demand. The study concludes that accounting for dynamic changes in O-D demand size and structure is crucial for plausibly evaluating the impact of traffic information provision. The findings highlight that frequent updates to route guidance not only improve operational metrics like travel time and speed but also alter the underlying travel demand by releasing latent trips. This approach provides a more realistic framework for real-time traffic operations and transport planning, moving beyond static demand assumptions to capture the complex feedback loop between information provision and user behavior.

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