Simplified cellular automaton model for city traffic

Simon, P.; Nagel, Kai · 1998 · OpenAlex-citations

DOI: 10.1103/physreve.58.1286

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 introduces a simplified cellular automaton (CA) model designed to simulate urban traffic flow efficiently, addressing the computational limitations of detailed microsimulations in large-scale contexts. The authors argue that while microscopic approaches are necessary to capture dynamic congestion effects and individual route choices, full-scale simulations are often too computationally intensive. To resolve this, the study focuses on modeling traffic bottlenecks—specifically intersections—using a single-lane CA framework combined with probabilistic transition rules, rather than simulating every vehicle interaction in high detail. The methodology employs the Nagel/Schreckenberg CA rule for vehicle movement on single-lane links, where vehicles accelerate, slow down based on gaps, and undergo stochastic braking. Capacity restrictions at intersections are modeled using "impurity sites" or probabilistic transitions ($p_{trans}$), which determine whether a vehicle passes through an intersection or is forced to stop. The authors systematically investigate three scheduling schemes for these transitions: a "random" traffic light (where the probability of passing equals the green time fraction), a "normal" traffic light (periodic green/red cycles), and a "Dirac" traffic light (instantaneous green phases). The model was implemented in the TRANSIMS Dallas/Fort Worth case study, simulating approximately 300,000 trips in a 5x5 mile area. Due to the single-lane constraint, the input data was subsampled by a factor of 0.154 to match the model’s maximum throughput of ~1200 vehicles/hour per lane. The results demonstrate that none of the scheduling schemes produce a linear relationship between the fraction of green time and throughput. The random traffic light model yields non-linear flow characteristics, with low transition probabilities resulting in compact traffic jams and flow approximated by $F \approx p_{noise} p_{trans} / (1 + p_{trans})$. The normal traffic light model shows an almost linear relationship but suffers from reduced fluidity. The Dirac traffic light generates the highest flow for a given green fraction due to "particle-hole attraction" effects in the parallel update scheme. In the Dallas simulation, the model successfully reproduced realistic traffic dynamics, including queue formation at link ends and increased travel times during rush hours, as evidenced by space-time diagrams and travel time statistics. The significance of this work lies in providing a computationally efficient alternative to complex microsimulations for large-scale transportation planning. By capturing essential dynamics at bottlenecks through simplified probabilistic rules, the model allows for systematic scenario evaluations and activity-based forecasting without the prohibitive computational costs of detailed simulations. The findings highlight that simple stochastic models can effectively represent macroscopic traffic phenomena, such as congestion spreading and peak-hour delays, making them viable tools for analyzing national or metropolitan transportation systems.

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.

• mental model of traffic