Optimizing traffic lights in a cellular automaton model for city traffic

Brockfeld, Elmar; Barlovic, Robert; Schadschneider, Andreas; Schreckenberg, Michael · 2001 · Crossref

DOI: 10.1103/physreve.64.056132

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 investigates the impact of global traffic light control strategies on vehicular flow within a cellular automaton (CA) model designed for city traffic networks. The research is motivated by the need to optimize urban traffic capacity, where existing street networks often exceed their limits. The authors utilize the Chowdhury-Schadschneider (ChSch) model, which combines the Biham-Middleton-Levine (BML) model for city intersections with the Nagel-Schreckenberg (NaSch) model for highway dynamics. This hybrid approach allows for a more detailed description of vehicle movement on streets between intersections, incorporating acceleration, braking, and stochastic driver behavior, unlike the original BML model which neglects inter-intersection dynamics. The study focuses on a square lattice network where all streets and intersections are treated equally, without dominant roads, to analyze how traffic light cycle times and switching strategies affect overall throughput. The experimental design involves simulating traffic on an $N \times N$ square lattice with periodic boundary conditions. Vehicles move according to NaSch rules on street segments of length $D-1$ between intersections, with a maximum velocity $v_{max}=5$ and a randomization parameter $p$. The authors first analyze a "synchronized strategy" where all traffic lights switch simultaneously with a fixed cycle time $T$. To derive optimal parameters, they reduce the complex network problem to a simpler one-dimensional case: a single street with one traffic light acting as a bottleneck. They develop a phenomenological approach based on microscopic traffic patterns, identifying specific scenarios where vehicle clusters either pass intersections efficiently or are blocked by red lights. This theoretical framework is validated against numerical simulations to determine the relationship between cycle times, network geometry, and mean flow. The results demonstrate that network capacity strongly depends on traffic light cycle times, with optimal periods determined by the distance between intersections. For synchronized lights, the flow exhibits oscillating behavior with distinct maxima and minima corresponding to specific cycle lengths. The authors derive heuristic formulas for these optimal times, showing good agreement with simulation data. They find that optimal cycle times allow vehicle clusters to traverse the distance between intersections without stopping, maximizing throughput. Furthermore, the study tests improved global strategies, including "green wave" controls with defined offsets and random switching strategies. These advanced strategies yield surprising results, with the green wave strategy providing significant improvements in throughput compared to synchronized switching. Random switching is also shown to create a flexible strategy less dependent on specific model parameters. The significance of this work lies in providing a theoretical framework for optimizing urban traffic control systems. By linking optimal traffic light timing to geometric network characteristics, the findings offer guidelines for adjusting real-world traffic signals to enhance flow. The avoidance of gridlock states through modified update rules ensures that the model remains applicable to realistic traffic densities. The study highlights that simple synchronized controls are suboptimal and that coordinated strategies like green waves can substantially improve network efficiency. These insights contribute to the broader field of traffic science by demonstrating how microscopic CA models can inform macroscopic control strategies, potentially aiding in the design of more efficient autonomous or adaptive traffic light 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.

• traffic density