A behavioral multi-agent model for road traffic simulation

Doniec, Arnaud; Mandiau, René; Piechowiak, Sylvain; Espié, Stéphane · 2008 · Crossref

DOI: 10.1016/j.engappai.2008.04.002

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 realistically simulating road traffic at intersections, a complex problem often simplified in existing tools using centralized schedulers or mathematical car-following laws. The authors argue that such approaches fail to capture the emergent, individual behaviors of human drivers, particularly regarding norm violations and conflict resolution. To overcome these limitations, the study proposes a behavioral multi-agent model within the ArchiSim simulation framework. The model aims to replicate realistic driver interactions by incorporating two key psychological facets: opportunistic behavior, where drivers selectively violate traffic rules based on context, and anticipatory abilities, where drivers predict future states to avoid gridlock. The methodology employs a multi-agent system where each simulated driver acts autonomously. The coordination mechanism decomposes complex intersection interactions into elementary binary conflicts (T-type or X-type) governed by priority relationships. The model integrates an opportunistic layer that modifies standard Highway Code priorities based on contextual factors such as speed differences, driver impatience, and vehicle positioning. For instance, an impatient driver may override a stop sign if they perceive a safe gap, governed by specific logical rules involving distance and acceleration calculations. Additionally, an anticipatory layer uses preventive anticipation via constraint processing. Each agent constructs a mental representation of the environment, modeling agents, relations, and temporal domains. The system evaluates potential actions by propagating constraints to detect undesirable states, such as gridlock, and eliminates actions that would lead to these outcomes. The model was validated by simulating traffic in a real intersection and comparing the resulting traffic flow with empirical measurements. The results demonstrated that the inclusion of opportunistic and anticipatory behaviors significantly improved the realism of the simulation. Specifically, the anticipatory mechanism effectively reduced the frequency of gridlock situations, which were prevalent in earlier versions of the model that lacked this feature. The simulated traffic flow closely matched real-world observations, confirming that the decentralized, behavior-based approach accurately reproduces the dynamic and often non-normative interactions found in actual traffic systems. The significance of this work lies in its contribution to microscopic traffic simulation by providing a more accurate representation of human driver behavior. By moving away from centralized scheduling and rigid mathematical models, the proposed multi-agent framework allows for the simulation of complex phenomena such as traffic backup and signal violations. This approach offers a flexible and generic tool for studying traffic dynamics, optimizing flow, and evaluating road infrastructure, ultimately supporting more realistic forecasting and control strategies for 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
• 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