EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Cognitive Agents in Urban Mobility: Integrating LLM Reasoning into Multi-Agent Simulations

Christian Calderón iD, Pasqual Martí iD, Jaume Jordán iD, Javier Palanca iD, Vicente Julian iD

Sensors · 12.09.2025 · 25 (18) · 5688

DOI: 10.3390/s25185688 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 12.09.2025. Verfasst von Christian Calderón und 4 weiteren Autor:innen. Beteiligte Einrichtung: Valencian Research Institute for Artificial Intelligence, Universitat Politècnica de València (UPV), Camino de Vera s/n, 46022 Valencia, Spain. Hauptthemen laut MEDLINE: Cognition. Die Arbeit wurde bisher 10-mal zitiert (Crossref) und nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Spanish Government, Secretaría de Estado de Digitalización e Inteligencia Artificial, Cátedra ENIA-UPV, Programme for the recruitment of doctors within the Spanish Science and Technology System at the Universitat Politècnica de València..

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Urban mobility systems face escalating challenges associated with sustainability, equity, and resilience, further compounded by environmental pressures. Traditional agent-based models (ABMs) often fail to capture cognitively rich, adaptive behaviors, limiting their ability to simulate realistic user responses to disruptions. In this work, we propose a cognitive agent architecture based on Large Language Models (LLMs), featuring multi-horizon memory-driven planning, reflection, and adaptation. Integrated into the SimFleet agent-based simulator with realistic sociodemographic profiles, the agents dynamically generate, adjust, and reflect upon travel plans across a 20-day simulation involving over 320 individuals. Experimental results reveal emergent adaptation patterns under both stable and disrupted transport conditions, and an ablation study under severe service disruption quantifies the contributions of short-term and long-term memory modules to memory-driven reasoning, demonstrating the potential of LLM-driven agents to enhance the realism, flexibility, and interpretability of urban mobility simulations.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

CognitionHumansComputer SimulationCities

Stichwörter der Autor:innen

adaptive behavioragent-based modelinglarge language modelsmobility disruptionsurban mobility simulation

Autor:innen und Institutionen

  • Valencian Research Institute for Artificial Intelligence, Universitat Politècnica de València (UPV), Camino de Vera s/n, 46022 Valencia, Spain.
  • Valencian Graduate School and Research Network of Artificial Intelligence (VALGRAI), 46022 Valencia, Spain.

ORCID-Profile: Christian Calderón, Pasqual Martí, Jaume Jordán, Javier Palanca, Vicente Julian

Förderung

  • Spanish Government
  • Secretaría de Estado de Digitalización e Inteligencia Artificial, Cátedra ENIA-UPV
  • Programme for the recruitment of doctors within the Spanish Science and Technology System at the Universitat Politècnica de València.
  • European Regional Development Fund
  • Programme for the recruitment of doctors within the Spanish Science and Technology System at the Universitat Politècnica de València

Zitieren

APA

Calderón, C., Martí, P., Jordán, J., Palanca, J., & Julian, V. (2025). Cognitive Agents in Urban Mobility: Integrating LLM Reasoning into Multi-Agent Simulations. Sensors, 25(18), 5688. https://doi.org/10.3390/s25185688

Vancouver

Calderón C, Martí P, Jordán J, Palanca J, Julian V. Cognitive Agents in Urban Mobility: Integrating LLM Reasoning into Multi-Agent Simulations. Sensors. 2025;25(18):5688. doi:10.3390/s25185688

BibTeX
@article{Caldern2025,
  author = {Calderón, Christian and Martí, Pasqual and Jordán, Jaume and Palanca, Javier and Julian, Vicente},
  title = {Cognitive Agents in Urban Mobility: Integrating LLM Reasoning into Multi-Agent Simulations},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {18},
  pages = {5688},
  doi = {10.3390/s25185688}
}
Kostenlos starten →