GAMS: Geospatial Knowledge Enhanced Agentic Framework for Urban Human Mobility Behavior Simulation

ACL ARR 2026 January Submission10937 Authors

Keywords: human behavior, mobility simulation, agent, geospatial knowledge

Abstract: Recently, researchers have explored leveraging the commonsense knowledge and reasoning capabilities of large language models to accelerate human mobility simulation. However, these methods suffer from several critical shortcomings, including inadequate modeling of urban spaces, poor integration with individual mobility patterns, and weak alignment with collective mobility distributions. To address these limitations, we propose \textbf{G}eospatial Knowledge Enhanced \textbf{A}gentic framework for \textbf{M}obility \textbf{S}imulation (GAMS), an agentic framework that leverages the language based geospatial foundation model to simulate human mobility in urban space. GAMS comprises three core modules: MobExtractor, which extracts template-based mobility profiles and synthesizes new ones during simulation; GeoGenerator, which generate trajectory points considering collective mobility knowledge and urban geospatial knowledge; and TrajEnhancer, which captures individual mobility regularities from real-world trajectories and refine to generate style-aligned synthetic trajectories. Experiments on two real-world datasets show that \ourmodel~achieves superior performance over various state-of-the-art methods, with more than a 17\% improvement by capturing diverse mobility regularities.

Paper Type: Long

Research Area: NLP Applications

Research Area Keywords: LLM Agents, applications, NLP tools for social analysis, NLP applications

Languages Studied: English

Submission Number: 10937