ActivityEditor: Learning to Synthesize Physically Valid Human Mobility

2026-04-07

AI Summary

This work addresses the challenge of modeling human mobility in data-scarce regions where historical trajectories are unavailable. To this end, the authors propose ActivityEditor, a zero-shot cross-regional trajectory generation framework based on a dual large language model (LLM) agent architecture. The approach synergistically combines two stages—intent generation and editing refinement—augmented with demographic priors, a reinforcement learning–driven reward mechanism enforcing physical constraints, and multi-dimensional plausibility validation. This study presents the first application of a dual-LLM agent paradigm to synthetic human mobility modeling. Experimental results demonstrate that ActivityEditor achieves superior zero-shot transfer performance across multiple urban areas, generating trajectories that significantly outperform existing methods in both statistical fidelity and physical validity.

Abstract

Human mobility modeling is indispensable for diverse urban applications. However, existing data-driven methods often suffer from data scarcity, limiting their applicability in regions where historical trajectories are unavailable or restricted. To bridge this gap, we propose \textbf{ActivityEditor}, a novel dual-LLM-agent framework designed for zero-shot cross-regional trajectory generation. Our framework decomposes the complex synthesis task into two collaborative stages. Specifically, an intention-based agent, which leverages demographic-driven priors to generate structured human intentions and coarse activity chains to ensure high-level socio-semantic coherence. These outputs are then refined by editor agent to obtain mobility trajectories through iteratively revisions that enforces human mobility law. This capability is acquired through reinforcement learning with multiple rewards grounded in real-world physical constraints, allowing the agent to internalize mobility regularities and ensure high-fidelity trajectory generation. Extensive experiments demonstrate that \textbf{ActivityEditor} achieves superior zero-shot performance when transferred across diverse urban contexts. It maintains high statistical fidelity and physical validity, providing a robust and highly generalizable solution for mobility simulation in data-scarce scenarios. Our code is available at: https://anonymous.4open.science/r/ActivityEditor-066B.

Problem

• human mobility modeling
• data scarcity
• zero-shot trajectory generation
• cross-regional transfer

Innovation

• zero-shot trajectory generation
• dual-LLM-agent framework
• human mobility modeling
• reinforcement learning
• physical validity