A Unified Model for Human Mobility Generation in Natural Disasters

TL;DR

This paper introduces UniDisMob, a universal model for human mobility in natural disasters that generalizes across different disaster types and cities using physics-informed prompts and meta-learning, outperforming existing methods.

Contribution

The paper presents a novel unified framework combining physics-guided alignment and meta-learning to generate human mobility patterns in diverse disaster scenarios and cities.

Findings

Achieves over 13% performance improvement on average.

Effectively generalizes to unseen disaster types.

Demonstrates robustness across multiple cities.

Abstract

Human mobility generation in disaster scenarios plays a vital role in resource allocation, emergency response, and rescue coordination. During disasters such as wildfires and hurricanes, human mobility patterns often deviate from their normal states, which makes the task more challenging. However, existing works usually rely on limited data from a single city or specific disaster, significantly restricting the model's generalization capability in new scenarios. In fact, disasters are highly sudden and unpredictable, and any city may encounter new types of disasters without prior experience. Therefore, we aim to develop a one-for-all model for mobility generation that can generalize to new disaster scenarios. However, building a universal framework faces two key challenges: 1) the diversity of disaster types and 2) the heterogeneity among different cities. In this work, we propose a unified model for human mobility generation in natural disasters (named UniDisMob). To enable cross-disaster generalization, we design physics-informed prompt and physics-guided alignment that leverage the underlying common patterns in mobility changes after different disasters to guide the generation process. To achieve cross-city generalization, we introduce a meta-learning framework that extracts universal patterns across multiple cities through shared parameters and captures city-specific features via private parameters. Extensive experiments across multiple cities and disaster scenarios demonstrate that our method significantly outperforms state-of-the-art baselines, achieving an average performance improvement exceeding 13%.