糖心传媒

Project Leads

Yitong Li

Team Members

, Sandeep Agrawal

Cluster: Governance of Migration and Cyber-physical Infrastructure

This project examines how housing, transportation, and essential services shape newcomer settlement and integration in the context of rapid demographic change and climate risks. Using AI, spatiotemporal analytics, and complex systems modeling, the project identifies infrastructure gaps, predicts settlement patterns, and evaluates targeted interventions to improve service accessibility, support newcomer integration, and strengthen community resilience.

This research is grounded in a complex adaptive systems framework that conceptualizes urban environments as coupled socio-infrastructure systems shaped by dynamic interactions among immigration behavior, built environment characteristics, and climate impacts. The project has three objectives:

1. To develop data-driven spatiotemporal models that characterize and predict the evolution of newcomer settlement patterns using multi-modal data
2. To identify infrastructure gaps through a socio-informed functionality metric that captures interactions among settlement patterns, infrastructure networks, and climate-related risks;
3. To design AI-driven scenario optimization models and an agentic AI platform to evaluate infrastructure intervention strategies that enhance service accessibility, newcomer integration, and community resilience.

1. How to identify and predict mismatches between dynamic newcomer needs and infrastructure capacity?
2. How do the interactions among settlement patterns, infrastructure systems, and climate-related risks lead to infrastructure gaps that constrain immigrant integration?
3. How to optimize infrastructure and service to enhance newcomer integration and community resilience?

The methodology has three tasks. First, multi-modal datasets on socio-demographic characteristics, infrastructure attributes, and environmental conditions will be integrated. AI-based computer vision and spatiotemporal modeling will be used to characterize and predict newcomer settlement patterns while addressing data limitations through imputation methods. Second, a socio-informed functionality metric will be designed using multi-layered socio-infrastructure networks to assess accessibility to housing and essential services. Climate hazard data will be incorporated to evaluate how disruptions affect accessibility and identify vulnerable areas where infrastructure gaps constrain newcomer integration. Third, reinforcement learning-based multi-objective optimization will be applied to identify infrastructure interventions that improve accessibility, resilience, and cost-effectiveness. An agentic AI platform will enable policymakers to compare scenarios and evaluate infrastructure investment strategies under changing demographic and climate conditions.

The project is in the planning stage.

Climate Change; Socio-Infrastructure Interdependency; AI for Infrastructure Planning; Infrastructure Accessibility.