Data-driven scalable pipeline using national agent-based models for real-time pandemic response and decision support

Area of Science:

• Computational epidemiology
• Public health informatics

Background:

• Pandemic preparedness requires robust simulation capabilities.
• Existing models often lack integration and automation for rapid response.

Purpose of the Study:

• To develop an integrated, data-driven operational pipeline for pandemic planning.
• To enhance the speed, reliability, and accuracy of national epidemic simulations.

Main Methods:

• An automated semantic-aware scheduling system for high-performance computing.
• A data pipeline for integrating national and county-level data.
• A digital twin of US social contact networks (288M individuals, 12.6B interactions).
• An extended parallel agent-based simulation model for epidemic dynamics.

Main Results:

• The pipeline enables 400 national simulation replicates in under 33 hours.
• Reduced human intervention leads to faster turnaround times.
• Increased reliability and accuracy in simulation results.

Conclusions:

• The developed pipeline significantly advances real-time epidemic sciences.
• It provides a powerful tool for federal and state-level pandemic planning and response.