USARPAC operates under contested logistics conditions across the Indo-Pacific, responding to large-scale humanitarian disasters while maintaining readiness for large-scale combat operations. Both missions draw from a centralized supply chain, across vast distances and often-degraded networks. Our models and framework provide a data-driven analytical capability that makes those tradeoffs visible to planners and complements senior leader decision-making.
What the work delivers
University of Arkansas' Dr. Rob Curry and doctoral student LTC Clay Woody designed a framework that addresses two operational questions:
- Where should critical supplies be pre-positioned before a crisis?
- How should those supplies move when routes, ports, or comms degrade?
The models use mixed-integer programming to balance disaster-driven demand against theater posture, redundancy, transportation limits, and infrastructure degradation. Scenario-based simulation stress-tests each network design across the full portfolio of futures USARPAC plans against, including most-likely and most-dangerous conditions.
The output is a prototype analytical decision support capability that complements senior leader decision-making with data-driven analysis. It identifies the supply nodes and routing strategies that perform consistently across both crisis and conflict scenarios. While still under development, the modeling approach and assumptions have been iteratively refined and validated with operational partners.
Operational partners
- USARPAC Chief Data Office (CDO)
- 8th Theater Sustainment Command
- USARPAC G5
- U.S. Military Academy Operations Research Center (ORCEN)
These partners have validated the models against real operational requirements through multiple iterations. The capability is now undergoing strategic-level testing and refinement to support future consideration and integration into planning and exercises.
Why this matters in theater
Faster crisis response. Tougher supply networks. Better positioning decisions made before the typhoon lands.