AI and the World's Hard Problems · Economics & social sciences · arXiv · 2026-10-26
A Dynamic Inventory Rationing Model for Last-Mile Delivery of Medication in Liberia's Community Health Program
The paper studies how to allocate scarce medications from a central warehouse to a health facility and the community health assistants it supervises in Liberia. It formulates the problem as a finite-horizon stochastic control model that balances referral costs, untreated patients and a demand-weighted fairness utility. Using the deep splitting method with neural networks, calibrated on malaria data from Lofa County, the policy cuts operating costs by about 6% relative to a fairness-only benchmark while retaining about 98% of its fairness utility.