Robust optimization for United States Department of Agriculture food aid bid allocations

• An MIP model developed for current one-round food aid bidding system used by USDA.
• Humanitarian logistics challenges addressed via robust-optimization.
• Box and ellipsoid based frameworks employed to model uncertainty.
• USDA demand, supplier and carrier bid pricing uncertainties addressed.
• Case study utilizing historical invoice data demonstrates our model applicability.

The U.S. Department of Agriculture (USDA) currently uses a bidding system to determine carriers and suppliers that would partner in providing food aid annually in response to global emergencies and famine. We mimic the USDA approach via a robust optimization model featuring box and ellipsoid uncertainty frameworks to account for uncertainties in demand, supplier and carrier bid prices. Through a case study utilizing historical invoice data, we demonstrate our model applicability in improving ocean carrier and food supplier bid pricing strategy and similar supply chain network optimization problems. Through a validation algorithm we demonstrate the value of our robust models.