Optimization Based Modeling for the Food Supply Chain's Resilience to Outbreaks

Abstract

Scant research focuses on the resiliency of food supply chain networks to outbreaks, despite the estimated 600 million global foodborne illnesses annually. Outbreaks that cross country, state and provincial lines are virulent due to the number of people they can affect and difficulty controlling them. Research is needed on food supply chain networks, which are not well-characterized in relation to foodborne illnesses or generally. This paper introduces the United States Food, Energy, and State Transportation (US-FEAST) model and demonstrates its applicability via analysis of a hypothetical demand shock resulting from multistate food contamination. US-FEAST is an optimization-based model across all fifty states with yearly timesteps to 2030. It is a framework integrating food system data from multiple individual data sources. To calibrate, we develop a bilevel optimization routine to generate synthetic, state-level data and provide estimates of otherwise unavailable data at the intersections of the food and transportation systems. The results of US-FEAST elucidate potential heterogenous state-level variations in response, regional changes in food flows, vulnerabilities in the supply chain, and implications for food system resilience. While the generated data and scenarios are not empirical evidence, they provide insights to aid in planning by projecting outcomes and intervention effects. Our results estimate a 23% beef production decrease and 4% price decrease provide a road map toward data needs for quantifying food system resilience to foodborne illness. US-FEAST and its framework may have global utility for studying food safety in national and international food supply chain networks.