Data-driven Planning in the Face of Supply Disruption in Global Agricultural Supply Chains

The intricacies of global food networks have been exacerbated by increased globalization, advances in farming/logistics technology, and a rising agricultural exchange between countries. Certain economies, especially regions with low agricultural yield, rely on food imports and are susceptible to food insecurity due to potential negative disruptions to the global food network. These rising complexities in global food networks result in increased dependencies between countries, rendering the overall network extremely vulnerable. Local disruptions to production levels could entirely cripple the food network and lead to longterm reduced food access worldwide. Understanding the impact of different disruptions and potential mitigation strategies at the country level on agricultural supply chains becomes important in the analysis of the global allocation of agricultural products. We model a stochastic resource allocation problem with nonlinear connectivity costs to capture trade dynamics between countries. We compare model recommendations to historical trade flow data including coffee import/export between countries, unveiling the value of centralized planning under potential disruption scenarios against the current practices.