Integrating food insecurity mapping and equitable resource allocation for foodbank redistribution in an end-to-end decision support framework

Identifying food insecure regions and effectively allocating resources are critical components for enhancing food security policies. The accuracy of local food insecurity solutions greatly depends on the quality of integration between the identification of and allocation to vulnerable populations. The literature pertaining to the quantitative modelling of food insecurity is fragmented and scattered across various disciplines, lacking a cohesive interdisciplinary approach. In this paper, we introduce a data-driven end-to-end decision support framework, called the food insecurity mapping and allocation (FIMA) framework, which employs a machine learning-based food insecurity index for quantifying the degree of food insecurity experienced by the population of a region. This index is integrated into a resource allocation model for ensuring equitable allocation of resources to vulnerable populations, thereby facilitating a more targeted approach towards food policy interventions. The proposed framework comprises three modelling components, namely a spatial analysis component, a machine learning component, and a resource allocation component, which are integrated to provide complete decision support in respect of food insecurity. The practical application of the FIMA framework is demonstrated in the form of a case study involving a foodbank in South Africa. The results showcase significant enhancements in resource allocation efficiencies, illustrating how insight derived from applying the framework can lead to better-informed decisions. By leveraging the power of analytical tools to guide interventions, the FIMA framework offers a significant advancement in the field of food security modelling, providing a robust tool for humanitarian organisations aiming to combat food insecurity more effectively. • FIMA is a novel end-to-end framework for food insecurity mitigation. • A novel index estimates food insecurity using machine learning algorithms. • A heuristic is employed to allocate surplus food, guided by the index. • A South African case shows improved equitability over historical allocations. • Simulation results confirm the robustness of the heuristic.