Data-driven Approach for Optimizing Food bank Operation

A food bank is a redistribution system that provides essential food and living supplies to marginalized groups, sourced from individuals and corporations. Recently, food banks have faced challenges due to management and strategic deficiencies, rather than limited supply or excessive demand. Therefore, we propose the reallocation of donations within the food bank system. This study aims to propose operational strategies through the reallocation of donations in food bank distribution, with the objective of improving effectiveness, efficiency, and equity in food distribution. Given the complexity of the problem, we present a machine learning-based two-stage solution framework that finds approximate solutions within a reasonable timeline. We conduct a simulation study on the 2022 Gyeonggi Metropolitan Food Bank dataset. We find that, while redistributing donated foods may increase routing costs, it can significantly enhance the operational efficiency and equity of the entire food bank system. Our research contribution lies in providing an effective evaluation method and objective function, as well as proposing an inventory-based machine learning algorithm to ensure the matching between complex inventory and demands, thereby enhancing the operational efficiency and timeliness of food bank operations.