Highlights

In Algorithmic Game Theory, fairness emerges as a pivotal theme, guiding research regarding equitable item allocation among agents, based on well established fairness criteria. Within a growing body of literature on online or dynamic fair division (e.x. [BKP+23]), my collaborators and I address the pressing issue of food insecurity and waste in our paper, “Automating Food Drop: The Power of Two Choices for Dynamic and Fair Food Allocation” (EC 2024). This work introduces a theoretically informed solution that enhances the distribution efforts of our non-profit collaborator in Indiana, the Indy Hunger Network (IHN) and its Food Drop program. Moving away from traditional, manual, and labor-intensive processes that led to uneven allocations, our cloud-based system, deployed on Amazon Web Services (AWS) and available as open-source software, significantly improves the management and distribution of food donations. In Figure 1 you can see a typical use case scenario of our platform. Our paper presents the algorithmic principles and empirical analysis that inform the design of our system. Through our theoretical analysis we establish guarantees for our algorithm’s performance and demonstrate the tightness of our work by showing that, even in simple scenarios, no other algorithm can offer better guarantees. This claim is further strengthened by comparative experiments, highlighted in Figure 2, where we can see that our algorithm (3rd image of Figure 2) achieves almost perfect food allocation in the state of California whereas other natural algorithms leave many areas severely under served. This research contributes to solving the technical problems associated with dynamic fair division and has a practical impact on reducing food waste and alleviating food insecurity. For more information regarding this project I refer the reader to the relevant press release [Kin24].