An Optimized Approach to Vaccine Clinic Placement
Connor Douglas, Aditya Krishnamachar, Vishesh Patel
Abstract:
Determining the optimal placement of stores, factories, and other industrial facilities has long been the topic of much scrutiny throughout the computer science and operations research communities. Indeed, this topic has been studied thoroughly under the name of the facility location problem. Yet, when profit is not a driving factor, there is little incentive to scrutinize these placements, which can often involve a costly process of modeling and optimizing over large datasets. As the COVID-19 pandemic rages on, we see the opportunity to leverage past research in this space for the public good. In this paper, we make two novel contributions. The first is that we put forth and implement a weighted k-median algorithm. This algorithm efficiently approximates the position of k clusters in order to minimize the weighted sum of distances to the nearest center for all points in a population. The second contribution is that we apply this model to inform vaccine clinic placement in the greater St. Louis area. Our work is broadly applicable to any population and comes at a time when vaccination rates are insufficiently low and the rate of contagion is high.