Designing Near-Optimal Spatial Vaccine Allocation Strategies
Talk, MIDAS Network Annual Meeting, Atlanta, Georgia
Talks and presentations
Vaccine Allocation Approximation Guarantees for Curbing Outbreaks
Talk, INFORMS Annual Meeting, Indianapolis, Indiana
We seek to develop near-optimal strategies for timely vaccine allocation to reduce the burden of pandemics. While preventative vaccine allocation is important, effective vaccine allocations in a non-preemptive setting (i.e., after the start of an outbreak) is crucial to stem further damage. We expect to attack these problems by taking advantage of the possible submodularity of the objective function for meta-population models. We expect the problem structure we expose to lead to insights that advance the scalability and accuracy of vaccine allocation algorithms. We explore various spatial and temporal settings for allocation, leading to more realistic problem formulations. Our experimental analysis is driven by the FRED and SafeGraph human mobility datasets, which we utilize through meta-population disease models.