Stochastic Programming Models: Community Health Pathways Scheduling and Optimal Vaccine Allocation
Abstract
The World Health Organization (WHO) designated the novel COrona VIrus (COVID-19) a global pandemic. This pandemic combined with the growing and aging population has created a crisis of unprecedented dimension regarding shortages in health care workforce. COVID-19 has
exacerbated care coordination by pushing the healthcare system to a limit in terms of rationing the allocation of scarce and valuable social and medical resources. Therefore, in this dissertation, we consider a Community Health Pathways HUB model to optimize resource scheduling and optimal vaccination allocation models to control the outbreaks.
Scheduling pathways involves uncertainty in resource availability because as human resources
may not report for work due to unforeseen circumstances such as delays in previous assignments. Similarly, the vaccination allocation problem involves uncertainty in COVID-19 characteristics, such as vaccine efficacy towards mutating variants, infectivity, and susceptibility. Stochastic programming is a framework for modeling optimization problems that involve data uncertainty. This dissertation considers three fundamental approaches of stochastic programming: 1) stochastic programming with recourse in which infeasibility is not allowed, only recourse/corrective actions with a certain cost; 2) chance-constrained programming in which infeasibility is allowed up to a certain probability; and 3) integrated chance-constrained programs that not only allow for infeasibility but also restrict it up to a certain threshold. The first approach is applied to a community health pathways scheduling problem, while the last two approaches are applied to optimal vaccine allocation under uncertainty for COVID-19.
Citation
Gong, Jiangyue (2022). Stochastic Programming Models: Community Health Pathways Scheduling and Optimal Vaccine Allocation. Doctoral dissertation, Texas A&M University. Available electronically from https : / /hdl .handle .net /1969 .1 /197139.