Dynamic hospital resource scheduling during pandemics with stochastic optimization

Abstract

The COVID-19 pandemic has highlighted the need to effectively manage hospital resources: ICU beds and ventilators. These resources are significant for sustaining life, especially in severe cases. Traditional deterministic models often fall short in addressing the uncertainties associated with patient inflows and resource availability. This paper develops a novel two-stage stochastic programming model which aims to dynamically allocate resources to deal with the variability of inpatient admissions. To this end, the scenarios are developed using Monte Carlo simulation based on the probabilities estimated from the historical data. The model is created in Python language and solved using the Gurobi optimizer in 0.05s, a large-scale scenario optimization analysis problem with 42 variables and 35 constraints. The KPIs show the highest utilization of ventilators at 66. 67% and the average reduction of 53.5 in the number of offers an ICU practical shortfall leading to better patient care and shorter wait times. This research presents a data-driven tool to enhance the decision-making process and the healthcare system's overall readiness to maintain its strategic reserves by implementing flexible staffing models to improve preparation for disasters such as the pandemic. Its stochastic optimization framework makes hospital resource allocation more efficient, offering a scalable, resilient solution for tackling future pandemic challenges.