Application of Distributionally Robust Optimization Markov Decision-Making Under Uncertainty in Scheduling of Multi-category Emergency Medical Materials

In the preliminary stages of public health emergencies, many regional public health systems do not have enough medical resources to address the needs that arise from the emergencies. Additionally, due to the rapid development of emergency situations, it is difficult to accurately understand all medical material demands. Thus, we develop a multi-category emergency medical materials robust scheduling method for multi-period, which continuously schedules the availability of multi-category emergency medical materials during times of uncertain demand. First, we developed a single-period Distributionally Robust Optimization (DRO) model to provide a powerful strategy for scheduling multi-category emergency medical materials with little information on material demand. In the DRO model, we prioritize medical material requirements into different categories, assuming only that the means and variances of information on the demand are available to seek an optimal implementation strategy in a single period. We then combine the DRO scheduling model with the Markov Decision Process (MDP) and extend it to the multi-period Distributionally Robust Optimization Markov Decision Process scheduling model (DRO-MDP). Our DRO-MDP model provides encouraging guidelines to solve the multi-period scheduling problem of multi-category emergency medical materials in uncertain situations. A simulated experiment is used to demonstrate the effectiveness of the proposed model. The simulation uses COVID-19 data from New Delhi, India in the spring of 2021. It is important to note that the model we propose can be easily generalized as a framework for any multi-category resource allocation problem with uncertain needs.