Decision-Dependent Distributionally Robust Markov Decision Process Method in Dynamic Epidemic Control

In this paper, we present a Distributionally Robust Markov Decision Process (DRMDP) approach for addressing the dynamic epidemic control problem. The Susceptible-Exposed-Infectious-Recovered (SEIR) model is widely used to represent stochastic spread of infectious diseases, such as COVID-19. While Processes (MDP) offers mathematical framework identifying optimal actions, vaccination and transmission-reducing intervention, combat disease spreading according SEIR model. However, uncertainties in these scenarios demand more robust that less reliant on error-prone assumptions. primary objective our study introduce new DRMDP allows an ambiguous distribution transition dynamics. Specifically, consider worst-case probabilities within decision-dependent ambiguity set. To overcome computational complexities associated with policy determination, propose efficient Real-Time Dynamic Programming (RTDP) algorithm capable computing policies based reformulated accurate, timely, scalable manner. Comparative analysis against classic MDP demonstrates achieves lower proportion infections susceptibilities at reduced cost.