OPTIMAL CONTROL OF EXTREME EPIDEMIC RISKS
- Gulf Journal of Mathematics : 1-10
Résumé
We propose a stochastic framework for the optimal control of epidemics, combining an SIR model with Conditional Value-at-Risk. The dependence between transmission and recovery rates is modeled using four copulas: Gaussian, Student, Archimax, and Vine. We establish theoretical results on existence, uniqueness, stability, and robustness of the optimal control. Simulations on COVID-19 data from Burkina Faso show that Archimax reduces the risk measure by 31% compared to the Gaussian copula. A complementary Hid-
den Markov analysis identifies three epidemic regimes with strong persistence, highlighting the value of modeling tail dependence for epidemic risk management.
Mots-clés
Optimal control, Risks measures, Copula, Markov chains, SIR model, COVID-19