MORO, Mme Maëlys (2024) Modèles PINNs pour l’élaboration de stratégies de contrôle optimal de modèles épidémiologiques structurés PRE - Research Project, ENSTA.

[img]
Preview
PDF
2032Kb

Abstract

The COVID-19 crisis has raised numerous questions about the risk of future pandemics. These are likely to become more frequent with climate change. Firstly, rising temperatures are expanding the habitats of disease-carrying animals such as mosquitoes and ticks, as well as extending their active periods. Additionally, the increasing proximity between humans and animals heightens the risk of new virus transmissions. In this context, extensive research is being conducted to model the spread of infectious diseases. Several models in which the population is divided into subgroups have already been developed. These are called compartmental models. However, these models may be incomplete as they do not always consider factors such as vaccination rates or mortality attributable to the disease. In this context, a new compartmental epidemiological model has been developed. This model includes seven compartments : Susceptible (S), Exposed (E), Infected (I), Hospitalized (H), Recovered (R), and Deceased (D). The evolution of the population in these compartments is governed by a set of partial differential equations. The aim of this work is to solve this system of equations using artificial intelligence methods. Subsequently, the optimal vaccination coverage (i.e., minimizing health costs) could be found.

Item Type:Thesis (PRE - Research Project)
Uncontrolled Keywords:Physics informed neural networks, Compartmental models in epidemiology, Optimal control strategies
Subjects:Information and Communication Sciences and Technologies
Mathematics and Applications
ID Code:10067
Deposited By:Maëlys MORO
Deposited On:02 sept. 2024 16:57
Dernière modification:02 sept. 2024 16:57

Repository Staff Only: item control page