Pandemic Simulation And Forecasting For An Empowered Policy-making: Convergence Of Machine Learning And Epidemiological Models

Luxembourg is entering the post-lockdown stage of the Covid-19 pandemic. It goes without saying that a key concern is the risk of triggering a new outbreak, due to over-permissive post-lockdown policies. However, understanding the propagation of the pandemic remains challenging, mainly because no existing model can accurately evaluate the individual contributions of the mitigation strategy (border control, school closure, open-air activities, retail activities...) on the reproduction rate of the disease. Moreover, current predictions lean on selected experts' opinions and on epidemiological models whose parameters are set arbitrarily. This impedes any reliable analysis and scheduling of proper post-lockdown measures. Therefore, the objective of PILOT is to develop a data-driven pandemic simulation and forecasting tools to support policymakers in designing safe and efficient exit strategies. Thereby, they will enable appropriate planning of those measures, allowing policymakers to answer practical questions such as: "How to prioritize and schedule the re-opening of major Luxembourg's employers?" or "which global exit strategies guarantee that the hospitalization rate never exceeds 25% of available beds?"