Outbreaks of infectious diseases can have a massive impact on public health and society. Figuring out how to deal with these outbreaks is tricky because there are a lot of different factors to consider. Current research mostly focuses on just one aspect, like how fast the disease spreads. But stopping an outbreak involves multiple different goals, including saving lives, keeping people healthy, but also minimizing economic impact, and downstream effects on people’s mental wellbeing. When introducing policy measures, we need to take an integrated approach taking all of these goals into account.
To help with future decision-making, we’ve developed (together wtih colleagues at VUB in Brussels and the National University of Ireland in Galway) a new approach using deep multi-objective reinforcement learning, which enables us to balance multiple goals. We’ve put this approach to the test by looking at different ways Belgium could have lifted lockdown measures during the COVID-19 pandemic. Our goal was to find strategies that would minimize both the number of COVID-19 cases and the impact of the measures on society. We’ve linked up a complicated decision-making process with a model of how COVID-19 spreads in Belgium and tried out different strategies.
Our research shows that using this approach can help us make better decisions about how to tackle outbreaks. It gives us a better understanding of how different strategies might play out and helps us find the best ways to keep people safe and healthy.
Congratulations to all researchers involved!
Read the full story: https://doi.org/10.1016/j.eswa.2024.123686
