A few months after the emergence of a new coronavirus on the side of China, the whole world hopes for a treatment or a vaccine which will make it possible to fight against the pandemic of Covid-19. In the meantime, only health policies seem able to limit the spread of SARS-CoV-2. It is still necessary to be able to assess its effectiveness. This is the objective of a partnership between the Development Research Institute (IRD) and EDF R&D: modeling the impact of such policies at the local level.
Game of Thrones, World War Z or the mythical Lord of the Rings. What these film productions have in common is the use of “synthetic crowds”. Animations made on a computer, but which faithfully reproduce the reality of crowd phenomena.“This is an example of what multi-agent systems (AMS) can bring,” explains Mathieu Schumann, an engineer with EDF’s research and development firm. This branch of l’ artificial intelligence aims to reproduce the behavior of a set of agents — processes, robots or human beings –, evolving in a certain environment, interacting with each other.
Many disciplines also exploit these technologies. So the l’energy or telecommunications sector. And it is these ADMs that are now at the heart of a partnership between EDF’s research and development research institute (IRD). The two entities agreed to combine their research resources to address the challenges of the fight against Covid-19 and to provide local decision-makers with a decision-making tool in this unprecedented crisis situation. While conventional epidemiology models allow us to project the evolution of population compartments (infected compartments, cured, deceased,etc.), the use of multi-agent modelling allows us to take into account the heterogeneity of the population and the characteristics of individuals (age,occupations, etc.), at several territorial scales.
EDF and the Institute for Research for Development (IRD) have decided to combine their research resources to address the challenges of the fight against Covid-19. © EDF
“For nearly 15 years, EDF’s research and development has been developing expertise in these multi-agent systems. This is how we develop solutions to simulate the daily life of people in the habitat, much like in the video game The Sims, to better anticipate energy consumption,” says Mathieu Schumann. consommations But here, it is the IRD — and more accurately its Vietnam-based Ummisco research unit — that has developed with its partners a model that allows — in conjunction with epidemiological models — to assess the risks of coronavirus spread at the local level.
It should be remembered that the IRD is a benchmark in agent-based modeling. d’agents. For fifteen years now, he has been developing an open source software suite for modeling and simulation called GAMA Platform®. And from the beginning of the coronavirus crisis, the Institute has designed,on this platform, simulator for disease propagation and health policy evaluation called COMOKIT — for Covid-19 Modeling Kit.
An initial simulation of a model to support local authorities’ decisions on containment measures to be adopted when a municipality is affected by Covid-19. © GAMA Platform®
The power of high-performance computing
Among the entry data of this simulator: statistical data such as the number of houses in the locality studied,the number of inhabitants per household,the distribution of these inhabitants of en age or the distribution of assets.“All available information that can help generate a virtual population that is representative of the real population,” explains Mathieu Schumann. “Then they are digital avatars that are modeled to run the simulation. We do not use any personal data.»
Explore an important set of settings in a short time
To get results quickly — in the coming days or weeks –, the simulator harnesses the computing power available to EDF’s research and development. Specifically,the GAIA High Performance Calculator (40,000 3.2 PFlops computing cores) that allows you to run parallel calculations on many processors to simulate the interactions of thousands of agents. An essential asset to explore an important set of parameters in very short times. This is of course sought in the context of the crisis that we are experiencing.
The COMOKIT simulator was designed as a generic simulator, which can adapt to any study case. It will explore multiple scenarios and thus define the most appropriate containment perimeter, the impact of school closures, the most efficient containment time or the most effective measures to implement.
A testing phase is underway in Vietnam. The aim is to extend these simulations to other countries such as Cameroon or Senegal where the IRD is also present. And contacts are already being made with French communities that might be interested in these simulation models.
Topic produced in collaboration with EDF