Fernando Peruani - Beyond Complex Network Models: Epidemic Models Based on Moving Agent Systems

Fernando Peruani's talk on "Beyond Complex Network Models: Epidemic Models Based on Moving Agent Systems" at Leibniz AI Lab workshop.

Epidemic models canonically assume one of the following supports on to top of which the spreading occurs: i) a well-mixed population, ii) a (regular) lattice, or iii) an underlying complex network over which the disease propagates. Is there anything else beyond these three supports, i.e. well-mixed populations, lattices, or complex networks? In this talk we will discuss one alternative: the use of systems of mobile agents, where the agents adopt different states (e.g. state S, I, R). The theoretical importance of these models is paramount: they interpolate between well-mixed populations and lattice models, and exhibit a behavior that share same similarities with -- despite it cannot be reduced to -- the one observed on dynamical complex networks. At the application level, the advantage of these models is that they allow to evaluate the impact of human mobility at scales smaller than large-scale transportation networks. We will use of these models to investigate different sources of fluctuations and assess how predictable is the evolution of an epidemics. And in particular, we will see how a vaccination can lead, counterintuitively, to an increase of infections. Refs: Peruani, Sibona, Phys. Rev. Lett. 100, 168103 (2008); Soft Matter 15, 497-503 (2019), and Marcolongo et al. preprint (2021)