Sebastien Billès: Joint evolution and scale separation

 In the field of artificial evolution, one often evolves a controller for a machine, and sometimes an embodiment for a simple controller.  In this study, we aim at understanding the joint evolution of those two elements, by using simple models with a defined "software" and "hardware" part, such as a neural system and a limb. This allows different levels of impact upon mutation, resulting in a fixation of the more impactful levels early in the simulation.

Paul Pineau: Collective Dynamics of Trail-Interacting Particles with Saturating Interactions

  
Trail-interacting particles leave footprints that locally modify their future motion. Unlike standard autochemotactic models, the footprints are permanent: they neither diffuse nor decay. We study a class of trail-mediated interactions that saturate with the local footprint density, reflecting a finite response to increasingly strong environmental signals.

Such a mechanism has recently been shown to describe the migration of a single fibroblast cell confined to a one-dimensional environment. A natural extension is to investigate the collective dynamics arising in the presence of multiple interacting cells. In this talk, I address this problem and present several exact statistical results obtained within the framework of self-interacting random walks.