Robin Gras

French

Robin Gras has created EcoSim, a freely available simulation of a predator-prey ecosystem. It can help better understand how human behaviour can affect ecosystems by predicting possible extinctions, the invasion of species or the spread of disease.
https://sites.google.com/site/ecosimgroup/research/ecosystem-simulation

This research is focused on the study of theoretical ecological questions using a simulation approach. We have designed EcoSim, a generic individual-based predator-prey evolving ecosystem simulation. One major and unique contribution is that this is the only simulation modelling the fact that individual behaviours affect evolution and speciation. Compute Canada infrastructure is used to run the simulation and also store and analyze the data generated. EcoSim manages hundreds of thousands of agents simultaneously. In a run, more than one billion agents can be born and several thousands species can emerge and become extinct. In addition to academic publications, this work has received extensive coverage in traditional and technical media, such as ACM TechNews, USA Today, Windsor Star, Metro, International Innovation, Computer Power User and C-Fax 1070 radio. The results of this research have the potential to impact Canadians by considering difficult theoretical ecological questions which, due to the spatial and temporal scale involved, can hardly be understood using only traditional field and laboratory approaches. This research will bring some new insights on the effect of human behaviour on ecosystems by predicting possible species extinction, the probability of a species invasion or the diffusion of diseases in populations. This research would not be possible without Compute Canada infrastructure because, for statistical purposes, many runs of the simulation with the same environmental conditions must be performed. Moreover, for each of the current six-plus projects on which we work, several different environmental conditions need to be tested, each of them requiring a huge amount of computational resources.

Top