Optimizing bioenergetic food web models of ecosystems using gamification and machine learning

Ecosystems are complex systems with many interdependent participants. Bioenergetic models of population dynamics help provide insight into specific aspects of ecosystem behavior. Due to the complex, nonlinear behavior of these models, and the large number of input parameters, it is difficult to parameterize them to correctly reflect realworld phenomena. We address this problem in the context of allometric trophic network models, a category of bioenergetic models based on food webs. Using custom simulation software to generate large numbers of simulated ecosystems, we apply machine learning to help navigate the large parameter space, revealing combinations of model parameters that result in sustaining ecosystems. Furthermore, we apply gamification to take advantage of human intuition, using the insights gained from the machine learning process to provide automatic guidance to players.