Bayesian Inference-Informed Agent-Based Modelling of Migratory Fish near Energy Infrastructure
Jennifer Gaskell (University of Glasgow)
Thursday 31st October, 2019 14:00-15:00 Maths 311B
Here, we develop an agent-based model coupled to a hydrodynamic model, to understand the movement of migratory fish near infrastructure. Unmeasurable parameters, such as the memory of a fish, are inferred through sequential Monte Carlo Approximate Bayesian Computation using trajectories from videos of small-scale flume experiments. The model is then validated on a larger scale in rivers, before being used to analyse the best infrastructure design, locations for new energy sites and advice for policy-makers. This work is part of the EPSRC Vaccinating the Nexus project focusing on water-energy-food nexus solutions.