At the moment some threads are given computationally intensive areas like cities, whereas others have predominantly sea or low population densities. Instead, we could rank grid cells by population size (as a lookup stored in the EpiSystem) and assign each thread every nth grid cell, for n total threads, distributing the work roughly evenly.
At the moment some threads are given computationally intensive areas like cities, whereas others have predominantly sea or low population densities. Instead, we could rank grid cells by population size (as a lookup stored in the
EpiSystem
) and assign each thread every nth grid cell, for n total threads, distributing the work roughly evenly.