What Can Honeybees Tell Us About Social Learning

We independently implemented and studied the nest-site selection model described in Passino, K. M., and Seeley, T. D., “Modeling and analysis of nest-site selection by honeybee swarms,” 2006, focusing on the default parameter values they obtained by field calibration. We focus on aspects of the model pertaining both to imitation and social learning and to the model as kind of metaheuristic. Among other things, we find that the model is robust to different parameterizations of social learning, but that at least a modicum of social learning is essential for successful nest-site selection (in the model). Regarding the model as a metaheuristic, we find that it robustly produces good but significantly non-optimal nest-site selections. Instead of a single-criterion metaheuristic, the algorithm is best seen as balancing three objectives: choose the best of the available sites in the neighborhood, make the choice quickly, minimize risk of failing to choose a site.