Animal groups represent magnificent archetypes of self-organized collective behavior. As such, they have attracted enormous interdisciplinary interest in the last years. From a mechanistic point of view, animal aggregations remind physical systems of particles or spins, where the individual constituents interact locally, giving rise to ordering at the global scale. This analogy has fostered important research, where numerical and theoretical approaches from physics have been applied to models of self-organized motion. In this paper, we discuss how the physics methodology may provide precious conceptual and technical instruments in empirical studies of collective animal behavior. We focus on three-dimensional groups, for which empirical data have been extremely scarce until recently, and describe novel experimental protocols that allow reconstructing aggregations of thousands of individuals. We show how an appropriate statistical analysis of these large-scale data allows inferring important information on the interactions between individuals in a group, a key issue in behavioral studies and a basic ingredient of theoretical models. To this aim, we revisit the approach we recently used on starling flocks, and apply it to a much larger data set, never analyzed before. The results confirm our previous findings and indicate that interactions between birds have a topological rather than metric nature, each individual interacting with a fixed number of neighbors irrespective of their distances.
An Algorithm for Global Optimization Inspired by Collective Animal Behavior
