Computational mechanisms affecting the efficiency of resource use in the honey bee swarm

Abstract

AbstractThe honey bee swarm carried out its best‐of‐N nest site computations by operating more than a dozen different information processing loops in parallel and by recruiting more resources to provide greater precision in loops evaluating the better quality sites. The positive feedback effects of recurrent recruitment by means of waggle dance signalling amplified the utilisation of the swarm's energy, memory and carrier resources. The relatively strong negative feedback effects of various attenuation mechanisms tended to reduce resource use and therefore counter‐balanced amplification for long enough for a meaningful nest site survey and quorum decision to be made. Some information processing mechanisms such as exploration tendency, waggle dance signalling, site non‐specific attenuation, noise reduction, independent site evaluation, energy efficient coding, mixed precision processing, self‐organising computation and quorum decision making were found to profoundly influence the efficiency of resource use. Significant insights were also gained into extended cognition, dark data processing, information quality and resource leakage. Finally, the energy cost of acquiring and processing sensory information was estimated.