A self-organizing model for task allocation via frequent task quitting and random walks in the honeybee

Social insect colonies are able to quickly redistribute their thousands of workers between tasks that vary strongly in space and time. How individuals collectively track spatial variability is particularly puzzling because bees have access only to local information. This work presents and tests a model showing how honeybees solve their fundamental within-nest spatial task-allocation problem. The algorithm, which is self-organizing and derived from empirical studies, couples two processes with opposing effects. Frequent task quitting, followed by patrols, during which bees are insensitive to task stimuli, serves to randomize individual location throughout the nest without reference to variation in task demand, while a foraging-for-work-like mechanism provides the opposing force of localizing individuals to areas of high task demand. This simple model is shown to generate sophisticated patterns of task allocation. It allocates bees to tasks in proportion to their demand, independent of their spatial distribution in the nest, and also reallocates labor in response to temporal changes in task demand. Finally, the model shows that task-allocation patterns at the colony level do not reflect colonies allocating particular individuals to tasks. In contrast, they reflect a dynamic equilibrium of workers switching between tasks and locations in the nest.