Strategies for Collective Adaptive Workload Distribution in Varying Work Conditions

Social insects such as ants robustly dig during collective nest construction. Individuals quickly adapt to changing work conditions, despite the complexity of cooperative digging and environmental constraints. In this work, we model how behavior changes if individual ants optimize an objective in a changing environment. The theory predicts that collective excavation and worker density change with tunnel length if individuals optimize their excavation behavior. To test the predictions, we studied the dynamics of confined ant excavation behavior in laboratory experiments and a collective robophysical model consisting of cohesive granular-media-excavating robots. We developed adaptive rules for robots to optimize successful pellet retrieval by modulating individual behavior. Our theory suggests a strategy by which individual behavior can drive adaptation of a group in changing work conditions through optimization of individual behavior. This may lead to insights into how social insects adapt to their environments; and how to develop similar robust and scalable strategies for task-oriented and constrained physical swarm systems.