From Honeybees to Robots and Back: Division of Labor based on Partitioning Social Inhibition Payam Zahadat * , Sibylle Hahshold, Ronald Thenius, Karl Crailsheim, Thomas Schmickl Artificial Life Lab of the Department of Zoology, Universit¨ atsplatz 2, Karl-Franzens University Graz, 8010 Graz, Austria Phone: +43 316 380 3982 {payam.zahadat, thomas.schmickl}@uni-graz.at Abstract In this paper, a distributed adaptive partitioning algorithm inspired by division of labor in honeybees is investigated for its applicability in a swarm of underwater robots in one hand and is qualitatively com- pared with the behaviour of honeybee colonies on the other hand. The algorithm, Partitioning Social Inhi- bition (PSI), is based on local interactions and uses a simple logic inspired from age-polyethism and task allocation in honeybee colonies. The algorithm is analyzed in simulation and is successfully applied here to partition a swarm of underwater robots into groups demonstrating its adaptivity to changes and appli- cability in real world systems. In a turn towards the inspiration origins of the algorithm, three honeybee colonies are then studied for age-polyethism behaviours and the results are contrasted with a simulated swarm running the PSI algorithm. Similar effects are detected in both the biological and simulated swarms suggesting biological plausibility of the mechanisms employed by the artificial system. 1 Introduction Social insects are promising sources of inspiration in the field of swarm intelligence and collective robotics. Such insects are highly capable of self-organization and self-regulation of their colonies and adaptation to changes in the internal and environmental conditions. One of the prominent characteristics of social insects is division of labor happening in, e.g., honeybee colonies [34, 14], wasps [37], termites [8] and ants [13, 20]. * corresponding author 1