A Consensus-Achieving Algorithm for Robot Swarms Inspired by Ant Collective Decision-Making

This paper proposes a novel consensus-achieving algorithm for robot swarms inspired by the collective decision-making of ants. The best-of- problem, a collective decision-making challenge, has attracted significant attention in the field of swarm robotics. In the best-of- problem, agents evaluate multiple candidate solutions, and all agents must achieve consensus on the optimal solution through local communication. The proposed method allows agent swarms to select the optimal solution from multiple candidates and achieve consensus through local communication. It uses the response threshold variance system and the collective decision-making mechanisms observed in ants. Additionally, the proposed method includes the majority synching bias to prevent swarm division and the multi-lineage model to evaluate the candidates based on the multiple evaluation indexes. This study evaluates the proposed method on the nest migration of the rock ants (Temnothorax albipennis) to investigate its effectiveness through simulation experiments. We demonstrate that the proposed method allows agent swarms to select the optimal solution from multiple competing candidates and achieve consensus through local communication without the swarm division.