The QS Algorithm: Emergent and Resilient Network Formation
Abstract
Swarms of simple, identical robots often rely on centralized control or predetermined communication networks, creating single points of failure that make them vulnerable to faults and network breakdown. These vulnerabilities have been observed in real-world scenarios, including communication failures following Hurricane Katrina and in DARPA’s Subterranean Challenge. This work introduces a decentralized algorithm that eliminates those weaknesses by generating communication networks without a fixed topology. The network dynamically grows and adapts as agents are added or lost, maintaining connectivity without global coordination. The algorithm is inspired by bacterial quorum sensing and uses autoinducer-like signals, detection thresholds, and inhibitory mechanisms to regulate collective behavior and shape network structure. This biologically inspired approach enables agents to form efficient communication pathways using only local interactions. This Quorum Sensing Algorithm was developed and tested in an AnyLogic simulation. Agents were initialized at random positions in a fixed area and allowed Poisson movement. Across swarm sizes from 100 to 1000 agents, the algorithm consistently produced networks with average minimum path lengths below 2.8, demonstrating efficient message propagation without parameter tuning. This capability enables robust communication in environments where traditional infrastructure fails such as underwater, in space, or in other resource-constrained settings with limited sensing and communication range.
The QS Algorithm: Emergent and Resilient Network Formation
Swarms of simple, identical robots often rely on centralized control or predetermined communication networks, creating single points of failure that make them vulnerable to faults and network breakdown. These vulnerabilities have been observed in real-world scenarios, including communication failures following Hurricane Katrina and in DARPA’s Subterranean Challenge. This work introduces a decentralized algorithm that eliminates those weaknesses by generating communication networks without a fixed topology. The network dynamically grows and adapts as agents are added or lost, maintaining connectivity without global coordination. The algorithm is inspired by bacterial quorum sensing and uses autoinducer-like signals, detection thresholds, and inhibitory mechanisms to regulate collective behavior and shape network structure. This biologically inspired approach enables agents to form efficient communication pathways using only local interactions. This Quorum Sensing Algorithm was developed and tested in an AnyLogic simulation. Agents were initialized at random positions in a fixed area and allowed Poisson movement. Across swarm sizes from 100 to 1000 agents, the algorithm consistently produced networks with average minimum path lengths below 2.8, demonstrating efficient message propagation without parameter tuning. This capability enables robust communication in environments where traditional infrastructure fails such as underwater, in space, or in other resource-constrained settings with limited sensing and communication range.