The impact of interaction models on the coherence of collective decision-making: A case study with simulated locusts

Abstract

A key aspect of collective systems resides in their ability to exhibit coherent behaviors, which demonstrate the system as a single unit. Such coherence is assumed to be robust under local interactions and high density of individuals. In this paper, we go beyond the local interactions and we investigate the coherence degree of a collective decision under different interaction models: (i) how this degree may get violated by massive loss of interaction links or high levels of individual noise, and (ii) how efficient each interaction model is in restoring a high degree of coherence. Our findings reveal that some of the interaction models facilitate a significant recovery of the coherence degree because their specific inter-connecting mechanisms lead to a better inference of the swarm opinion. Our results are validated using physics-based simulations of a locust robotic swarm.