Agreeing to disagree: synergies between particle swarm optimisation and complex networks

Abstract

Due to its numerous applications in fields like systems biology, medicine, technology, engineering, or social sciences, the new science of complex networks (CNs) has become extremely popular over the last couple of decades. Particle swarm optimisation (PSO) and CN science have significant common ground, as they both deal with a set of agents (particles for PSOs, vertices for CNs) which interact according to an underlying topology. Moreover, one of the most important branches of CN science is represented by social networks, while PSO was originally conceived for optimisation through the simulation of social behaviour, thus further emphasising the strong tie between the two scientific fields. A prominent problem in CNs is the opinion dynamics in social networks. To this end, opinion interaction models are tested in order to verify if they can recreate a realistic social behaviour; a ubiquitous feature of such reallife social behaviours is persistent opinion disagreement. One of the most important characteristics of a social network which fosters disagreement is that it contains stubborn agents, namely vertices which never change their opinion. We discuss models of disagreement, such as that with stubborn agents and tolerance-based models, then offer a new perspective in exploring the frontiers between network science and PSO.