Social Interactions Matter: Is Grey Wolf Optimizer a Particle Swarm Optimization Variation?

Many swarm-based algorithms are proposed using different inspirations from nature with the fact that they perform better than older versions. At the same time, some can resemble similar computational performances regardless of their inspirations. To understand the mechanisms of such similarities, recent works have analyzed and compared swarm-based algorithms via a network based on the information flow shared collectively. Here, we modeled networks of the social behavior of GWO (from wolves) and PSO (from birds) algorithms to investigate the extent of their similarities considering their temporal dynamics. To make sure that both algorithms had similar communication principles, we also designed the KBest topology for PSO that mimics the GWO communication. Using metrics from Network Science such as the Portrait Divergence, Local and Global Connectivity, our results showed that GWO can have different temporal signatures than PSO regardless of using a similar communication topology. Thus, we show that GWO is probably not just a variation of PSO.