Vortex Swarm Optimization: New Metaheuristic Algorithm

Abstract

This paper proposes a new swarm intelligence algorithm called vortex swarm optimization, inspired from vortex collective behavior where particles move around a common known center, they cover all the directions so they play as alarm system where any individual at any direction see the target (resources, predator…etc.) it informs other individuals to take an action. This type of collective behavior was seen in many creatures like (Bats and fishes). They use this type of collective behavior for feeding, access to resources, moving, information transmission, decision making, and protection from predators. This type is different from straight-ahead collective behavior (polarized collective behavior) where individuals move forward. The majority of swarm intelligence algorithms inspired their techniques from straight-ahead collective behavior, but vortex collective behavior had a few pieces of research so the importance of this research appears. The algorithm performs the two basics functions that any searching algorithm had (exploitation and exploration) by circling phase the algorithm will know the direction of the promising region, By swimming phase the algorithm will go to this region (exploration)and by attacking phase individuals will reduce their interaction range and reduce their velocity to get much closer to each other and best region found so far (exploitation), Algorithm skip local minimum by eliminating some individuals to a new regions to explore them and to change vortex direction to explore new directions. The algorithm was tested over ten benchmark functions and it competes with some other powerful algorithms.