Multiple cooperating swarms for non-linear function optimization

Abstract

This paper investigates a new approach applied to particle swarm optimization. The paper addresses the idea of having multiple swarms searching for a solution while cooperating with each other by exchanging their best solutions. The experiments show that this approach behaves in a way that is dependent on the function being optimized. They also show that changing the synchronization period (number of generations) after which the swarms cooperate with each other has a great effcct on the obtained solution quality.