Scatter PSO - A more effective form of Particle Swarm Optimization
Peng-Yeng Yin, Fred Glover, Manuel Laguna, Jiaxian Zhu · 2007
A fertile complementarity exists between scatter search (SS) and particle swarm optimization (PSO). Shared and contrasting principles underlying these methods provide a fertile basis for combining them to create a hybrid method. We identify a specific hybrid, Scatter PSO, giving rise to two variants that prove more effective than the constriction factor model of PSO. Applied to finding global minima for continuous nonlinear functions, Scatter PSO not only is able to obtain better solutions to a widely used set of benchmark functions, but also proves more robust under a variety of experimental conditions.