Gaussian-distributed Particle Swarm Optimization: A novel Gaussian Particle Swarm Optimization
Joon-Woo Lee, Ju-Jang Lee · 2013
Particle Swarm Optimization (PSO) is a metaheuristic widely used for optimization, which is inspired by social behavior of bird flocking or fish schooling. The PSO algorithm, however, generally have several parameters that need to be properly set before using the algorithm. The choice of PSO parameters is known that it has considerable influence on optimization performance. There have therefore been many studies for setting PSO parameters. Among them, Gaussian PSO (GPSO) was proposed, which was based on the Gaussian distribution and had improved the convergence ability of PSO without the parameter tuning. This paper proposes a novel Gaussian-based PSO, called Gaussian-Distributed PSO (GDPSO), which was developed through a new approach unlike GPSO. The GDPSO also do not need the parameter tuning like GPSO, and it especially had better performance and value to solve the high dimensional or difficult problems than GPSO in the result of the comparative simulation on several well-known benchmark functions.