A Fractal Evolutionary Particle Swarm Optimizer

Abstract

A Fractal Evolutionary Particle Swarm Optimization (FEPSO) is proposed based on the classical particle swarm optimization (PSO) algorithm. FEPSO applies the fractal Brownian motion model used to describe the irregular movement characteristics to simulate the optimization process varying in unknown mode, and include the implied trends to go to the global optimum. This will help the individual to escape from searching optimum too randomly and precociously. Compared with the classical PSO algorithm, each particle contains a fractal evolutionary phase in FEPSO. In this phase, each particle simulates a fractal Brownian motion with an estimated Hurst parameter to search the optimal solution in each sub dimensional space, and update correspond sub location. The simulation experiments show that this algorithm has a robust global search ability for most standard composite test functions. Its optimization ability performs much better than most recently proposed improved algorithm based on PSO.