A Family Particle Swarm Optimization Based on Tree Structure

For the independence of particles in original particle swarm optimization (PSO), we had proposed the family PSO (FPSO) previously. To study the relationship between families and set up effective management to the families in the FPSO, a family tree was introduced into the FPSO. Based on the proposed tree structure of FPSO (Tree-FPSO) different tasks could be conveniently assigned to different families according to their places in the family tree. Simulations for six benchmark functions demonstrated the advantage and effectiveness of Tree-FPSO. The experiments also showed that families playing different roles were very important to improve the performance at the convergence speed and this made the PSO higher efficiency.