Hierarchical Particle Swarm Optimization for Optimization Problems

In this paper, a hierarchical particle swarm optimization (HPSO) is proposed to improve the premature convergence in the PSO approach. In the proposed HPSO approach, all particles are arranged in a regular tree structure and move up or down in the tree based on their fitness value. For the velocity update of each particle, it depends on the position of each particle in the tree. Besides, a mutation operator is added into the proposed HPSO approach. Consequently, the diversity of the population increases so that the HPSO approach can improve the premature convergence in the PSO approach. Finally, several benchmark functions for optimization problems are utilized to illustrate the effectiveness of the proposed HPSO approach to improving the premature convergence.