Self-Adaptive Crossover Particle Swarm Optimizer for Multi-dimension Functions Optimization

Based on analyzing that solution diversity can be improved by bringing crossover operation into particle swarm optimization, crossover particle swarm optimizer is put forward and applied to optimize multi-dimension benchmark functions. Outcomes testify that crossover OPS can achieve better performances than other current mended PSOs, and cost less CPU time. Four self-adaptive probability models are adopted to adjust the crossover probability based on particle swarm optimization convergence model. Results and convergence rate of the four models are compared and analyzed finally.