伪随机数质量对简单粒子群优化算法性能的影响

On the Impact of Pseudo-random Number Quality on Simple Particle Swarm Optimization Performance

伪随机数对粒子群优化算法的性能影响主要体现在算法种群多样性上。低质量的伪随机数会导致粒子群优化算法的性能出现不稳定的现象,通过对几种典型伪随机数的分析比较之后得出,粒子编码长度和伪随机数的周期的相互作用才是导致算法不稳定的原因。相关实验也验证了这一结果。

Pseudo-random number on the performance of PSO algorithm is mainly reflected in the diversity of the population.Low-quality pseudo-random number will lead to the performance of PSO instability phenomenon,through the typical pseudo-random number drawn after analysis and comparison of the particle and pseudo-random number code length interaction cycle is the result algorithm for the cause of instability.Experiments also confirmed this result.