摘要
针对差分进化算法在运行后期收敛速度慢和容易陷入局部最优的不足,提出一种混合聚类的粒子群差分进化算法.利用一步K-均值聚类算法改进粒子群优化算法的速度更新,使用线性递减的选择概率将改进后的粒子群算法与差分进化算法相融合,并在一定条件下对种群中部分较差个体进行重置.对9个典型测试函数的数值试验和与其他三种进化算法的比较结果表明:所提算法收敛速度快,寻优能力强并且鲁棒性好.
A hybrid differential evolution algorithm is introduced because the basic differential evolution algorithm has disadvantages of low convergence speed and local optimum.Firstly,K-means cluster algrithm was used to modify the velocity updating formula of the particle swarm optimization algorithm. Then the modified particle swarm optimization algorithm was combined with the differential evolution algorithm by means of a linear decreasing selective probability.Finally some individuals with poor performance were reset under certain conditions.Numerical experiments on nine typical benchmark functions illustrate that the proposed algorithm is fast in convergence speed,strong in search ability and good in robustness.
出处
《西安工业大学学报》
CAS
2016年第5期357-364,共8页
Journal of Xi’an Technological University
基金
国家自然科学基金(61273311
61173094)
关键词
K-均值聚类
混合算法
差分进化
粒子群优化
种群重置
K-means clusters
hybrid algorithms
differential evolution
particle swarm optimization
population reset
