改进粒子群算法的多峰值优化研究

Improved Particle Swarm Optimization for Multi-peak Problems

粒子群优化算法对于多维函数的最优解搜索存在前期易陷入局部最优,后期收敛速度缓慢的问题.将改进的k-中心点聚类分析与PSO相结合提出了一种混合粒子群算法KM-PSO,用于多峰值问题的优化.在算法中,利用k-中心点聚类分析方法将粒子群划分成若干个子群,结合PSO的隐含并行搜索的优势增强了寻优性能.不仅增加了粒子间的信息交换,抑制了早熟收敛,还提高了全局寻优速度和计算精度.仿真实验结果表明,KM-PS0性能优于基本粒子群优化算法.

The optimal solution for multi-dimensional search function,the Particle Swarm Optimization（PSO） easily trapped into local optimum,and convergence speed is slow later.The PSO algorithm employing improved k-medoids clustering analysis algorithm（KM-PSO） is proposed.In KM-PSO,the current particles is firstly divided into multi sub-population by improved k-medoids clustering,and PSO with the advantage of the implicit parallel search enhanced optimization performance.It not only exchanges more information among particle,restrains the tendency of premature,but also increases the converging rate and accuracy.Through the experimental results of both analysis and comparison to prove that KM-PSO is superior to original particle swarm optimization algorithm.