摘要
提出了一种新的基于粒子群和模拟退火的聚类算法。每个粒子作为聚类问题的一个可行解组成粒子群,粒子的位置由聚类中心向量表示。为避免粒子群陷入局部最优解,结合聚类问题的实际特点,提出了利用模拟退火的概率突跳性的两个解决方案。实验结果表明,新算法增强了全空间的搜索能力,性能优于粒子群算法和传统的K-means算法,具有较好的收敛性,是一种有效的聚类算法。
A new clustering algorithm is proposed based on particle swarm optimization and simulated annealing.The particle swarm is composed of particles,and each particle is a possible solution of the clustering problem,the position of the particle is represented by cluster center vector.To escape from local optimum,two solutions are proposed using the probabilistic jumping property of simulated annealing algorithm combined with the clustering problem.The experimental results on different datasets show that the new algorithm has enhanced the global search ability,has better performance than particle swarm optimization and K-means algorithm,has better global convergence,and it is an effective clustering algorithm.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第35期139-141,191,共4页
Computer Engineering and Applications
基金
国家自然科学基金No.60874075
聊城大学科研基金(No.X071021)~~
关键词
粒子群优化
模拟退火
聚类
particle swarm optimization
simulated annealing
clustering
