布谷鸟搜索
聚类分析
差异进化
混乱的
计算机科学
粒子群优化
局部搜索(优化)
算法
水准点(测量)
数学优化
局部最优
数学
人工智能
大地测量学
地理
作者
Saida Ishak Boushaki,Nadjet Kamel,Omar Bendjeghaba
标识
DOI:10.1016/j.eswa.2017.12.001
摘要
This paper presents a new quantum chaotic cuckoo search algorithm (QCCS) for data clustering. Recent researches show the superiority of cuckoo search (CS) over traditional meta-heuristic algorithms for clustering problems. Unfortunately, all the cuckoos have identical search behaviours that may lead the algorithm to converge to local optima. Also, the convergence rate is sensitive to initial centroids seeds that are randomly generated. Therefore, the main contribution of this paper is to extend the CS capabilities using nonhomogeneous update inspired by the quantum theory in order to tackle the cuckoo search clustering problem in terms of global search ability. Also, the randomness at the beginning step is replaced by the chaotic map in order to make the search procedure more efficient and improve the convergence speed. In addition, an effective strategy is developed to well manage the boundaries. The experimental results on six famous real-life datasets show the significant superiority of the proposed QCCS over eight recent well known algorithms including, genetic quantum cuckoo search, hybrid cuckoo search and differential evolution, hybrid K-means and improved cuckoo search, standard cuckoo search, quantum particle swarm optimization, differential evolution, hybrid K-means chaotic particle swarm optimization and genetic algorithm for all benchmark datasets in terms of internal and external clustering quality.
科研通智能强力驱动
Strongly Powered by AbleSci AI