摘要
首先分析蚁群聚类算法,并指出其存在的问题;然后给出传统的蚁群聚类算法在船舶电网云数据聚类的实现流程,针对算法中存在的问题,提出利用惯性因子、随机初始化等方式改进和优化算法对船舶电网故障进行诊断;最后通过实验进行说明,优化后的蚁群聚类算法与K-mean算法、粒子群K-mean算法相比具有较好的收敛性。
Firstly,this paper analyzed the ant colony clustering algorithm,and points out the problems. Then give the traditional ant colony clustering algorithm implementation process in ship power cloud data clustering. According to the algorithm problems,proposed the use of inertia factor,random initialization and other ways to improve and optimize the algorithm for ship power system fault diagnosis.Finally,experimental results showed optimized ant colony clustering algorithm was better convergence than K-mean algorithm,particle swarm K-mean algorithm.
出处
《舰船科学技术》
北大核心
2016年第4X期19-21,共3页
Ship Science and Technology
关键词
蚁群算法
船舶故障诊断
云数据聚类
ant colony algorithm
fault diagnosis of ships
cloud data clustering
