作　　者： ; ; ; ;

机构地区： 中山大学地理科学与规划学院水资源与环境研究中心

出　　处： 《水利学报》 2009年第6期743-748,755,共7页

摘　　要： 为了克服模糊C-均值聚类(FCM)算法依赖初值的缺点,引入人工鱼群算法(AFS)建立一种新的聚类算法,应用于洪水分类研究。该算法将聚类中心看作食物源,通过样本抽样产生初始鱼群,利用人工鱼群算法能全局寻优和快速收敛的特点,得到一个较优的初始聚类结果,再使用FCM算法进行局部搜索,以避免因初值选取不当,而有可能陷入局部最小的缺陷。该方法应用于对西江流域洪水资料的分析结果表明,新算法具有比FCM算法更好的性能表现,使得到的分类结果更加准确合理。 In order to overcome the shortcoming of depending on starting value of the Fuzzy C-Mean Clustering （FCM） algorithm, the Artificial Fish Swarm Algorithm （AFSA） is introduced to combine with this algorithm to establish a new method for flood classification. The AFSA takes the cluster center as food source and produces the initial fish swarm through sampling. The advantages of global optimization searching and rapid convergence of AFSA are utilized to find a superior initial cluster result. On this basis the FCM is applied to carry out the local search for the final optimization, by which the problem of falling into partial smallest value due to improper starting value possibly happened to FCM algorithm can be avoided. The proposed method is applied to analyze the flood data of Xijiang River. The result verifies that it gives more accurate and reasonable classification than that obtained by simplex FCM algorithm.

关 键 词： 人工鱼群算法 模糊 均值聚类算法 洪水分类

领　　域： [天文地球] [水利工程] [天文地球]