针对无网格并行计算的基于模糊均值聚类的分区方法  

Partition method based on fuzzy C-mean for meshless parallel computation

作　　者：王琥[1] 李恩颖[1] 李光耀[1] 

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《计算机应用研究》2007年第12期103-105,共3页Application Research of Computers

基　　金：国家自然科学基金重点资助项目(60635020);国家杰出青年基金资助项目

摘　　要：针对国际上普遍采用的基于回归几何二分的分区方法的缺陷,提出了基于模糊均值聚类的均匀模糊均值聚类分区算法。采用该方法对不同类型的无网格几何模型进行了分区,并根据分区信息对三维实体模型进行了并行计算。通过与基于回归几何二分法的比较,充分验证该算法的有效性和可行性。This paper suggested the uniform fuzzy C-mean partition method based on the fuzzy C-mean theory according to the shortcoming of the most popular partition method recursive coordinate bisection method. Also applied this partition method for different type geometry meshless models, and implemented the partition results of 3D meshless method via proposed method for parallel meshless simulation. Compared the partition results and computation efficiency by uniform fuzzy C-mean with the recursive coordinate bisection method, it proved the suggested method was portable and efficiency method for meshless partition.

关 键 词：并行计算 分区 均匀模糊均值聚类 无网格 回归几何二分法 

分 类 号：TP338.6[自动化与计算机技术—计算机系统结构]