A Fuzzy Clustering Optimization Method for Task Decomposition at Dynamic Alliance Layer
-
摘要: 分析了动态企业联盟对任务分解的需求,提出了一种企业层任务向联盟层内子项目任务聚合的模糊聚合算法。通过对功能任务模块间的联系程度按照0~9赋值,根据各模块间的联系密切程度赋值,形成各模块间关系的邻接矩阵,并设置不同的阈值对邻接矩阵进行模糊聚类;综合各任务模块内的聚合度和模块之间的分离度为适应度函数,并以此为评判指标,计算在各阈值下的适应度函数值。以适应度函数最大值时的阀值为最优聚类阀值,最终得到任务在联盟层内最佳模块分解方案。Abstract: This paper analyses the needs for dynamic alliance to do its task decomposition and then proposes a fuzzy clustering algorithm which clusters the task from enterprise layer to subprojects in alliance layer.It assigns values to the degrees of closeness of relationship among task modules according to the numbers 1 through 9 T and then creates the adjacency matrices of the modules.It then sets different thresholds to perform the fuzzy clustering of the adjacency matrices.It uses the clustering degree among task modules and the separation degree between them as the values of adaptiveness function to calculate the value of the adaptiveness function at each threshold value.It finally obtains the best scheme of task module decomposition with the best clustering threshold value when the value of the adaptiveness function is maximum.
-
Key words:
- dynamic alliance /
- task decomposition /
- fuzzy clustering /
- task module
-
[1] 高新波.模糊聚类分析及应用[M]. 西安:西安电子科技大学出版社,2004 [2] Mali K.Clustering and its validation in a symbolic framework[J]. Pattern Recognition Letters,2003,24:2367~2376 [3] 何清.模糊聚类分析理论与应用研究进展[J]. 模糊系统与数学,1998,12(2):89~94 [4] Zadeh L A.Fuzzy sets[J]. Information and Control,1965,(8) [5] 杨海成,廖文和.基于知识的三维CAD技术及应用[M]. 北京:科学出版社,2005
点击查看大图
计量
- 文章访问数: 275
- HTML全文浏览量: 25
- PDF下载量: 4
- 被引次数: 0