网格资源模糊聚类查找的研究

Study for grid resources search on fuzzy clustering

为了解决动态网格环境中资源查找的难题,提出了基于特征加权模糊K-原型聚类的网格资源查找算法。该算法根据资源请求对各维资源关心程度的不同,用特征加权模糊K-原型聚类算法对数值型、类属型并存的混合型网格资源节点集合进行划分。然后根据资源的静态数值特征与类属特征,确定与资源请求属性特征值最相似的类簇。最后综合资源的动态数值特征选择最优的资源节点。模拟实验的结果表明,与其他同类算法比较,算法能提高资源查找的查准率、鲁棒性和降低平均响应时间。

In order to solve the problem of resources search in dynamic grid environment,this paper proposes a search algorithm for grid resources based on feature weighted fuzzy K-prototypes algorithm.The algorithm partitions mixed grid resource node sets that include numerical and categorical resources based on different weight of resource request and feature weighted fuzzy K- prototypes algorithm.Then,according to static numerical feature and categorical feature of resources,the algorithm fixes the most similar resource cluster with feature eigenvalue of resource request.Finally,the algorithm selects optimal resource node through dynamic numerical feature of resources.Results of simulation experiment show that the method can not only improve precision and robustness but also reduce mean response time of resources search compared with other methods.