Research On Grid Resource Fuzzy Clustering

The technique of the Grid is the further development to the distributed network and realizes to share common resources at the higher level, which organizes the heterogeneous resources that belong to different administrative domains flexibly and effectively. In that way,People can not only share the information of web pages, but also share all the soft or hard resources, such as computing resources, storage resources,network resources and so on.In this paper, with the goal of improving the efficiency of resource discovery, the resource discovery method is researched thoroughly,.And in-depth analysis of the research situation, this paper presents a new fuzzy clustering method.Fuzzy C-Means cluster algorithm (FCM) is a widespread and sensitive algorithm, but it also has many shortcomings as the sensibility to noisy data, it usually leads to local minimum, and it depends on initial value, especially in the instance of large amounts of clustering objects. This paper proposed a hybrid FCM algorithm based on PSO and the trigger of GA to maximize the clustering of grid resources. By setting the trigger condition, avoided the potential local iteration, and simplified the complexity of the algorithm a certain extent.To verify the effectiveness and the availability of the proposed method, the simulation experiments were carried out.Through comparative analysis with the traditional algorithms, simulation results indicate that the method improves the running speed and accuracy of clustering, and can be well applied to the grid resource clustering, thereby the preliminary work of grid resource scheduling will be optimized.At last the concepts of this paper are summarized and future works are proposed.