Network Operation Situation Awareness Based On Adaptive And Weighted Fuzzy Clustering

Today,network has been closely related to people's work and life.Network operation situation awareness means to analyze the current network operation status and predict the variation trend of operation status,so the network administrator can make corresponding countermeasures when the network will deteriorate.The thesis focuses on the research of network situation awareness model and assessment method.The thesis proposes a network operation situation awareness model which combines the improved evaluation method to analyze the network operation data to obtain the network operation situation value.The research content and innovation points of this thesis are as follows:1.Establish the network operation situation awareness model.The thesis refers to the model of JDL,Endsley and OODA,proposes the network operation situation awareness model.The model could be mainly divided into four layers: network management layer,rule layer,situation decision-making layer and situation prediction layer.2.In the thesis,fuzzy clustering is adopted as the evaluation method of network operation situation awareness.In view of the two problems of fuzzy clustering method(FCM),which is sensitive to the initial clustering center and the clustering number,the thesis proposes the corresponding solutions.Finally,an improved fuzzy clustering algorithm based on attribute weighting is proposed.1)The sample density is used to set the initial clustering center,because the clustering center is the sample point where most samples are concentrated.2)The thesis proposes an improved fuzzy clustering effectiveness index,namely NL index.The simulation results of adaptive fuzzy clustering algorithm(AFCM)show that this index is superior to the L index.That is,the maximum value is obtained at the optimal number of clusters with a high probability.3)An improved weighted fuzzy clustering algorithm is proposed.Aiming at the classification accuracy problem of traditional fuzzy clustering algorithm,a fuzzy clustering algorithm based on attribute weight weighting was proposed.The difference in the calculation of attribute weight in this thesis is PCA is used to analyze the characteristic values of the attributes of samples,and then the attribute weight can beobtained by mapping.3.Network operation situation simulation experiment.Using the Abilene network network as a simulation environment,PCA-FCM algorithm is used to get the operation state and draw the diagram of the operating state of the network running for a week.