Research Of Complex Network Evolution Model And Node Importance

Many real networks such as Internet networks,social networks and biological networks and so on, which can be abstracted into complex networks.These networks have small word effect and scale-free behavior, It is many researchers'driving force to study the relationship between these topological features,which makes the complex network become a hot research topic.Based on preferential attachment producing mechanism of BA model,many improved evolution models are proposed by scholars, which can not only capture the dynamic nature of network formation,but also has important meaning for rationality of actual networks designing and structural characteristics.In this paper,we mainly study the complex network evolution models and nodes importance,and the main work is as following.1.We summarize complex network evolution models and related theories of complex network.2.In allusion to the shortage of preference connection mechanism of BA model, this paper proposes two.improved network evolution models.The first model is based on the improved PageRank, which uses the improved algorithm PageRank to replace degree as nodes connected priority mechanism in BA model,and improved algorithm PageRank considers a more realistic browsing behavior.Computer simulation results show improved network model not only meets the power-law degree distribution,but also it's average path length and clustering coefficient are better than original BA network evolution model.therefore,the improved network model is proved to be more reasonable.The other model is a hybrid model,which considers nodes importance can not only reflect by the importance of neighbor nodes,and should have connection with itself attribution.It uses corresponding nodes importance assessment method replace degree as nodes connected priority mechanism in BA model.Experiment results show that compared with the BA network evolution model,the hybrid model's clustering coefficient smaller than BA model,and the average path length longer,when the network reaches a fixed size,3.The essence of the network evolution model is that network formation mechanism is based on node importance.To make better network evolution model,it study the assessment method of the node importance in-depth.This paper proposes two integrated measuring methods of the node importance including adjustable weighted measuring method and based on the equivalence class integrated measuring method.Adjustable weighted measuring method comprehensive considers three commonly methods of measuring nodes importance-degree,centrality, betweeness, Experimental results show that compared to three common methods,adjustable weighted measuring method is more integrated and more comprehensive to measuring the node importance.Based on the equivalence class integrated measuring method of nodes importance that uses the most stringent partial order relationship to comprehensive evaluate node importance,which is base on common measuring method of nodes importance.In the strict partial order circumstances,it gives equivalence classes results in the network correspongding set.Experimental results demonstrated that the nodes importance evaluate algorithm based on equivalence class seemed more reasonable than the classical algorithm PageRank.It solves that BA model uses degree to evaluate node importance,which making the network evolution is more advantageous.