BA Improved Scale-free Network Model And Optimization Of Network Path

Complex network theory can be used to describe a large number of phenomena in brain neural network communication.The study found that the brain network structure of many species has typical complex network characteristics.For example,the degree distribution of the network obeys the power law distribution,and the network topology has the characteristics of high clustering coefficient and short path length in the small world network.In complex networks,the concepts of growth networks and prioritized connections were introduced in the BA scale-free network model,making it become one of the most commonly used models in complex networks.However,in the BA network,the degree of each new node is the same.The priority connection probability is linear with the original node value.The new nodes cannot generate new connections and the clustering characteristics is poor.These shortcomings limit the scope of application of the BA network in the brain network.According to the characteristics of information transmission between neurons in biological neural networks,this paper improves the above disadvantages of BA scale-free network.First,the concept of node importance was introduced in this paper,indicating the importance of the node in the network,and the importance of all nodes in the network obeys the power law distribution.According to the importance of each node,the degree value of the node and the weight of the edge connected to the node(referred to herein as the connection strength)can be calculated.Therefore,the degree of each node newly joined to the network varies.At the same time,the concept of connection cost was proposed,which measures the cost of information passing through the edge.Secondly,a BA improved network model was proposed in this paper,and two improvements to the priority connection probability was introduced at the same time.One is that the priority connection probability is nonlinearly positively correlated with the degree of the original node of the network,and the other is that the priority connection probability is affected by the original node.The degree of the relationship between the degree and the distance between the old and new nodes is a positive correlation with the degree and a negative correlation with the distance.In this paper,the influence of nonlinear parameters on the network degree distribution in these two improved methods is analyzed,and the range of parameter values is confirmed.That makes the network degree distribution obey the power law distribution.In addition,this paper analyzes the static characteristics of the BA improved networks in terms of distance distribution,connection cost,clustering coefficient and average path length.The results show that when the priority connection probability is affected by the dual value of the original node's degree and the distance between the old and new nodes,the network has the characteristics of the small world network including high clustering coefficient and short path length.The network nodes tend to communicate with each other by shorter path which will cause low connection costs.Finally,based on the four indicators of network efficiency,degree,feature vector and median,this paper calculates the value of each node in the BA improved network,and then identifies the central node of the network.Then based on the Dijkstra shortest path algorithm based on the minimum connection cost,a path optimization algorithm is proposed.By establishing a new path between the original nodes of the network,the path optimization is performed on the shortest path that satisfies the condition,including no center optimization and central optimization.The simulation results show that this path optimization algorithm has lower connection cost for communication between nodes than the Dijkstra shortest path algorithm.