Research On Network Topology Segmentation Algorithm Based On Dumbbell Structure

For the complex network,the research on the network topology is the foundation of the network study.It can help to build the complicated network model and explore the complex network by constantly exploring the inherent topology of the real complex network as the network is large and complex in the real life.Based on the self-similarity of complex networks,this thesis presents a new type of topology named-dumbbell structure.The dumbbell structure is defined as the number of nodes in the structure that are neighbors of each other and only two nodes are greater than or equal to 3.Based on the definition of dumbbell structure,a simple dumbbell network growth model is constructed,and a network topology partitioning algorithm based on dumbbell structure is proposed,which divides the network into different dumbbells.Then we choose the small-world network and the scale-free network to verify the dumbbell segmentation algorithm.The experimentally generated network proves that the real network with small world and the real network with scale-free can be divided into different dumbbells.The entire network can be divided into dumbbell structure subnets.In the dumbbell topology,the master node and the center side play more important roles,by improving the performance of the main node and the center side,the performance and quality of the network can be improved significantly.The algorithm divide the network into various levels of dumbbells,with different importance.In the network structure the higher level a dumbbell has,the more important it is.With limited network resources,you can give priority to improve the performance of high-level dumbbells to make better use of network resources.