Research On Visualization Technology Of Large Scale Time-Varying Network Structure

With the rapid increase of the scale of the visual network,the complexity of the network connection structure and nodes increases.At the same time,the sparseness of the network connection and the complex space-time evolution process make the traditional network visualization analysis method difficult to meet the needs of complex computer network structure analysis.In response to this problem,this paper has conducted in-depth research on the visualization technology of large-scale time-varying network structure.Firstly,this paper introduces the visualization of network structure from two aspects of the basic theory and related technologies of large-scale network visualization.On this basis,a FMFR(Fast Multilevel Algorithm Based on FR)layout algorithm that can clearly reflect the structural relationship between nodes and quickly lay out a beautiful large-scale network map is proposed.And then,in the large-scale network data,a visualization model called IGP(Initialization,Grade,Partition)is proposed to draw large-scale dynamic graphs.The superiority of the model is proved by experiments.The main contributions of this paper are reflected in the following three aspects:(1)Based on the FR(Fruchterman and Reingold)layout algorithm,an FMFR algorithm for quickly plotting large-scale network graphs is proposed.The algorithm uses a multilevel method as the framework of the algorithm,and optimizes the repulsive force calculation in the FR algorithm,which reduces the time complexity of the FR algorithm from O(|V|~2)to O(|V|log|V|).At the same time,in order to accelerate the overall efficiency of the algorithm,new methods such as centroid algorithm and constraint normalization are also used in the FMFR algorithm.In this paper,the FMFR algorithm is compared with several classical large-scale network graph drawing algorithms.The experimental results also prove the effectiveness of the algorithm.(2)Based on the FMFR algorithm proposed in this paper,an IGP model is designed and implemented to draw large-scale time-varying network graphs.Firstly,the model uses the FMFR algorithm to generate an initial static graph and initially partition all nodes;Secondly,the concept of"grade"is introduced in the dynamic graph update process to improve the stability of the graph structure,and the node with higher grade is more difficult to move;Finally,the layout is optimized by a simplified force-directed algorithm.In this paper,the average displacement and total energy are used as the evaluation layout standard.The experimental results show that compared with other algorithms of the same type,the IGP model can obtain better layout effect and using less time for layout.(3)Based on the above algorithm,this paper designs and implements a visual prototype system for large-scale time-varying network structure analysis.Several large-scale network graph layout algorithms and some common visualization interaction techniques(eg.navigation,filtering,data tagging,coordinated multiviews,etc.)are integrated in the prototype system.On the one hand,the prototype system can help users discover the structural relationship between large-scale time-varying network data and its specific structural features,and on the other hand,the prototype system helps users to explore and mine hidden information inside the network.