Research And Application Of Network Topology Layout Algorithm In Special-shaped Area

As one of the branches of information visualization,the network topology visualization technology which has played an active role in people's management of network and analysis of network performance,has enhanced people's understanding of the structure and relationships of various networks such as the Internet and social networks from the visual point of view.As the core of topology visualization,network topology layout algorithms have always been an important research direction.Due to the diversity of network data sources,the layout of topology is no longer limited to the performance of traditional logical view.Researches on the layout of various specific constraints should be focused.With the development of Geographic Information System,the layout algorithms within the multi-shaped regions resulting from the combination of geographic information and topology have become new research propositions.The traditional layout method only deals with the logical relationships,and it cannot adapt well to the topological display based on geographical information which mainly reflects in its poor layout space occupancy.On the other hand,with the increase of the layout areas,the visual clutter problem caused by the edges in topology visualization becomes more and more serious.Although simplifying edges through abstract nodes can alleviate the adverse effects,it does lose more details.In this case,edge bundling can take into account the details and compressing the visual space to some extent.Traditional edge bundling methods mostly deal with the geometric features of the edges.For the network layout,the display of topology information will be relatively ignored.Based on these above,this thesis makes the following tasks from the perspective of node layout and edge compression layout:1.To solve the problem of node layout within multi-shaped regions,a node layout method is proposed for the special-shaped areas according to the traditional FR model.The algorithm firstly uses the Label Propagation algorithm to do community detection and determines multi-zone constraints for community layouts through additional geographic information.Then it adds attractive forces of regional vertices based on the FR model to move nodes.In the end it uses multi-layered areas to attract nodes to do second adjustments.After that above,the resulting network topology layout can reflect the constraints of the real geographical information and keep neat.2.As to the problem of visual clutter caused by the increasing number of edges,this thesis proposed a hierarchical edge bundling algorithm based on the node layout.The edges are divided into two layers according to the different topological information.In the interior of the region,the quick edge bundling which uses special data structures referring to the traditional edge-to-edge geometric feature is considered,and the community relationship attribute is added to the edge bundling between the regions.This not only finishes the job of space compression with details retaining,but also highlights the performance of topological relationships among regions.3.A network topology visualization system in special-shaped areas is designed and implemented according to the different processing of nodes and edges described above.The thesis completes the system design and implementation in detail from the perspectives of design background,implementation modules,and system functions verification tested by real data.This thesis studies the network topology of specific constraints from the perspective of nodes and edges,and shows the final good visual effects.It has a positive effect on the display of network structures.