Research On Complex Network Centrality And Applications In Disaster Broadcasting

Discrepancies among nodes in complex network are obvious.How to evaluate importance of nodes in complex network has become a vital problem in complex network research,especially in application fields.This paper focuses on network centrality,and has done some concrete researches from both theoretical and practical perspectives.Compared with traditional centrality measures,this paper strengthens the combination of practical requires and network topology characters to determine a node's importance,rather than topology structure only.We also propose a match method based on analysis of centrality differences and application requires helping user to find a proper centrality measure according to real requires.As to emergency chain reaction characteristics,we introduce modeling theory of complex network to set up a general framework to describe how a single emergency event evolves to a regional disaster chain or network.From different network hierarchies,the same emergency event may have various properties,which can be simulated through our disaster diffusion model and regional key event can also be detected.Theoretically,this paper is based on recent theory of complex network.Starting from centrality definition,different centrality measures have been discussed.There basic categories of node relation problem,short-path problem and network flow problem are proposed to describe the difference of centrality measure.On the other hand,practical network requires are concluded to node region,information transparency and node dynamics.Based on bilateral research on theory and application requires we set up a match model,which can give some hints to solve the problem of different measures,applied in the same network brought out different results.What is more,we also discuss the possibility of how to integrate different measures to a single need.A general framework is founded to describe emergency evolution modes,such as single emergency,disaster chain and disaster network.This method can make all connection among various emergencies within a region clear,and also help administrator to realize which event maybe the potential disaster.Practically,this paper insists finding a proper way to apply centrality theory to real network.Firstly,we extract real network requires and analyze how topology determine node importance.We strengthen the positive aspect of network need to redefine node importance. As considering to different hierarchies play different roles in some broadcast mechanism.We propose a node sub-tree method to quantify hierarchy influence,and also introduce a compensation function to describe nodes' differences.Meanwhile,we simulate disaster diffusion on infrastructure network such as power grid and water pile line system.Through screening simulation results,we can recognize which node is the hub.