Application Research On Survivability Measure Of Directed Weighted Network

Anti-destroying ability has always been a hot topic in the study of complex network theory. In view of different study fields and different levels, anti-destroying ability has different specific meaning. In the past, we mainly studied anti-destroying measure algorithm of network, but its application research is less. This thesis mainly studies the specific meaning and application prospects on survivability measure of directed weighted network in both China’s civil aviation network and text network.The research content of this thesis mainly includes: data acquisition, network structure, complex features, anti-destroying ability analysis and application. The specific research contents are as follows:(1) The structure, evolution model and anti-destroying of civil aviation network has researched. First, the simulation software Pajek and the civil aviation network data acquisition and processing are amply introduced; Again, the degree distribution and node centrality of the civil aviation network are analyzed; Then, the evolution model of the civil aviation network are discussed based on the Monte Carlo simulation; Finally, the civil aviation network anti-destroying ability is analyzed. The results show that the greater betweenness of the node when received attacks, the greater the influence of network. So we must consciously avoid large transferring quantities of navigable city in the planning of the civil aviation network.(2) The structure, complexity and anti-destroying ability of text network have studied. First of all, relevant technology of language network which based on complex network theory are introduced; then, we put forward the method of text directed weighted network based on word(vocabulary); finally, this thesis makes statistical analysis for the directed weighted text networks, and analyses text network anti-destroying ability from two different angle: the first, this thesis evaluate the importance of text nodes based on the idea of Rank algorithm, and study text network anti-destroying ability about nodes of different levels; another, change trends of text network anti-destroying ability are studied from the article length change angle. Research shows that different nodes of sets have different effect on the performance of text network invulnerability, each text co-occurrence networks has scale-free property with the length increase of both Chinese and English articles, but the change tendency of the parameters in text co-occurrence networks and anti-damage properties are different, revealing some commonalities and differences in both Chinese and English writing. Obviously, the nodes position in the topology layer of text network is more important, the probability that it becomes text keyword is greater; therefore, we can extract the text keywords from text network invulnerability angle.