Event Space-Correlation Analysis Algorithm Based on Ant Colony Optimization
Historical disaster events are taken as a case for space-correlation analysis, three-dimensional disasters space-time complex network are modeled and chain relationship of disaster nodes are mined by looking for similar space vector in network. Then trans
- PDF / 277,805 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 89 Downloads / 222 Views
Abstract. Historical disaster events are taken as a case for space-correlation analysis, three-dimensional disasters space-time complex network are modeled and chain relationship of disaster nodes are mined by looking for similar space vector in network. Then transformed the vector discover problem into a path optimization problem and solved by using ant colony algorithm, where the pheromone parameter in the process of optimal-path finding is concerned as the algorithm result, in order to solve the problem of path competition which existed when only to solve the optimal path. Experimental results of MATLAB show that this method has high accuracy and practicality. Keywords: Complex network Space correlation Colony Optimization Pheromone
Disaster chain
Ant
1 Introduction More and more facts have shown that drought, floods, earthquake and other disaster events often do not exist in isolation, but have some connections. Such as, after 8.5-magnitude earthquake in Indonesia Sumatra in March 2005, major flood struck China’s Pearl River Basin in July, and after the 8.3-magnitude earthquake in Sumatra in September 2007, snow and ice storms struck the south of China in early 2008, because of this similar chain-like cognate phenomenon of disasters, the disaster chain and disaster evolution mechanism [1–4] have been got more and more attention, many researchers of our country believe that the occurrence of the Sumatra earthquake is directly related to the above two disasters of our country, “(earthquake causes) the evaporation of sea water causes clouds and rain, they become flood in summer, become winter rain or snow in winter” [5–7], and proposed the disaster’s homology, chain nature and rhythmicity which reflected by the relationship between disasters, has a high value for the prediction of disasters. However, the relationship between disasters is not often instantaneous and self-evident, but is a large span of geographical and time, so the analysis and excavation of the relationship between disasters is a very complex problem. Now the disaster correlation studies are mainly statistical analysis of disaster records, however, due to the space migration and time delayed etc. complex characteristics of the disaster correlation, it is difficult to effectively extract the association rules, thus the efficiency and credibility are low. Literature [8, 9] was based on time series similarity matching, achieved the correlation analysis for the earthquake areas, © Springer International Publishing Switzerland 2016 D.-S. Huang et al. (Eds.): ICIC 2016, Part I, LNCS 9771, pp. 563–570, 2016. DOI: 10.1007/978-3-319-42291-6_56
564
S. Liu et al.
and achieved very good results, but the model can only obtain generally relevance degree, but cannot determine the causal relationship between the regions, namely, it shows the law of the region A cause the earthquake of region B, or shows the law of the region B cause the earthquake of region A, or both A and B are the areas of frequent earthquake, they shows a fake matching. In view
Data Loading...
