Research On Multiple Sources Localization Methods In Complex Networks For Security Incident

Communication in complex network can depict the spread process of a large number of real phenomena such as diseases,rumors and etc.It is an important research direction of network dynamics and has attracted the attention of a large number of scholars.Source localization in complex network is the reverse process of spread and a new challenging research direction.The problem of multiple sources localization in complex network is more complicated and relevant research has great theoretical significance and practical value.Security incidents as traction,this paper abstracts the problem of source localization of security incident into the problem of source localization in complex network,and focuses on the theory,method and application of multiple sources localization in network,develops a new method in multiple sources localization at last.Specifically,it includes the following three aspects:Firstly,from the perspective of time information,the paper defines the problem of multiple sources localization in complex network and then proposes a multi-sources localization method based on the time information of observers.Based on network topology and partial timestamps,through simulating the reverse diffusion process in complex network,this paper uses grouping and set covering to locate the multiple sources.The experiments on synthetic networks and empirical networks show that the accuracy of sources localization of this method is significantly better than the benchmark algorithms,and it has good robustness and anti-noise.Secondly,from the perspective of state information,this paper defines the problem of multiple sources localization in complex network and then proposes a multi-sources localization method based on the state information of observers.Based on the given potential concentration label of each node in complex networks,this paper inverts the initial state of network and finds the nodes with maximal of potential concentration label as the sources.The experiments on synthetic networks and empirical networks show that the accuracy of sources localization of this method is significantly better than the benchmark algorithms,and it has good robustness and anti-noise.Thirdly,two kinds of multi-sources localization methods based on time information and state information are respectively applied to urban space network.The model of urban space network was designed and constructed.Based on the spread process of urban space network,the multi-sources localization methods based on time information and state information were applied to urban space network,respectively,and the multiple sources localization experiments were carried out.The experiment results show that the two kinds of multi-sources localization methods are effective and have good anti-noise performance.